pdfs.semanticscholar.org · journal of physiology volume 582, issue 2,july 2007,pp. 471-524...

Journal of Physiology

Volume 582, Issue 2,July 2007,pp. 471-524

PERSPECTIVES R. Alberto Travagli

The nucleus tractus solitarius: an integrative centre with ‘task-matching’ capabilities J Physiol 2007 582: 471. First Published online on May 31, 2007

D. A. Giussani Hypoxia, fetal growth and early origins of disease: the Andean curse on the Conquistadors J Physiol 2007 582: 472. First Published online on June 7, 2007

Stefan Kääb Variety is the spice of life: searching for the substrates of regional myocardial electrical properties J Physiol 2007 582: 473. First Published online on May 31, 2007

R. McNeill Alexander Flat and bouncy walking J Physiol 2007 582: 474. First Published online on May 17, 2007

Mark W. J. Strachan Physiological responses to hypoglycaemia – not all ‘just in the head' J Physiol 2007 582: 475-476. First Published online on June 21, 2007

Jonathan D. Kaunitz NKCC1: tales from the dark side of the crypt J Physiol 2007 582: 477. First Published online on May 24, 2007

JOURNAL CLUB Sachin Makani and Edward Zagha

Out of the cleft: the source and target of extra-synaptic glutamate in the CA1 region of the hippocampus J Physiol 2007 582: 479-480. First Published online on June 7, 2007

TOPICAL REVIEW Klaus I. Matthaei

Genetically manipulated mice: a powerful tool with unsuspected caveats J Physiol 2007 582: 481-488. First Published online on May 10, 2007

CELLULA Shingo Hotta, Kozo Morimura, Susumu Ohya, Katsuhiko Muraki, Hiroshi Takeshima, and Yuji Imaizumi

Ryanodine receptor type 2 deficiency changes excitation–contraction coupling and membrane potential in urinary bladder smooth muscle J Physiol 2007 582: 489-506. First Published online on March 15, 2007

Amy Reynolds, Alyson Parris, Luke A. Evans, Susanne Lindqvist, Paul Sharp, Michael Lewis, Richard Tighe, and Mark R. Williams

Dynamic and differential regulation of NKCC1 by calcium and cAMP in the native human colonic epithelium J Physiol 2007 582: 507-524. First Published online on May 3, 2007

NEUROSCIENCE Gee Hee Kim, Shuji Suzuki, and Kenro Kanda

Age-related physiological and morphological changes of muscle spindles in rats J Physiol 2007 582: 525-538. First Published online on May 10, 2007

Jin Bong Park, Silvia Skalska, Sookjin Son, and Javier E. Stern Dual GABAA receptor-mediated inhibition in rat presympathetic paraventricular nucleus neurons J Physiol 2007 582: 539-551. First Published online on May 10, 2007

Jackie W. Lu, Victor B. Fenik, Jennifer L. Branconi, Graziella L. Mann, Irma Rukhadze, and Leszek Kubin

Disinhibition of perifornical hypothalamic neurones activates noradrenergic neurones and blocks pontine carbachol-induced REM sleep-like episodes in rats J Physiol 2007 582: 553-567. First Published online on May 10, 2007

Erika D. Eggers, Maureen A. McCall, and Peter D. Lukasiewicz Presynaptic inhibition differentially shapes transmission in distinct circuits in the mouse retina J Physiol 2007 582: 569-582. First Published online on April 26, 2007

G. B. Awatramani, J. D. Boyd, K. R. Delaney, and T. H. Murphy Effective release rates at single rat Schaffer collateral–CA1 synapses during sustained theta-burst activity revealed by optical imaging J Physiol 2007 582: 583-595. First Published online on April 26, 2007

Dmitriy Fayuk and Jerrel L. Yakel Dendritic Ca2+ signalling due to activation of 7-containing nicotinic acetylcholine receptors in rat hippocampal neurons J Physiol 2007 582: 597-611. First Published online on May 17, 2007

T. W. Bailey, S. M. Hermes, K. L. Whittier, S. A. Aicher, and M. C. Andresen A-type potassium channels differentially tune afferent pathways from rat solitary tract nucleus to caudal ventrolateral medulla or paraventricular hypothalamus J Physiol 2007 582: 613-628. First Published online on May 17, 2007

Erin M. Johnson, Ethan T. Craig, and Hermes H. Yeh TrkB is necessary for pruning at the climbing fibre–Purkinje cell synapse in the developing murine cerebellum J Physiol 2007 582: 629-646. First Published online on April 26, 2007

Carlos R. Cassanello and Vincent P. Ferrera Computing vector differences using a gain field-like mechanism in monkey frontal eye field J Physiol 2007 582: 647-664. First Published online on May 17, 2007

CARDIOVASCULAR Natalia Igosheva, Paul D. Taylor, Lucilla Poston, and Vivette Glover

Prenatal stress in the rat results in increased blood pressure responsiveness to stress and enhanced arterial reactivity to neuropeptide Y in adulthood J Physiol 2007 582: 665-674. First Published online on May 10, 2007

Nathalie Gaborit, Sabrina Le Bouter, Viktoria Szuts, Andras Varro, Denis Escande, Stanley Nattel, and Sophie Demolombe

Regional and tissue specific transcript signatures of ion channel genes in the non-diseased human heart J Physiol 2007 582: 675-693. First Published online on May 3, 2007

Regis R. Lamberts, Nazha Hamdani, Tenoedj W. Soekhoe, Nicky M. Boontje, Ruud Zaremba, Lori A. Walker, Pieter P. de Tombe, Jolanda van der Velden, and Ger J. M. Stienen

Frequency-dependent myofilament Ca2+ desensitization in failing rat myocardium J Physiol 2007 582: 695-709. First Published online on May 3, 2007

Simon McMullan, Ann K. Goodchild, and Paul M. Pilowsky Circulating angiotensin II attenuates the sympathetic baroreflex by reducing the barosensitivity of medullary cardiovascular neurones in the rat J Physiol 2007 582: 711-722. First Published online on March 15, 2007

Vladimir Ivancev, Ivan Palada, Zoran Valic, Ante Obad, Darija Bakovic, Niki M. Dietz, Michael J. Joyner, and Zeljko Dujic

Cerebrovascular reactivity to hypercapnia is unimpaired in breath-hold divers J Physiol 2007 582: 723-730. First Published online on April 5, 2007

R. F. Kelly and H. M. Snow Characteristics of the response of the iliac artery to wall shear stress in the anaesthetized pig J Physiol 2007 582: 731-743. First Published online on April 5, 2007

Nicole M. Rummery, James A. Brock, Poungrat Pakdeechote, Vera Ralevic, and William R. Dunn

ATP is the predominant sympathetic neurotransmitter in rat mesenteric arteries at high pressure J Physiol 2007 582: 745-754. First Published online on May 17, 2007

RESPIRATORY Stéphane N. Glénet, Claire De Bisschop, Frederic Vargas, and Hervé J. P. Guénard

Deciphering the nitric oxide to carbon monoxide lung transfer ratio: physiological implications J Physiol 2007 582: 767-775. First Published online on May 10, 2007

Nadia Randrianarison, Brigitte Escoubet, Chrystophe Ferreira, Alexandre Fontayne, Nicole Fowler-Jaeger, Christine Clerici, Edith Hummler, Bernard C. Rossier, and Carole Planès

-Liddle mutation of the epithelial sodium channel increases alveolar fluid clearance and reduces the severity of hydrostatic pulmonary oedema in mice J Physiol 2007 582: 777-788. First Published online on April 12, 2007

ALIMENTARY Hiroe Yanagida, Kenton M. Sanders, and Sean M. Ward

Inactivation of inducible nitric oxide synthase protects intestinal pacemaker cells from postoperative damage J Physiol 2007 582: 755-765. First Published online on May 17, 2007

SKELETAL MUSCLE AND EXERCISE Firas Massaad, Thierry M. Lejeune, and Christine Detrembleur

The up and down bobbing of human walking: a compromise between muscle work and efficiency J Physiol 2007 582: 789-799. First Published online on April 26, 2007

Patrick T. Fueger, Candice Y. Li, Julio E. Ayala, Jane Shearer, Deanna P. Bracy, Maureen J. Charron, Jeffrey N. Rottman, and David H. Wasserman

Glucose kinetics and exercise tolerance in mice lacking the GLUT4 glucose transporter J Physiol 2007 582: 801-812. First Published online on May 10, 2007

Satoshi Fujita, Hans C. Dreyer, Micah J. Drummond, Erin L. Glynn, Jerson G. Cadenas, Fumiaki Yoshizawa, Elena Volpi, and Blake B. Rasmussen

Nutrient signalling in the regulation of human muscle protein synthesis J Physiol 2007 582: 813-823. First Published online on May 3, 2007

Gordon L. Warren, Mukesh Summan, Xin Gao, Rebecca Chapman, Tracy Hulderman, and Petia P. Simeonova

Mechanisms of skeletal muscle injury and repair revealed by gene expression studies in mouse models J Physiol 2007 582: 825-841. First Published online on May 3, 2007

Carlo Cifelli, François Bourassa, Louise Gariépy, Krystyna Banas, Maria Benkhalti, and Jean-Marc Renaud

KATP channel deficiency in mouse flexor digitorum brevis causes fibre damage and impairs Ca2+ release and force development during fatigue in vitro J Physiol 2007 582: 843-857. First Published online on May 17, 2007

INTEGRATIVE Denham S. Ward, William A. Voter, and Suzanne Karan

The effects of hypo- and hyperglycaemia on the hypoxic ventilatory response in humans J Physiol 2007 582: 859-869. First Published online on May 3, 2007

Jane K. Cleal, Paul Brownbill, Keith M. Godfrey, John M. Jackson, Alan A. Jackson, Colin P. Sibley, Mark A. Hanson, and Rohan M. Lewis

Modification of fetal plasma amino acid composition by placental amino acid exchangers in vitro J Physiol 2007 582: 871-882. First Published online on May 3, 2007

Stacy Zamudio, Lucrecia Postigo, Nicholas P. Illsley, Carmelo Rodriguez, Gladys Heredia, Michael Brimacombe, Lourdes Echalar, Tatiana Torricos, Wilma Tellez, Ivan Maldonado, Elfride Balanza, Tatiana Alvarez, Julio Ameller, and Enrique Vargas

Maternal oxygen delivery is not related to altitude- and ancestry-associated differences in human fetal growth J Physiol 2007 582: 883-895. First Published online on May 17, 2007

J Physiol 582.2 (2007) p 471 471

PERSPECT IVES

The nucleus tractus solitarius: anintegrative centre with‘task-matching’ capabilities

R. Alberto Travagli

Neuroscience Pennington Biomedical

Research Center-LSU System, 6400 Perkins,

Road, Baton Rouge, LA, 70808, USA

Email: [email protected]

Over the years many of our neuro-

science colleagues have questioned our

reasons for studying ‘boring’ brainstem

autonomic circuits. Their assumption was

that the neurons comprising these reflex

pathways were simply relay stations with

no ‘personality’, i.e. the response of the

neuronal population ‘in toto’ was already

pre-determined resulting in a generic,

stereotyped and predictable outcome. All

integrative capabilities for modulation

of cardiorespiratory or gastrointestinal

reflexive functions were to be determined by

‘higher’ neuronal systems such as the hypo-

thalamus or the enteric nervous system.

An increasingly large amount of evidence

confuting this diminutive concept of brain-

stem reflexive circuits, though, is starting to

emerge.

Neurons of the nucleus tractus solitarius

(NTS) receive vagal sensory information

from cardiorespiratory and sub-

diaphragmatic organs of the gastrointestinal

tract. NTS neurons comprise many inter-

mixed cellular types encompassing a

vast array of neurochemical phenotypes

scattered throughout the various NTS sub-

nuclei with no apparent organization. This

type of apparent dissociated organization

makes it hard to identify with certainty

the visceral inputs and function(s) of

specific neuronal populations and only

an approximate guess can be made with

regards to their target organ.

In this issue of The Journal of Physiology,

Bailey and colleagues (Bailey et al.

2007) combine neuroanatomical tracing

techniques with an electrophysiological

approach in horizontal brainstem slice

preparations to identify and characterize

NTS neurons that project selectively to

either the paraventricular nucleus of the

hypothalamus (PVN) or the caudal ventro-

lateral medulla (CVLM). This experimental

strategy allows the correlation between long

distance connections (PVN and CVLM) and

the ‘black box’ of the local networks of the

NTS. In their work, Bailey and colleagues

report that identified NTS neurons

projecting to PVN can be distinguished

from those neurons projecting to CVLM

based on the presence of the fast-transient

IA-like potassium current, on their larger

soma size and their more complex dendritic

morphology. Conversely, CVLM-projecting

NTS neurons had a more faithful response

to stimulation of the tractus solitarius. The

paper by Bailey et al. is just the most recent

work in a series of publications from this

group that aim to provide evidence of the

exquisite organization of these pathways and

networks in the control of cardiovascular

functions (Bailey et al. 2002; Jin et al. 2004).

Our group is conducting research with a

very similar focus, although concentrating

on the circuits controlling gastrointestinal

homeostatic functions (Browning et al.

2006; Browning & Travagli, 2006). The

overall idea driving the work of these groups

is that brainstem circuits are specialized

or tuned at multiple levels, starting from

the neuronal membrane itself and working

up from the local network connections

to the pathways projecting to very distant

areas. The biophysical, pharmacological

and synaptic diversity in the neuronal

organization, although confusing (we are

just at the beginning of the discovery

phase), may actually represent meaningful

patterns of specialization or segregated

lines of specificity, where synaptic inputs

and membrane characteristics offer a level

of potential mechanistic redundancy that

allows adjacent neurons to ‘recognize’ and

reinforce each other’s common pathway

and goal.

This type of cellular organization implies

a ‘task matching’ capability even within

the brainstem neurons of the NTS. In

fact, although small in size, the NTS is

devoted to the integration of vital cardiac,

respiratory and gastrointestinal functions,

whose demands vary greatly both in terms

of timing as well as duration of response.

Whilst we can afford to have a gastro-

intestinal response occurring with a delay

of a few seconds (or minutes), such

an untimely response would certainly be

incompatible with life when applied to the

control of the baroreflex or respiration.

This type of cellular organization within

the NTS should force us to be extremely

cautious in the physiological interpretation

of generic manipulations that analyse a

single outcome such as cFos expression,

calcium oscillations, blood pressure or

gastric motility responses to micro-

injections. A multifaceted approach that

includes the electrophysiological analysis

of cellular mechanisms is needed to put

in place the pieces of the puzzle of that

beautiful ‘task-matching’ black box of the

NTS.

References

Bailey TW, Hermes SM, Whittier KL, Aicher SA

& Andresen MC (2007). J Physiol 582,

613–628.

Bailey TW, Jin YH, Doyle MW & Andresen MC

(2002). J Neurosci 22, 8230–8237.

Browning KN & Travagli RA (2006). Auton

Neurosci 126–127, 2–8.

Browning KN, Zheng Z, Gettys TW & Travagli

RA (2006). J Physiol 575, 761–776.

Jin YH, Bailey TW, Li BY, Schild JH & Andresen

MC (2004). J Neurosci 24, 4709–4717.

C© 2007 The Author. Journal compilation C© 2007 The Physiological Society DOI: 10.1113/jphysiol.2007.137091

http://jp.physoc.org

472 J Physiol 582.2 (2007) p 472

PERSPECT IVES

Hypoxia, fetal growth and earlyorigins of disease: the Andeancurse on the Conquistadors

D. A. Giussani

Physiology, Development & Neuroscience,

University of Cambridge, Cambridge

CB2 3EG, UK


The genetic potential for fetal growth

is influenced by the available nutrient

and oxygen supply to the unborn child.

However, the contribution of each of these

components, and in particular of fetal

oxygenation, to the control of prenatal

growth has been difficult to isolate. This

is partly because all conditions associated

with fetal hypoxia, either in health or

disease, are also affected by reductions

in fetal nutrition. For instance, maternal

chronic experimental hypoxia in rats

slows fetal growth but it also reduces

maternal food intake (De Grauw et al.

1986). In humans, birth weight is reduced

with increasing altitude (Moore et al.

2004) and in sea level pregnancies with

placental insufficiency (Barker, 1998).

Because most high altitude populations are

impoverished, and placental insufficiency

decreases nutrient and oxygen transfer to

the baby, the extent to which the reduction

in fetal growth under these conditions

is governed by fetal under-nutrition

or under-oxygenation, again, remains

uncertain. A study by Zamudio et al.

(2007) in this issue of The Journal of

Physiology is an attempt to determine in

humans the relative importance of genetic,

developmental and environmental causes

for high altitude-associated intrauterine

growth restriction in a country uniquely

placed to address the question.

Bolivia lies in the heart of South America

and is conveniently split by the Andean

Cordillera into two halves, residing at

extremely different altitudes. La Paz is

to the West of the country standing at

approximately 4000 m above sea level. In

marked contrast, as the country spans into

the Amazon to the East, there are several

sea level regions including Santa Cruz, the

second largest Bolivian city. Bolivia is also

made up of striking economically divergent

populations with varied ethnicity. A large

component of the inhabitants are of Andean

origin with a high percentage of poverty.

Inhabitants of European ancestry maintain a

standard of living similar to that in affluent

populations in the UK (Mapa de pobreza,

1995).

Previous studies have shown that

pregnancy at high altitude restricts fetal

growth but, rather counter-intuitively,

babies born from low-income families in La

Paz are born heavier than babies born from

high-income highland families (Giussani

et al. 2001). This is because low-income

families in La Paz are mostly Aymara

Indians, and prolonged high altitude

residence ancestry develops a protection

against the effects on fetal growth of prenatal

hypoxia, the longest resident population

experiencing the least decline, while the

shortest historical residents, the greatest

decline (Julian et al. 2007). Importantly,

the level of protection is so strong that it

overpowers the effect of possible under-

nutrition on fetal growth, even in highly

impoverished Andean mothers. Since its

introduction, the physiology underlying

this Andean curse on the Conquistadors has

gathered increasing attention.

The study of Zamudio et al. (2007)

is an exceptional contribution to this

knowledge, as it brings together teams

of scientific and clinical experts to assess

oxygen delivery at extreme altitude in

La Paz and at sea level in Santa Cruz,

in European and Andean pregnancies,

with historical residence quantified using

single nucleotide polymorphisms. The

study tested the hypothesis that greater

maternal uteroplacental oxygen delivery

would explain increased fetal growth at

altitude in Andeans versus Europeans.

They conclude that genetically mediated

96

100

104

108

112

116

Systolic blood

pressure(mmHg)

European Andean

*

Figure 1. Arterial blood pressure in sealevel and high altitude pregnancySystolic blood pressure (mean + S.E.M.) inpregnant women at term of Europeanancestry at sea level (open columns, n = 44)or high altitude (filled columns, n = 37) andin pregnant women at term of Andeanancestry at sea level (n = 46) or high altitude(n = 33). ∗P < 0.05, sea level versus altitude.

differences in maternal oxygen delivery do

not explain the Andean advantage. Rather,

the mechanism underlying this protection is

likely to reside within the feto-placental unit,

perhaps due to differences in fetal substrate

utilization.

Embedded in the wealth of data is another

important point. Since slow fetal growth

is associated with an increased risk of

heart diseases in later life (Barker, 1998),

it is of interest whether the prevalence of

cardiovascular disease is increased at high

altitude relative to sea level, and whether

developmental hypoxia alone, independent

of the maternal nutritional status, can

programme an early origin of disease.

However, the negligible number of studies

in which basal arterial blood pressure

was measured in adult residents rather

than climbers at high altitude report

conflicting data, inconsistencies that may

be related to the altitude residence ancestry

of the individuals under study (Giussani,

2006). Zamudio et al. (2007) show

that systolic blood pressure is markedly

elevated in term pregnant women at high

altitude relative to sea level, however, the

increment is greatly diminished in Andean

women (Fig. 1). These results, albeit in

pregnancy, not only support the concept

that developmental hypoxia may trigger

an increased susceptibility to hypertension

in adulthood, but they highlight that the

Andean curse on the Conquistadors spans

from the physiology of the unborn child to

disease in later life.

References

Barker DJP (1998). Mothers, Babies, and Disease

in Later Life. Churchill Livingstone,

Edinburgh.

De Grauw TJ, Myers R & Scott WJ (1986). Biol

Neonate 49, 85–89.

Giussani DA (2006). In Developmental Origins of

Health and Disease, ed. Gluckman PD &

Hanson MA, pp. 178–190. Cambridge

University Press.

Giussani DA, Phillips PS, Anstee S & Barker DJ

(2001). Ped Res 49, 490–494.

Julian CG, Vargas E, Armaza JF, Wilson MJ,

Niermeyer S & Moore LG (2007). Arch Dis

Child Fetal Neonatal (in press).

Mapa de pobreza (1995). Ministerio de Desarrollo

Humano. Republica de Bolivia, 2nd edn.

Moore LG, Shriver M, Bemis L, Hickler B,

Wilson M, Brutsaert T, Parra E & Vargas E

(2004). Placenta 25 (Suppl. A), S60–S71.

Zamudio S, Postigo L, Illsey NP et al. (2007).

J Physiol 582, 883–895.



J Physiol 582.2 (2007) p 473 473

PERSPECT IVES

Variety is the spice of life:searching for the substrates ofregional myocardial electricalproperties

Stefan Kaab

Department of Medicine I, University

Hospital Grosshadern, Ludwig-Maximilians-

University, Munich, Germany


Variety and diversity as well as regional

specificity of gene expression are the basis

of normal cardiac function. Transcriptional

regulation has been demonstrated to

be an essential compensatory response

to structural and functional changes in

the heart. In order to understand trans-

criptional regulation in defined disease

states it is essential to have a detailed

knowledge of expression patterns in

non-diseased myocardial tissue.

In this issue of The Journal of Physiology

Gaborit et al. (2007) give a detailed analysis

of differentially expressed ion channel genes

controlling regional diversity of myocardial

electrical properties in non-diseased human

hearts. High-throughput real-time RT-PCR

in a hypothesis driven approach allowed

for a comprehensive description of 79

ion channels and related genes in specific

regions of the human heart.

Differential expression of a variety of genes

is giving clues to the molecular substrates

controlling distinct myocardial electrical

properties of specific regions in the heart

and has been demonstrated in anatomically

distinct regions of the heart such as atrial

versus ventricular myocardium (Barth

et al. 2005) or ventricular endocardium

versus ventricular epicardium (Rosati et al.

2001). Genome wide expression studies

in heart failure resulted in a description

of deregulated gene clusters that can

be attributed and sorted by functional

groups such as metabolism, inflammation

and signal transduction. While regional

differences in deregulation of ion channels

in normal and failing heart has been noted

(Ellinghaus et al. 2005; Kaab et al. 2004), this

study, in an unprecedented way, investigates

a comprehensive set of ion channel genes

and related genes in atria, ventricular

epicardium and endocardium as well as in

Purkinje fibres. The authors present and

discuss their data in the context of previous

findings and highlight novel findings of

potential functional significance. Among

the wealth of data, the detailed analysis of

the Purkinje fibre system deserves special

attention. The participation of Nav1.7 in

TTX-sensitive Purkinje fibre INa, and the

importance of low-level Na+,K+-ATPase

and high-level inositol-trisphosphate

receptor expression in the Purkinje system

warrant further functional studies with

respect to arrhythmogenesis. These data

add to the discussion on Ca2+ signalling in

the Purkinje system that was substantiated

by the recent discovery of the molecular

substrates of T- and L-type Ca2+ currents in

Purkinje fibres in dog (Rosati et al. 2007).

In comparison to previous pan-genomic

approaches on regional specific gene

expression in the heart this quantitative

RT-PCR based study allowed for the

detection of low-abundance mRNAs and

unmasked differences in expression levels

to a greater extent than pan-genomic arrays

(Kaab et al. 2004; Ellinghaus et al. 2005).

These data provide an important reference

with a focus on low abundance ion channel

genes for other studies aiming at:

genes exclusively expressed in atrial or

ventricular tissue;

genes controlling electrical properties that

may play an important functional role in

arrhythmogenesis;

transcriptional deregulation contributing to

a remodelling process in defined disease

states in the human heart (e.g. hypertrophy,

heart failure).

To link deregulation of ion channel genes

to other potentially more global gene

deregulation processes in the course of

specific disease states more information

especially on transcription factors involved

will be needed.

More sophisticated and meticulous

dissecting methods as well as a combination

with functional studies have recently led to

the detailed characterization of ion channel

expression in rabbit atria and sinus node,

linking the variety of gene expression on

a microscopic level to specific electrical

properties that guarantee physiological

excitation and conduction (Tellez et al.

2006).

This holds out the promise that we will

see a detailed analysis and understanding of

the transcriptional control of key anatomical

substrates of excitation and conduction such

as sinus node, atrio-ventricular node, and

crista terminalis in the human heart in the

future.

References

Barth AS, Merk S, Arnoldi E, Zwermann L,

Kloos P, Gebauer M, Steinmeyer K, Bleich M,

Kaab S, Pfeufer A, Uberfuhr P, Dugas M,

Steinbeck G & Nabauer M (2005). Pflugers

Arch 450, 201–208.

Ellinghaus P, Scheubel RJ, Dobrev D, Ravens U,

Holtz J, Huetter J, Nielsch U & Morawietz H

(2005). J Thorac Cardiovasc Surg 129,

1383–1390.

Gaborit N, LeBouter S, Szuts V, Varro A,

Escande D, Nattel S & Demolombe S (2007).

J Physiol 582, 675–693.

Kaab S, Barth AS, Margerie D, Dugas M,

Gebauer M, Zwermann L, Merk S, Pfeufer A,

Steinmeyer K, Bleich M, Kreuzer E, Steinbeck

G & Nabauer M (2004). J Mol Med 82,

308–316.

Rosati B, Dun W, Hirose M, Boyden PA &

McKinnon D (2007). J Physiol 579, 465–471.

Rosati B, Pan Z, Lypen S, Wang HS, Cohen I,

Dixon JE & McKinnon D (2001). J Physiol

533, 119–125.

Tellez JO, Dobrzynski H, Greener ID, Graham

GM, Laing E, Honjo H, Hubbard SJ, Boyett

MR & Billeter R (2006). Circ Res 99,

1384–1393.

C© 2007 The Authors. Journal compilation C© 2007 The Physiological Society DOI: 10.1113/jphysiol.2007.136929


474 J Physiol 582.2 (2007) p 474

PERSPECT IVES

Flat and bouncy walking

R. McNeill Alexander

Institute of Integrative and Comparative

Biology, University of Leeds, Leeds LS2 9JT,

UK


Walking is one of the major items in human

energy budgets. Researchers in the 1950s

found that it accounted for 20% of a typical

clerk’s weekly energy consumption, and

27% of a miner’s. Fifty years later, many of us

walk less, but walking remains a substantial

energy cost. It seems reasonable to expect

people to adjust their gaits to keep the cost

of moving around as low as possible (see for

example Alexander, 2002).

Measurements of oxygen consumption

have shown that we do indeed adjust many

features of gait for economy of energy.

We break into a run at the speed (about

2 m s−1) at which running becomes more

economical than walking. At any particular

walking speed, we spontaneously adjust our

strides to the length that minimizes energy

consumption at that speed. We do not set

our left and right feet down along the same

line as if on a tightrope, nor well out to either

side of the body in the manner of toddlers,

drunkards and sailors on a rolling ship;

the intermediate strategy that we actually

adopt minimizes energy costs (Donelan et al.

2001).


Massaad et al. (2007) report a study of

unusually flat or bouncy human gaits. We

bob up and down in normal walking because

we keep each leg almost straight while its

foot is on the ground. Accordingly, we rise

as the supporting leg becomes vertical and

descend again towards the end of the step,

so as to move forward along a series of

arcs of circles. The force the foot exerts

on the ground stays in line with the leg,

so we slow down as we rise and speed

up as we fall. Kinetic energy is converted

to gravitational potential energy and back

again in the manner of a pendulum. In

principle, no work is required until the other

foot lands on the ground and we set out on a

new arc. At that stage, the downward motion

of the body has to be halted, and muscular

work is needed to start it moving upwards on

the new arc. This work would be avoided if

we did not bob up and down, but each knee

would have to bend and re-extend while its

foot was on the ground, and work would

be needed for that. Analysis of a simple

theoretical model (Srinivasan & Ruina,

2006) concluded that work requirements

would be minimized by keeping leg length

constant, and so bobbing up and down. A

more elaborate musculoskeletal model that

predicted metabolic energy costs similarly

showed a normal bobbing gait as optimal

(Sellers et al. 2003).

Massaad et al. (2007) measured the

mechanical work and oxygen used in

flat, normal and bouncy walking. Despite

the indication of Srinivasan & Ruina’s

model, that normal walking would

minimize work, Massaad et al. found no

significant difference in work requirements

between flat and normal walking. Oxygen

consumption in flat walking, however,

was up to double the rate for normal

walking at the same speed, as also found

by Ortega & Farley (2005). Thus, muscle

efficiency was remarkably low, possibly

due to the need for high forces in the

quadriceps muscles when the knee was

bent at mid-stance (the metabolic rate of

an active muscle depends on the force it is

exerting as well as on the rate at which it

is doing work). This explanation has been

offered for the high metabolic energy cost

of the bent-legged bipedal walking style of

chimpanzees (Carey & Crompton, 2005).

In bouncy walking, both work and oxygen

consumption were higher than for normal

walking, and efficiency was close to normal.

There are problems in research that

depends on comparisons of metabolic

and mechanical energy costs, between

normal and unusual patterns of movement.

Estimates of total muscular work are subject

to uncertainties regarding co-contraction

of antagonistic muscles, and regarding

energy transfer between body segments.

Unaccustomed gaits may be performed

less skilfully than normal ones, increasing

oxygen consumption, even when (as in this

case) time has been allowed for practice. The

conclusion of this study, nevertheless, seems

clear; the bounciness of our walk may look

like wasteful exuberance, but it does save

energy.

References

Alexander RMcN (2002). Am J Human Biol 14,

641–648.

Carey TS & Crompton RH (2005). J Human Evol

48, 25–44.

Donelan JM, Kram R & Kuo AD (2001). Proc

Biol Sci 268, 1985–1992.

Massaad F, Lejeune TM & Detrembleur C

(2007). J Physiol 582, 789–799.

Ortega JD & Farley CT (2005). J Appl Physiol 99,

2099–2107.

Sellers WL, Dennis LA & Crompton RH (2003).

J Exp Biol 206, 1127–1136.

Srinivasan M & Ruina A (2006). Nature 439,

72–75.



J Physiol 582.2 (2007) pp 475–476 475

PERSPECT IVES

Physiological responses tohypoglycaemia – not all ‘just inthe head’

Mark W. J. Strachan

Metabolic Unit, Western General Hospital,

Crewe Road, Edinburgh EH4 2XU, Scotland,

UK


The human brain has enormous metabolic

requirements that are almost completely

provided by the oxidation of glucose.

However, because it is unable to synthesize

or store glucose, the brain is reliant on the

cerebral circulation to provide a constant

supply of its primary source of energy. Acute

hypoglycaemia quickly causes energy failure

in cerebral neurons and this is manifest by

the onset of neuroglycopaenic symptoms,

such as poor concentration, drowsiness

and reduced co-ordination. As blood

glucose concentrations fall further, cognitive

impairment and confusion develop and

Figure 1. Major components of the physiological responses to acute hypoglycaemia in humansReproduced from Heller (2003).

ultimately seizures, coma and permanent

neurological deficit occur.

Because of the dependence of the central

nervous system on glucose, multiple

mechanisms have evolved to maintain

glucose homeostasis (Fig. 1). A fall in blood

glucose is detected by glucose sensors within

the brain, located mainly in the ventro-

medial nuclei of the hypothalamus, and

the hepatic portal system. Activation of

these glucose sensors instigates a cascade of

responses to raise blood glucose (King &

Macdonald, 1999). These responses include

the following.

1 Release of counterregulatory hormones

that antagonize insulin action and suppress

its endogenous secretion.

2 Stimulation of the autonomic nervous

system (principally the sympathetic

division), which not only promotes

counterregulation, but also induces

important haemodynamic and other

end-organ effects.

3 Generation of warning symptoms that

alert the individual to the development

of hypoglycaemia and the need to take

corrective action.

The haemodynamic changes include

increased heart rate, cardiac output and

systolic blood pressure. There are also

significant increases in regional blood

flow, not only to the brain, but also to

other organs (notably the liver and muscle)

that can increase substrate delivery to the

brain.

It is well established that the carotid bodies

activate the autonomic nervous system and

increase ventilation in response to falling

oxygen or rising carbon dioxide tensions.

There are also in vivo data from dogs that

suggest that the carotid bodies play a role

in the sensing of hypoglycaemia and in

the initiation of the subsequent counter-

regulatory response (Koyama et al. 2000).

In vitro, several studies have demonstrated

that carotid bodies are responsive to low

glucose concentrations and that ventilatory

responses are additive to those obtained

during hypoxia or hypercapnia (Zhang et al.

2007). In one study in rats, hypoglycaemia



476 Perspectives J Physiol 582.2

increased spontaneous ventilation and this

effect was abolished if the carotid sinus

nerves were sectioned; however, in vitro

chemoreceptor discharge frequency was

not altered by low glucose concentrations

leading the authors to conclude that glucose

was being sensed indirectly by the carotid

bodies (Bin-Jaliah et al. 2004).

The study by Ward et al. (2007) reported

in this issue of the The Journal of

Physiology examined the effects of acute

hypo- and hyperglycaemia on the hypoxic

ventilatory response in humans. When

blood glucose concentration was dropped

to 2.8 mmol l−1, using a hyperinsulinaemic

glucose clamp technique, in 11 healthy

volunteers, there was a 54% increase in

isocapnic ventilation and a 108% increase in

the hypoxic ventilatory response. There was

a predictable elevation in counterregulatory

hormones. Clearly there are several plausible

mechanisms to explain this response. In

keeping with the animal and in vitro studies

described above, there could be direct

(or indirect) sensing of hypoglycaemia

by the carotid bodies. Alternatively, there

could be direct stimulation of the central

nervous system respiratory centre by

counterregulatory hormones or via direct

projections from glucose-sensing neurons

in the hypothalamus or even in the tractus

solitarius itself. Whatever the underlying

mechanism, the physiological response

makes biological ‘sense’ in that there is an

attempt by the body to ensure that oxygen

tensions in tissues are optimized at a time

when the concentrations of glucose are

suboptimal. This ensures that whatever

glucose is available for metabolism is

oxidized, to release its maximum potential

energy yield, rather than undergoing

anaerobic metabolism.

Intriguingly, Ward et al. (2007) observed

that acute hyperglycaemia (blood glucose

concentration of 11.2 mmol l−1) was also

associated with a mild increase in the

hypoxic ventilatory response. This was

not in keeping with in vitro data.

However, it must be remembered that

hypoglycaemia and hyperglycaemia are not

simple opposites. Hypoglycaemia poses a

direct threat to the viability of physiological

systems, whereas transient hyperglycaemia

does not. Therefore, we should not always

expect to see mirror image physiological

responses to the two conditions. The

physiological mechanism and value of

the ventilatory response to hyperglycaemia

await elucidation.

Acute hypoglycaemia is rare in healthy

humans, but continues to blight the lives of

people with insulin-treated diabetes more

than 80 years since the discovery of insulin.

Studies like those of Ward et al. (2007)

will not suddenly transform the lives of

people with diabetes, but they add to our

incremental knowledge of the physiology

of hypoglycaemia. Studies of hypoglycaemia

are also of relevance in understanding

the physiology of exercise and so further

investigations in this area are certainly

warranted.

References

Bin-Jaliah I, Maskell PD & Kumar P (2004).

Indirect sensing of insulin-induced

hypoglycaemia by the carotid body in the rat.

J Physiol 556, 225–266.

Heller SR (2003). In Textbook of Diabetes, 3rd

edn, ed. Pickup JC & Williams G. Blackwell

Science, Oxford. p. 33.1–33.19.

King P & Macdonald IA (1999). In Frier BM &

Fisher BM, ed. Hypoglycaemia in Clinical

Diabetes. John Wiley & Sons,

Chichester. p. 1–27.

Koyama Y, Coker RH, Stone EE, Lacy DB,

Jabbour K, Williams PE & Wasserman DH

(2000). Diabetes 49, 1342–1442.

Ward DS, Voter WA & Karan S (2007). J Physiol

582, 859–869.

Zhang M, Buttigieg J & Nurse CA (2007).

J Physiol 578, 735–750.

C© 2007 The Author. Journal compilation C© 2007 The Physiological Society


J Physiol 582.2 (2007) p 477 477

PERSPECT IVES

NKCC1: tales from the dark sideof the crypt

Jonathan D. Kaunitz1,2,3,4

1Greater Los Angeles Veteran Affairs

Healthcare System, WLAVA Medical

Center, Los Angeles, CA 90073, USA and2Division of Digestive Diseases, 3CURE:

Digestive Diseases Research Center and4Department of Medicine, UCLA School of

Medicine, Los Angeles, CA 90024, USA


Intestinal electrolyte secretion consist

of several fundamental components: a

basolateral uptake pathway, identified as

the Na+–K+–Cl− cotransporter (NKCC1)

(Payne et al. 1995) and an apical secretory

pathway. NKCCs are highly conserved

across organ and species, serving essential

ion secretory and absorptive functions.

Distinct regulatory mechanisms are

present for apical, and for basolateral

(‘dark side’) transporters. For example,

in the colonic crypt-derived cell line

HT-29, Slotki et al. (1993) found that

prolonged incubation with the cAMP

activator forskolin increased K+ uptake

attributable to cotransporter activity

independently of new protein synthesis,

leading them to speculate that NKCC

is regulated by membrane insertion, in

contrast to the new channel synthesis

observed for CFTR. In HT-29 cells in

response to phorbol esters, another group

demonstrated a rapid increase followed by

a slow decline in cotransporter-associated

K+ uptake, mirrored by a decline in

[3H]bumetanide binding as a surrogate

for plasma membrane NKCC expression.

The change in Bmax with unchanged

affinity suggested that regulation was via

cotransporter internalization (Franklin

et al. 1989). Indeed, control of insertion of

the related renal NKCC paralogue NKCC2 is

currently accepted as the primary means of

regulating cotransporter function (Mount,

2006).

Although intestinal Cl− secretory function

is traditionally ascribed to the crypts,

cotransporter function, measured by cell

volume measurements in response to

shrinkage or secretagogues in intact crypts

and in vesicles prepared from the baso-

lateral membrane of surface and crypt

cells, was not only intact in the upper

villous and surface, but appeared to

be of increased activity (Diener, 1994;

Wiener & van Os, 1989). In contrast,

Flemmer et al. (2002), using an antibody

recognizing phosphorylated NKCC, noted

NKCC activation by adrenergic agonists

only in the colonic crypt.


Reynolds et al. (2007) have used crypts

prepared from human rectal biopsies to

study regulation of NKCC1 function.

Using a combination of fluorescently

labelled NKCC1, Na+,K+-ATPase, and

acetylcholine receptor antibodies and

dynamic digital microimaging, the authors

monitored NKCC1 localization, and

measured cellular Ca2+ with fura-2 and

pHi in response to NH4+ with BCECF. In

a series of visually stunning images, the

authors documented patterns of NKCC1

distribution in human rectal biopsies,

demonstrating the propagation of a Ca2+

wave in response to acetylcholine. Secretion

was also inferred from measurements

of crypt width and the migration of

calcein-labelled cells. In response to acetyl-

choline, crypt widening and acceleration of

the plateau-phase pHi decrease indicated

NKCC1 activation accompanied by fluid

secretion. NKCC1 activity decreased after

30 min of ACh exposure, accompanied

by a striking redistribution of NKCC1

immunofluorescence to the apical pole.

With 4 h of ACh stimulation, NKCC1

re-inserted into the basolateral membrane.

In contrast, prolonged forskolin exposure

gradually increased NKCC1 activity while

maintaining its presence in the basolateral

membrane.

This paper is remarkable in several

respects: it reports an adaptation of the

isolated colonic crypt preparation, first

described 20 years ago for fluorescence

measurements (Kaunitz, 1988), as a

useful means of studying clinical tissue.

Rather than use surrogate systems, the

investigators have shown that intact organs

obtained clinically can yield remarkable

results. Furthermore, although the authors

did not devise any single technique,

the combination of measurements of

crypt width, pHi in response to NH4+,

cellular Ca2+, immunolocalization, and

measurements of calcein-labelled cell

movement represents how relatively

straightforward methods, when applied

judiciously, carefully and thoughtfully,

can amplify the overall power of their

observations.

The authors indeed confirmed that

as in the nephron, colonic NKCC1

activity in response to Ca2+ signalling is

regulated via plasma membrane insertion

and re-expression, and also provided a

direct and plausible mechanism for the

secretory synergy observed with combined

cAMP and Ca2+ signals, and also the

inhibitory effect of epidermal growth factor

receptor/MAP kinase transactivation by

elevation of cellular Ca2+. The results

build on the observations summarized

above, yielding fresh insights into intestinal

NKCC1 regulation. I hope that other

investigators follow the lead of this group in

the study of intestinal transport regulation.

References

Diener M (1994). Pflugers Arch 426, 462–464.

Flemmer AW, Gimenez I, Dowd BF, Darman RB

& Forbush B (2002). J Biol Chem 277,

37551–37558.

Franklin CC, Turner JT & Kim HD (1989). J Biol

Chem 264, 6667–6673.

Kaunitz JD (1988). Am J Physiol Gastrointest

Liver Physiol 254, G502–G512.

Mount DB (2006). Am J Physiol Renal Physiol

290, F606–F607.

Payne JA, Xu JC, Haas M, Lytle CY, Ward D &

Forbush B III (1995). J Biol Chem 270,

17977–17985.

Reynolds A, Parris A, Evans LA, Lindqvist S,

Sharp P, Lewis M, Tighe R & Williams MR

(2007). J Physiol 582, 507–524.

Slotki IN, Breuer WV, Greger R & Cabantchik ZI

(1993). Am J Physiol Cell Physiol 264,

C857–C865.

Wiener H & van Os CH (1989). J Membr Biol

110, 163–174.



J Physiol 582.2 (2007) pp 479–480 479

JOURNAL CLUB

Out of the cleft: the source andtarget of extra-synapticglutamate in the CA1 region ofthe hippocampus

Sachin Makani and Edward Zagha

New York University School of Medicine,

Department of Neuroscience and

Physiology, 550 First Avenue, New York,

NY 10016, USA


Glutamate is the central nervous system’s

principal excitatory transmitter, and its

levels in the brain’s extracellular space are

under tight spatial and temporal regulation.

During evoked vesicular release, glutamate

levels can peak in the millimolar range

within the synaptic cleft. Active transport,

enzymatic degradation and diffusion all

work to reduce this level so as to avert the

continual activation or desensitization of

receptors, and give postsynaptic neurons a

high signal-to-noise input. However, it has

been shown that a much lower, nanomolar

to micromolar ambient concentration of

glutamate remains within the interstitium.

It is therefore possible that such a tonic level

may be affecting cellular excitation. Indeed,

even such low quantities of the amino acid

are sufficient to activate NMDA receptors

(NMDARs). The potential tonic activation

of NMDARs is of particular interest as these

channels have unique electrical properties

and are involved in intracellular cascades

implicated in a wide range of cellular

processes, from learning and memory to

excitotoxicity.

Sah et al. (1989) were the first to

show that NMDARs could be active

without synaptic stimulation. Studying rat

hippocampal slices, they recorded from

CA1 pyramidal neurons in whole-cell

configuration under voltage clamp at

−35 mV. After the addition of APV

(d-AP5), a competitive NMDAR antagonist,

the group noted a substantial (hundreds

of picoamps) net outward shift in the

holding current. This effect was magnesium

and voltage dependent, hallmarks of an

NMDAR-mediated conductance. Notably,

they also observed that APV reversibly

reduced action potential output in response

to current injection, suggesting a role for the

tonic conductance in enhancing neuronal

excitability.

This Journal Club article focuses on The

Journal of Physiology publication by Le

Meur et al. (2007), who recently added

to this area of interest. Voltage clamping

CA1 pyramidal neurons at +40 mV, they

observed a net inward shift in holding

current after the addition of APV, and an

I–V curve of the APV-sensitive component

typical of NMDARs. Importantly, the

change in current after APV was only

tens of picoamps, an order of magnitude

less than that seen by Sah et al. (1989).

In different cells, they superfused NBQX

and LY341495 and saw no change in the

tonic current, suggesting that the observed

change is dependent on NMDARs and not

other ionotropic or metabotropic glutamate

receptors. If ambient glutamate was indeed

activating this NMDAR-dependent current,

then increasing its concentration should

have the opposite effect of adding APV.

Indeed, after blocking glutamate uptake

with TBOA, there was an immediate net

outward shift in the holding current. Inter-

estingly, in experiments performed in the

presence of TBOA, blockers of AMPARs did

cause a small effect, and this was amplified by

cyclothiazide, a drug that inhibits AMPAR

desensitization. This suggested that in their

slice recording conditions, the regulation

of ambient glutamate levels and receptor

desensitization both act to prevent the

tonic activation of non-NMDA glutamate

receptors.

A tonic current mediated by glutamate

receptors contrasts sharply with the

standard dogma of glutamate signalling,

which involves evoked, vesicular release

from one neuron to another within a highly

structured synapse. While this ambient

glutamate could theoretically originate

via spillover from high-frequency synaptic

activity, work subsequent to Sah et al.

(1989) has shown that it is instead mediated

by neighbouring glia. Le Meur et al. (2007)

therefore focused on confirming this glial

origin in their experimental preparation.

They first treated slices with bafilomycin A1

to prevent vesicular release. In these slices,

there was no difference from control slices

in the change in holding current induced

by TBOA or MK-801, a non-competitive

NMDAR antagonist. This confirmed that

the glutamate source was indeed not

synaptic. In the next experiment, they

used L-methionine sulfoximine (MSO) to

inhibit glutamine synthetase, a glial enzyme

responsible for converting glutamate

to glutamine. This treatment has been

shown to drastically increase glutamate

concentrations in glial cells. While the

amplitude of the basal tonic current was

unchanged, currents induced by TBOA

were doubled in slices first treated with

MSO. These data support the notion that

the ambient glutamate is of glial origin,

and underscore the idea that glia play a far

greater role than a simple ‘support network’

for neurons. A number of the above

experiments have been done previously and

similar results were obtained (Angulo et al.

2004).

If the tonic conductance is not activated

by vesicular release, one could hypothesize

that the ambient glutamate was activating

extrasynaptic NMDARs, specifically. Le

Meur et al. (2007) first tested this idea with

pharmacological tools to block particular

NMDAR subunits, as such an approach

has been used previously to discriminate

synaptic from extrasynaptic NMDA (Fellin

et al. 2004) as well as GABAA receptors.

EPSC amplitude and holding current

were compared in the presence of these

blockers. However, these studies did not

clearly differentiate between the NMDAR

subunits mediating synaptic versus tonic

currents. PPDA, an NMDAR2C- and

D-preferring blocker, inhibited the tonic

slightly more than the synaptic current,

but infenprodil (an NMDAR2B antagonist)

inhibited neither, and NVP-AAM077

(an NMDAR2A-preferring antagonist)

inhibited them equally.

In a key experiment, they then turned to the

use-dependent general NMDAR antagonist,

MK-801. After recording control EPSCs at

+40 mV in the presence of NBQX, they

stopped synaptic stimulation and washed in

MK-801. Stimulation was resumed, and the

first EPSC after MK-801 superfusion was

compared with the last one before it. The

holding current was monitored during the

same time frame. If the NMDARs mediating

the tonic current were extrasynaptic, one

would expect to see a change in the holding

current during MK-801 wash-in (as these

receptors would be open), and less of

a change in the synaptic current (these

receptors would have been closed, owing

to a lack of glutamate in the cleft). This is

indeed what Le Meur et al. (2007) observed.



480 Journal Club J Physiol 582.2

While the tonic current was approximately

halved, the peak amplitudes of the EPSCs

before and after wash-in did not differ.

There was, however, a decrease in the

decay time constant of the EPSC after

MK-801 addition, which is consistent with

the use-dependent blocker shortening the

mean channel open-time, demonstrating

the efficacy of MK-801 at these synapses.

While the idea that extrasynaptic NMDARs

are mediating the tonic current has been

suggested previously (Fellin et al. 2004), this

open channel block experiment elegantly

revealed the existence of separate tonic and

phasic populations of glutamate receptors.

With mounting evidence for ambient

glutamate present both in vitro and in

vivo, many questions remain regarding the

signalling roles of extrasynaptic receptor

activation. This publication focused on

a tonic glutamate-mediated current.

However, Angulo et al. (2004) and others

have described transient currents of

the order of hundreds of milliseconds

which were also suggested to be due to

glial glutamate release. This issue of the

dynamics of extra-synaptic glutamate is

critical in determining its potential roles

in neuronal signalling. Such observations

argue for the importance of understanding

glial physiology with respect to the spatial

and temporal variations in extracellular

glutamate concentration, including

mechanisms and modulation of release and

reuptake.

Equally intriguing questions remain

regarding the postsynaptic effects of

ambient glutamate. NMDAR activation may

affect neuronal activity on multiple time

scales. The small tonic NMDAR component

did not affect the overall input–output

relationship of CA1 neurons in this

study, which contrasts with measurements

of Sah et al. (1989). However, tonic

NMDAR activation may be particularly

important in shaping synaptic inputs.

As NMDARs are expressed in dendritic

membranes, they are poised to interact

with local synaptic inputs prior to somatic

integration. Moreover, the magnesium

block of NMDARs results in negative slope

conductance profiles at voltages near and

below action potential generation, which

means that depolarization from a nearby

EPSP will boost the amount of inward

currents flowing through open NMDARs.

The effects of such boosting may be EPSP

amplification or possibly regenerative

dendritic spiking. Furthermore, there have

been reports that glial-dependent NMDAR

activation may be highly temporally

correlated in nearby neurons, suggesting a

possible role of synchronous modulation

of excitability across multiple neuronal

elements (Fellin et al. 2004; Angulo et al.

2004).

At longer time scales, NMDAR activation

by ambient glutamate may result in

important intracellular signalling due to

the influx of Ca2+. NMDAR-mediated Ca2+

influx may act locally to modulate ion

channels or biochemical processes, as well

as distally by affecting signalling cascades

and gene transcription. Hardingham et al.

(2002) further suggest that activation of

extrasynaptic NMDARs, unlike synaptic

receptors, cause cAMP-response element

binding protein (CREB) dephosphorylation

and mitochondrial dysfunction and cell

death. These dramatic findings highlight the

importance of current studies like the one

discussed in this Journal Club article, as

glutamate activity outside the synapse may

be a critical parameter in the function of

neuronal circuits.

References

Angulo MC, Kozlov AS, Charpak S & Audinat E

(2004). Glutamate released from glial cells

synchronizes neuronal activity in the

hippocampus. J Neuroscience 24, 6920–6927.

Fellin T, Pascual O, Gobbo S, Pozzan T, Haydon

PG & Carmignoto G (2004). Neuronal

synchrony mediated by astrocytic glutamate

through activation of extrasynaptic NMDA

receptors. Neuron 43, 729–743.

Hardingham GE, Fukunaga Y & Bading H

(2002). Extrasynaptic NMDARs oppose

synaptic NMDARs by triggering CREB

shut-off and cell death pathways. Nat Neurosci

5, 405–414.

Le Meur K, Galante M, Angulo MC & Audinat E

(2007). Tonic activation by ambient glutamate

of non-synaptic origin in the rat

hippocampus. J Physiol 580, 373–383.

Sah P, Hestrin S & Nicoll RA (1989). Tonic

activation of NMDA receptors by ambient

glutamate enhances excitability of neurons.

Science 246, 815–818.

Acknowledgements

We would like to thank Dr Mitchell Chesler for

helpful comments.

C© 2007 The Authors. Journal compilation C© 2007 The Physiological Society


J Physiol 582.2 (2007) pp 481–488 481

TOPICAL REVIEW

Genetically manipulated mice: a powerful toolwith unsuspected caveats

Klaus I. Matthaei

Gene Targeting Laboratory, The John Curtin School of Medical Research, Building 131, Garran Road, The Australian National University,

Canberra, ACT 0200, Australia

Although genetic manipulations in mice have provided a powerful tool for investigating gene

function in vivo, recent studies have uncovered a number of developmental phenomena that

complicate the attribution of phenotype to the specific genetic change. A more realistic approach

has been to modulate gene expression and function in a temporal and tissue-specific manner. The

most common of these methods, the CreLoxP and tetracycline response systems, are surveyed

here and their recently identified shortcomings discussed, along with a less well known system

based on the E. coli lac operon and modified for use in mammals. The potential for further

complications in interpretation due to hitherto unexpected epigenetic effects involving transfer

of RNA or protein in oocytes or sperm is also explored. Given these problems we reiterate the

necessity for the use of completely reversible methods that will allow each experimental group of

animals to act as their own control. Using these methods with a number of specific modifications

to eliminate non-specific effects from random insertion sites and inducer molecules, the full

potential of genetic manipulation studies should be realized.

(Received 18 April 2007; accepted after revision 10 May 2007; first published online 10 May 2007)

Corresponding author K. I. Matthaei: Gene Targeting Laboratory, The John Curtin School of Medical Research,

GPO Box 334, Canberra City, ACT 0200, Australia. Email: [email protected]

Genetic manipulations in mice have provided apowerful tool for investigating gene function in vivo.However, complications in the interpretation of past andpresent studies have arisen following the observation thatresults can often vary substantially depending on thestrain of mice in which the manipulation is made (seeMatthaei, 2004 for a more extensive discussion). Otherunexpected results of genetic deletion studies, such asembryonic lethality (see for example Faast et al. 2001for H2AZ) and coordinate or compensatory regulationof other related genes (for example Lim et al. 2004;Blackburn et al. 2006 for GST zeta), provide furthercomplications to delineating the specific role of a genein a particular tissue type. More recent methods, whichenable both tissue-specific and temporal regulation ofgene expression, now provide a more realistic, if morecomplicated, approach to understanding gene function.However, unanticipated complications of these methodsare now being unearthed. This brief review will criticallysurvey the pitfalls of the currently available methods andpoint the way forward in this now seemingly unfathomablearea.

Problems with transgenic mice due to randomintegration

The generation of the first transgenic mouse by injection ofa DNA construct directly into the pronucleus of a fertilizedsingle cell mouse embryo (Palmiter et al. 1982) has resultedin the generation of thousands of such strains. Importantly,the site at which the DNA is integrated is random, asare the number of copies of the transgene. Although theexpression of the construct is faithful for the promoter, onmany occasions it may also be significantly influenced bythe local environment at the integration site (the ‘position’effect). This can lead to the promiscuous expression ofthe transgene (often referred to as ‘leakiness’), due tomodification of the specificity of the promoter, or at timesto a more severe phenotype, due to disruption of anunknown gene by insertion of the transgene (insertionalmutagenesis). A number of different ‘founder’ animalswith different copy numbers and different integrationsites must therefore be assessed in order to determinethe correct/faithful expression of each transgene.Surprisingly, in one example, 24 different foundersresulted in 24 different expression patterns (Feng et al.



482 K. I. Matthaei J Physiol 582.2

2000) making it impossible to determine which pattern wascorrect.

One way to overcome this problem is to use a ‘knock-in’procedure where the transgene is introduced into a specificlocus using homologous recombination in embryonicstem cells. In this situation a modified gene can beexpressed under its own promoter (see for example Yanget al. 2006), or a transgene can be expressed from aubiquitous promoter, such as Rosa26 (Soriano, 1999). Analternative approach to improve promoter specificity is touse larger DNA fragments, which can be manipulated inbacterial (BAC) or yeast (YAC) artificial chromosomes, todrive gene expression. These larger fragments can avoidthe position effect since they contain more of the localenvironment of the promoter thereby ensuring its abilityto function normally. These fragments can vary in lengthup to 250 kb as opposed to the 1–2 kb commonly used fortransgenic constructs and they result in faithful controlof reporter gene expression (see for example Moreiraet al. 2006). Indeed a major initiative using this approachestimates that 85% of BAC transgenic constructs expressreproducibly in multiple transgenic lines (Gong et al.2003), thereby reducing the need to generate and assessmore than a few lines to identify faithful expression.However, it is not clear whether the presence of othergenes, transcription factors, microRNAs or control regionsin these fragments may compromise results. In the BACstudies mentioned above, the readout was expression ofa reporter gene (EGFP) and not the expression of afunctional protein such as a particular cytokine that canproduce specific effects. It would be interesting to comparethe phenotype of the interleukin(IL)-5 transgenic mousethat uses a minimal promoter (Dent et al. 1990) witha mouse expressing IL-5 from a BAC which would alsocontain a number of other cytokines due to their very closeproximity, including IL-3, IL-4, IL-13 and granulocytemacrophage colony stimulating factor (Lee & Young, 1989;McKenzie et al. 1999; Avery et al. 2004). This approach,without careful analysis, may therefore not be as widely

pCAGG B-galactosidase.STOPLoxP LoxP No EGFP

X

pCAGG LoxP EGFP

FloxStop

delLoxP. EGFP

Cre Recombinase

Figure 1. Deletion of intra-LoxP DNA by Cre recombinaseMice transgenic for the ‘β-galactosidase Floxed stop’ EGFP construct (called FloxStop) express β-galactosidase andnot EGFP. When crossed with a Cre mouse the doubly transgenic mice no longer express β-galactosidase due toits deletion (called delLoxP) and the mice are green due to the removal of the stop and de novo EGFP expression.

applicable as first thought and the ‘knock-in’ procedure,although more complex, may still be the method of choice.Again, caution must be taken since the ‘knock-in’ willproduce an insertional deletion of the gene controlled bythe promoter to be used (see for example Gazit et al. 2006).Moreover, even in the heterozygous state the mice canhave a phenotype if the gene exhibits haploinsufficiency(deletion of one functional allele resulting in modifiedgene expression). However, the insertional deletion can beavoided if the construct uses an internal ribosome entrysite (IRES) followed by the transgene and these are placedbetween the stop codon and the polyadenylation signalof the target gene. In this case a bicistronic transcriptis made containing both genes that are then translatedinto two separate proteins (see for example Rees et al.1996), resulting in the expression of the transgene withoutdisrupting the endogenous gene.

However, the potential ‘leakiness’ of conventionaltransgenic mice complicates most current studiesinvolving genetic manipulations, as we will see in thefollowing sections.

Tissue and temporal specific control of gene function

CreLoxP. The deletion of a gene results in embryoniclethality if there is an absolute requirement for thatgene during development. This lethality can be avoidedif the gene is deleted later in life or if it is deletedin a tissue-specific manner particularly if the gene hasmultiple functions. The most common method to obtaintissue-specific control of gene deletion is the CreLoxPsystem modified from bacteriophage. LoxP sites are shortDNA sequences that are recognized by a specific DNArecombinase enzyme called Cre (causes recombination)that deletes any DNA between the two sequences (seeFig. 1). An exon in the gene of interest is flanked by theseLoxP sites and the modified gene is introduced into itscorrect location in embryonic stem cells by homologousrecombination. The stem cells are injected into mouse



J Physiol 582.2 Caveats of genetically manipulated mice 483

blastocysts and mice generated (for a fuller descriptionof chimæra generation see Matthaei, 2004) in which thegene is ‘floxed’ but fully functional until it is inactivated bythe Cre recombinase due to the removal of the exon. Thisis achieved by breeding the ‘floxed’ mice with a mousetransgenic for the Cre recombinase usually under thecontrol of a tissue-specific promoter (first described in Guet al. 1994, but see also Rajewsky et al. 1996; Kwan, 2002).The result is deletion of the gene in a specific tissue.

A major problem with this method is the ‘position’effect as described above, whereby Cre is expressednon-specifically in other tissues, or the promoter drivingCre is active early during development so that the geneis ablated at the wrong time or in the wrong tissue.An improvement is to provide temporal control of Creexpression through the use of an inducible promoter thathas been modified to be ‘off’ until the mice are treated withthe inducer. This has been achieved using the syntheticsteroid RU486 (first described in Kellendonk et al. 1996)in low doses to prevent anti-progesterone effects in mice,such as abortion, although side-effects in other tissues,such as in neurons (Hendry et al. 1987) may continue tobe a problem. In some systems these side-effects may beavoided by the topical application rather than ingestionof RU486 (see, for example, the activation of Keratin.Cre,where the effect is limited mostly to the skin (Zhou et al.2002)). Other Cre inducible systems using Tamoxifen havealso been established (Zheng et al. 2002); however, againthere are potential side-effects due to the use of this drug.A further difficulty, that has never been addressed, is thatthe mammalian genome contains many ‘pseudo’ LoxPsites (Thyagarajan et al. 2000) that potentially could berecombined by Cre recombinase with totally unknowneffects. Such a consideration requires the inclusion of aCre-transgenic mouse amongst the control wild-type mice(see also the use of littermates as control below).

Notwithstanding the various drawbacks inherent inthe use of conventional transgenic mice, the majorfailing with the Cre recombinase system is that it is notreversible.

Reversible tissue and temporal specific controlof gene over-expression

Tetracycline response system. Reversible control of geneexpression has for some time been possible withthe tetracycline response system (Kistner et al. 1996).In this procedure, one mouse strain is made toexpress a tetracycline-responsive transactivator protein,tTa (tet-off) or rtTa (tet-on), usually under the control ofa tissue-specific promoter. A second mouse expresses thegene of interest under the control of a minimal promoterthat requires the transactivator protein (tTA or rtTA) andtetracycline. Doubly transgenic mice for both transgenes

can then be regulated by the addition (tet-on) orremoval (tet-off) of tetracycline in the drinking waterallowing control of the gene of interest in a tissue- andtemporal-specific manner.

However, this system again is based on conventionaltransgenic mice and for the reasons given above canalso suffer from promiscuous expression. A particularproblem occurs due to the lack of complete silence ofthe minimal promoter when the transactivator protein isabsent, hence the system is never fully ‘off’. More recently,a major improvement for the rtTA system was developedinvolving a second protein (tTS) that is a fusion of thetransactivator protein and the KRAB-AB silencing domainof the Kid-1 protein (Zhu et al. 2002). tTS binds tothe minimal promoter in the absence of tetracycline andensures complete silencing of the promoter (Zhu et al.2002). In the presence of tetracycline, tTS dissociates fromthe promoter allowing rtTa to bind and drive transcriptionallowing better temporal control. However, it is importantto note that tetracycline is an antibiotic that could affectthe bacterial flora of the experimental mice necessitatingmultiple control groups. Moreover the recent findingthat the rtTA protein itself can produce emphysema-likesymptoms in the lungs of mice is a further complication(Sisson et al. 2006). Whether the tTS protein also causessimilar symptoms in the lung or indeed if there is a generalproblem of expressing rtTA or tTS protein in differenttissues needs clarification.

In spite of these potential problems, the tetracyclinesystem has enabled tissue and temporal control of genefunction.

LacO/LacIR system. Recently, an alternative, reversible andinducible expression system based on the lac operonof E. coli was developed for use in mammals (Croninet al. 2001). In E. coli, the lac operon functions by arepression mechanism involving the production of aninhibitor protein LacIR, that binds to regulatory sites lacOin the promoter and turns off transcription of the genesrequired for lactose metabolism. By adding lactose, theLacIR protein changes its binding affinity for the lacOsequences, dissociates and transcription of the lac genescan occur. E. coli uses this system to tightly control thegenes required for the utilization of lactose and it iscompletely reversible. In a proof of principle, Scrable andcolleagues were able to modify the tyrosinase promoterwith LacO sites and successfully switch coat colour on andoff at will when IPTG (an analogue of lactose) was addedto the drinking water of mice that also expressed the LacIR

protein (Cronin et al. 2001). Tyrosinase activity couldalso be regulated during discrete periods of embryogenesis(Cronin et al. 2003).

In a subsequent study, similar control of gene expressionduring embryogenesis was demonstrated in vivo fora luciferase reporter gene under the control of the




ubiquitous Huntingtin promoter with introduced LacOsites (HuntingtinLac O-luciferase) (Ryan & Scrable, 2004).In this study it was shown that luciferase activity couldbe regulated in live offspring in utero by the addition ofIPTG to the drinking water of the pregnant dam (Ryan &Scrable, 2004).

In order to study the regulation of this genein different tissues we obtained mice expressing theHuntingtinLac O-luciferase (HDOSluc) reporter and miceexpressing the LacIR repressor protein from ProfessorScrable. By crossing these mice we have observedrepression of luciferase expression by the LacIR proteinto less than 10% in a range of different tissues (Fig. 2A andMatthaei et al. 2006). Importantly, addition of IPTG to thedrinking water produced de-repression (re-activation) ofluciferase expression within 48 h to 80% of normal activity(Fig. 2B for kidney). This level of repression is similar tothat achieved with siRNA (Carmell et al. 2003) but withoutside-effects in non-target tissues (Jackson et al. 2003). Wetherefore see tight reversible regulation of a transgene usinga common compound (sugar) that should have minimalside-effects. This system warrants intensive investigationsince it possibly provides the best method of controllinggene overexpression or deletion (see below).

Epigenetic complications for geneticallymanipulated mice

Epigenetics is the transmission of information from a cellor multicellular organism to its descendants without thatinformation being encoded in the nucleotide sequence.Evidence is now emerging that epigenetic changes canoccur due to the transmission of RNA or protein in theoocyte or sperm from transgenic parents to non-transgenicoffspring. We were first alerted to this problem when wewere crossing a ‘β-galactosidase Stop floxed EGFP’ mouse(see Fig. 1) with Cre mice controlled by the Tnap promoter

Figure 2. Reversible regulation of gene activity by the LacO/LacIR systemA, luciferase activity in different tissues from HuntingtinLacO-luciferase (HDOSluc) mice and repression of activity inHDOSluc mice doubly transgenic for LacIR. B, de-repression of luciferase activity at 24 and 48 h after addition of10 mM IPTG in the drinking water.

(Lomeli et al. 2000). In this cross we expected to see theCre delete the β-galStop and allow expression of EGFP(Figs 1 and 3A). Indeed, when the Cre transgenic parentwas the male we did see faithful transmission of the Cretransgene to the offspring and deletion of our floxed genein these mice (Cre/delLoxP, Fig. 3A). However, when theCre-carrying parent was the female, we saw deletion ofour floxed gene in all offspring, even in those that had notinherited Cre (WT/delLoxP, Fig. 3B). This phenomenonhas also been observed by others. Tissue-specific deletionoccurred when the Cre transgene controlled by the keratinK5 promoter was carried by the male parent, but deletionin all offspring when the Cre transgene was carried by thefemale parent (Ramirez et al. 2004), including those notinheriting Cre. Moreover, when the Cre parent was femalethe deletion was no longer tissue specific but occurred inall tissues of the offspring, suggesting deletion very earlyduring embryogenesis (Ramirez et al. 2004). These datasuggest that RNA or Cre protein may be present in alloocytes of a Cre transgenic female and this is transmittedto non-transgenic offspring causing generalized deletion.Moreover, since this phenomenon has been seen fortwo different promoters driving Cre, suggests that theeffect may be more general. In this case, non-transgeniclittermates would not constitute an appropriate control.

Direct evidence that transgenic RNA may causephenotypic changes in non-transgenic offspring hasalso been demonstrated recently. Rassoulzadegan andcolleagues observed that male mice carrying the Kit tm1Alf

transgene had spotted white feet and tail (Fig. 4). Whenthese mice were crossed with wild-type females theyproduced offspring that all had the spotted white feet andtail phenotype, including those that had not inherited theKit tm1Alf transgene (Rassoulzadegan et al. 2006; see alsoFig. 4). Moreover they were able to show that injection ofKit tm1Alf RNA into single cell mouse embryos producedoffspring that all had the spotted white feet and tail




WT/Cre Cre/delLoxP WT/delLoxP

X

FloxStop/FloxStop

B

WT/Cre FloxStop/FloxStop

X

Cre/delLoxP WT/FloxStop

A

Figure 3. Epigenetic modification of offspring by transfer of Cre recombinase in oocytesA, male wild-type Cre-positive mice crossed with FloxStop/FloxStop females produce offspring that are all positivefor FloxStop and half are also positive for Cre. Only the Cre-positive mice are green due to the deletion of the stop(delLoxP). B, female wild-type Cre-positive mice crossed with FloxStop/FloxStop males produce offspring that areall positive for Floxtop and half are also positive for Cre. However, all the offspring are green due to the deletionof the stop (delLoxP) even in mice not positive for Cre (circled). The phenotype is due to the transfer of Cre proteinfrom the female into the oocyte where it deletes the Floxed DNA.

phenotype (Rassoulzadegan et al. 2006). This suggests thatKit tm1Alf RNA from transgenic males can be transferredin the sperm to non-transgenic offspring inducing thephenotype in the absence of the transgene itself, acomplication not previously recognized.

The two sets of data given above have profoundimplications since they show that the use of non-transgeniclittermates as controls may be inappropriate withoutcareful comparison with wild-type mice. Mostimportantly the interpretation of data from all transgenicexperiments may not be as simple as first thought if

X

offspring

+/+

Wild type

KittmAlf/+

White-spottedfeet & tail

KittmAlf/+

White-spotted feet & tail

+/+

White-spotted feet & tail

Figure 4. Epigenetic modification of offspring by transfer of RNA in spermMale KittmAlf/+ mice with white spotted feet and tail tips crossed with wild-type mice produce offspring that allhave the white spotted phenotype, even the wild-type mice not carrying the mutant gene (circled). The phenotypeis the result of KittmAlf RNA transfer in the sperm from the KittmAlf male to the wild-type mice.

transgene products can produce epigenetic effects withoutevidence of the presence of the transgene in the genome.

The way forward

Future gene-deficient mice. The ideal system to test theeffects of a gene deletion should incorporate completereversibility and no non-specific side-effects. Currentlynone of the commonly used methods satisfy theserequirements since they rely on the use of transgenicmice with their inherent limitations. Our solution to




this dilemma is to insert LacO sites into the promoterof our gene of interest by homologous recombinationlike those employed to introduce LoxP sites. By crossingthese LacO mice with the ubiquitous LacIR-expressingmice, generalized gene expression can be regulated bythe addition and removal of IPTG in the drinking water.Moreover, this system will also allow the regulation ofmutated forms of genes of interest. In this case themutation as well as the LacO sites will be introducedinto the germline by homologous recombination, thenregulated by LacIR and IPTG as above. Tissue-specificregulation of the gene can be obtained by expression ofthe LacIR protein driven from tissue-specific promoters.However, to avoid the problem of position effect,‘knock-in’ methods of LacIR into tissue-specific promotersusing homologous recombination should be implementedusing IRES constructs as indicated above.

Future transgenic mice. The ideal expression of atransgene should be totally tissue specific, not modifythe host genome or be modified by the host genome,and be completely reversible. All current methods apartfrom site-specific insertions (see above) potentially sufferfrom modification by the local environment. Moreover, theonly commonly used reversible system using tetracyclinealso suffers from side-effects from the tTA protein andpossibly from the use of the antibiotic (see above). Toovercome these problems, we are currently trialing anew method in which we are integrating our transgenesof interest into episomes (D. Rangasamy & K. I.Matthaei, unpublished observations). Episomes arecircular, self-replicating DNAs often of viral origin thatare stably and faithfully inherited during mitosis but arenot integrated into the chromosomes of mammalian cells.Moreover, since they are not integrated, episomes carryinga transgene should not only avoid the ‘position’ effectfrom neighbouring genes but also not cause insertionalmutation of an unknown gene thereby making them apossibly ideal transgene carrier.

Naturally occurring episomes are large since theyrequire the production of an array of viral proteins toallow faithful replication and segregation into daughtercells (herpesvirus for example is 172 kb; Wade-Martinset al. 2000). This makes the genetic manipulation ofepisomes difficult. However, recently it was found thatonly one of the multiple proteins is necessary forepisomal replication, the scaffold/matrix region of thehuman β-interferon gene cluster (Piechaczek et al. 1999)resulting in a plasmid of only 6.7 kb. Moreover thisepisome avoids insertional mutagenesis and methylation(Jenke et al. 2004) suggesting that it is an ideal trans-gene carrier. It is our intention to use the LacO/LacIR

control and episomal expression of the transgene togenerate transgenic mice that have fully reversible,tissue-specific and temporal control of transgene

expression without modification from the host cell. Ifsuccessful, this will be a major breakthrough for thegeneration of transgenic mice.

Conclusion

It is not my intention to suggest that all geneticallymodified mouse experiments are flawed or that usefulinformation has not been generated from geneticallyaltered mice. I have also limited my discussion tosituations where tissue and temporal control of geneexpression is required. Simple deletions of gene functionwill continue to have a place in some situations as willtransgenic mice using minimal promoters. However, it isimperative to understand the limitations of these systemsso that appropriate controls can be included and correctconclusions can be made. It is for this reason that Hagg(1999) should be congratulated. He was one of the first topoint out that there was a problem with mixed geneticbackgrounds in gene deletion experiments. Hagg andcolleagues had published a specific phenotype of thedeletion of the p75 nerve growth factor receptor (Van derZee et al. 1996) but later discovered that the backgroundstrains were responsible for the phenotype and not thegene deletion. The senior author therefore corrected theliterature by retracting the original result (Hagg, 1999).Many other published data may be similarly affected bypreviously unforseen problems. My intention here is toalert the reader to such possibilities in order to make thebest use of a powerful tool which can be used to betterunderstand gene expression and function and therebyhealth and disease. At the same time, discussions like thesemight stimulate the reader to re-assess their data in the lightof some of the complications presented here, particularlywhen ‘inexplicable’ results are obtained.

In conclusion, the potential to use completely reversiblemethods like the LacO/LacIR system will allow eachexperimental group of animals to be their own controland should help detect epigenetic (or other) influencesin studies of both gene overexpression and gene deletionthereby providing exciting new developments for thefuture.

References

Avery S, Rothwell L, Degen WD, Schijns VE, Young J, KaufmanJ & Kaiser P (2004). Characterization of the firstnonmammalian T2 cytokine gene cluster: the clustercontains functional single-copy genes for IL-3, IL-4, IL-13,and GM-CSF, a gene for IL-5 that appears to be apseudogene, and a gene encoding another cytokineliketranscript, KK34. J Interferon Cytokine Res 24, 600–610.

Blackburn AC, Matthaei KI, Lim C, Taylor MC, Cappello JY,Hayes JD, Anders MW & Board PG (2006). Deficiency ofglutathione transferase zeta causes oxidative stress andactivation of antioxidant response pathways. Mol Pharmacol69, 650–657.




Carmell MA, Zhang L, Conklin DS, Hannon GJ & RosenquistTA (2003). Germline transmission of RNAi in mice. NatStruct Biol 10, 91–92.

Cronin CA, Gluba W & Scrable H (2001). The lac operator-repressor system is functional in the mouse. Genes Dev 15,1506–1517.

Cronin CA, Ryan AB, Talley EM & Scrable H (2003).Tyrosinase expression during neuroblast divisions affectslater pathfinding by retinal ganglion cells. J Neurosci 23,11692–11697.

Dent LA, Strath M, Mellor AL & Sanderson CJ (1990).Eosinophilia in transgenic mice expressing interleukin 5.J Exp Med 172, 1425–1431.

Faast R, Thonglairoam V, Schulz TC, Beall J, Wells JR, TaylorH, Matthaei K, Rathjen PD, Tremethick DJ & Lyons I (2001).Histone variant H2A.Z is required for early mammaliandevelopment. Curr Biol 11, 1183–1187.

Feng G, Mellor RH, Bernstein M, Keller-Peck C, Nguyen QT,Wallace M, Nerbonne JM, Lichtman JW & Sanes JR (2000).Imaging neuronal subsets in transgenic mice expressingmultiple spectral variants of GFP. Neuron 28, 41–51.

Gazit R, Gruda R, Elboim M, Arnon TI, Katz G, Achdout H,Hanna J, Qimron U, Landau G, Greenbaum E, Zakay-RonesZ, Porgador A & Mandelboim O (2006). Lethal influenzainfection in the absence of the natural killer cell receptorgene Ncr1. Nat Immunol 7, 517–523.

Gong S, Zheng C, Doughty ML, Losos K, Didkovsky N,Schambra UB, Nowak NJ, Joyner A, Leblanc G, Hatten ME &Heintz N (2003). A gene expression atlas of the centralnervous system based on bacterial artificial chromosomes.Nature 425, 917–925.

Gu H, Marth JD, Orban PC, Mossmann H & Rajewsky K(1994). Deletion of a DNA polymerase beta gene segment inT cells using cell type-specific gene targeting. Science 265,103–106.

Hagg T (1999). Neuronal cell death: retraction. Science 285,340.

Hendry IA, Hill CE & McLennan IS (1987). RU38486 blocksthe steroid regulation of transmitter choice in cultured ratsympathetic ganglia. Brain Res 402, 264–268.

Jackson AL, Bartz SR, Schelter J, Kobayashi SV, Burchard J,Mao M, Li B, Cavet G & Linsley PS (2003). Expressionprofiling reveals off-target gene regulation by RNAi. NatBiotechnol 21, 635–637.

Jenke AC, Scinteie MF, Stehle IM & Lipps HJ (2004).Expression of a transgene encoded on a non-viral episomalvector is not subject to epigenetic silencing by cytosinemethylation. Mol Biol Rep 31, 85–90.

Kellendonk C, Tronche F, Monaghan AP, Angrand PO, StewartF & Schutz G (1996). Regulation of Cre recombinase activityby the synthetic steroid RU 486. Nucl Acids Res 24,1404–1411.

Kistner A, Gossen M, Zimmermann F, Jerecic J, Ullmer C,Lubbert H & Bujard H (1996). Doxycycline-mediatedquantitative and tissue-specific control of gene expression intransgenic mice. Proc Natl Acad Sci U S A 93, 10933–10938.

Kwan KM (2002). Conditional alleles in mice: practicalconsiderations for tissue-specific knockouts. Genesis 32,49–62.

Lee JS & Young IG (1989). Fine-structure mapping of themurine IL-3 and GM-CSF genes by pulsed-field gelelectrophoresis and molecular cloning. Genomics 5, 359–362.

Lim CEL, Matthaei KI, Blackburn AC, Davis RP, Dahlstrom JE,Koina ME, Anders MW & Board PG (2004). Mice deficientin glutathione transferase zeta/maleylacetoacetate isomeraseexhibit a range of pathological changes and elevatedexpression of alpha, mu, and pi class glutathionetransferases. Am J Pathol 165, 679–693.

Lomeli H, Ramos-Mejia V, Gertsenstein M, Lobe CG & Nagy A(2000). Targeted insertion of Cre recombinase into the TNAPgene: excision in primordial germ cells. Genesis 26, 116–117.

McKenzie GJ, Fallon PG, Emson CL, Grencis RK & McKenzieAN (1999). Simultaneous disruption of interleukin (IL)-4and IL-13 defines individual roles in T helper cell type 2-mediated responses. J Exp Med 189, 1565–1572.

Matthaei KI (2004). Caveats of gene targeted and transgenicmice. In Handbook of Stem Cells, ed. Lanza R, Gearhart J,Hogan B, Melton DW, Pederson R, Thomson J & West W,pp. 589–598. Elsevier Inc., San Diego, CA, USA.

Matthaei KI, Lam KE, Grayson TH, Scrable H & Hill CE(2006). Tissue and temporal control of gene function in vivo.Transgenic Res 15, 777.

Moreira PN, Pozueta J, Giraldo P, Gutierrez-Adan A &Montoliu L (2006). Generation of yeast artificialchromosome transgenic mice by intracytoplasmic sperminjection. Methods Mol Biol 349, 151–161.

Palmiter RD, Brinster RL, Hammer RE, Trumbauer ME,Rosenfeld MG, Birnberg NC & Evans RM (1982). Dramaticgrowth of mice that develop from eggs microinjected withmetallothionein-growth hormone fusion genes. Nature 300,611–615.

Piechaczek C, Fetzer C, Baiker A, Bode J & Lipps HJ (1999). Avector based on the SV40 origin of replication andchromosomal S/MARs replicates episomally in CHO cells.Nucl Acids Res 27, 426–428.

Rajewsky K, Gu H, Kuhn R, Betz UA, Muller W, Roes J &Schwenk F (1996). Conditional gene targeting. J Clin Invest98, 600–603.

Ramirez A, Page A, Gandarillas A, Zanet J, Pibre S, Vidal M,Tusell L, Genesca A, Whitaker DA, Melton DW & Jorcano JL(2004). A keratin K5Cre transgenic line appropriate fortissue-specific or generalized Cre-mediated recombination.Genesis 39, 52–57.

Rassoulzadegan M, Grandjean V, Gounon P, Vincent S, Gillot I& Cuzin F (2006). RNA-mediated non-mendelianinheritance of an epigenetic change in the mouse. Nature441, 469–474.

Rees S, Coote J, Stables J, Goodson S, Harris S & Lee MG(1996). Bicistronic vector for the creation of stablemammalian cell lines that predisposes all antibiotic-resistantcells to express recombinant protein. Biotechniques 20,102–104, 106, 108–110.

Ryan A & Scrable H (2004). Visualization of the dynamics ofgene expression in the living mouse. Mol Imaging 3, 33–42.

Sisson TH, Hansen JM, Shah M, Du Hanson KEM, Ling T,Simon RH & Christensen PJ (2006). Expression of the reversetetracycline-transactivator gene causes emphysema-likechanges in mice. Am J Respir Cell Mol Biol 34, 552–560.




Soriano P (1999). Generalized lacZ expression with theROSA26 Cre reporter strain. Nat Genet 21, 70–71.

Thyagarajan B, Guimaraes MJ, Groth AC & Calos MP (2000).Mammalian genomes contain active recombinaserecognition sites. Gene 244, 47–54.

Van der Zee CE, Ross GM, Riopelle RJ & Hagg T (1996).Survival of cholinergic forebrain neurons in developingp75NGFR-deficient mice. Science 274, 1729–1732.

Wade-Martins R, White RE, Kimura H, Cook PR & James MR(2000). Stable correction of a genetic deficiency in humancells by an episome carrying a 115 kb genomic transgene.Nat Biotechnol 18, 1311–1314.

Yang T, Riehl J, Esteve E, Matthaei KI, Goth S, Allen PD, PessahIN & Lopez JR (2006). Pharmacologic and functionalcharacterization of malignant hyperthermia in the R163CRyR1 knock-in mouse. Anesthesiology 105, 1164–1175.

Zheng B, Zhang Z, Black CM, de Crombrugghe B & Denton CP(2002). Ligand-dependent genetic recombination infibroblasts: a potentially powerful technique for investigatinggene function in fibrosis. Am J Pathol 160, 1609–1617.

Zhou Z, Wang D, Wang XJ & Roop DR (2002). In uteroactivation of K5.CrePR1 induces gene deletion. Genesis 32,191–192.

Zhu Z, Zheng T, Lee CG, Homer RJ & Elias JA (2002).Tetracycline-controlled transcriptional regulation systems:advances and application in transgenic animal modeling.Semin Cell Dev Biol 13, 121–128.

Acknowledgements

I wish to thank Caryl Hill for constructive criticism and

many helpful discussions on the content as well as careful

editing of the manuscript. I also thank Heidi Scrable for

the HuntingtinLac O-luciferase (HDOSluc) reporter and the

LacIR- expressing mice as well as candid discussions about the

LacO/LacIR system. I finally thank Martin Goulding for use of

the N/ZEG reporter mice in Fig. 3.



J Physiol 582.2 (2007) pp 489–506 489

Ryanodine receptor type 2 deficiency changes excitation–contraction coupling and membrane potential in urinarybladder smooth muscle

Shingo Hotta1, Kozo Morimura1, Susumu Ohya1, Katsuhiko Muraki1,2, Hiroshi Takeshima3

and Yuji Imaizumi1

1Department of Molecular and Cellular Pharmacology Graduate School of Pharmaceutical Science, Nagoya City University, Nagoya, Japan2Cell Signalling & Ion Channel Research Group, Cellular Pharmacology, School of Pharmacy, Aichigakuin University, Nagoya, Japan3Department of Biological Chemistry, Kyoto University Graduate School of Pharmaceutical Sciences, Kyoto, Japan

The possibility that the ryanodine receptor type 2 (RyR2) can function as the major Ca2+-induced

Ca2+ release (CICR) channel in excitation–contraction (E-C) coupling was examined in smooth

muscle cells (SMCs) isolated from urinary bladder (UB) of RyR2 heterozygous KO mice

(RyR2+/−). RyR2 mRNA expression in UB from RyR2+/− was much lower than that in wild-type

(RyR2+/+). In single UBSMCs from RyR2+/+, membrane depolarization under voltage clamp

initially induced several local Ca2+ transients (hot spots) in peripheral areas of the cell. Then, Ca2+

waves spread from Ca2+ hot spots to other areas of the myocyte. The number of Ca2+ hot spots

elicited by a short depolarization (< 20 ms) in UBSMCs of RyR2+/− was significantly smaller

than in those of RyR2+/+. The force development induced either by direct electrical stimulation or

by 10 μM acetylcholine in tissue segments of RyR2+/− was smaller than and comparable to those

in RyR2+/+, respectively. The frequency of spontaneous transient outward currents in single

myocytes and the membrane depolarization by 1 μM paxilline in tissue segments from RyR2+/−

were significantly lower and smaller than those in RyR2+/+, respectively. The urination frequency

and volume per voiding in RyR2+/− were significantly increased and reduced, respectively,

compared with RyR2+/+. In conclusion, RyR2 plays a crucial role in the regulation of CICR

during E-C coupling and also in the regulation of resting membrane potential, presumably via

the modulation of Ca2+-dependent K+ channel activity in UBSMCs and, thereby, has a pivotal

role in the control of bladder activity.

(Resubmitted 11 February 2007; accepted after revision 8 March 2007; first published online 15 March 2007)

Corresponding author Y. Imaizumi: Department of Molecular and Cellular Pharmacology, Graduate School of

Pharmaceutical Science, Nagoya City University, 3-1 Tanabedori, Mizuhoku, Nagoya 467-8603, Japan.


The ryanodine receptor (RyR) is one of the major Ca2+

release channels in endo- and sarcoplasmic reticulum (ERand SR). Previous cDNA cloning studies have definedthree subtypes of RyR (RyR1, RyR2 and RyR3) that areencoded by distinct genes in vertebrates (Meissner, 1994;Laporte et al. 2004). RyR1 is expressed abundantly inskeletal muscle cells (Takeshima et al. 1989). RyR2 isfound predominantly in cardiac muscle cells, althoughit is expressed at moderate levels in most excitable cells(Nakai et al. 1990; Otsu et al. 1990). RyR3 is expressed atlow levels in a wide variety of cell types including mostexcitable cells and certain non-excitable cells (Gianniniet al. 1992; Hakamata et al. 1992). Among these three

This paper has online supplemental material.

types of RyRs, RyR2 is expressed in cardiac muscle,smooth muscle and brain, and is considered to be thechannel mainly responsible for the Ca2+-induced Ca2+

release (CICR) mechanism (Fabiato, 1985; Iino, 1989). Incardiac ventricular myocytes, an action potential evokesopening of voltage-dependent Ca2+ channels (VDCCs)and Ca2+ influx into myoplasm, which activates RyR2in junctional SR to release Ca2+ (Fabiato, 1983). RyR2is considered to be an essential molecule for CICR inthe excitation–contraction coupling in cardiac ventricularmyocytes. RyR2 gene deletion in mice (RyR2−/−) resultsin embryonic lethality in homozygotes. Specifically, theprogeny die at approximately embryonic day 10 as a resultof cardiac arrest, which is presumably due to SR Ca2+ over-loading and resulting dysfunction of intracellular organellein cardiac myocytes (Takeshima et al. 1998).



490 S. Hotta and others J Physiol 582.2

In contrast to cardiac myocytes, in many types of smoothmuscles (SMs), inositol 1,4,5-trisphosphate (IP3)-inducedCa2+ release (IICR) by formation of IP3 via GTP bindingprotein coupled receptor stimulation and the subsequentactivation of phospholipase C is more common thanCICR. The contribution of CICR to E-C coupling widelyvaries between SM types. CICR making a substantialcontribution has been reported only in highly excitableSMs such as urinary bladder (UB) (Ganitkevich &Isenberg, 1992; Imaizumi et al. 1998; Hashitani et al. 2000)and vas deferens (Imaizumi et al. 1998), but not portal vein(Kamishima & McCarron, 1996). It has been suggested,however, that the coupling between VDCC and RyR isrelatively weak even in UBSM cells (UBSMCs) of the rabbit(Collier et al. 2000). There are also reports suggesting thatCICR may not be involved effectively in E-C coupling inguinea-pig UB (Herrera et al. 2000; Hashitani & Brading,2003) and rat and human uterus (Taggart & Wray, 1998;Kupittayanant et al. 2002). In a previous study, we havedemonstrated that CICR through RyR is essential for E-Ccoupling triggered by an evoked action potential in mouseUBSMCs. We also showed that the cross talk of CICR withIICR by IP3 formation may not be involved in the E-Ccoupling under these conditions (Morimura et al. 2006).

In many types of SMs, the expression levels of RyR2 aremuch lower than that in cardiac myocytes (Nakai et al.1990) and the expression of RyR3 has been suggested to belower than but occasionally comparable to that of RyR2(Chamber et al. 1999; Sanders, 2001). Subtype-specificcontributions of RyR2 and/or RyR3 to CICR during E-Ccoupling in SMs are, however, not well characterized. Thesubstantial contribution of RyR2 to Ca2+ spark generationhas been suggested only by the indirect evidence thatCa2+ sparks were altered in SM myocytes from FKBP12.6deficient mice (Wang et al. 2004). In addition, it hasbeen reported that the Ca2+ mobilization in UBSMmyocytes from RyR3 homozygous KO mice is notsignificantly different from that in wild-type mice (Jiet al. 2004). In contrast, Ca2+ spark/spontaneous trans-ient outward current (STOC) frequency was significantlyincreased in cerebral artery SMCs of RyR3 KO mice (Lohnet al. 2001). Consistent with the latter finding in RyR3−/−,it has been suggested recently that an alternative splicevariant of RyR3, which works as a dominant negativeconstruct, is predominantly expressed in SM tissues,including mouse UB (Dabertrand et al. 2006).

The present study was undertaken to elucidate thefunctional significance of RyR2 in E-C coupling in UBSMusing RyR2 heterozygous KO mice (RyR2+/−), in whichthe functional expression of RyR2 may be substantiallyreduced. The present results provide direct evidence for anobligatory role of RyR2 in E-C coupling, and also stronglysuggest that RyR2 activity can regulate resting membranepotential in UBSM through modulation of Ca2+ activatedK+ channel activity.

Methods

PCR Genotyping of RyR2+/− mice

Generation of the RyR2 knockout mice has beenreported previously (Takeshima et al. 1998). To determinethe genotypes of the mutant mice, the polymerasechain reaction (PCR) was carried out using primersfrom the genomic sequence: forward primer (Ex1-P6D,25 mer: GAGCCCCTAGAACATCCTGGTTAGC) andreverse primer (AInt-668, 25 mer: GCACCCTGG-GGGCAGCCTTCTCAGC) (Takeshima et al. 1998).Amplified DNAs were analysed on 1.5% agarose gels.

RNA extraction and RT-PCR

Eight- to ten-week-old male and female mice wereanaesthetized with ether and killed by decapitation. Allexperiments were carried out in accordance with theguiding principles for the care and use of laboratoryanimals of The Science and International Affairs Bureauof the Japanese Ministry of Education, Culture, Sports,Science and Technology, and also with the approval of theethics committee at Nagoya City University. Total RNAswere extracted from homogenates of aorta, brain, colon,diaphragm, heart, ileum, stomach, urinary bladder, uterusor vas deferens by the acid guanidium thiocyanate–phenolmethod following digestion with RNase-free DNase,and RT was performed with a Gibco BRL protocol aspreviously reported (Ohya et al. 1997). The designedprimers are shown in Table 1. The amplification profilesfor these primer pairs were as follow: 95◦C for 10 min toactivate the AmpliTaq polymerase, then 32 cycles of 9◦Cfor 15 s and 60◦C for 1 min, performed in a GeneAmp 2400thermal cycler (PE Applied Biosystems, USA).

Quantitative PCR

Real-time quantitative PCR was performed with the use ofSYBR Green Chemistry on an ABI 7000 sequence detector(Applied Biosystems, Foster City, CA, USA). Standardcurves were generated for the constitutively expressedglyceraldehyde-3-phosphate dehydrogenase (GAPDH)from regression analysis of the mean values of RT-PCRsfor the log10 diluted cDNA. Unknown quantities relativeto the standard curve for a particular set of primers werecalculated, yielding transcriptional quantification of geneproducts relative to the endogenous standard (GAPDH).Each cDNA sample was tested in triplicate. The specificprimers used in this series of experiments are listed inTable 1.

Cell isolation

Male mice between 8 and 10 weeks old were used inthese experiments. Single smooth muscle cells (SMCs)were enzymatically isolated from the urinary bladder



J Physiol 582.2 RyR2 deficiency in smooth muscle E-C coupling 491

Table 1. Oligonucleotide sequence of primers used for quantitative PCR

Primer Product GenBankPrimer sequence site length (bp) accession no.

RyR1 (+): 5′-ATTACAGAGCAGCCCGAGGAT-3′ 450–470 113 X83932(−): 5′-AGAACCTTCCGCTTGACAAACT-3′ 562–541

RyR2 (+): 5′-CTTCGATGTTGGCCTTCAAGAG-3′ 432–453 102 NM 023868(−): 5′-AGAACCTTCCGCTTGACAAACT-3′ 533–512

RyR3 (+): 5′-GGCCAAGAACATCAGAGTGACTAA-3′ 385–408 101 AF111166(−): 5′-TCACTTCTGCCCTGTCAGTTTC-3′ 485–464

VDCCα1c (+): 5′-ACCTGGAACGAGTGGAGTATCTCTT-3′ 473–497 114 NM 009781(−): 5′-TCCAACCATTGCGGAGGTAA-3′ 586–567

VDCCβ2 (+): 5′-CAGGGTTCTCAAGGTGATCAAAG-3′ 1477–1499 110 NM 023116(−): 5′-GAGGAACGGTGTTGGGATTTT-3′ 1586–1566

VDCCβ3 (+): 5′-CTCCCATCATCGTCTTTGTCAA-3′ 913–934 117 NM 007581(−): 5′-GCTTATCGTACGCCATCATCTG-3′ 1029–1008

BKα (+): 5′-GCATTGGTGCCCTCGTAATATAC-3′ 1308–1330 105 NM 010610(−): 5′-CGTTGAAAGCCATGTCGATCT-3′ 1412–1392

SK2 (+): 5′-AACCACCGCAGATGTGGATATT-3′ 732–753 103 AA692872(−): 5′-GGCATCGGTGAAAAGTTTGC-3′ 834–815

SK4/IK (+): 5′-CGTGCACAACTTCATGATGGA-3′ 885–905 106 AF072884(−): 5′-TTCCTTCGAGTGTGCTTGTAGTACA-3′ 990–966

JP2 (+): 5′-AAGAAGGGCCGTAAGGAAGT-3′ 2378–2397 106 AB024447(−): 5′-GGCCGATGTTCAGCAAGATC-3′ 2483–2464

SERCA2 (+): 5′-AGTTCATCCGCTACCTCATCTCA-3′ 2483–2505 119 AJ223584(−): 5′-CACCAGATTGACCCAGAGTAACTG-3′ 2601–2578

FKBP1a (+): 5′-ACTAGGCAAGCAGGAGGTGA-3′ 247–266 104 NM 008019(−): 5′-CTCCATAGGCATAGTCTGAGGAGAT-3′ 350–326

FKBP1b (+): 5′-GAGACGGAAGGACATTCCCTAAG-3′ 32–54 70 NM 016863(−): 5′-CCCTTTTGAAGCATTCCTGTGT-3′ 101–80

InsP3R1 (+) 5′-GGACCGGACAATGGAACAGAT-3′ 6928–6948 101 NM 010585(−) 5′-CATCCCGCTCTGTGGTGTAAT-3′ 7008–7028

Sorcin (+) 5′-TGGACAGGACGGACAAATTGA-3′ 84–104 101 NM 025618(−) 5′-GGCGACAAGTCTCCAGGTTAAA-3′ 184–163

GAPDH (+): 5′-CATGGCCTTCCGTGTTCCT-3′ 730–748 104 M32599(−): 5′-CCTGCTTCACCACCTTCTTGA-3′ 833–813

+, sense; –, antisense.

(UB) using the previously described method (Imaizumiet al. 1989) with slight modifications. In brief, mice wereanaesthetized with ether and killed by decapitation. UBwas dissected out and freed from other tissues in Ca2+-freeKrebs solution. The tissue was immersed in Ca2+-freeKrebs solution for 30–60 min in a test tube at 37◦C.Subsequently, the solution was replaced with Ca2+-freeKrebs solution containing 0.2–0.3% collagenase (Amanoenzyme, Nagoya, Japan). After 10–15 min treatment, thesolution was replaced with Ca2+-free and collagenase-freeKrebs solution. The tissue was then gently trituratedusing a glass pipette to isolate cells. At the start of eachexperiment, a few drops of the cell suspension wereplaced in a recording chamber. The bath was continuouslyperfused with Hepes-buffered solution at a flow rate of5 ml min−1.

Electrical recording and data analysis

Whole-cell voltage clamp was applied to single cellswith patch pipettes using a CEZ-2400 (Nihon Kohden,Japan) amplifier. The pipette resistance ranged from 2to 5 M�, when filled with pipette solution. The sealresistance was approximately 30 G�. Data were storedand analysed using menu-drive software as previouslyreported (Imaizumi et al. 1996). Membrane currentswere digitized using a pulse coding modulator (PCM)recording system (PCM-501ES; Sony, Tokyo, Japan) andstored on video tape. Data on tapes were replayed onto apersonal computer using data acquisition software. Dataanalysis was done on a computer using software (Cell-Soft)developed at the University of Calgary, Canada. Leakagecurrents at potentials positive to −60 mV were subtracted




on the computer, assuming a linear relationship betweencurrent and voltage in the range of −90 mV to −60 mV. Allexperiments were done at room temperature (23 ± 1◦C).

For intracellular recordings, thin strips (3 × 3 mm) ofUB were removed from the mucosal layer of a bladder.A strip was pinned to the bottom in a 1 ml chamber,perfused with Hepes-buffered Krebs solution gassed with95% O2–5% CO2 at a rate of 2–4 ml min−1 and keptat 36 ± 1◦C and pH 7.4. The transmembrane potentialwas measured with conventional glass microelectrodes,having resistance of 35–50 M�, and amplified by a highinput impedance amplifier with capacitance neutralization(MEZ-7200, Nihon Kohden, Tokyo, Japan) for monitoringon a dual-beam storage oscilloscope (VC-10, NihonKohden). The transmembrane potential was recordedcontinuously using a pen recorder (FBR-251 A, TOAElectrics, Ltd, Tokyo, Japan). During experimentalmanoeuvres, Hepes-buffered Krebs solution was used asthe external solution.

Measurement of Ca2+ signal from single cells

Ca2+ images were obtained using a fast laser-scanningconfocal microscope (RCM 8000; Nikon, Japan) and ratio3software (Nikon, Japan) in the same manner as reportedpreviously (Imaizumi et al. 1998; Ohi et al. 2001). Amyocyte was loaded with 100 μm fluo-4 or fluo-5 bydiffusion from the patch pipette. Excitation light of488 nm from an argon ion laser was delivered through awater-immersion objective (Nikon Fluo ×40, 1.15 NA).Emission light of > 515 nm was detected by a photo-multiplier. Fluorescence intensity (F) in a selected area wasmeasured as an average from pixels included in the area.It took 33 ms to scan one frame (512 × 512 pixels). Using1/2 and 1/4 band scan modes, frames that correspondingto areas of 170 × 55 μm or 170 × 27.5 μm were obtainedevery 16.5 and 8.25 ms, respectively. The resolution ofthe microscope was approximately 0.4 × 0.3 × 1.2 μm (x,y and z) based on the measurement using fluorescentbeads having diameters of 300, 500, 1000 and 1500 nm(sicastar-green F, Nacalai Tesque, Kyoto, Japan). Theconfocal plane through the cell was usually set, at a positionwhere the width of the cell was largest, 2–3 μm fromits lowest point. Recordings were started at 3 min afterrupturing the patch membrane to allow sufficient time forfluo-4 to diffuse into the myocyte.

In separate experiments, some measurements ofCa2+ signals were also performed using fura-2. SingleUBSMs were loaded with 10 μm fura-2 AM in standardHepes-buffered solution for 20 min at room temperature(23 ± 1◦C). Measurements of fura-2 fluorescence wereperformed with an Argus 50/CA imaging system(Hamamatsu Photonics, Japan). The frequency of imageacquisition was constant at one image every 1 s. Theintensity of emission fluorescence> 510 nm was measured

to the alternate excitation (340 nm and 380 nm). Theexperiments were carried out at room temperature.

Measurements of contractility from urinary bladdertissue strips

Measurements of contractility from UB tissue strips werecarried out as reported previously (Morimura et al. 2006).In brief, 8- to 12-week old male mice were anaesthetizedwith ether and killed by decapitation, and the UB wasremoved and placed in Ca2+-free Krebs solution. Thebladder was then cut open. The detrusor muscle wasisolated as small strips (0.8–1.2 mm wide and 5–6 mmlong). Each strip was placed in a tissue bath (∼6 mlin volume) containing aerated Krebs solution with 95%O2 and 5% CO2 and kept at 37◦C. One end of thestrip was pinned to the chamber bottom while the otherend was connected to a force-displacement transducerfor measurement of isometric contractility. Strips werestretched to approximately 1 mN of tension. To applyelectrical field stimulation, platinum stimulating electro-des were placed along a tissue in both sides. Electricalfield stimulation protocols are shown in the inset ofFig. 4A. To suppress the contractile component due totransmitter release from nerve endings in the bladderstrips, all experiments using tissue strips were conducted inthe presence of the following neurotransmitter antagonists(μm): 1 atropine, 1 phentolamine, 1 propranolol, 1tetrodotoxin and 10 suramin.

Solution

Standard Krebs solution was made daily and contained(mm): 112 NaCl, 4.7 KCl, 2.2 CaCl2, 1.2 MgCl2 25NaHCO3, 1.2 KH2PO4, 14 glucose. Ca2+-free Krebssolution was prepared by omitting Ca2+ from standardKrebs solution. Ca2+- and Mg2+-free Hanks’ solutioncontained (mm): 137 NaCl, 5.4 KCl, 0.17 Na2HPO4,0.44 KH2PO4, 4.2 NaHCO3, 5.6 glucose. Standard andmodified Krebs solutions and Ca2+- and Mg2+-free Hanks’solution were aerated with 95% O2–5% CO2 to obtainpH 7.4. For measurements of transmembrane potentialwith conventional glass microelectrodes, Hepes-bufferedKrebs solution having the following composition wasused as the external solution (mm): 120 NaCl, 4.8 KCl,1.2 CaCl2, 1.3 MgSO4, 12.6 NaHCO3, 1.2 KH2PO4, 5.8glucose, 10 Hepes and aerated with 95% O2–5% CO2

to obtain pH 7.4. For electrical recording from isolatedmyocytes, Hepes-buffered solution having the followingcomposition was used as the external solution (mm):137 NaCl, 5.9 KCl, 2.2 CaCl2, 1.2 MgCl2, 14 glucose,10 Hepes, and pH was adjusted to 7.4 with NaOH.In some experiments, K+ currents were blocked byuse of an external solution, in which 31 mm KCl wasreplaced by 30 mm tetraethylammonium chloride and1 mm 4-aminopyridine-HCl.




The standard pipette solution for membrane currentrecording contained (mm): 140 KCl, 4 MgCl2, 10 Hepes,5 Na2ATP, 0.05 EGTA and pH was adjusted to 7.2with KOH. When recording only Ca2+ channel currents,the pipette solution contained (mm): 120 CsCl, 20tetraethylammonium chloride, 1 MgCl2, 10 Hepes, 5EGTA, 2 Na2ATP, and pH was adjusted to 7.2 with CsOH.To measure whole K+ currents including Ca2+ activatedK+ currents, a pipette solution of following compositionwas used (mm): 140 KCl, 1 MgCl2, 6.1 CaCl2, 10 Hepes,2 Na2ATP, 10 EGTA, and the pH of the solution wasadjusted to 7.2 with KOH. The pCa of the solution wascalculated to be 6.5. When measuring both fluo-4 signalsand membrane currents, the pipette solution having thefollowing composition was used (mm): 140 KCl, 1 MgCl2,10 Hepes, 2 Na2ATP, 0.1 fluo-4, and pH was adjusted to7.2 with KOH.

Urination patterns

Female mice of 10–14 weeks old were used in theseexperiments. After stopping the supply of water for a day,water (1 ml) was supplied to mice by the probe. Mice werethen placed in standard cages for 1 h with the beddingreplaced by filter paper (Advantec, Japan). Urine spots werephotographed under UV light (Meredith et al. 2004).

Materials

The sources of pharmacological agents were asfollows: CdCl2, tetraethylammonium chloride (TEA),4-aminopyridine (4-AP), ryanodine, caffeine: Wako PureChemical Industries, Osaka, Japan; acetylcholine (ACh),paxilline: Sigma Chemical Co., St Louis, MO, USA; EGTA,Hepes: Dojin, Kumamoto, Japan); fluo-4, fluo-5, fura-2AM: Molecular Probes, Eugene, OR, USA.

Statistical analysis

Data are expressed as the mean ± s.e.m. in the text.Statistical significance between two or among multiplegroups was examined using Student’s t test or Tukey’stest after an F-test or one-way ANOVA, respectively.Significance is expressed in the figures by asterisks:∗P < 0.05, ∗∗P < 0.01.

Results

Changes in mRNA expression levels of RyR subtypesin RyR2+/− animals

The mRNA expression levels of RyR1, RyR2 and RyR3were determined based on real-time PCR in various tissuetypes of RyR2+/+ and RyR2+/− (Fig. 1, SupplementalTable). As expected, RyR1 mRNA is expressed abundantly

in skeletal muscle (diaphragm), but at much lower levelin other tissues. There was no difference in RyR1 mRNAexpression levels in skeletal muscle between RyR2+/+ andRyR2+/− progeny (Fig. 1A). The analysis also confirmedthat RyR2 mRNA was expressed predominantly in heartand moderately in SM tissues (Fig. 1B). The RyR2 mRNAexpression in heart of RyR2+/− was not significantlydifferent from that of RyR2+/+. The RyR2 mRNAexpression examined in eight SM tissues was significantlyreduced in RyR2+/− animals as compared to that inRyR2+/+, except for the colon and uterus. RyR3 mRNAwas expressed at low levels in a wide variety of tissues,but the levels were relatively high in urinary bladder (UB),vas deferens and uterus (Fig. 1C). In uterus, RyR3 mRNAexpression was comparable to that of RyR2. The differencein RyR3 mRNA expression levels between RyR2+/+ andRyR2+/− was not significant in any of the tissues examined.The decrease in RyR2 mRNA expression in RyR2+/− wasthe most extensive in UB (by 72%) (Fig. 1D).

Preliminary study by the Western blotting analysis usinganti-RyR antibody, which identifies all three RyR isoforms,suggested that the expression of RyR protein in UB fromRyR2+/− was substantially lower than that from RyR2+/+

(Supplemental Fig. 1). Based on these observations, thepresent study was focused on the difference in functionalcontribution of RyR to E-C coupling in UBSM fromRyR2+/− and RyR2+/+.

Local Ca2+ events and activation of membranecurrents during depolarization in UBSMCs

Confocal images of local changes in intracellularCa2+ concentration ([Ca2+]i) and membrane currentsduring depolarization were simultaneously recorded fromUBSMCs isolated from RyR2+/+ and RyR2+/−. UBSMCswere loaded with 100 μm fluo-4 by diffusion fromrecording pipettes during whole cell voltage clamp.Myocytes were depolarized from −60 to 0 mV for 50 ms.Figure 2A consists of fluorescent images obtained every8.25 ms from UBSMCs of RyR2+/+ and RyR2+/−. InUBSMC from RyR2+/+, the rise of [Ca2+]i duringdepolarization appeared first as local Ca2+ elevations(Ca2+ hot spots) in well-defined intracellular locations.Subsequently these changes spread to other parts ofcell as has been reported previously (Ohi et al. 2001;Morimura et al. 2006). In the image obtained 26.9 msafter depolarization, Ca2+ hot spots were observed in fiveseparate areas in a single confocal plane. This rise of [Ca2+]i

then slowly spread from the spots to the entire intra-cellar area forming a Ca2+ wave even after repolarization(Morimura et al. 2006). In contrast, at 26.9 ms, in UBSMCfrom RyR2+/−, Ca2+ hot spots were detected (4 hot spotsin Fig. 2A) but less Ca2+ wave extension was observed.

Figure 2B shows time courses of [Ca2+]i changes (redand green circles) measured from two Ca2+ hot spots,




which are indicated by red and green arrowheads inFig. 2A, respectively. The average signal in the whole cellarea is also denoted (blue circles). The rise of local andglobal [Ca2+]i in RyR2+/− appeared to be smaller thanthat in RyR2+/+. Figure 2B also shows membrane currents(black dots) elicited by depolarization from −60 to 0 mVfor 50 ms. The outward current elicited by depolarizationin RyR+/− was smaller than that in RyR+/+, while the initialinward current appeared to be similar between RyR+/− andRyR+/+. The major part of the outward current recordedat 0 mV (by over 80%) was inhibited by addition of 1 μm

paxilline or 100 nm iberiotoxin (not shown) (Morimuraet al. 2006). These findings indicate that BK channelcurrent is responsible for the major part of this outward

Figure 1. The mRNA expression of RyR subtypes in various tissues from RyR2+/+ and RyR2+/−

Quantitative analyses for RyR subtypes were performed by real-time PCR. RyR1 (A), RyR2 (B) and RyR3 (C) mRNAexpression levels were compared between RyR2+/+ and RyR2+/−. A, n = 4, except for portal vein and aorta ofRyR2+/− (n = 3); B, n = 4, except for urinary bladder of RyR2+/+ (n = 8) and portal vein and aorta of RyR2+/−(n = 3); C, n = 4, except for portal vein and aorta of RyR2+/− (n = 3). D, comparison of mRNA expression levelsof RyR1–3 in urinary bladder. n = 4, except for RyR2 of RyR2+/+. ∗P < 0.05 versus RyR2+/+.

current as has been reported previously (Imaizumi et al.1998). The cell capacitance of UBSMCs was not differentbetween RyR2+/+ (44.4 ± 2.33 pF, n = 20) and RyR2+/−

(45.5 ± 1.92 pF, n = 19; P > 0.05).Figure 2C summarizes the [Ca2+]i results (F/F0)

measured at Ca2+ hot spots and in whole cell area inRyR2+/+ and RyR2+/− UBSM cells. The [Ca2+]i rise inhot spots in RyR2+/− was significantly smaller than that inRyR2+/+ (P < 0.01). The global [Ca2+]i rise in RyR2+/−

was also significantly smaller than that in RyR2+/+

(P < 0.01). Figure 2D denotes summarized results whichdescribe the number of Ca2+ hot spots per single confocalimage obtained after 18.7 ms depolarization typicallyshown in Fig. 2A. Ca2+ hot spots in RyR2+/− was




significantly smaller than that in RyR2+/+ (P < 0.05).Moreover, the averaged peak amplitude of outward currentin RyR2+/− was significantly smaller than that of RyR2+/+

(P < 0.05) (Fig. 2E).Simultaneous recordings of Ca2+ images and

voltage-dependent Ca2+ (VDCC) currents were alsoperformed in the presence of 30 mm tetraethyammonium(TEA) and 1 mm 4-aminopyridine (4-AP) to block K+

currents (Supplemental Fig. 2). Cells were depolarizedfrom −60 to 0 mV for 30 ms in the same manner asFig. 2. Results were roughly the same as those shown inFig. 2. [Ca2+]i rises in hot spots (F/F0: 2.80 ± 0.211,n = 16) and whole cell area (2.08 ± 0.307, n = 5)in RyR2+/− were significantly smaller than those inRyR2+/+, respectively (4.29 ± 0.157, n = 31, P < 0.001;3.31 ± 0.370, n = 6, P < 0.05). The number of Ca2+

hot spots in images at 18.7 ms in RyR2+/− (3.2 ± 0.20,n = 5) was significantly smaller than that in RyR2+/+

(5.0 ± 0.63, n = 6, P < 0.05). The current density of peakVDCC current at 0 mV was comparable between RyR2+/+

(−7.06 ± 0.784, n = 6) and RyR2+/− (−6.65 ± 1.28,n = 5; P > 0.05).

We also measured the density of VDCC currents andBK channel currents in RyR2+/+ and RyR2+/− UBSM cellsunder the conditions where [Ca2+]i was strongly buffered.In these protocols, myocytes were depolarized for 200 msfrom a holding potential of −60 mV to test potentialsin 10 mV steps. To measure VDCC currents, outwardcurrents were blocked by Cs+ in the pipette solution,which also included 5 mm EGTA (see Methods). Figure 3Ashows the relationship between the current density andtest potentials for VDCC. The peak VDCC was measuredat +10 mV. Note that there was no difference in the currentdensity between RyR2+/+ and RyR2+/− at any potentialsexamined. The shape of VDCC measured at +10 mV inRyR2+/− was also comparable to that in RyR2+/+; theratio of VDCC current amplitude at the end versus thatat the peak was 0.642 ± 0.025 (n = 20) in RyR2+/+ and0.618 ± 0.019 (n = 19, P > 0.05) in RyR2+/−.

To measure the density of BK channel current, outwardcurrents were recoded under the conditions, where [Ca2+]i

was fixed at pCa 6.5 by Ca2+-EGTA buffer and the Ca2+

influx was blocked by 0.1 mm Cd2+. Figure 3B showsthe outward K+ currents under these conditions. TheI–V relationships were obtained before and after BKchannel currents were blocked by 1 μm paxilline. Figure 3Bdemonstrates I–V relationships of BK channel currentdensity as the current component sensitive to paxilline.There was no difference in the current density betweenRyR2+/+ and RyR2+/− at any potentials examined.

Contribution of RyR2 to contraction

Contractions induced by direct electrical stimulation weremeasured from UBSM strips of RyR2+/+ and RyR2+/−

in the presence of various neurotransmitter antagonists.Figure 4A shows representative measurement of contra-ctions induced with electrical field stimulation in fourseparate conditions (the inset table). The amplitude ofcontraction depended on the stimulation conditions andcould be summarized as follows: (1) the contractionresulting from a train of 10 pulses (10 ms in duration) andapplied at 200 ms intervals > (2) a train of 10 pulses (3 msin duration) and 200 ms in interval > (3) a single pulse at10 ms in duration > (4) a single pulse at 3 ms in duration.Addition of 100 μm ryanodine transiently increased thetone and reduced the amplitude of contraction in bothRyR2+/+ and RyR2+/−. The increased tone returned tothe baseline level with time, whereas the contractions byelectrical stimulation remained small even after washoutof ryanodine (not shown).

When contraction was induced by single 3 ms pulse,the contraction in RyR2+/− was significantly smaller thanthat in RyR2+/+ (P < 0.01) (Fig. 4B). This contractionwas markedly reduced by ryanodine in tissues fromRyR2+/+, but not changed significantly in RyR2+/−. In thepresence of ryanodine, the contraction amplitude due to asingle 3 ms stimulation was comparable between RyR2+/+

and RyR2+/−. When the contraction was induced by thetrain of 10 pulses (10 ms in duration), the amplitudein RyR2+/− was not significantly different from that ofRyR2+/+ (P > 0.05) (Fig. 4C). It is also notable that theaddition of 100 μm ryanodine did not significantly changethe amplitude in both RyR2+/+ and RyR2+/− under theseconditions.

The sensitivity of these contractions to 100 μm

ryanodine is summarized in Fig. 4D. Note that theamplitude of contraction following a single 3 ms pulsewas significantly decreased by ryanodine in both RyR2+/+

and RyR2+/− but the rate of decrease was much smallerin RyR2+/− than in RyR2+/+ (P < 0.01). In contrast, thecontraction induced by the train of 10 pulses (10 ms induration) was not significantly changed by ryanodineand the results were comparable between RyR2+/+ andRyR2+/− (P > 0.05).

To compare the maximum contractile responsesin RyR2+/+ and RyR2+/−, 10 μm ACh was appliedin each preparation at the beginning of experiment.No significant difference was observed betweenRyR2+/+ (3.1 ± 0.15 mN mg−1, n = 16) and RyR2+/−

(2.8 ± 0.20 mN mg−1, n = 13; P > 0.05).

Sensitivity to caffeine

Ca2+ release by caffeine from SR through RyR wasexamined in UBSMCs from RyR2+/− and RyR2+/+. Ca2+

fluorescence intensity ratios based on fura-2 (F340/F380)were measured in single UBSMCs (Fig. 5A). When caffeinewas added sequentially at concentrations from 0.3 to




Figure 2. Ca2+ hot spots observed following depolarization in UBSM cells from RyR2+/+ and RyR2+/−

Membrane currents and Ca2+ events were simultaneously monitored in single UBSM cells isolated from RyR2+/+and RyR2+/− under voltage clamp. Ca2+ images were obtained using fluo-4 and laser scanning confocal micro-scopy. A, representative Ca2+ images during and following depolarization from −60 to 0 mV for 50 ms in RyR2+/+and RyR2+/− UBSM cells. Voltage clamp depolarization started at 0 ms. The edge of cells is indicated by the whiteline. Arrows indicated hot spots. The Ca2+ hot spots were identified and measured in the images as follows (Ohiet al. 2001). (1) The averaged fluorescence intensity ratio (F/F0) in a cluster of pixels, which includes neighbouring4 pixels or more, was larger than 2.0. (2) The F/F0 in a hot spot was measured as the average at pixels in a circularspot of 1.3 μm in diameter. (3) The area of the hot spot spreads and the F/F0 in it increased or maintained thehigh level over 100 ms. B, changes in F/F0 in the small area of hot spots ‘a’ (red) and ‘b’ (green) were measured




50 mm, the rise of Ca2+ fluorescence intensity due tocaffeine was increased in a concentration-dependentmanner in both RyR2+/− and RyR2+/+. The changes inratio by caffeine in RyR2+/− were not significantly differentfrom those in RyR2+/+ at any concentrations of caffeineexamined (Fig. 5B). In Fig. 5C, these data are normalizedto the maximum in each cell and the caffeine concentrationeliciting half-maximum rise of the ratio was obtained ineach cell as the EC50. The EC50 of caffeine in RyR2+/+

(2.9 ± 0.43 mm, n = 18–22) was significantly lower thanthat in RyR2+/− (5.2 ± 0.54 mm, n = 16–23; P < 0.01)(Fig. 5C).

Spontaneous transient outward currents and restingmembrane potential

Spontaneous transient outward currents (STOCs) weremeasured under whole-cell voltage clamp in UBSMCsfrom RyR2+/+ and RyR2+/− at holding potentials in therange from −60 to −30 mV at 10 mV step. Figure 6Ashows representative recordings at −30 mV. Note thatthe frequency of STOCs with the peak amplitude over20 pA at −30 mV in RyR2+/− was significantly lowerthan that in RyR2+/+ (P < 0.05) (Fig. 6B). In contrast,the averaged amplitude of STOCs was not significantlydifferent between RyR2+/+ and RyR2+/− at any potentialsexamined (Fig. 6C). The amplitude histogram of STOCs(Fig. 6D) indicates that the distribution of STOCs withsmall amplitude between 20 and 30 pA was significantlylarger in RyR2+/− than in RyR2+/+ (P < 0.05). IntegratedSTOCs per 1 s from base line at −30 mV in RyR2+/−

was significantly smaller than that of RyR2+/+ (P < 0.05)(Fig. 6E).

The resting membrane potentials were measuredfrom tissue preparations using conventional glass micro-electrodes in the presence of 1 μm atropine. The restingmembrane potential in RyR2+/− (−44.7 ± 0.74 mV,n = 11) was slightly but significantly more depolarizedthan that in RyR2+/+ (−46.8 ± 0.68 mV, n = 10;P < 0.05) (Supplemental Fig. 3). In the presence of 1 μm

paxilline, there was no difference in the membranepotentials between RyR2+/+ (−41.5 ± 0.90 mV, n = 11)and RyR2+/− (−41.9 ± 0.90 mV, n = 10; P > 0.05). Theextent of depolarization by paxilline in RyR2+/−

(3.3 ± 0.41 mV, n = 11) was significantly smaller than thatin RyR2+/+ (5.0 ± 0.57 mV, n = 10; P < 0.05).

from two Ca2+ hot spots indicated by red and green arrowheads in A, respectively. The data are plotted againsttime. F0 was the average fluorescence intensity before depolarization. F/F0 ratios measured as the average fromwhole cell area (blue) were also plotted. The black dots show membrane currents under whole cell voltage clamp.C, summarized data of [Ca2+]i increases detected as peak F/F0 at Ca2+ hot spots and in whole cell areas. Datawere obtained from experiments shown in A. The numerals in parentheses indicate number of cells examined. D,number of Ca2+ hot spots per cell. The Ca2+ hot spots were measured in one confocal plane obtained 18.7 msafter the start of depolarization in the experiments shown in A. The numerals in parentheses indicate number ofcells examined. E, peak amplitude of outward currents activated by depolarization from −60 to 0 mV for 50 msas shown in A. The numerals in parentheses indicate number of cells examined. ∗P < 0.05, ∗∗P < 0.01 versusRyR2+/+ in C, D and E.

Figure 3. Current density of VDCC currents and BK channelcurrents in UBSM cells from RyR2+/+ and RyR2+/–

A, the voltage dependence of VDCC current density was determinedas I–V relationships. By replacement of K+ in the pipette fillingsolution with Cs+, K+ currents were suppressed, and VDCC currentswere obtained as the 100 μM Cd2+ sensitive current component.Intracellular Ca2+ was buffered with 5 mM EGTA in the pipettesolution (no addition of Ca2+). Cells were depolarized for 150 ms from−60 mV to test potentials by 10 mV step. The average of peak currentdensity in UBSMCs of RyR2+/+ and RyR2+/− UBSM cells was plottedagainst test potentials, to which cells were depolarized from −60 mV(RyR2+/+, �, n = 20; RyR2+/−, •, n = 19). Note that there was nosignificant difference in current density between RyR2+/+ andRyR2+/− at any potentials examined. Inset shows the current traces at+ 10 mV. B, the I–V relationships for BK currents were obtained underfixed [Ca2+]i. The [Ca2+]i in the pipette solution was adjusted at pCa6.5 with 10 mM EGTA and suitable Ca2+. Ca2+ influx was blocked by100 μM Cd2+. Cells were depolarized for 150 ms from −60 to testpotentials by 10 mV step. The inset shows outward currents at −40,−20, 0, +20, +40 and +60 mV. BK channel currents were measuredas 1 μM paxilline sensitive current component. The relationshipsbetween the density of BK current component were plotted againsttest potentials (RyR2+/+, �, n = 4; RyR2+/−, •, n = 4). There was nosignificant difference in current density between RyR2+/+ andRyR2+/− at any potentials examined.




Figure 4. Measurement of contractions and effects of 100 μm ryanodineA, representative traces of contractions recorded from UBSM strips. Inset shows four sets of conditions, (a)–(d), forelectrical stimulation. In the presence of various neurotransmitter antagonists (1 μM atropine, 1 μM phentolamin,1 μM propranolol, 1 μM TTX, 10 μM suramin), contractions were induced by electrical field stimulation in UBSMstrips every 90 s. Changes in contraction magnitude by 100 μM ryanodine were compared between RyR2+/+ andRyR2+/−. B, summarized data based on contractions evoked by single 3 ms pulse corresponding to the stimulationcondition (a) in the inset of A, before (control: left two columns) and after the addition of 100 μM ryanodine(right two columns). Open and filled columns indicate RyR2+/+ and RyR2+/−, respectively. Data were collected from




Figure 5. Difference in caffeine sensitivitybetween RyR2+/+ and RyR2+/– UBSM cellsThe rise of [Ca2+]i elevated by caffeine in UBSMCs wascompared between RyR2+/+ and RyR2+/−. UBSM cellswere loaded by fura-2 AM, and Ca2+ fluorescenceintensity ratio (F340/F380) was measured. A, caffeine wasadded sequentially at concentrations from 0.3 to50 mM. B, summarized data showing the rise of fura-2ratio (�ratio) activated at each concentration ofcaffeine. C, the relationships between concentrations ofcaffeine and the relative �ratio. The data shown in Bwere replotted by taking the maximum �ratio as unityin each cell. The numerals in parentheses indicate thenumber of cell examined.

mRNA expression of molecules regulating Ca2+

mobilization in subcellular microdomain

The mRNA expression levels of molecules which wererelated to Ca2+ mobilization in SMCs were comparedquantitatively in UB from RyR2+/+ and RyR2+/−. Thisanalysis included: junctophilin type 2 (JP2), VDCC(α1c, β2, β3), Ca2+ activated K+ channel (BK, SK2,SK4), SR/ER Ca2+-ATPase (SERCA2), IP3 receptor type1(IP3R1), FK506 binding protein (FKBP1a, 1b) andSorcin.

JP is a protein that is a component of the junctionalcomplex between the plasma membrane and ER/SR(Nishi et al. 2000; Takeshima et al. 2000; Moriguchiet al. 2006). In the present study, JP2 mRNA expressionlevel in UB was much larger than that in heart. JP2expression level in brain was negligible. JP2 mRNAexpression levels in heart and UB were not differentbetween RyR2+/+ and RyR2+/− (P > 0.05) (Fig. 7A).Based on our analyses of mRNA expression of VDCCsubunits, α1c, β2 and β3 subunits in UB, significantdifference between RyR2+/+ and RyR2+/− was onlyobserved in β3 subunit. The difference was relativelysmall (P < 0.05) (Fig. 7B). The mRNA expression ofCa2+ activated K+ channels (BK, SK2, SK4/IK), SERCA2and IP3R mRNA was not significantly different between

experiments typically shown in A. C, summarized data based on contractions evoked by the train of 10 pulses (10 msin duration) corresponding to (d), before (control) and after the addition of 100 μM ryanodine. D, summarizeddata based on the sensitivity of contraction to 100 μM ryanodine. The component of contraction susceptible toryanodine is shown as a percentage of the contraction before ryanodine application. In B–C, the developed forcehas been normalized to the tissue weight in each preparation (mN mg−1). In B–D, the numerals in parenthesesindicate number of preparations examined. ∗P < 0.05, ∗∗P < 0.01 versus RyR2+/+ or control.

RyR2+/+ and RyR2+/− UB (Fig. 7C–E). FKBP1a (FKBP12) mRNA expression level in RyR2+/− UB was smallerthan that in RyR2+/+ (P < 0.05), but the difference wassmall (Fig. 7F). There was no significant difference ofFKBP1b (FKBP12.6) mRNA expression between RyR+/+

and RyR2+/−. Sorcin is a molecular component, whichmay modulates Ca2+ release through RyR2 in cardiacmyocytes and SM as well (Farrell et al. 2003). Theexpression of sorcin mRNA was not significantly differentbetween RyR2+/+ and RyR2+/− (Fig. 7G).

Urination patterns in RyR2+/+ and RyR2+/–

The possibility that RyR2 deficiency may affect urinationactivity was examined in RyR2+/− in comparison withRyR2+/+. Urination patterns of freely moving female micefor 1 h after 1 ml water intake were recorded as urinespots on the filter paper (Fig. 8A). The averaged area ofurine spots from RyR2+/− was significantly smaller thanthat from RyR2+/+. In contrast, the number of spots ofRyR2+/− was significantly larger than that of RyR2+/+. Thetotal area of spots was comparable between them. Theseresults indicate that smaller volume of urine was excretedmore frequently from RyR2+/− than in RyR2+/+, while thetotal urine volume was comparable.




Discussion

Our results demonstrate that RyR2 deficiency in UBSMCsresults in a markedly reduced contribution of CICR to E-Ccoupling in this tissue. In addition, a reduced contributionof BK channel current to the resting membrane potentialwas observed, presumably due to the reduction of Ca2+

spark frequency and related decrease in transient outwardcurrent. In combination, these results provide directevidence for an obligatory role of RyR2 as an essentialmolecular component of CICR in E-C coupling and alsoas one of the key regulators of resting membrane potentialin SMs.

Evaluation of RyR2+/– UBSM is a tool to analyse thephysiological roles of RyR2 in SMs

Because RyR2 homozygous KO mice (RyR2−/−) die atapproximately embryonic day 10 as a consequence of

Figure 6. Spontaneous transient outwardcurrents (STOCs) are reduced in UBSM cellsof RyR2+/−

STOCs in UBSMCs from RyR2+/+ and RyR2+/−were measured at holding potentials from −60to −30 mV at 10 mV step under whole-cellpatch clamp. A, representative recordings ofSTOCs at −30 mV in RyR2+/+ and RyR2+/−UBSMCs. B and C, summarized data offrequency (B) and amplitude of STOCs (C) at−30 mV in UBSMCs from RyR2+/+ ( �) andRyR2+/− (•). D, distribution histogram of STOCevents against the amplitude in each 10 pAbins over 20 pA. Original data were obtained ata holding potential of −30 mV in UBSMCsfrom RyR2+/+ ( �) and RyR2+/− (•). E, chargedisplacement due to STOCs at a holdingpotential of −30 mV. B–E, numerals inparentheses indicate the number of cellsexamined. ∗P < 0.05 versus RyR2+/+.

cardiac arrest (Takeshima et al. 1998), it is difficultto study physiological roles of RyR2 in E-C couplingin SMs in these animals. Another plausible approachinvolves the application of siRNA directed against RyR2to cultured SMC. However, it has been reported that RyR2expression levels in myocytes changes markedly duringprimary culture (Park et al. 1998). In addition, thesecultured SMCs may not be suitable for measurementof contraction, which is essential for evaluating E-Ccoupling.

In the present study, the marked decrease in RyR2mRNA expression and also RyR protein expression in UBfrom RyR2+/− provide strong motivation to attemptingto identify the underlying molecule mechanism. Theresults obtained by evaluating RyR functions by measuringCa2+ hot spots, STOCs, the sensitivity to caffeine andtwitch contraction induced by direct electrical stimulation,all suggest a significant reduction of RyR function. Incontrast, mRNA analyses suggest that the expression




of molecules known to be essential for local Ca2+

regulation, such as L-type VDCC α1C and β2 subunits,BK channel α and β1 subunits, SK2, SK4, SERCA2,IP3R1, FKBP12.6 (FKBP1b), sorcin and junctophilin2(JP2), was not changed in UB of RyR2+/− in comparisonwith RyR2+/+. FKBP12 (FKBP1a) and VDCC β3 subunitswere slightly but significantly reduced. Among four typesof JPs, which are component proteins of the junctionalcomplexes between the plasma membrane and ER/SR,JP2 is expressed preferentially in cardiac muscle, smoothmuscle and brain (Takeshima et al. 2000; Nishi et al. 2000).Specifically our finding that JP2 is abundantly expressed inUBSM suggests the existence of ‘tight coupling’ betweenRyR and VDCC in mouse UBSMC regardless of RyR2+/−

and RyR2+/+. Importantly, neither RyR1 nor RyR3 mRNAexpression levels were changed in RyR2+/− UB. ThePCR primer designed for RyR3 in this study did notdistinguish RyR3 splice variants which have been identifiedrecently (Jiang et al. 2003; Dabertrand et al. 2006). Takentogether, our results suggested that RyR2 is selectively andmarkedly reduced in UB of RyR2+/−. This finding providean important opportunity: UBSMC from RyR2+/− isconsidered to be an excellent tool to analyse the physio-logical roles of RyR2 in SMs.

Figure 7. The mRNA expression of moleculesrelated to the regulation of Ca2+ mobilization insubcellular microdomainsQuantitative analyses of mRNA expression of severalmolecules were performed using real-time PCR. A, themRNA expression of junctophilin type 2 (JP2) (n = 4). B,the mRNA expression of voltage-dependent Ca2+channel (VDCC) α1c, β2 and β3 subunits in UB. n = 4,except for β2 and β3 of RyR2+/+ (n = 6). C, the mRNAexpression of Ca2+ activated K+ channel subtypes (BK,SK2 and SK4) in UB. n = 4, except for BK in RyR2+/+(n = 6). D, the mRNA expression of Ca2+ pump on SRmembrane, SERCA2. n = 4 and 8 for RyR2+/+ andRyR2+/−, respectively. E, the mRNA expression of InsP3

receptor. n = 3 and 4 for RyR2+/+ and RyR2+/−,respectively. F and G, the mRNA expression of FK506binding protein (FKBP1a, 1b) and sorcin in UB,respectively. n = 4 for each. B and F, ∗P < 0.05 versusRyR2+/+.

Significant contribution of RyR2 to CICR and E-Ccoupling in UBSM

There is some controversy about the extent to whichCICR contributes to E-C coupling in electrically activeSMs. Phenomena corresponding to CICR have beendemonstrated in UBSMC of the guinea-pig (Ganitkevich& Isenberg, 1992). Subsequently, however, it has beenreported that CICR does not contribute to the SM E-Ccoupling in portal vein (Kamishima & McCarron, 1996).Another study also concluded that there is only a verylimited contribution of CICR to E-C coupling in UB tissuepreparations of the guinea-pig (Herrera et al. 2000). Atpresent, the function of Ca2+ release through RyR inUBSMCs of the guinea-pig has been suggested to be theactivation of BK channels but not the contractile system.Kotlikoff and colleagues have proposed ‘loose coupling’between VDCC in plasmalemma and RyR in SR for CICRin rabbit UBSMCs (Collier et al. 2000; Kotlikoff, 2003). Wehave recently shown that CICR in UBSMC of the mouseis almost completely blocked by ryanodine and that CICRis essential for E-C coupling triggered by a single actionpotential (Morimura et al. 2006). CICR occurs in twosteps during E-C coupling, which is triggered by an actionpotential; Ca2+ influx during an action potential increases




[Ca2+]i significantly, and this initiates CICR in discretehot spot sites via functional coupling between VDCCs andryanodine receptors (Ohi et al. 2001). In the second step,which also involves Ca2+ release, Ca2+ waves slowly spreadto other Ca2+ store sites in mouse UB; the Ca2+ sourcefor twitch contraction induced by an action potential ismainly attributable to CICR following Ca2+ influx throughVDCC. We acknowledge the possibility that there might bespecies difference between mouse and guinea-pig/rabbit.The contribution of CICR versus that of Ca2+ influx as theCa2+ source for contraction in E-C coupling may be largerin mouse UB.

In the present study, both the number of Ca2+ hotspots and the increase in [Ca2+]i in the spots in theearly stage of depolarization (< 20 ms) were significantlysmaller in UBSMCs of RyR2+/− than in those of RyR2+/+.

Figure 8. Urinary bladder activity of RyR2+/+ and RyR2+/–

A, representative urination pattern of freely moving female micevisualized on filter papers under UV light from RyR2+/+ and RyR2+/−.B, average area of a urine spot quantified from filter paper. C, averagenumber of urine spots quantified from filter paper. D, average area oftotal urine spots quantified from filter paper. B–D, numerals inparentheses indicate the number of animals. ∗P < 0.05, ∗∗P < 0.01versus RyR2+/+.

Correspondingly, the activation of BK channel currentunder these conditions was also smaller in RyR2+/− thanRyR2+/+. Although the deletion of RyR1 results in thedecrease in VDCC activity in skeletal muscle (Fleig et al.1996), the density of VDCC was not changed in RyR2+/−

in the present study. The density of BK channel current wasalso not changed in RyR2+/− UBSMCs, in which [Ca2+]i

was fixed at pCa 6.5 (not shown). Therefore, the smalleractivation of BK channel current upon depolarization inUBSC from RyR2+/− than that from RyR+/+ is presumablydue to the decreased function of Ca2+ hot spots in RyR2+/−

UBSMCs.The twitch contractions induced by single 3 ms

stimulation in UB tissue segments of RyR2+/− weresignificantly smaller than those of RyR2+/+. The decreasein amplitude of these twitch contractions by 100 μm

ryanodine was over 60% in RyR2+/+ but only 37%in RyR2+/−, indicating that the contribution of CICRvia RyR2 to the contraction is significantly smaller inRyR2+/−. The results for twitch contractions induced bytrain stimulation (10 pulse 10 ms in duration) were notsignificantly different between RyR2+/− and RyR2+/+. Inaddition, this pattern of response was not affected byryanodine, indicating that Ca2+ influx through VDCC isthe predominant sources of Ca2+ for contraction underthe severe stimulation conditions. Taken together, thesefindings strongly suggest that RyR2 is the essential andcentral molecule responsible for CICR elicited by a singleaction potential in mouse UBSMCs.

Regulation of resting membrane potential by RyR2

STOCs were first identified in 1986 in intestinal SMs(Benham & Bolton, 1986). They have been shown tobe due to a burst of openings of BK channels followingspontaneous local Ca2+ release (Bolton & Imaizumi,1996), which has been identified as Ca2+ sparks (Nelsonet al. 1995; Laporte et al. 2004). The relationship betweenCa2+ sparks and the regulation of resting membranepotential by BK channel activity has been well studiedin arterial SMs (Knot et al. 1998; Imaizumi et al.1999; Jaggar et al. 2000) and also in UBSMCs of theguinea-pig (Ohi et al. 2001). A well defined functionalorganization between three molecular components, RyR,BK channels and VDCC, in subcellular microdomains hasbeen suggested; a spontaneous Ca2+ spark in the superficialarea activates 10–100 BK channels and induces membranehyperpolarization, which reduces Ca2+ channel activityand causes relaxation (Perez et al. 1999; ZhuGe et al.1999; Furstenau et al. 2000). Several lines of evidencehave supported this hypothesis. The decreased activityof BK channels following BKβ1 subunit deletion causesarterial hypertension due to a slightly depolarized restingmembrane potential and results in increased activity of




VDCC in arterial SMCs (Pluger et al. 2000). Bladderinstability reported in BKα subunit KO mice may bedue to this mechanism (Meredith et al. 2004). In thisstudy, it is shown directly that the deficiency of RyR2resulted in the decrease in STOC frequency and integratedSTOCs in RyR2+/− UBSMCs in comparison with those inRyR2+/+. Correspondingly, we found more depolarizedresting membrane potential and smaller depolarizationby paxiline in UBSM of RyR2+/− than those in RyR2+/−.These results provide further support for the hypothesisthat RyR2 is responsible also for the regulation ofresting BK channel activity in UBSMCs. Based on thiscombination of findings, we conclude that spontaneousCa2+ release through RyR2 in junctional SR as Ca2+ sparksactivates BK channels in the junction and regulates restingmembrane potential.

Contribution of RyR2 versus RyR3 in UBSMCs

In canine cardiac Purkinje fibre myocytes, in whichthe T-tubule system is not well developed, substantialfunctional roles of RyR3, as well as RyR2, have beenreported in the generation of Ca2+ sparks and Ca2+ wave-lets (Stuyvers et al. 2005). In contrast, detailed informationabout the functional significance of RyR3 in SMs is limited.The functional roles of RyR2 versus RyR3, which is alsoexpressed in UBSM, for Ca2+ spark generation has beenshown indirectly using RyR3 homozygous KO mice andalso FKBP12.6 homozygous KO mice (Ji et al. 2005). SinceFKBP12.6 selectively interacts with RyR2 and reduces itsactivity (Marx et al. 2001), the increase in Ca2+ sparkfrequency and Ca2+ wave speed in FKBP12.6−/− UB isthought to be due to the enhanced activity of RyR2 byFKBP12.6 deficiency (Ji et al. 2005). On the other hand,Ca2+ sparks and Ca2+ waves in UB of RyR3−/− were verysimilar to those in RyR3+/+ UB. Therefore, it has beensuggested that RyR3 may not play a significant role inthe generation of Ca2+ sparks in UBSM. In contrast, anincreased frequency of Ca2+ sparks has been reported incerebral artery SMC of RyR3 null mice (Lohn et al. 2001).Moreover, novel alternative splice variants of RyR3 havebeen isolated from SMs (Chen et al. 1997; Jiang et al. 2003;Dabertrand et al. 2006). It has been suggested recentlythat an alternative splice valiant of RyR3, which worksas dominant negative, is predominantly expressed in SMtissues, particularly in uterus (Dabertrand et al. 2006).More importantly, RyR3 including the dominant negativetype can form a heterotetramer with RyR2 (Jiang et al.2003). It is somewhat puzzling that the dominant negativetype variant of RyR3 appears to be extensively expressedeven in mouse UB (Dabertrand et al. 2006). This findingis not consistent with the lack of changes in Ca2+ sparksin RyR3−/− UBSMCs (Ji et al. 2003). If the dominantnegative splice variant of RyR3 is significantly expressed in

mouse UBSMCs as reported, then the functional changesin RyR2+/− UBSMCs reported here may possibly be dueto both the deficiency of RyR2 per se and the increasedratio of dominant negative RyR3 versus RyR2. In thepresent study, however, we show that the maximum riseof caffeine-induced [Ca2+]i in UBSMCs of RyR2+/− iscomparable to that in RyR2+/+, while the sensitivity tocaffeine was slightly but significantly reduced in RyR2+/−.This set of results is not completely consistent with theassumption that the dominant negative splice variantof RyR3 functions in a dominant negative fashion inUBSMCs of RyR2+/−, since this variant is insensitive tocaffeine (Jiang et al. 2003; Dabertrand et al. 2006).

Potential roles of RyR2 in the regulation of urinarybladder activity

It is notable that the frequency of urination and the volumeper voiding were significantly increased and reduced,respectively, in RyR2+/− in comparison with those inRyR2+/+. Interestingly, it has been reported that a modelof detrusor instability following outlet obstruction resultsin down-regulation of RyR expression (Jiang et al. 2005).A line of evidence supporting the functional significanceof BK channels in the regulation of urination activityhas been greatly accumulated (Christ & Hodges, 2006).It has been shown that mice genetically lacking the BKchannel α subunit demonstrate a marked increase inurination frequency corresponding to an overactivebladder (Meredith et al. 2004). The deficiency of negativefeedback mechanism for the control of [Ca2+]i, in whichthe activation of BK channels plays the central role, isconsidered to be responsible for the enhanced contra-ctility of bladder smooth muscle in BK channel KO mice.In this respect, both the decreased STOC frequency andthe depolarized resulting membrane in UBSMCs fromRyR2+/− fit with the theory of a central role of BK channelsin the regulatory mechanism of bladder function. If it is thecase, RyR2 deficiency may result in an overactive bladder.In contrast, the contraction induced by direct electricalstimulation was rather reduced in UB of RyR2+/−, whenthe stimulation conditions were moderate. Therefore, itcannot be concluded whether the changes in urinationpattern in RyR2+/− is due to over-contractility mediatedby the reduced activity of BK channels or is simply dueto a smaller contribution of CICR to the contraction forvoiding. Alternatively, it cannot be ruled out that RyR2deficiency may result in lower nervous activity to triggerthe voiding.

Conclusion

Our findings demonstrate that a down-regulation ofRyR2 in UBSMCs can reduce both the contribution




of spontaneous Ca2+ release through RyR2 to restingmembrane potential and the essential signalling involvedin CICR due to depolarization. Thus, RyR2 plays anessential role in CICR for the regulation of E-C couplinginduced by an action potential and also in the regulationof resting membrane potential, presumably via themodulation of Ca2+ dependent K+ channel activity inUBSM. It became clear in this study that RyR2 aswell as the BK channel (Christ & Hodges, 2006) has apivotal role in the control of urinary bladder activity. Thebroader physiological questions related to the influence ofRyR2 deficiency in bladder function in RyR2+/− and thephysiological roles of RyR3 in UBSM remain to be definedand determined.

References

Benham CD & Bolton TB (1986). Spontaneous transientoutward currents in single visceral and vascular smoothmuscle cells of the rabbit. J Physiol 381, 385–406.

Bolton TB & Imaizumi Y (1996). Spontaneous transientoutward currents in smooth muscle cells. Cell Calcium 20,141–152.

Chambers P, Neal DE & Gillespie JI (1999). Ryanodinereceptors in human bladder smooth muscle. Exp Physiol 84,41–46.

Chen SR, Li X, Ebisawa K & Zhang L (1997). Functionalcharacterization of the recombinant type 3 Ca2+ releasechannel (ryanodine receptor) expressed in HEK293 cells.J Biol Chem 272, 24234–24246.

Christ GJ & Hodges S (2006). Molecular mechanisms ofdetrusor and corporal myocyte contraction: identifyingtargets of pharmacotherapy of bladder and erectiledysfunction. Br J Pharmacol 147, S41–S55.

Collier ML, Ji G, Wang YX & Kotlikoff MI (2000).Calcium-induced calcium release in smooth muscle: loosecoupling between the action potential and calcium release.J Gen Physiol 115, 653–662.

Dabertrand F, Morel JL, Sorrentino V, Mironneau J,Mironneau C & Macrez N (2006). Modulation of calciumsignalling by dominant negative splice variant of ryanodinereceptor subtype 3 in native smooth muscle cells. CellCalcium 40, 11–21.

Fabiato A (1983). Calcium-induced release of calcium from thecardiac sarcoplasmic reticulum. Am J Physiol Cell Physiol245, C1–C14.

Fabiato A (1985). Simulated calcium current can both causecalcium loading in and trigger calcium release from thesarcoplasmic reticulum of a skinned canine cardiac Purkinjecell. J Gen Physiol 85, 291–320.

Farrell EF, Antaramian A, Rueda A, Gomez AM & Valdivia HH(2003). Sorcin inhibits calcium release and modulatesexcitation-contraction coupling in the heart. J Biol Chem278, 34660–34666.

Fleig A, Takeshima H & Penner R (1996). Absence of Ca2+current facilitation in skeletal muscle of transgenic micelacking the type 1 ryanodine receptor. J Physiol 496,339–345.

Furstenau M, Lohn M, Ried C, Luft FC, Haller H & Gollasch M(2000). Calcium sparks in human coronary artery smoothmuscle cells resolved by confocal imaging. J Hypertens 18,1215–1222.

Ganitkevich VY & Isenberg G (1992). Contribution of Ca2+-induced Ca2+ release to the [Ca2+]i transients in myocytesfrom guinea-pig urinary bladder. J Physiol 458, 119–137.

Giannini G, Clementi E, Ceci R, Marziali G & Sorrentino V(1992). Expression of a ryanodine receptor-Ca2+ channelthat is regulated by TGF-β. Science 257, 91–94.

Hakamata Y, Nakai J, Takeshima H & Imoto K (1992). Primarystructure and distribution of a novel ryanodine receptor/calcium release channel from rabbit brain. FEBS Lett 312,229–235.

Hashitani H & Brading AF (2003). Ionic basis for the regulationof spontaneous excitation in detrusor smooth muscle cells ofthe guinea-pig urinary bladder. Br J Pharmacol 140, 159–169.

Hashitani H, Bramich NJ & Hirst GD (2000). Mechanisms ofexcitatory neuromuscular transmission in the guinea-pigurinary bladder. J Physiol 524, 565–579.

Herrera GM, Heppner TJ & Nelson MT (2000). Regulation ofurinary bladder smooth muscle contractions by ryanodinereceptors and BK and SK channels. Am J Physiol Regul IntegrComp Physiol 279, R60–R68.

Iino M (1989). Calcium-induced calcium release mechanism inguinea pig taenia caeci. J Gen Physiol 94, 363–383.

Imaizumi Y, Henmi S, Uyama Y, Atsuki K, Torii Y, Ohizumi Y& Watanabe M (1996). Characteristics of Ca2+ release foractivation of K+ current and contractile system in somesmooth muscles. Am J Physiol Cell Physiol 271, C772–C782.

Imaizumi Y, Muraki K & Watanabe M (1989). Ionic currents insingle smooth muscle cells from the ureter of the guinea-pig.J Physiol 411, 131–159.

Imaizumi Y, Ohi Y, Yamamura H, Ohya S, Muraki K &Watanabe M (1999). Ca2+ spark as a regulator of ion channelactivity. Jpn J Pharmacol 80, 1–8.

Imaizumi Y, Torii Y, Ohi Y, Nagano N, Atsuki K, Yamamura H,Muraki K, Watanabe M & Bolton TB (1998). Ca2+ imagesand K+ current during depolarization in smooth musclecells of the guinea-pig vas deferens and urinary bladder.J Physiol 510, 705–719.

Jaggar JH, Porter VA, Lederer WJ & Nelson MT (2000).Calcium sparks in smooth muscle. Am J Physiol Cell Physiol278, C235–C256.

Ji G, Feldman ME, Greene KS, Sorrentino V, Xin HB &Kotlikoff MI (2004). RYR2 proteins contribute to theformation of Ca2+ sparks in smooth muscle. J Gen Physiol123, 377–386.

Jiang HH, Song B, Lu GS, Wen QJ & Jin XY (2005). Loss ofryanodine receptor calcium-release channel expressionassociated with overactive urinary bladder smooth musclecontractions in a detrusor instability model. BJU Int 96,428–433.

Jiang D, Xiao B, Li X & Chen SR (2003). Smooth muscle tissuesexpress a major dominant negative splice variant of the type3 Ca2+ release channel (ryanodine receptor). J Biol Chem278, 4763–4769.

Kamishima T & McCarron JG (1996). Depolarization-evokedincreases in cytosolic calcium concentration in isolatedsmooth muscle cells of rat portal vein. J Physiol 492, 61–74.

C© 2007 The Authors. Journal compilation C© 2007 The Physiological Societyat



Knot HJ, Standen NB & Nelson MT (1998). Ryanodinereceptors regulate arterial diameter and wall [Ca2+] incerebralarteries of rat via Ca2+-dependent K+ channels.J Physiol 508, 211–221.

Kotlikoff MI (2003). Calcium-induced calcium release insmooth muscle: the case for loose coupling. Prog Biophys MolBiol 83, 171–191.

Kupittayanant S, Luckas MJ & Wray S (2002). Effect ofinhibiting the sarcoplasmic reticulum on spontaneous andoxytocin-induced contractions of human myometrium.Br J Obst Gynaecol 109, 289–296.

Laporte R, Hui A & Laher I (2004). Pharmacologicalmodulation of sarcoplasmic reticulum function in smoothmuscle. Pharmacol Rev 56, 439–513.

Lohn M, Jessner W, Furstenau M, Wellner M, Sorrentino V,Haller H, Luft FC & Gollasch M (2001). Regulation ofcalcium sparks and spontaneous transient outward currentsby RyR3 in arterial vascular smooth muscle cells. Circ Res 89,1051–1057.

Marx SO, Gaburjakova J, Gaburjakova M, Henrikson C,Ondrias K & Marks AR (2001). Coupled gating betweencardiac calcium release channels (ryanodine receptors). CircRes 88, 1151–1158.

Meissner G (1994). Ryanodine receptor/Ca2+ release channelsand their regulation by endogenous effectors. Annu RevPhysiol 56, 485–508.

Meredith AL, Thorneloe KS, Werner ME, Nelson MT & AldrichRW (2004). Overactive bladder and incontinence in theabsence of the BK large conductance Ca2+-activated K+channel. J Biol Chem 279, 36746–36752.

Moriguchi S, Nishi M, Komazaki S, Sakagami H, Miyazaki T,Masumiya H, Saito SY, Watanabe M, Kondo H, Yawo H,Fukunaga K & Takeshima H (2006). Functional uncouplingbetween Ca2+ release and afterhyperpolarization inmutanthippocampal neurons lacking junctophilins. ProcNatl Acad Sci U S A 103, 10811–10816.

Morimura K, Ohi Y, Yamamura H, Ohya S, Muraki K &Imaizumi Y (2006). A two step Ca2+ intracellular releaseunderlies excitation-contraction coupling in mouse urinarybladder myocytes. Am J Physiol Cell Physiol 290, C388–C403.

Nakai J, Imagawa T, Hakamata Y, Shigekawa M, Takeshima H& Numa S (1990). Primary structure and functionalexpression from cDNA of the cardiac ryanodinereceptor/calcium release channel. FEBS Lett 271, 169–177.

Nelson MT, Cheng H, Rubart M, Santana LF, Bonev AD, KnotHJ & Lederer WJ (1995). Relaxation of arterial smoothmuscle by calcium sparks. Science 270, 633–637.

Nishi M, Mizushima A, Nakagawara K & Takeshima H (2000).Characterization of human junctophilin subtype genes.Biochem Biophys Res Commun 273, 920–927.

Ohi Y, Yamamura H, Nagano N, Ohya S, Muraki K, WatanabeM & Imaizumi Y (2001). Local Ca2+ transients anddistribution of BK channels and ryanodine receptors insmooth muscle cells of guinea-pig vas deferens and urinarybladder. J Physiol 534, 313–326.

Ohya S, Tanaka M, Oku T, Asai Y, Watanabe M, Giles WR &Imaizumi Y (1997). Molecular cloning and tissuedistribution of an alternatively spliced variant of an A-typeK+ channel α-subunit, Kv4.3 in the rat. FEBS Lett 420,47–53.

Otsu K, Willard HF, Khanna VK, Zorzato F, Green NM &MacLennan DH (1990). Molecular cloning of cDNAencoding the Ca2+ release channel (ryanodine receptor) ofrabbit cardiac muscle sarcoplasmic reticulum. J Biol Chem265, 13472–13483.

Park MY, Park WJ & Kim DH (1998). Expression ofexcitation-contraction coupling proteins during muscledifferentiation. Mol Cell 8, 513–517.

Perez GJ, Bonev AD, Patlak JB & Nelson MT (1999).Functional coupling of ryanodine receptors to KCa channelsin smooth muscle cells from rat cerebral arteries. J GenPhysiol 113, 229–238.

Pluger S, Faulhaber J, Furstenau M, Lohn M, Waldschutz R,Gollasch M, Haller H, Luft FC, Ehmke H & Pongs O (2000).Mice with disrupted BK channel β1 subunit gene featureabnormal Ca2+ spark/STOC coupling and elevated bloodpressure. Circ Res 87, e53–e60.

Sanders KM (2001). Mechanisms of calcium handling insmooth muscles. J Appl Physiol 91, 1438–1449.

Stuyvers BD, Dun W, Matkovich S, Sorrentino V, Boyden PA &ter Keurs HE (2005). Ca2+ sparks and waves in caninepurkinje cells: a triple layered system of Ca2+ activation. CircRes 97, 35–43.

Taggart MJ & Wray S (1998). Contribution of sarcoplasmicreticular calcium to smooth muscle contractile activation:gestational dependence in isolated rat uterus. J Physiol 511,133–144.

Takeshima H, Komazaki S, Hirose K, Nishi M, Noda T & IinoM (1998). Embryonic lethality and abnormal cardiacmyocytes in mice lacking ryanodine receptor type 2. EMBO J17, 3309–3316.

Takeshima H, Komazaki S, Nishi M, Iino M & Kangawa K(2000). Junctophilins: a novel family of junctionalmembrane complex proteins. Mol Cell 6, 11–22.

Takeshima H, Nishimura S, Matsumoto T, Ishida H, KangawaK, Minamino N, Matsuo H, Ueda M, Hanaoka M, Hirose T& Numa S (1989). Primary structure and expression fromcomplementary DNA of skeletal muscle ryanodine receptor.Nature 339, 439–445.

Wang YX, Zheng YM, Mei QB, Wang QS, Collier ML, FleischerS, Xin HB & Kotlikoff MI (2004). FKBP12.6 and cADPRregulation of Ca2+ release in smooth muscle cells.Am J Physiol Cell Physiol 286, C538–C546.

ZhuGe R, Tuft RA, Fogarty KE, Bellve K, Fay FS & Walsh JV Jr(1999). The influence of sarcoplasmic reticulum Ca2+concentration on Ca2+ sparks and spontaneous transientoutward currents in single smooth muscle cells. J Gen Physiol113, 215–228.

Acknowledgements

This work was supported by a Grant-in-Aid for Scientific

Research on Priority Areas (18059029) from The Ministry of

Education, Culture, Sports, Science and Technology and by a

Grant-in-Aid for Scientific Research (B) (17390045) from the

Japan Society for the Promotion of Science to Y.I. This works

was also supported by a Grant-in-Aid for Research on Health

Sciences focusing on Drug Innovation from the Japan Health




Sciences Foundation to Y.I. We thank Dr W. R. Giles (University

of Calgary, Calgary, Canada) for providing data acquisition

and analysis programs and also for his critical reading of this

manuscript.

Supplemental material

Online supplemental material for this paper can be accessed at:

http://jp.physoc.org/cgi/content/full/jphysiol.2007.130302/DC1

and

http://www.blackwell-synergy.com/doi/suppl/10.1113/jphysiol.

2007.130302



J Physiol 582.2 (2007) pp 507–524 507

Dynamic and differential regulation of NKCC1 by calciumand cAMP in the native human colonic epithelium

Amy Reynolds1, Alyson Parris1, Luke A. Evans1,2, Susanne Lindqvist1, Paul Sharp1, Michael Lewis2,

Richard Tighe3 and Mark R. Williams1

1School of Biological Sciences, University of East Anglia, Norwich, Norfolk NR4 7TJ, UK2Department of General Surgery, Norfolk and Norwich University Hospital, Norwich, Norfolk NR4 7UY, UK3Department of Gastroenterology, Norfolk and Norwich University Hospital, Norwich, Norfolk NR4 7UY, UK

The capacity of the intestine to secrete fluid is dependent on the basolateral Na+–K+–2Cl−

co-transporter (NKCC1). Given that cAMP and Ca2+ signals promote sustained and transient

episodes of fluid secretion, respectively, this study investigated the differential regulation

of functional NKCC1 membrane expression in the native human colonic epithelium.

Tissue sections and colonic crypts were obtained from sigmoid rectal biopsy tissue

samples. Cellular location of NKCC1, Na+–K+-ATPase, M3 muscarinic acetylcholine receptor

(M3AChR) and lysosomes was examined by immunolabelling techniques. NKCC1 activity

(i.e. bumetanide-sensitive NH4+ uptake), intracellular Ca2+ and cell volume were assessed

by 2′,7′-bis(2-carboxyethyl)-5-(and-6)-carboxyfluorescein (BCECF), Fura-2 and differential

interference contrast/calcein imaging. Unstimulated NKCC1 was expressed on basolateral

membranes and exhibited a topological expression gradient, predominant at the crypt base.

Cholinergic Ca2+ signals initiated at the crypt base and spread along the crypt axis. In response,

NKCC1 underwent a Ca2+-dependent 4 h cycle of recruitment to basolateral membranes,

activation, internalization, degradation and re-expression. Internalization was prevented by the

epidermal growth factor receptor kinase inhibitor tyrphostin-AG1478, and re-expression was

prohibited by the protein synthesis inhibitor cylcoheximide; the lysosome inhibitor chloroquine

promoted accumulation of NKCC1 vesicles. NKCC1 internalization and re-expression were

accompanied by secretory volume decrease and bumetanide-sensitive regulatory volume

increase, respectively. In contrast, forskolin (i.e. cAMP elevation)-stimulated NKCC1 activity was

sustained, and membrane expression and cell volume remained constant. Co-stimulation with

forskolin and acetylcholine promoted dramatic recruitment of NKCC1 to basolateral membranes

and prolonged the cycle of co-transporter activation, internalization and re-expression. In

conclusion, persistent NKCC1 activation by cAMP is constrained by a Ca2+-dependent cycle

of co-transporter internalization, degradation and re-expression; this is a novel mechanism to

limit intestinal fluid loss.

(Received 6 February 2007; accepted after revision 26 April 2007; first published online 5 May 2007)

Corresponding author M. Williams: School of Biological Sciences, University of East Anglia, Norwich, Norfolk NR4

7TJ, UK. Email: [email protected]

The regulation of transepithelial fluid transport is ofutmost interest because excessive fluid secretion isassociated with numerous intestinal diseases, includingenteric infections and inflammatory bowel disease. Undernormal circumstances, constitutive fluid absorptionpredominates over fluid secretion and serves to limitfluid loss from the body. Stimulated fluid secretion isrequired to flush the crypt lumen of noxious substancesand coat the surface epithelium with hydrated mucous(Barrett & Keely, 2000; Matthews, 2002; Geibel, 2005).Physiological regulation stems from neurohormonal and

neuroimmune pathways which exert tight control overfluid secretion via modulation of epithelial cell Ca2+ andcyclic nucleotide levels. In diarrhoea, these regulatorypathways are hijacked by pathophysiological stimulisuch as bacterial enterotoxins (e.g. CTx, STa), viruses(e.g. rotavirus), bile acids or inflammatory mediators.As a result, fluid secretion is stimulated, while fluidabsorption and barrier function are compromised (Field,2003; Turner, 2006). A molecular understanding of theseprocesses is beginning to yield antidiarrhoea strategies(e.g. Clayburgh et al. 2006; Geibel et al. 2006; Sonawane



508 A. Reynolds and others J Physiol 582.2

et al. 2006). In the case of intestinal fluid secretion,active transcellular Cl− transport provides the osmoticimpetus for passive fluid flow across the polarizedepithelium; Cl− efflux across the apical membrane ismediated by the cystic fibrosis transmembrane regulator(CFTR) Cl− channel, and basolateral Cl− uptake ismediated by a Na+–K+–2Cl− co-transporter, NKCC1.Classically, CFTR has been considered the primary site ofregulation. However, basolateral transport pathways canindependently regulate transcellular Cl− transport andthus the capacity for fluid secretion/diarrhoea. In fact,NKCC1 is emerging as a central integrator of cellularsignals that determine the secretory status of the intestinalepithelium (Matthews, 2002). Differential regulationof NKCC1 by Ca2+ and cAMP (adenosine 3′,5′-cyclicmonophosphate) is of particular interest given thestriking differences in fluid secretion promoted by theseintracellular messengers.

cAMP-dependent secretagogues (e.g. vasoactiveintestinal peptide or forskolin) generate a gradual,sustained secretory response compared with theremarkably fast and transient fluid secretion elicitedby Ca2+-dependent secretagogues (e.g. acetylcholine)(Dharmsathaphorn & Pandol, 1986; Vajanaphanich et al.1995; Mall et al. 1998). Barrett and colleagues havedemonstrated that Ca2+-mediated transactivation ofEGFR/MAP (epidermal growth factor receptor/mitogenactivated protein) kinase pathways in T84 cells play acentral role in the negative regulation of Ca2+-mediatedCl− secretion (Keely et al. 1998, 2000; McCole et al. 2002;Keely & Barrett, 2003). Other inhibitory signals that mayact downstream of, or in parallel to, EGFR include PKC(protein kinase C) activation (Kachintorn et al. 1992;Matthews et al. 1993). Significantly, the ‘braking’ influenceof such anti-secretory signals (see Keely et al. 1998) canpersist beyond the transient period of Ca2+-mediated Cl−

secretion. For example, direct activation of PKC uncouplesthe secretory machinery of the cell from activation bysubsequent elevations in intracellular Ca2+ (Kachintornet al. 1992) or cAMP (Matthews et al. 1993). Intriguingly,simultaneous elevation of cell Ca2+ and cAMP elicitsa synergistic secretory response (Dharmsathaphorn &Pandol, 1986; Vajanaphanich et al. 1995; Mall et al. 1998).These distinctive secretory responses are determined bythe complex interplay between the second messengersand their respective effects on the secretory machinery ofthe cell. In this regard, numerous questions are raised asto how these signalling outputs differentially regulate theactivity of key membrane transporters, such as NKCC1,at the cellular and tissue level.

NKCC1 is a bumetanide-sensitive member of a family ofcation–Cl− co-transporters (reviewed in Haas & Forbush,2000). Bumetanide inhibits stimulated Cl− secretion(Dharmsathaphorn & Pandol, 1986; Mall et al. 2000) andNKCC1 knockout mice exhibit reduced levels of intestinal

fluid secretion (Flagella et al. 1999). Co-transporteractivity is regulated by changes in cell volume andintracellular Cl− which modulate its phosphorylationstatus via a recently characterized macromolecularcomplex (Gimenez, 2006). New direct evidence suggeststhat secretagogues also regulate NKCC1 activity viamodulation of basolateral membrane expression. PKCactivation in T84 cells stimulated rapid NKCC1 inter-nalization from the basolateral membrane (Farokhzadet al. 1999; Del Castillo et al. 2005). Given the distinctivesecretory responses invoked by Ca2+ and/or cAMP, itis of prime interest to study the regulation of NKCC1internalization by these second messengers. Moreover,to gain insight into the clinical significance of thisphenomenon, it is vital that regulation of NKCC1internalization is examined in the native human colonicepithelium.

Investigation of transporter expression, trafficking andactivity in the native intestinal epithelium has beenhampered by the short-term viability of epithelial tissue exvivo. In the present study we have utilized a 3D crypt modelof the native human colonic epithelium that maintainsits tissue morphology and cellular polarity in culture.Fluorescence imaging techniques were used to studythe consequences of Ca2+ (Lindqvist et al. 1998, 2002)and/or cAMP signals to NKCC1 membrane expression andactivity at the cellular and tissue level. Our observationsdemonstrate that cholinergic Ca2+ signals promote anovel cycle of co-transporter activation, internalization,degradation and re-expression. Furthermore, Ca2+ andcAMP signals, alone or in combination, differentiallyregulate NKCC1 membrane expression levels to influenceco-transporter activity.

Methods

Tissue samples and human colonic crypt isolation

This study was performed in accordance with theNorwich District Ethics Committee, Norfolk and NorwichUniversity Hospital, UK. Rectosigmoid biospy specimenswere obtained from patients undergoing exploratorycolonoscopy and subsequently found to have no apparentpathology of the large intestine (age range 25–93 years).Human colonic crypts were isolated in a similar fashionto that previously described (Lindqvist et al. 1998, 2002).Briefly, the biopsies were washed in Hepes-bufferedsaline (HBS): (mm) NaCl 140, KCl 5, Hepes(N-2-hydroxyethylpiperazine-N ′-2-ethanesulphonicacid) 10, d-glucose 5.5, Na2HPO4 1, MgCl2 0.5,CaCl2 1, and placed in HBS, which was devoid ofboth Ca2+ and Mg2+, and supplemented with EDTA(diaminoethanetetraacetic acid disodium salt) (1 mm),for 1 h at room temperature. Crypts were liberated byvigorous shaking and were fixed to non-fluorescent



J Physiol 582.2 NKCC1 trafficking in human colonic crypts 509

glass coverslips coated with type 1 collagen (Sigma, UK).Crypts were then placed in standard serum-free tissueculture conditions for 0–48 h prior to experimentation:Dulbecco’s modifed Eagles’s medium, streptomycin(5 μg ml−1) and penicillin (5000 units ml−1), in ahumidified 95% air, 5% CO2 incubator at 37◦C. All tissueculture media and supplements were from Invitrogen,UK.

Fluorescence imaging of intracellular Ca2+ and pH

Isolated colonic crypts were loaded with eitherFura-2/AM (3 μm, 45 min) or BCECF/AM(2′,7′-bis (carboxyethyl)-5,6-carboxyfluoresceinacetoxymethylester) (5 μm, 45 min) to monitor Ca2+

i

or pHi as previously described (Williams et al. 1992;Lindqvist et al. 1998, 2002). The loaded specimen wasplaced in an experimental chamber located on the stageof an inverted epifuorescence microscope (×40 or ×20neofuor Nikon objective). Crypts were continuouslyperifused with HBS (37◦C). The exchange half-time of thebathing solution within the perfusion chamber (volume200 μl) was 2 s. Experimental solutions were administeredvia a two-way tap. Fluorescence ratio data are presentedin the form of pseudocolour images or as traces, wherethe average ratio values within regions of interest (ROIs)located along the crypt axis are plotted with respect totime.

NKCC1 activity

NKCC1 activation levels were assessed by the rate ofplateau phase acidification associated with influx ofNH4

+, a substrate for the co-transporter (Heitzmannet al. 2000; Bachmann et al. 2003). BCECF-loaded cryptswere pulsed with NH4Cl (40 mm, 3 min) in the presenceor absence of secretagogues. An initial intracellularalkalization (i.e. influx of weak base) was followedby a secretagogue-stimulated acidification (i.e. entry ofNH4

+), the rate of which was fitted by linear regression.Removal of NH4Cl from the bathing medium produceda classical ‘intracellular acid load’. Where used, inhibitorswere pre-incubated for 30 min.

Cell volume and luminal flow experiments

To monitor crypt cell dimensions, calcein fluorescence(10 μm, 30 min; Calbiochem, Merck Bioscience, UK)and differential interference contrast (DIC) images wereacquired using the argon laser line of a Zeiss 510 Metaconfocal fluorescence microscope. Crypt dimensions wereanalysed using Axiovision 4.1 software (Carl Zeiss Ltd).In addition to cells lining the crypt axis, a calcein-positivecell was occasionally observed within the crypt lumen. To

demonstrate secretagogue-induced fluid flow in the cryptlumen, the position of the luminal cell was tracked withrespect to time following stimulation with acetylcholine.

Immunocytochemistry and immunohistochemistry

The immunolabelling procedure has been describedelsewhere (Lindqvist et al. 2002). Briefly, cryosectionsor isolated crypts fixed in 4% paraformaldehyde weresubjected to antigen retrieval: a 5 min incubation with1% sodium dodecyl sulphate (SDS) (BDH). Sectionswere permeabilized with either Triton X-100 (0.5% w/vPBS, 30 min) or methanol (−20◦C, 10 min). Non-specificbinding sites were blocked with 10% goat serum and 1%bovine serum albumin, for 2 h, and washed with PBS. Theprimary antibodies used in this study were (all at 1:200dilution): rabbit polyclonal anti-M3 muscarinic receptorantibody (Research & Diagnostic Antibodies, Benicia,CA, USA); mouse monoclonal anti-Na–K-ATPase α-1,clone C464.6 antibody (Upstate, NY, USA); mouse mono-clonal anti-LAMP-1 (lysosome-associated membraneprotein-1 antibody (Developmental Studies HybridomaBank, University of Iowa, Baltimore, USA) to labellysosomes. Three different primary antibodies (a generousgift of Professor Chris Lytle, University of California,Riverside, CA, USA) were used to label NKCC1; T84and TEFS-2 are rabbit polyclonal antibodies directedagainst the C terminus of human NKCC1, and NT isa rabbit polyclonal antisera against the N terminus ofhuman NKCC1 (McDaniel et al. 2005). All presentedexperiments have been conducted with NT andcorroborated by use of TEFS2 and/or T84 antibodies(data not shown). Alexafluor-conjugated goat and rabbitsecondary antibodies (Invitrogen, UK) were used inconjuction with confocal microscopy (Zeiss 510-Meta).Non-specific labelling was determined to be negligible byomitting the primary antibody or substituting it with theappropriate immunoglobulin (e.g. Figures 1B and 2B).

Immunocytochemistry image analysis. Relative amountsof labelled protein were analysed with the 4D cellimaging Volocity software (Improvision, Coventry, UK).Background corrected fluorescence intensity levels weremeasured by placing discrete regions of interest on thebasal membrane, the lateral membrane and at the apicalpole. For each crypt at least three representative cells wereanalysed at each of the base, mid and upper crypt regions;a minimum of three crypts from at least two patientscontributed to the mean fluorescence intensity reading ineach case.

Statistics

Data are expressed as means ± s.e.m. (n is the numberof crypts derived from N patients). Differences amonggroups were determined using one-way ANOVA and


at UNIV OF ARKANSAS on July 16, 2007 jp.physoc.orgDownloaded from



Figure 1. NKCC1 expression patterns in the human colonicepitheliumA, topological intensity gradient for NKCC1 immunolabelling (filledarrow, crypt base; open arrow, surface epithelium; ∗crypt lumen). B,cross section (×63) of lower crypt region (filled arrow, basalmembrane; open arrow, apical membrane). Neg, a typical negativecontrol obtained by omitting primary NT antibody or replacing it withIgG immunoglobulin. C–E, three types of expression pattern commonlyfound in a single tissue sample. C, crypt base exhibiting basal (e.g.open arrow) and lateral (filled arrow) labelling (×63). D, vesicularlabelling in perinuclear region of adjacent crypts (e.g. filled arrow). E,loss of basolateral labelling and presence of immunoreactivity at apicalpole (e.g. filled arrow). White scale bar, 30 μm; N = 4 patients; PI,propidium iodide. Similar labelling patterns were obtained with theT84 and TEFS2 NKCC1 polyclonal antibodies (data not shown).

the Tukey’s post hoc method of multiple comparisons.Differences between two groups were determined using theMann–Whitney rank-sum test. P < 0.05 was consideredstatistically significant.

Results

Topology and polarity of NKCC1 expression

To gain an initial insight into the contribution ofNKCC1 membrane expression dynamics to co-transporterregulation in the native human colonic epithelium,normal biopsy tissue samples were examined by immuno-histochemistry. The NT polyclonal NKCC1 antibody(McDaniel et al. 2005) exhibited three different labellingpatterns in the same clinical biopsy sample (Fig. 1).Consistent with previous observations of the mammalianintestine, NKCC1 labelling predominated on basal andlateral membranes, and this was more intense in thelower crypt regions (Strong et al. 1994; Flemmer et al.2002) (Fig. 1A–C). In addition to this expression pattern,high magnification confocal microscopy revealed groupsof adjacent crypts that exhibited perinuclear labelling ofa vesicular nature at the basal pole of the cell (Fig. 1D).The third type of pattern was conspicuous by the absenceof NKCC1 on basolateral membranes and the presenceof labelling at the apical pole (Fig. 1E). To investigatethe direct influence of secretagogues and their attendantsecond messengers on co-transporter expression, NKCC1location was studied in isolated human colonic crypts.

Isolated crypts, the functional units of the humancolonic epithelium, maintained their flask-likemorphology and physiological viability for a numberof days in culture (Supplemental Fig. 1). Immuno-labelling procedures using the NT polyclonal antibodydemonstrated basolateral membrane NKCC1 labellingof isolated crypts. A prominent expression gradientfor NKCC1 labelling along the crypt-axis was evident(Fig. 2A–C), which was in marked contrast to theexpression profile of the housekeeping basolateralmembrane protein Na+–K+-ATPase (Fig. 2D and E).Prior to investigating the Ca2+-dependent regulation ofNKCC1 expression, we sought to characterize cholinergicCa2+ signals in isolated human colonic crypts.

Cholinergic excitation–secretion coupling

Pharmacological studies in our laboratory havedetermined the functional expression of the M3

muscarinic receptor subtype at the base of rodent andhuman colonic crypts (Lindqvist et al. 1998, 2002)Immunolabelling of isolated human colonic cryptsrevealed pronounced expression of M3AChR on basalmembranes which progressively diminished towardsthe crypt surface (Fig. 3A–C). Addition of acetylcholine




(10 μm) to Fura-2-loaded crypts initiated a Ca2+ signal atthe human crypt base which propagated along the entirecrypt axis (Fig. 3D and F) (Lindqvist et al. 1998, 2002).Ca2+ mobilization initiated in the apical pole of cells

Figure 2. NKCC1 expression in isolated human colonic cryptsA, NKCC1 expression gradient along the crypt axis (filled arrow, crypt-base; open arrow, crypt surface). B, basal andlateral membrane NKCC1 expression (open arrow, basal; filled arrow, lateral). Neg, typical negative control labellingobtained by omitting primary NT antibody or replacing it with IgG immunoglobulin. C, NKCC1 immunoreactivityfluorescence intensity level along the crypt axis (∗significant difference at the P < 0.001 level compared with thecrypt base or lower mid-crypt region). D and E, Na+–K+-ATPase expression along the crypt axis was relativelyconstant; &, a small, but significant difference (P < 0.001) on comparison of Na+–K+-ATPase labelling intensityto mid-crypt levels. Inserts are representative curvilinear intensity profiles for a single colonic crypt. n = 3 cryptsderived from N = 2 patients in each case. White scale bar, 30 μm. The same NKCC1 labelling pattern was obtainedwith the T84 and TEFS2 polyclonal antibodies (data not shown).

located at the very base of the crypt, and this was followedby intracellular signal propagation to the basal pole ofthe cell (Fig. 3E). The kinetics of the Ca2+ response werebiphasic in nature (Fig. 3G); in the continued presence




of the agonist the stimulated intracellular Ca2+ levelsfell gradually, but remained elevated above baseline(Fig. 3G). Ca2+ signals evoked by repetitive stimulation(10 μm for 2 min at 30 min intervals) were reproduciblein nature (Fig. 3H) and were abolished by TMB-8(3,4,5-trimethoxybenzoic acid-8-(diethylamino) octylester; 100 μM), an inhibitor of intracellular Ca2+ release(Palade et al. 1989) (Fig. 3I). The observed spatiotemporalcharacteristics of cholinergic Ca2+ signals pointed to acentral role in co-ordinating excitation–secretion couplingin colonic crypts.

Cholinergic stimulated fluid flow was monitoredindirectly by the movement of calcein-loaded discardedcells or particle matter along the lumen of isolated humancolonic crypts (Halm & Halm, 1999) (Fig. 4A). Theadvent of acetylcholine-stimulated fluid flow suggestedthat NKCC1 had been stimulated. NKCC1 activation levelswere assessed by the rate of plateau phase acidificationassociated with influx of the NH4

+, a substrate forthe co-transporter (Heitzmann et al. 2000; Bachmannet al. 2003). Acetylcholine stimulated the rate of plateauphase acidification (i.e. rate of NKCC1-mediated NH4

+

influx) by nearly sixfold, and this was abolished by theNKCC1 inhibitor bumetanide (100 μm; Fig. 4B and C).Bumetanide did not influence the slight plateau phaseacidification exhibited under non-stimulated conditions(data not shown). TMB-8 (100 μm) also abolished acetyl-choline stimulation of the plateau phase acidification rate(Fig. 4D and E); the intracellular Ca2+ antagonist did notinfluence the slight plateau phase acidification exhibitedunder non-stimulated conditions (data not shown). Acutestimulation with acetylcholine also promoted increasedNKCC1 labelling on basal and lateral membranes (Fig. 4Fand G). The increase in diffuse labelling at the apical polewas not significant (Fig. 4F and G).

Acetylcholine promotes selective NKCC1internalization

Cholinergic stimulated fluid flow in the crypt lumenwas accompanied by a reduction in crypt cell volume(Fig. 5A–5C). The outer crypt width and crypt lengthremained constant over a similar time frame: crypt width,77.2 ± 4.1 μm (t = 0) and 76.5 ± 3.6 μm (t = 30 min,

Figure 3. Cholinergic crypt Ca2+ signallingA, M3 muscarinic acetylcholine receptor (M3AChR) expression gradient along the crypt axis (×20). B, highmagnification of crypt cell M3AChR immunoreactivity (×63 objective; N = 5, n = 10). C, fluorescence intensitylevels for M3AChR immunoreactivity. The insert is a representative curvilinear intensity profile for a single coloniccrypt; ∗ and &, significant difference at P < 0.001 on comparison of M3AChR labelling intensity to the crypt baseand mid-crypt region, respectively. D, acetylcholine (10 μM)-stimulated colonic crypt Ca2+ wave (×20 objective,time in seconds, Fura-2 ratio pseudocolour scale; N > 30). D, acetylcholine (10 μM) stimulated colonic crypt Ca2+wave (×40, time in seconds, Fura-2 ratio pseudocolour scale; N > 50). F–I, Fura-2 ratio traces for specified regionsof interest along crypt axis (the black bar indicates the presence of acetylcholine, 10 μM). White scale bar, 30 μm.

P = 0.886); crypt length, 214.5 ± 6.2 μm (t = 0) and218.0 ± 8.4 μm (t = 30 min, P = 1.00). NKCC1 activitywas downregulated (i.e. a reduction in basal ion uptake)during the acetylcholine-stimulated SVD (secretoryvolume decrease) (Fig. 5D and E). Examination ofNKCC1 expression revealed the formation of punctatevesicular-like fluorescence at basal and lateral membranesin the colonic crypt base following 10 min acetylcholinestimulation (Fig. 5F). At 30 min, NKCC1 labelling wasdramatically internalized from basolateral cell membranesand re-located to vesicles at the apical pole of the cell(Fig. 5G and H). Acetylcholine is thought to be a labilemolecule in vivo and it is unlikely that crypt cells will beexposed to the neurotransmitter for long periods; briefexposure to acetylcholine (10 μm, 3 min) was sufficient toinduce internalization of NKCC1 (Fig. 5H). Significantly,the housekeeping Na+–K+-ATPase protein was notsubject to internalization following cholinergicstimulation (Fig. 5I).

Ca2+-mediated EGFR transactivation has a recognizedrole in the negative regulation of cholinergic-stimulatedchloride secretion in the T84 cell line (Keely et al.1998, 2000; McCole et al. 2002). It was thus of majorinterest to pharmacologically assess the contributionmade by this pathway to acetylcholine-stimulated NKCC1internalization. Pre-incubation (30 min) of humancolonic crypts with the Ca2+ release inhibitor TMB-8(100 μm) or a EGFR kinase inhibitor (Levitzki & Gazit,1995), tyrphostin AG1478 (1 μm), prevented NKCC1internalization (Fig. 6). The intracellular Ca2+ chelatorBAPTA-AM (25 μm) prevented acetylcholine-stimulatedNKCC1 internalization and treatment with the Ca2+

ionophore ionomycin (5 μm) mimicked the actions ofacetylcholine (data not shown).

Internalized NKCC1 is targeted to lysosomes, andprotein synthesis is required for co-transporterre-expression

The long-term fate of internalized NKCC1 was initiallystudied by extending the time course of immunolabellingexperiments. Both membrane and vesicular NKCC1labelling were barely detectable at 1 h post acetylcholinestimulation (Fig. 7A and B). At 4 h post-stimulation,




NKCC1 expression at basal and lateral membraneswas restored with no evidence of vesicular labelling.Cycloheximide (100 μg ml−1), a potent inhibitor ofprotein synthesis, prevented NKCC1 re-expression(Fig. 7B and C). Significantly, at 15 min post-stimulation,vesicles positive for the lysosomal marker LAMP-1 wererecruited from the apical pole towards basal membranes,and internalized NKCC1 colocalized with LAMP-1labelling (Fig. 7D and E). Lysosomal degradation ofinternalized NKCC1 is supported by the observationthat, in presence of the lysosomotropic agent chloroquine(100 μm) (Janda et al. 2006), internalized NKCC1accumulated in LAMP-1-positive vesicles 1 h post acetyl-choline stimulation (Fig. 7F).

Figure 4. Cholinergic activation of relative NKCC1 activityA, acetylcholine (ACh, 10 μM) stimulated the movement of a calcein-labelled cell located in the crypt lumen(arrow head, time in seconds; n = 3, N = 3). B, ACh stimulation of relative NKCC1 activity; measuredas the bumetanide-sensitive plateau phase acidification rate in response to an NH4Cl challenge. BCECF(2′,7′-bis(2-carboxyethyl)-5-(and-6)-carboxyfluorescein) ratio traces from three separate crypts are illustrated. Thedashed line indicates a linear fit to plateau phase acidification in each case. The black bar indicates the presenceof NH4Cl (control, 40 mM, 3 min) and/or ACh (10 μM, 3 min); crypts treated with bumetanide (100 μM) werepreincubated for 30 min. C, summary of plateau phase acidification rates normalized to control values. D and E,TMB-8 (100 μM) abolished ACh-stimulated plateau phase acidification. For each group, BCECF experiments wereconducted on n = 3–24 crypts derived from N = 2–17 patients. F and G, enhanced basolateral membrane labellingof NKCC1 following acute stimulation with ACh (10 μM; n = 2, n = 3). White scale bar, 30 μm; black bar, 3 min.

Following lysosomal degradation, functional re-expression of NKCC1 (Fig. 8A) conferred on crypt cellsthe ability to undergo bumetanide-sensitive regulatoryvolume increase (RVI; Fig. 8B). Functionally restoredNKCC1 expression could be subjected to further roundsof cholinergic-stimulated secretory volume decrease(Fig. 8C).

Differential regulation of NKCC1 expressionand activity

Forskolin is a potent activator of cAMP-dependent fluidsecretion in the intestinal epithelium (Geibel et al. 2006).




Accordingly, forskolin stimulated the rate of plateauphase acidification over fourfold in a bumetanide-sensitivemanner (Fig. 9A and B). However, crypt cell volumeremained relatively constant over 4 h (Fig. 9C), in markedcontrast to the SVD stimulated by acetylcholine (Fig. 5C).Similarly, NKCC1 activity did not diminish and was40% greater than the initial stimulated level (Fig. 9D).The sustained levels of NKCC1 activity in the continuedpresence of forskolin are supported by maintainedexpression of basolateral membrane NKCC1 and a distinctlack of vesicles containing internalized NKCC1 (Fig. 9Eand F).

Colonic crypts in vivo are subject to combinatorialstimulation with Ca2+-dependent and cAMP-dependentsecretagogues. Co-stimulation of isolated colonic cryptswith acetylcholine and forskolin produced a delayed,but ultimately greater, SVD than that stimulated byacetylcholine alone (Fig. 10A). The NKCC1-mediatedplateau phase acidification rate was enhanced sevenfoldby acetylcholine and forskolin co-stimulation (Fig. 10Band C); this level of activity was sustained for 30 min(Fig. 10D). It should be noted that on addition of NH4Cl,the BCECF ratio does not reach a steady-state valuebefore the onset of pronounced NKCC1-mediated plateauphase acidification. Consequently, the measured rate isoffset by the continued influx of NH3 and is thus anunderestimate of NKCC1 relative activity under theseconditions. Unexpectedly, co-stimulation promoted atwofold increase in NKCC1 labelling intensity of the basalmembrane over 30 min (Fig. 9E), when vesicles eventuallyformed (Fig. 10E). Although delayed, the co-transporterwas rapidly internalized and degraded between 30 minand 1 h (Fig. 10E). Full NKCC1 basolateral membraneexpression was restored at 6 h post-stimulation (Fig. 10E).Relative NKCC1 expression levels over time followingco-stimulation are summarized in Fig. 10F .

Discussion

Intestinal fluid secretion is stimulated by a diversenumber of secretagogues that are commonly coupledto the generation of intracellular Ca2+ or cyclic

Figure 5. ACh-stimulated NKCC1 internalizationSecretory volume decrease monitored by (A) differential interference contrast and (B) fluorescence calcein time-lapseimaging. Note ACh-stimulated widening of the inner crypt lumen (filled arrow), while the outer crypt dimensionremained constant (see text for details). C, analysis of apico-basal dimension with respect to time (i.e. extent ofopen arrow in A). The insert illustrates widening of the lumen along the entire crypt axis (time is given in min; N = 5,n = 5). D and E, ACh (10 μM) stimulation of plateau phase acidification was inhibited at 30 min post-stimulation(N = 2, n = 3); the black bar indicates the presence of NH4Cl (control, 40 mM, 3 min); ACh (10 μM, 3 min) waseither added coincidentally with NH4Cl (middle trace) or 30 min beforehand (right trace). Punctate perinuclearNKCC1 immunolabelling at 10 min ACh treatment (F) and internalization to the apical pole at 30 min (G). H,image analysis of NKCC1 immunoreactivity at 30 min post-stimulation with ACh (10 μM) for either 3 min or theentire 30 min (N = 3, n = 5). I, Na+–K+-ATPase immunolabelling was not internalized following ACh stimulation(10 μM, 30 min; N = 3, n = 5). White scale bar, 30 μm.

nucleotides. These second messengers promote transientand sustained episodes of fluid secretion, respectively,and in combination they act synergistically. An under-standing of how the secretory machinery of theintestinal epithelium is differentially modulated by thesesecond messengers is central to the development ofnovel therapeutic strategies for conditions such assecretory diarrhoea. Intestinal fluid secretion is driven bytransepithelial Cl− transport, which involves the concertedaction of apical CFTR and basolateral NKCC1. AlthoughCFTR has been a major focus of attention, NKCC1 isnow emerging as an integrator of secretagogue signalsthat determine levels of fluid secretion (Haas & Forbush,2000; Matthews, 2002). The co-transporter is regulatedby phosphorylation, [Cl−]i, pHi and cell volume, and wasmost recently shown to be rapidly internalized followingstimulation by carbachol (Del Castillo et al. 2005). Thisobservation pointed to NKCC1 trafficking as a possiblemeans by which Ca2+ and cAMP could differentiallymodulate co-transporter activity. To test this hypothesiswe investigated the functional membrane expression ofNKCC1 in a novel ex vivo 3D tissue culture model of thenative human colonic epithelium. This study has shown forthe first time that NKCC1 is dynamically and differentiallyregulated by trafficking events in response to cholinergicCa2+ signals and cAMP, alone or in combination. Theobserved NKCC1 trafficking events are consistent with amajor role in negatively regulating Ca2+ mediated fluidsecretion, promoting sustained fluid secretion by elevatedcAMP, and curtailing the synergistic response invoked byco-stimulation.

NKCC1 was expressed on basolateral membranes ofisolated crypts in the unstimulated state, which wassimilar to the first type of labelling pattern described inclinical tissue sections (cf. Figs 1C and 2B). A topologicalNKCC1 expression gradient was also apparent inimmunolabelling of rat colon and in situ hybridizationof human intestine (Strong et al. 1994; Flemmer et al.2002). Electrophysiological findings in rat isolated cryptssupport the notion that the lower region of the crypt is thepredominant site of secretion (Welsh et al. 1982;Greger et al. 1997). Other transporters such as the




Na+–K+-ATPase did not exhibit an expression gradientalong the crypt axis.

Acetylcholine stimulation of fluid flow in the cryptlumen and the spatial characteristics of the cholinergicCa2+ wave are in keeping with a well-established rolefor intracellular Ca2+ in excitation–secretion coupling(Hardcastle et al. 1984). At the cellular level, activationof M3AChR located on basal membranes initiated a Ca2+

signal at the opposite pole of the cell; this phenomenonwas first described in pancreatic acinar cells (Ashby et al.2003). The spread of a Ca2+ wave from the apical poleacross each crypt cell is consistent with a role for Ca2+

in co-ordinating transepithelial Cl− transport. Althoughthe contribution of Ca2+-sensitive apical Cl− channelsis unclear, bumetanide-sensitive NKCC1 activity (relativeincrease in plateau phase acidification rate) was stimulatedby acetylcholine and inhibited by blocking Ca2+ release. At

Figure 6. Dependence of NKCC1 internalization on Ca2+ and EGFR signalsA, and B, TMB-8 (100 μM, 30 min preincubation), an inhibitor of Ca2+ release (Palade et al. 1989), blockedACh (10 μM, 30 min)-induced NKCC1 internalization (2 ≤ N ≤ 3, 3 ≤ n ≤ 7). The EGFR kinase inhibitor tyrphostinAG1478 (1 μM) also inhibited ACh-stimulated NKCC1 internalization (N = 2, n = 5). White scale bar, 30 μm;∗significant difference (P < 0.001) compared with all other treatments.

the tissue level, M3AChR expression exhibited a similartopological expression gradient to NKCC1 along thecrypt-axis. The high levels of M3AChR expression at thecrypt base account for the site of cholinergic Ca2+ signalinitiation in this region, i.e. increased sensitivity to acetyl-choline (Lindqvist et al. 1998), and presumably NKCC1co-expression optimizes unidirectional flushing of thecrypt lumen.

NKCC1 stimulation following acute treatment withacetylcholine was accompanied by an increase intransporter expression at basal and lateral membranes.There is a large body of evidence from polarizedMDCK cells describing the participation ofa subapical compartment (also known as therecycling endosome) during endocytotic recyclingas well as biosynthetic sorting of basolateral(and apical) proteins, to which the trans-Golgi




Figure 7. NKCC1 internalization, degradation and re-expressionA and B, recovery time course of basolateral membrane NKCC1 expression (4 h). B and C, cycloheximide (CHX,100 μg ml−1, 2 h) did not affect Ach-induced NKCC1 internalization (30 min), but prevented co-transporterre-expression (4 h). D, Ach (10 μM, 15 min) promoted vesicular labelling of lysosome-associated membraneprotein-1 (LAMP-1; red) along the crypt axis (e.g. cf. open arrows in upper and lower panels) and mobilizedLAMP-1 labelling to the basal pole (e.g. compare filled arrows in upper and lower panels), where early signs ofvesicular NKCC1 were evident at the crypt base (e.g. dashed arrows). E, enlarged view of crypt base demonstratingcolocalization of NKCC1 vesicles with LAMP-1 labelling. Comparison of open and closed arrows emphasizes theshift of LAMP-1 labelling from the apical to basal pole (N = 3, n = 4). F, the lysosome inhibitor chloroquine(CHQ, 100 μM) promoted accumulation of internalized NKCC1 in LAMP-1-positive vesicles post Ach (10 μM, 1 h)stimulation.




network can also contribute (Kreitzer et al. 2003; Muth& Caplan, 2003; Hoekstra et al. 2004). Thus it ispossible that following acute acetylcholine stimulation,propagating Ca2+ signals promoted basolateral traffickingof NKCC1 from an organizing centre located at theapical pole, i.e. the site of Ca2+ release. This studyfound no evidence of net NKCC1 membrane insertionduring stimulation with forskolin alone, even thoughthe stimulated activity of the co-transporter increasedby 40% during 30 min stimulation, as was the casein T84 cells (D’Andrea et al. 1996). The increase indiffuse NKCC1 labelling at the apical pole following acuteforskolin treatment suggests that cAMP may modulatetrafficking activity in the apical organizing centre.However, this only translated into increased membranousNKCC1 levels following co-stimulation with forskolinand acetylcholine, thereby implicating a requirement ofintracellular Ca2+ mobilization for net NKCC1 membraneinsertion. A role for propagating intracellular Ca2+ wavesin directing NKCC1 insertion and basolateral traffickingmay be investigated in the future by real-time imaging

Figure 8. Functional NKCC1 expression is restored at 4 h following initial cholinergic cycle ofco-transporter internalization, degradation and re-expressionA, following initial ACh (10 μM) treatment for 3 h, re-stimulation at 4 h stimulated bumetanide (Bum)-sensitiveNH4

+ uptake (i.e. acidification rate; N = 2, n = 3). B, NKCC1 activity was not required for ACh-induced secretoryvolume decrease (SVD), but functional re-expression of NKCC1 was necessary for regulatory volume increase(RVI) (Bum, 100 μM). C, cholinergic secretory volume decrease was reproducible following initial SVD/RVI cycle(experimental conditions as for A). ∗ and &, significant difference (P < 0.001).

of NKCC1-GFP (green fluorescent protein) expressed inisolated crypts (Reynolds A, Parris A & Williams MRunpublished).

Rapid downregulation of acetylcholine-stimulatedNKCC1 activity was first indicated by morphometricobservations of secretory cell volume decrease. Onerationale for the observed reduction in crypt cell volumewas that the basolateral ion uptake mechanism, i.e.NKCC1 activity, was rapidly downregulated, favouringthe loss of ions via an apical efflux pathway, leadingto fluid loss. NKCC1 activity was not required forSVD decrease itself (i.e. bumetanide-insensitive), but wasrequired for the subsequent regulatory volume increase.The time course of secretory cell volume decrease and RVIpredicted that NKCC1 activity would shut downwithin 30 min and be restored 4 h later (Fig. 8B).Dramatic internalization of NKCC1 stimulated byacetylcholine and subsequent re-expression at baso-lateral membranes followed the expected time course.However, the nature of the internalized NKCC1 vesiclesis presently unclear. For other types of transporter,




Figure 9. Effect of forskolin on functional expression of NKCC1A and B, forskolin (FSk, 10 μM)-stimulated NKCC1 activity. C, SVD was not stimulated by FSk. D, NKCC1 activity wassignificantly enhanced 30 min post FSk stimulation (expressed as fold increase over FSk-stimulated co-transporteractivity at 1 min); ACh data derived from Fig. 5C are shown for comparison (∗significant difference in NKCC1activity at 30 min compared with co-transporter activity at 1 min; N = 2, n = 3; P < 0.001). E and F, NKCC1membrane labelling intensity was unaffected by FSk treatment (0–4 h), but acute stimulation (5 min) increasedimmunoreactivity at the apical pole (N = 4, n = 3; P < 0.05).




e.g. CFTR (Bradbury et al. 1999), the non-gastricH+–K+-ATPase ATPAL1 (Reinhardt et al. 2000),aquaporin-2 (Brown, 2003) and the Na+–H+ exchanger(Chow et al. 1999), endocytosis occurs via clathrin-coatedvesicles, which continue to traffick between late end-

Figure 10. Consequences of co-stimulation with ACh and FSk to functional NKCC1 expressionA, combinatorial stimulation induced a delayed, secretory volume decrease that was eventually greater than thatinduced by ACh treatment alone (P < 0.001; N = 2, n = 3). B and C, co-stimulation increased Bum-sensitive, NH4

+uptake (i.e. acidification rate; N = 4, n = 8). D, NKCC1 activity (expressed as fold increase over FSk-stimulatedco-transporter activity at 1 min) was sustained for at least 30 min following co-stimulation with FSk and ACh(relative rates of acidification at 30 min and 1 min post-stimulation for either agonist alone are shown forcomparison). E and F, NKCC1 labelling intensity increased at basal membranes and less dramatically so at lateralmembranes over 15 min co-stimulation; NKCC1 vesicles formed at 30 min (arrow heads); cellular co-transporterexpression was reduced at 1 h and was fully re-expressed at 6 h (E and F); ∗P < 0.001, &P < 0.05; white scale bar,30 μm.

osomes and the plasma membrane, or traverse the sub-apical compartment and merge with lysosomes. Thepresent study demonstrated that virtually all NKCC1immunolabelling disappeared at 1 h and its re-appearance4 h later was dependent on protein synthesis. Moreover,




internalized NKCC1 co-localized with a lysosomal markerthat had been dramatically mobilized to the basal polefollowing acute acetylcholine stimulation. In the presenceof a lysosomal inhibitor, accumulation of internalizedNKCC1 into lysosomal vesicles at the apical pole, indicatesthat NKCC1 is internalized and degraded, most likelyvia a lysosomal pathway. Notwithstanding the currentpharmacological, physiological and immunocytochemicalevidence, this notion awaits biochemical confirmation.In T84 cells, biotinylation assays demonstrated thatcarbachol-stimulated NKCC1 internalization was notdegraded, but was instead re-cycled back to the baso-lateral membrane during a 30 min period. Immuno-labelling of lateral membranes was evident for at least3 h (Del Castillo et al. 2005), confirming the continuouspresence of membranous NKCC1 in the presence ofcarbachol. However, NKCC1 was targeted for degradationin T84 cells by chronic phorbol ester activation of PKC,but via a proteosomal route (Del Castillo et al. 2005).Activation of different PKC isoforms in human cryptsversus T84 cells may explain the different fates of NKCC1following M3AChR activation. Thus, in human coloniccrypts, cholinergic stimulation promotes cycles of NKCC1insertion, activation, internalization, degradation andfunctional re-expression over a 4 h period.

Pharmacological and genetic evidence has shownthat loss of NKCC1 activity inhibits intestinal fluid/Cl−

secretion (Dharmsathaphorn & Pandol, 1986; Flagellaet al. 1999; Mall et al. 2000). It follows that NKCC1internalization provides a mechanism by which Ca2+-mediated fluid secretion is attenuated. Presumably, thisrepresents a compromise between the physiologicalpressures exerted on the intestinal epithelium, i.e.stimulation of sufficient fluid secretion to flush thecrypt lumen of noxious substances, while preservingthe hydrated state of the tissue and the body bylimiting fluid loss. In contrast, forskolin maintainedfunctional membrane expression levels of NKCC1 forat least 4 h, which is consistent with sustaining asecretory response. Initially, this was also the case inresponse to co-stimulation with forskolin and acetyl-choline. Dramatic recruitment of NKCC1 to the basalmembrane was accompanied by sustained stimulationof NKCC1 activity for at least 30 min. The presenceof a secretory volume decrease during this period ofsustained NKCC1 activation indicates that an apicalion efflux pathway(s) is hyperactivated under theseconditions. NKCC1 internalization was delayed, as was thesubsequent re-expression of NKCC1 (ca 6 h). Theincreased loss of fluid associated with co-stimulationwould be compensated by the longer recovery period.Paradoxically, although Ca2+ mobilization stimulates fluidsecretion initially, Ca2+-mediated NKCC1 internalization

may limit net fluid loss stimulated by cholera toxin (Geibelet al. 2006).

The signals regulating NKCC1 membrane expressionrepresent a plausible therapeutic target in conditions suchas secretory diarrhoea. Work on T84 cells has identifiedthe EGFR signalling pathway (Keely et al. 1998, 2000;Chow et al. 2000; McCole et al. 2002; Keely & Barrett,2003) and PKC-ε (Song et al. 2001; Del Castillo et al.2005) as prime candidates in the negative regulation ofCa2+-mediated fluid secretion. Carbachol stimulation ofNKCC1 internalization in T84 cells is PKC-ε dependentand, in the present study, a pharmacological inhibitorof the EGFR kinase inhibited NKCC1 internalization inhuman colonic crypts. The extensive work of Barrett andcolleagues has characterized the M3AChR–Ca2+–EGFRtransactivation pathway and the respective contributionof downstream MAP kinase and PI3 kinase signalling tonegative regulation of Ca2+-dependent fluid secretion(Keely et al. 1998, 2000; McCole et al. 2002; Keely &Barrett, 2003). The relative effects of these pathwayson acute NKCC1 membrane expression remain to beinvestigated. Interestingly, EGFR transactivation by cAMPcoupled receptor activation promoted fluid secretionin T84 cells (Bertelsen et al. 2004). EGFR may thushave a dual influence on NKCC1 membrane expression,i.e. Ca2+-dependent EGFR transactivation promotesNKCC1 internalization, whereas cAMP-dependent EGFRtransactivation may stabilize NKCC1 membraneexpression.

In summary, this study demonstrates the central rolethat regulation of NKCC1 membrane expression plays indetermining co-transporter activity in the native humancolonic epithelium. Cholinergic Ca2+ signals promotethe recruitment and activation of basolateral membraneNKCC1 along the crypt-axis. Activated NKCC1 undergoesa Ca2+-dependent cycle of internalization, degradationand re-expression. Persistent activation of NKCC1 byelevated cAMP is constrained by this Ca2+-dependent cycleof events. Although, these findings provide a novel insightinto the mechanisms that may serve to limit intestinalfluid loss, direct measurements of intestinal fluid secretionare required to confirm the regulatory role of NKCC1internalization. Investigation of the nature and status ofNKCC1 trafficking signals in secretory diarrhoea will shedfurther light on the molecular regulation of intestinal fluidsecretion in health and disease.

References

Ashby MC, Camello-Almaraz C, Gerasimenko OV, PetersenOH & Tepikin AV (2003). Long distance communicationbetween muscarinic receptors and Ca2+ release channelsrevealed by carbachol uncaging in cell-attached patchpipette. J Biol Chem 278, 20860–20864.




Bachmann O, Wuchner K, Rossmann H, Leipziger J,Osikowska B, Colledge WH, Ratcliff R, Evans MJ, Gregor M& Seidler U (2003). Expression and regulation of theNa+–K+–2Cl− cotransporter NKCC1 in the normal andCFTR-deficient murine colon. J Physiol 549,525–536.

Barrett KE & Keely SJ (2000). Chloride secretion by theintestinal epithelium: molecular basis and regulatory aspects.Annu Rev Physiol 62, 535–572.

Bertelsen LS, Eckmann L & Barrett KE (2004). Prolongedinterferon-gamma exposure decreases ion transport,NKCC1, and Na+–K+-ATPase expression in humanintestinal xenografts in vivo. Am J Physiol Gastrointest LiverPhysiol 286, G157–G165.

Bradbury NA, Clark JA, Watkins SC, Widnell CC, Smith H & St& Bridges RJ (1999). Characterization of the internalizationpathways for the cystic fibrosis transmembrane conductanceregulator. Am J Physiol Lung Cell Mol Physiol 276,L659–L668.

Brown D (2003). The ins and outs of aquaporin-2 trafficking.Am J Physiol Renal Physiol 284, F893–F901.

Chow CW, Khurana S, Woodside M, Grinstein S & Orlowski J(1999). The epithelial Na+/H+ exchanger, NHE3, isinternalized through a clathrin-mediated pathway. J BiolChem 274, 37551–37558.

Chow JY, Uribe JM & Barrett KE (2000). A role for proteinkinase cepsilon in the inhibitory effect of epidermal growthfactor on calcium-stimulated chloride secretion in humancolonic epithelial cells. J Biol Chem 275,21169–21176.

Clayburgh DR, Musch MW, Leitges M, Fu YX & Turner JR(2006). Coordinated epithelial NHE3 inhibition and barrierdysfunction are required for TNF-mediated diarrhea in vivo.J Clin Invest 116, 2682–2694.

D’Andrea L, Lytle C, Matthews JB, Hofman P, Forbush B 3rd &Madara JL (1996). Na+–K+–2Cl− cotransporter (NKCC) ofintestinal epithelial cells. Surface expression in response tocAMP. J Biol Chem 271, 28969–28976.

Del Castillo IC, Fedor-Chaiken M, Song JC, Starlinger V, Yoo J,Matlin KS & Matthews JB (2005). Dynamic regulation ofNa+–K+–2Cl− cotransporter surface expression by PKC-ε inCl−-secretory epithelia. Am J Physiol Cell Physiol 289,C1332–C1342.

Dharmsathaphorn K & Pandol SJ (1986). Mechanism ofchloride secretion induced by carbachol in a colonicepithelial cell line. J Clin Invest 77, 348–354.

Farokhzad OC, Sagar GD, Mun EC, Sicklick JK, Lotz M, SmithJA, Song JC, O’Brien TC, Sharma CP, Kinane TB, Hodin RA& Matthews JB (1999). Protein kinase C activationdownregulates the expression and function of the basolateralNa+–K+–2Cl− cotransporter. J Cell Physiol 181, 489–498.

Field M (2003). Intestinal ion transport and thepathophysiology of diarrhea. J Clin Invest 111,931–943.

Flagella M, Clarke LL, Miller ML, Erway LC, Giannella RA,Andringa A, Gawenis LR, Kramer J, Duffy JJ, Doetschman T,Lorenz JN, Yamoah EN, Cardell EL & Shull GE (1999). Micelacking the basolateral Na+–K+–2Cl− cotransporter haveimpaired epithelial chloride secretion and are profoundlydeaf. J Biol Chem 274, 26946–26955.

Flemmer AW, Gimenez I, Dowd BF, Darman RB & Forbush B(2002). Activation of the Na+–K+–2Cl− cotransporterNKCC1 detected with a phospho-specific antibody. J BiolChem 277, 37551–37558.

Geibel JP (2005). Secretion and absorption by colonic crypts.Annu Rev Physiol 67, 471–490.

Geibel J, Sritharan K, Geibel R, Geibel P, Persing JS, Seeger A,Roepke TK, Deichstetter M, Prinz C, Cheng SX, Martin D &Hebert SC (2006). Calcium-sensing receptor abrogatessecretagogue-induced increases in intestinal net fluidsecretion by enhancing cyclic nucleotide destruction. ProcNatl Acad Sci U S A 103, 9390–9397.

Gimenez I (2006). Molecular mechanisms and regulation offurosemide-sensitive Na+–K+–2Cl− cotransporters. CurrOpin Nephrol Hypertens 15, 517–523.

Greger R, Bleich M, Leipziger J, Ecke D, Mall M & KunzelmannK (1997). Regulation of ion transport in colonic crypts. NewsPhysiol Sci 12, 62–66.

Haas M, Forbush B 3rd. (2000). The Na+–K+–2Cl−cotransporter of secretory epithelia. Annu Rev Physiol 62,515–534.

Halm DR & Halm ST (1999). Secretagogue response of gobletcells and columnar cells in human colonic crypts.Am J Physiol Cell Physiol 277, C501–C522.

Hardcastle J, Hardcastle PT & Noble JM (1984). Theinvolvement of calcium in the intestinal response tosecretagogues in the rat. J Physiol 355, 465–478.

Heitzmann D, Warth R, Bleich M, Henger A, Nitschke R &Greger R (2000). Regulation of the Na+–K+–2Cl−cotransporter in isolated rat colon crypts. Pflugers Arch 439,378–384.

Hoekstra D, Tyteca D & van Ijzendoorn SC (2004). Thesubapical compartment: a traffic center in membranepolarity development. J Cell Sci 117, 2183–2192.

Janda E, Nevolo M, Lehmann K, Downward J, Beug H &Grieco M (2006). Raf plus TGFβ-dependent EMT is initiatedby endocytosis and lysosomal degradation of E-cadherin.Oncogene 25, 7117–7130.

Kachintorn U, Vongkovit P, Vajanaphanich M, Dinh S, BarrettKE & Dharmsathaphorn K (1992). Dual effects of a phorbolester on calcium-dependent chloride secretion by T84epithelial cells. Am J Physiol Cell Physiol 262, C15–C22.

Keely SJ & Barrett KE (2003). p38 mitogen-activated proteinkinase inhibits calcium-dependent chloride secretion in T84colonic epithelial cells. Am J Physiol Cell Physiol 284,C339–C348.

Keely SJ, Calandrella SO & Barrett KE (2000).Carbachol-stimulated transactivation of epidermal growthfactor receptor and mitogen-activated protein kinase inT(84) cells is mediated by intracellular Ca2+, PYK-2, and p60(src). J Biol Chem 275, 12619–12625.

Keely SJ, Uribe JM & Barrett KE (1998). Carbachol stimulatestransactivation of epidermal growth factor receptor andmitogen-activated protein kinase in T84 cells. Implicationsfor carbachol-stimulated chloride secretion. J Biol Chem 273,27111–27117.

Kreitzer G, Schmoranzer J, Low SH, Li X, Gan Y, Weimbs T,Simon SM & Rodriguez-Boulan E (2003). Three-dimensional analysis of post-Golgi carrier exocytosis inepithelial cells. Nat Cell Biol 5, 126–136.




Levitzki A & Gazit A (1995). Tyrosine kinase inhibition: anapproach to drug development. Science 267,1782–1788.

Lindqvist S, Hernon J, Sharp P, Johns N, Addison S, Watson M,Tighe R, Greer S, Mackay J, Rhodes M, Lewis M, StebbingsW, Speakman C, Evangelista S, Johnson I & Williams M(2002). The colon-selective spasmolytic otilonium bromideinhibits muscarinic M(3) receptor-coupled calcium signalsin isolated human colonic crypts. Br J Pharmacol 137,1134–1142.

Lindqvist SM, Sharp P, Johnson IT, Satoh Y & Williams MR(1998). Acetylcholine-induced calcium signaling along therat colonic crypt axis. Gastroenterology 115,1131–1143.

Mall M, Bleich M, Schurlein M, Kuhr J, Seydewitz HH, BrandisM, Greger R & Kunzelmann K (1998). Cholinergic ionsecretion in human colon requires coactivation by cAMP.Am J Physiol Gastrointest Liver Physiol 275,G1274–G1281.

Mall M, Wissner A, Seydewitz HH, Kuehr J, Brandis M, GregerR & Kunzelmann K (2000). Defective cholinergic Cl−secretion and detection of K+ secretion in rectal biopsiesfrom cystic fibrosis patients. Am J Physiol Gastrointest LiverPhysiol 278, G617–G624.

Matthews JB (2002). Molecular regulation of Na+–K+–2Cl−cotransporter (NKCC1) and epithelial chloride secretion.World J Surg 26, 826–830.

Matthews JB, Awtrey CS, Hecht G, Tally KJ, Thompson RS &Madara JL (1993). Phorbol ester sequentially downregulatescAMP-regulated basolateral and apical Cl− transportpathways in T84 cells. Am J Physiol Cell Physiol 265,C1109–C1117.

McCole DF, Keely SJ, Coffey RJ & Barrett KE (2002).Transactivation of the epidermal growth factor receptor incolonic epithelial cells by carbachol requires extracellularrelease of transforming growth factor-alpha. J Biol Chem277, 42603–42612.

McDaniel N, Pace AJ, Spiegel S, Engelhardt R, Koller BH,Seidler U & Lytle C (2005). Role of Na+–K+–2Cl−cotransporter-1 in gastric secretion of nonacidic fluid andpepsinogen. Am J Physiol Gastrointest Liver Physiol 289,G550–G560.

Muth TR & Caplan MJ (2003). Transport protein trafficking inpolarized cells. Annu Rev Cell Dev Biol 19, 333–366.

Palade P, Dettbarn C, Alderson B & Volpe P (1989).Pharmacologic differentiation betweeninositol-1,4,5-trisphosphate-induced Ca2+ release and Ca2+-or caffeine-induced Ca2+ release from intracellularmembrane systems. Mol Pharmacol 36, 673–680.

Reinhardt J, Grishin AV, Oberleithner H & Caplan MJ (2000).Differential localization of human nongastricH+–K+-ATPase ATP1AL1 in polarized renal epithelial cells.Am J Physiol Renal Physiol 279, F417–F425.

Sonawane ND, Hu J, Muanprasat C & Verkman AS (2006).Luminally active, nonabsorbable CFTR inhibitors aspotential therapy to reduce intestinal fluid loss in cholera.FASEB J 20, 130–132.

Song JC, Hanson CM, Tsai V, Farokhzad OC, Lotz M &Matthews JB (2001). Regulation of epithelial transport andbarrier function by distinct protein kinase C isoforms.Am J Physiol Cell Physiol 281, C649–C661.

Strong TV, Boehm K & Collins FS (1994). Localization of cysticfibrosis transmembrane conductance regulator mRNA in thehuman gastrointestinal tract by in situ hybridization. J ClinInvest 93, 347–354.

Turner JR (2006). Molecular basis of epithelial barrierregulation: from basic mechanisms to clinical application.Am J Pathol 169, 1901–1909.

Vajanaphanich M, Schultz C, Tsien RY, Traynor-Kaplan AE,Pandol SJ & Barrett KE (1995). Cross-talk between calciumand cAMP-dependent intracellular signaling pathways.Implications for synergistic secretion in T84 colonicepithelial cells and rat pancreatic acinar cells. J Clin Invest 96,386–393.

Welsh MJ, Smith PL, Fromm M & Frizzell RA (1982). Cryptsare the site of intestinal fluid and electrolyte secretion.Science 218, 1219–1221.

Williams MR, Duncan G, Croghan PC, Riach R & Webb SF(1992). pH regulation in tissue-cultured bovine lensepithelial cells. J Membr Biol 129, 179–187.

Acknowledgements

The authors are indebted to the efforts of the staff in the

Gastroenterology and General Surgery Departments at the

Norfolk and Norwich University Hospital. We are also grateful

to the Biotechnology and Biological Sciences Research Council

(A.R., Committee PhD studentship), Humane Research Trust

(A.P., PhD studentship) and the Boston Leukaemia and Cancer

Research Fund (L.A.E., MD training grant) for financial support.

Professor Chris Lytle is thanked for generous provision of the

NKCC1 antibodies. Dr Paul Thomas provided expert help with

bioimaging. This work is dedicated to the memory of Professor

George Duncan (former PhD supervisor and mentor to M.R.W.).

Authors’ present addresses

P. Sharp: School of Biomedical and Health Sciences, Department

of Nutrition and Dietetics, King’s College, Franklin Wilkins

Building, 150 Stamford Street, London SE1 9NH, UK.

S. Lindqvist: School of Medicine, University of East Anglia,

Norwich, Norfolk NR4 7TJ, UK.

Supplemental material

Online supplemental material for this paper can be accessed at:

http://jp.physoc.org/cgi/content/full/jphysiol.2007.129718/DC1

and

http://www.blackwell-synergy.com/doi/suppl/10.1113/

jphysiol.2007.129718



pdfs.semanticscholar.org · journal of physiology volume 582, issue 2,july 2007,pp. 471-524...

Documents