Corrections

CELL BIOLOGYCorrection for “Spinster is required for autophagic lysosomereformation and mTOR reactivation following starvation,”by Yueguang Rong, Christina McPhee, Shuangshen Deng, LeiHuang, Lilian Chen, Mei Liu, Kirsten Tracy, Eric H. Baehreck,Li Yu, and Michael J. Lenardo, which appeared in issue 19, May10, 2011, of Proc Natl Acad Sci USA (108:7826–7831; first pub-lished April 25, 2011; 10.1073/pnas.1013800108).The authors note that, due to a printer’s error, the author name

Eric H. Baehreck should have appeared as Eric H. Baehrecke.Additionally, the author name Christina McPhee should haveappeared as Christina K. McPhee. The corrected author lineappears below. The online version has been corrected.

Yueguang Rong, Christina K. McPhee, ShuangshenDeng, Lei Huang, Lilian Chen, Mei Liu, Kirsten Tracy,Eric H. Baehrecke, Li Yu, and Michael J. Lenardo

www.pnas.org/cgi/doi/10.1073/pnas.1108410108

DEVELOPMENTAL BIOLOGYCorrection for “Tissue-specific roles of Axin2 in the inhibitionand activation of Wnt signaling in the mouse embryo,” by LihuiQian, James P. Mahaffey, Heather L. Alcorn, and KathrynV. Anderson, which appeared in issue 21, May 24, 2011 of ProcNatl Acad Sci USA (108:8692–8697; first published May 9,2011; 10.1073/pnas.1100328108).The authors note that on page 8693, left column, first para-

graph, line 9, “T-to-C” should instead appear as “T-to-A.” Theauthors note that on page 8693, right column, first paragraph,lines 3 and 4, “Axin2canp/Axinnull mice (n=15)” should insteadappear as “Axin2canp/Axin2null mice (n=15).”

www.pnas.org/cgi/doi/10.1073/pnas.1108478108

NEUROSCIENCECorrection for “Antidepressant effects of selective serotoninreuptake inhibitors (SSRIs) are attenuated by antiinflammatorydrugs in mice and humans,” by Jennifer L. Warner-Schmidt,Kimberly E. Vanover, Emily Y. Chen, John J. Marshall, and PaulGreengard, which appeared in issue 22, May 31, 2011, of ProcNatl Acad Sci USA (108:9262–9267; first published April 25,2011; 10.1073/pnas.1104836108).The authors note that the following acknowledgment was

omitted from the article: “This work was supported by NIHGrant MH090963.”

www.pnas.org/cgi/doi/10.1073/pnas.1109215108

www.pnas.org PNAS | July 5, 2011 | vol. 108 | no. 27 | 11297–11298

CORR

ECTIONS

Dow

nloa

ded

by g

uest

on

Aug

ust 2

5, 2

020

Dow

nloa

ded

by g

uest

on

Aug

ust 2

5, 2

020

Dow

nloa

ded

by g

uest

on

Aug

ust 2

5, 2

020

Dow

nloa

ded

by g

uest

on

Aug

ust 2

5, 2

020

Dow

nloa

ded

by g

uest

on

Aug

ust 2

5, 2

020

Dow

nloa

ded

by g

uest

on

Aug

ust 2

5, 2

020

Dow

nloa

ded

by g

uest

on

Aug

ust 2

5, 2

020

Dow

nloa

ded

by g

uest

on

Aug

ust 2

5, 2

020

NEUROSCIENCECorrection for “Functional agonism of insect odorant receptorion channels,” by Patrick L. Jones, Gregory M. Pask, DavidC. Rinker, and Laurence J. Zwiebel, which appeared in issue 21,May, 24, 2011, of Proc Natl Acad Sci USA (108:8821–8825; firstpublished May 9, 2011; 10.1073/pnas.1102425108).The authors note that Figures 2, 3, and 4 appeared incorrectly.

The corrected figures and their legends appear below.

www.pnas.org/cgi/doi/10.1073/pnas.1108343108

Fig. 4. 8-Br-cAMP and 8-Br-cGMP did not elicit currents in AgOrco orAgOrco+AgOr10 cells. (A) Representative trace from whole-cell recordingsfrom cells expressing AgOrco with application of 8-Br-cAMP, 8-Br-cGMP, andVUAA1. (B) Representative trace from cells expressing AgOrco+AgOr10 withapplication of 8-Br-cAMP, 8-Br-cGMP, BA, and VUAA1. (C) Representativetrace from cells expressing rCNGA2 with application of 8-Br-cGMP. Holdingpotentials for all recordings were −60 mV. (D) Histogram of normalizedcurrents from cyclic nucleotide and control responses (n = 4). All currentsnormalized to VUAA1 responses.

Fig. 2. RR blocks inward currents of AgOrco alone and in complex. (A–C)Representative traces of RR-blocked inward currents in AgOrco (A) andAgOrco+AgOr10 (B and C) cells. Holding potential was −60 mV for A–C. (D)Analysis of RR blockage of VUAA1 and BA-induced currents from A (n = 5),B (n = 5), and C (n = 4).

Fig. 3. AgOrco is a functional channel and responds to VUAA1 in outside-out membrane patches. (A) Single-channel recording from an outside-outexcised patch pulled from a cell-expressing AgOrco. (B–D) Expansions oftrace A before (B), during (C), and after (D) a 5-s application of −4.0 logMVUAA1. All-point current histograms of trace expansions are on the rightsides of B–D. Excised membrane patch was held at −60 mV.

11298 | www.pnas.org

Dow

nloa

ded

by g

uest

on

Aug

ust 2

5, 2

020

Spinster is required for autophagic lysosomereformation and mTOR reactivationfollowing starvationYueguang Ronga,1, Christina K. McPheeb,c,1, Shuangshen Denga,1, Lei Huanga, Lilian Chend, Mei Liua, Kirsten Tracyc,Eric H. Baehreckec, Li Yua,2, and Michael J. Lenardoe

aState Key Laboratory of Biomembrane and Membrane Biotechnology, School of Life Science, Tsinghua University, Beijing 100084, China; cDepartment ofCancer Biology, University of Massachusetts Medical School, Worcester, MA 01605; bDepartment of Molecular and Cellular Biology, Harvard University,Cambridge, MA 02138; dCollege of Biological Sciences, China Agricultural University, Beijing 100193, China; and eLaboratory of Immunology, NationalInstitute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD 20892

Edited by David M. Sabatini, Whitehead Institute, Cambridge, MA, and accepted by the Editorial Board April 4, 2011 (received for review September 14, 2010)

Autophagy is a conserved cellular process to degrade and recyclecytoplasmic components. During autophagy, lysosomes fuse withan autophagosome to form an autolysosome. Sequestered com-ponents are degraded by lysosomal hydrolases and presumablyreleased into the cytosol by lysosomal efflux permeases. Followingstarvation-induced autophagy, lysosome homeostasis is restoredby autophagic lysosome reformation (ALR) requiring activation ofthe “target of rapamycin” (TOR) kinase. Spinster (Spin) encodesa putative lysosomal efflux permeasewith the hallmarks of a sugartransporter. Drosophila spin mutants accumulate lysosomal carbo-hydrates and enlarged lysosomes. Here we show that defects inspin lead to the accumulation of enlarged autolysosomes. We findthat spin is essential for mTOR reactivation and lysosome reforma-tion following prolonged starvation. Further, we demonstrate thatthe sugar transporter activity of Spin is essential for ALR.

lysosomal storage disease

During autophagy, lysosomes fuse with autophagosomes toform autolysosomes, where contents are degraded by lyso-

somal hydrolases and released by lysosomal efflux transporters(1). The autophagic/lysosomal pathway is critical to cellular ho-meostasis. Defects in autophagy lead to the accumulation ofdamaged organelles, misfolded proteins, and toxic metabolites,and are associated with neurodegeneration and other abnor-malities (1, 2). Defects in specific lysosomal hydrolysis have beenimplicated in lysosomal storage disorders (LSDs; ref. 2). Lossof lysosomal protease activity can lead to the accumulation ofundigested material, as well as neurodegenerative disease. Inaddition, defective efflux of lysosomal contents by lysosomaltransporters can lead to accumulation of lysosomal substratesand defective lysosomal function (3).Lysosomal efflux transporters are a family of lysosomal mem-

brane proteins required for the export of lysosomal degradationproducts, such as amino acids and monosaccharides (4). A subsetof lysosomal storage diseases has been linked to mutations foundin lysosomal efflux transporters. For example, defects in Sialin,a sialic acid transporter, leads to sialic acid storage diseases(SASD), and defects in the lysosome Arginine transporter leadto Juvenile Batten Disease (5, 6). Spinster (Spin) (also known asbenchwarmer) is a late endosomal/lysosomal membrane proteinwith the amino acid sequence of a lysosomal sugar carrier in themajor facilitator superfamily (4, 7). Spin is a transmembrane pro-tein containing 8–12 transmembrane domains (8). In Drosophila,hypomorphic mutations in spin lead to decreased adult life span,defects in courtship behavior, accumulation of autoflourescentpigments, and neurodegeneration (5, 8, 9). Drosophila spin mu-tants also exhibit neuromuscular synaptic overgrowth (8) andenhanced tau-mediated toxicity (4). In zebrafish, loss of thespin homolog not really started (nrs) leads to embryonic lethalitycharacterized by the accumulation of opaque substances in the

yolk (9). Interestingly, Drosophila spin mutants exhibit endocyticdefects, as well as widespread accumulation of lysosomal car-bohydrates and enlarged lysosomes (4). Little is known, however,about the mechanism leading to the accumulation of enlargedlysosomes in spin mutants.ALR is an evolutionarily conserved lysosome regeneration

cycle that governs nutrient sensing and lysosome homeostasisfollowing starvation-induced autophagy (10). In response to star-vation, mTOR is inhibited, leading to the induction of autophagy.After prolonged starvation, however, mTOR is reactivated.Upon mTOR reactivation, tubules extrude from autolysosomalmembranes and give rise to vesicles that ultimately mature intofunctional lysosomes (10). The degradation of autophagic cargois required for mTOR reactivation after starvation, and inhibit-ing mTOR reactivation leads to the accumulation of enlargedautolysosomes. In addition, ALR requires the dissociation of thesmall GTPase Rab7 from autolysosomes, and overexpression ofconstitutively active Rab7 results in the accumulation of enlargedautolysosomes (10).Here we report that loss of spin leads to the accumulation of

enlarged autolysosomes that fail to degrade their contents inboth mammalian cells and Drosophila. We show that spin is re-quired for mTOR reactivation and lysosome reformation fol-lowing prolonged starvation. Interestingly, reactivation of mTORsignaling after starvation is sufficient to induce lysosome refor-mation even in the context of decreased spin function. Impor-tantly, we find that the sugar transporter activity of Spin isessential for ALR. Our findings elucidate the role of this lyso-somal efflux transporter in ALR and reveal its contributionto LSDs.

ResultsMammalian Spin Colocalizes with the Lysosomal Membrane MarkerLamp1. We used RNAi to screen a collection of permeases, andidentified Spinster (Spin) as a candidate regulator of autophagiclysosome reformation. spin encodes a protein with the hallmarksof a sugar transporter in the major facilitator superfamily relatedto the arabinose efflux permease (4, 7). Spin has been localizedto the late endosome/lysosome in Drosophila and zebrafish (4, 9).In mammalian cells, Spin has been reported to localize to mi-

Author contributions: Y.R., C.K.M., S.D., E.H.B., L.Y., and M.J.L. designed research; Y.R.,C.K.M., S.D., L.H., L.C., M.L., K.T., and L.Y. performed research; Y.R., C.K.M., S.D., L.H., L.C.,M.L., K.T., E.H.B., L.Y., and M.J.L. analyzed data; and L.Y. and M.J.L. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission. D.M.S. is a guest editor invited by the EditorialBoard.1Y.R., C.K.M., and S.D. contributed equally to this manuscript.2To whom correspondence should be addressed. E-mail: liyulab@mail.tsinghua.edu.cn.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1013800108/-/DCSupplemental.

7826–7831 | PNAS | May 10, 2011 | vol. 108 | no. 19 www.pnas.org/cgi/doi/10.1073/pnas.1013800108

tochondria (11). We expressed Spin-GFP in normal rat kidney(NRK) cells and found that Spin-GFP largely colocalizes withthe acidic compartment dye Lysotracker, but not the mitochon-drial marker Mitotracker (Fig. 1A). Furthermore, we foundthat Spin-GFP specifically colocalizes with the late endosomal/lysosomal membrane marker lysosomal-associated membraneprotein 1 (Lamp1; Fig. 1B). These data suggest that mammalianSpin localizes to Lamp1-positive compartments, and like itsfly and zebrafish counterparts, is a late endosomal/lysosomalprotein.

Defects in spin Lead to Accumulation of Enlarged Lamp1-PositiveCompartments. We next generated Spin knockdown cells (Fig.S1). When cultured in nutrient-rich conditions, these cells ex-hibited grossly normal morphology of Lamp1-positive structures,although we noted that a fraction of these cells exhibit a slight

enlargement and subtle increase in the perinuclear localizationof Lamp1-positive structures (Fig. 1C). However, we found thatserum withdrawal led to the dramatic enlargement of Lamp1-positive structures in spin knockdown cells which differed strik-ingly from control cells (Fig. 1C). To test the specificity of thisphenotype, we knocked down rat spin by expressing an RNAiconstruct that fails to bind the human spin sequence due to asingle nucleotide mismatch (Fig. S2), and overexpressed humanSpin in these knockdown cells. We found that whereas 94% ofspin knockdown cells exhibited enlarged Lamp1-positive com-partments, only 12% of cells overexpressing human Spin exhibitthis phenotype indicating that the human sequence rescues thelysosomal defect (Fig. 1 D and E). This result suggests that thephenotype is due to a specific Spin depletion effect and not anoff-target effect.Significantly, we observed similar effects in vivo in the fatbody

of Drosophila spin mutants, which is a nutrient storage and mo-bilization organ akin to the mammalian liver. Upon starvation,the fatbody cells of spin mutants accumulate enlarged Lamp1-GFP-marked structures compared with fatbody cells of un-starved control animals (Fig. 1F). Together, these data indicatethat defects in spin lead to the accumulation of enlarged Lamp1-positive structures in vivo.

Decreased spin Function Causes Enlarged Autolysosomes. We foundthat the enlargement of Lamp1-positive structures upon loss ofspin is starvation-dependent. We therefore tested whether thisphenotype is dependent upon starvation-induced autophagy. Wecoexpressed Spin-YFP and cyan fluorescent protein (CFP)-taggedmicrotubule-associated light chain 3 (CFP-LC3), a marker ofautophagosomes, in NRK cells. We found that, after 4 h of star-vation, Spin-YFP (Red) localized to discrete ring-like structuressurrounding CFP-LC3 marked autophagosomes (Fig. 2A).Cotreatment of cells with the autophagy inhibitor 3-methyl adenine(3-MA) completely inhibited the enlargement of Lamp1-positivestructures upon starvation (Fig. 2B and Fig. S3 A–C). Moreover,concomitant knockdown of spin and beclin1, a gene required forautophagy and also known asAtg6, reduced the percentage of cellsexhibiting enlarged Lamp1-positive compartments after starvationfrom 89% to 38% (Fig. 2 C–E). These data demonstrate that thestarvation-induced enlargement of Lamp1-positive compartmentsupon spin knockdown is autophagy-dependent.We next investigated whether the enlarged Lamp1-positive

structures are autolysosomes. We knocked down spin in NRKcells stably expressing CFP-LC3 and Lamp1-YFP, and foundthat the accumulated Lamp1-positive structures are LC3 positive(Fig. S4 A and B). Consistent with the idea that autolysosomesaccumulate upon starvation, TEM analysis revealed that single-membrane-bound vesicles containing undigested cytoplasmiccontents accumulate upon starvation in control and spin-knock-down cells (Fig. S4C). However, spin knockdown led to theaccumulation of larger, more distended single-membrane struc-tures than those found in control cells, which is most evident 10 hafter starvation (Fig. S4D). Similarly, TEM revealed that largeautolysosomes with cargo accumulate in vivo in Drosophila fatbody (Fig. S5). We have shown that, in NRK cells, the formationof autolysosomes peaks around 4 h after starvation, and at 10 hafter starvation, most autolysosomes have disappeared (10).However, the TEM analyses indicate that, in spin-knockdowncells, autolysosomes are long lasting, and that 10 h after starvation,significant numbers of autolysosomes persist (Fig. S4 C and D).These data indicate that cells with decreased spin function accu-mulate enlarged autolysosomes upon starvation.

spin Is Required for ALR Following Starvation. Lysosome numbersare restored by ALR following starvation-induced autophagy(10). To determine whether ALR is defective upon spin knock-down, we coexpressed LC3 and Lamp1 in NRK cells and ana-

A C

B

D

F

E

Fig. 1. Spin colocalizes with the lysosomal membrane marker Lamp1, anddefects in spin lead to enlargement of Lamp1-positive compartments. (A)NRK cells were transfected with Spin-GFP (green) and stained for Mito-tracker (red, Left) or Lysotracker (red, Right). (Scale bar, 5 μm.) (B) NRK cellswere transfected with Spin-GFP and Lamp1-Cherry. (Scale bar 5 μm.) (C) NRKcells were transfected with nonspecific (NS)- or spin-RNAi. After 2 d, cellswere again transfected with NS- or spin-RNAi and Lamp1-YFP; 16 h aftertransfection, cells were starved for 0 or 10 h and observed by confocal mi-croscopy. (Scale bar, 5 μm.) (D) NRK cells were transfected with nonspecific(NS)-RNAi or RNAi against spin. Two days after transfection, cells wereretransfected with RNAi and either hSpin-CFP (Rescue)/Lamp1-YFP or Vector/Lamp1-YFP. Twenty-four hours after the second transfection, cells werestarved for 10 h and then observed by confocal microscopy. (E) Cells from Dwere assessed in a blind fashion for rescue of the enlarged lysosome phe-notype after starvation for 10 h and quantified. One hundred cells werecounted. Error bars represent s.d. from more than three independentexperiments. (F) Control Drosophila expressing tub-Lamp1-GFP versustransheterozygous spin10403/spinK09905 (spin mutant) larvae expressing tub-Lamp1-GFP were either fed or starved for 12 h, and lysosomes in fatbodycells were observed by fluorescent microscopy. Blue: Höescht. Green: Lamp1-GFP. (Scale bars, 20 μm.)

Rong et al. PNAS | May 10, 2011 | vol. 108 | no. 19 | 7827

CELL

BIOLO

GY

lyzed the kinetics of autophagosome formation and lysosomenumber during starvation. Similar to our previous results (10),we found that after 4 h of starvation, LC3 and Lamp1 largelycolocalize, indicating the formation of LC3 and Lamp1-positiveautolysosomes and the consumption of most of the lysosomes inthe cell into autolysosomes (Fig. 3A). By 12 h of starvation,however, LC3 is diffuse indicating that autophagy has ended andaggregated LC3 has been degraded (Fig. 3A). In addition, after12 h of starvation in control cells, Lamp1-positive, LC3-negativelysosomes recover in number to 80% of the levels found innonstarved control cells (Fig. 3B). We found that, upon spinknockdown, LC3 and Lamp1 colocalize after 4 h of starvation(Fig. 3B). By contrast, in spin knockdown cells LC3 and Lamp1colocalize even after 12 h of starvation, and lysosome numberfails to recover in these cells (Fig. 3 A and B).

During ALR, Lamp1-positive tubules extrude from autolyso-somal membranes after 8 h of starvation, and give rise to Lamp1-positive and LC3-negative vesicles that ultimately mature intofunctional lysosomes (10). We expressed CFP-LC3 and Lamp1-YFP in NRK cells and performed 3D reconstructions of Z-stacked scanning confocal images of control and spin knockdowncells after 8 h of starvation. We found that whereas Lamp1-YFPtubules were abundant in control cells, they were reduced in spinknockdown cells (Fig. 3C). Thus, spin knockdown prevents ALR.We previously reported that ALR requires the dissociation

of the small GTPase Rab7 from autolysosomes, and that over-expression of constitutively active Rab7 prevents ALR and re-sults in the accumulation of enlarged autolysosomes (10). We

B

C

A

0

20

40

60

80

100

NS Spin Spin/Bec

% c

ells

wit

h e

nla

rged

lyso

som

es

NS Bec

Beclin-1

Actin

NS

NS

Spin

Spin

Spin+Bec

Spin+Bec

0h 0h 0h

10h 10h 10h

Lamp1

Lamp1

D E

Control

3-MA

Lamp1

10h

10h

Lamp1

Spin + +

SpinLC3

4h

Fig. 2. Enlargement of lysosomes upon spin knockdown is autophagy de-pendent. (A) NRK cells were transfected with CFP-LC3 (pseudocolored asgreen) and Spin-YFP (pseudocolored as red), and 24 h after transfection, cellswere starved for 4 h. (Inset) A close-up of the discrete ring structure of Spin-YFP surrounding the LC3 assembly. (Scale bars, 5 μm.) (B) NRK cells weretransfected with nonspecific (NS)- or spin-RNAi. After 2 d, cells were againtransfected with NS- or spin-RNAi and Lamp1-YFP; 16 h after transfection,cells were starved for 10 h with or without the addition of 1 mM 3-MA, andobserved by confocal microscopy. (Scale bars, 5 μm.) (C) NRK cells weretransfected with nonspecific (NS)-RNAi, spin-RNAi, or spin-RNAi plus Beclin 1(Bec)-RNAi. After 2 d, cells were again transfected with RNAi and Lamp1-YFP; 16 h after transfection, cells were starved for 0 or 10 h and observed byconfocal microscopy. (Scale bars, 5 μm.) (D) Western blot for the Beclin-1protein and Actin control for a representative sample of cells transfectedwith NS-RNAi plus spin-RNAi, or spin-RNAi plus beclin 1-RNAi. (E) Cellsexpressing nonspecific (NS)-RNAi, spin-RNAi, or spin-RNAi plus Beclin 1 (Bec)-RNAi were quantified in a blind fashion for the presence of enlarged lyso-somes after 10 h of starvation. One hundred cells were counted. Error barsindicate the SD from more than three independent experiments.

NS

Spin

12h

12hMerge

Merge

Lamp1

Lamp1

Rab7

Rab7

A

0102030405060708090

100

1 2 3 4

系列1

系列2

NS

Spin

100

80

40

60

20

00h 4h 8h 12h

% L

amp

1 p

osi

tive

stru

ctu

reB

D

LC3Lamp1

LC3Lamp1

NS

Spin

0h 4h 8h 12h

0h 4h 8h 12h

NS Spin8h 8h

LC3Lamp1Lysotracker

LC3Lamp1Lysotracker

C

Time (h)

LC3 Lamp1

NS

Spin

0h 4h 8h 12h

0h 4h 8h 12h

NS

Spin

0h 4h 8h 12h

0h 4h 8h 12h

Lamp2

Rab7

NS Spin

Rat

io o

f si

gn

al i

nte

nsi

ty

bet

wee

n R

ab7

and

Lam

p2

0

1

2

3

4

5

NS 12h Spin 12hNS Spin

E

Fig. 3. spin is required for autophagic lysosome reformation. (A) NRK-CFP-LC3 (pseudocolored as green)-Lamp1-YFP (pseudocolored as red) stable cellswere transfected with nonspecific (NS)- or spin-RNAi and starved for 0, 4, 8,or 12 h. (Scale bar, 5 μm.) (B) Cells from (A) were assessed in a blind fashionfor change in number of Lamp1-positive structures relative to nonstarvedcells. Fifty cells were counted per timepoint. Error bars represent S.D. derivedfrom multiple experiments. (C) Nonspecific (NS)- or spin-RNAi-transfectedNRK-CFP-LC3-Lamp1-YFP cells were starved for 8 h. Cells were then Z-stackscanned at 0.5-μm intervals, and images were collapsed to construct 3Dmodels using IMRIS. Blue, LC3; Green, Lamp1; Red, Lysotracker. (Scale bars, 5μm.) (D) NRK cells were transfected with nonspecific (NS) or spin-RNAi andafter 2 d, cells were transfected with NS or spin-RNAi, CFP-Rab7, and Lamp1-YFP. Twenty-four hours after transfection, cells were starved for 0 or 12 hbefore imaging. (Scale bars, 5 μm.) (E) NS or spin-RNAi-transfected NRK cellswere starved for 12 h, autolysosome/lysosome were purified, and the pro-tein levels of Rab7 and Lamp2 were analyzed by Western blot. The signalintensity of Rab7 and Lamp2 were measured by IPP and the histogram showsthe ratio between Rab7 and Lamp2.

7828 | www.pnas.org/cgi/doi/10.1073/pnas.1013800108 Rong et al.

investigated the localization of CFP-Rab7 and Lamp1-YFP after12 h of starvation in control and spin knockdown cells. We foundthat, unlike in control cells, CFP-Rab7 colocalizes with Lamp1-YFP in spin knockdown cells even after 12 h of starvation, andthe percentage of cells with large persistent CFP-Rab7 vesiclesis dramatically increased in spin knockdown cells (Fig. 3D).Furthermore, the amount of endogenous Rab7 on lysosomal/autolysosomal membranes is increased in starved spin knock-down cells (Fig. 3E) Together, these data indicate that spinknockdown leads to defects in several markers of ALR.

Sugar Transporter Activity of Spin Is Essential for ALR. Spin encodesa sugar transporter, and an E to K mutation in the 217 positionleads to a loss of function in the sugar transporter activity in flies(4). Consistent with this observation, we found that carbohy-drates accumulate in the autolysosomes of spin knockdown NRKcells following prolonged starvation as indicated by the accu-mulation of periodic acid Schiff staining material (Fig. 4A). Wenext studied the E217 residue, which is conserved in spin acrossspecies and presumed to be functionally required (Fig. 4B).Hence, we tested whether the sugar transporter activity is re-quired for autophagic lysosome reformation by comparing therescue efficiency of wild-type spin plasmid with the E217K mu-tant human spin plasmid. Whereas expression of wild-type Spinrescued ALR, expression of mutant Spin E217K failed to rescuethe accumulation of giant Lamp1-positive structures and ALR,indicating the critical role of this residue in the human ortholog(Fig. 4C).

Degradative Capacity of Autolysosomes, but Not Lysosomes, IsImpaired by spin Knockdown. LC3 that is internalized upon for-mation of autolysosomes is subject to degradation. We noticedthat, after 10 h starvation, LC3 is retained in the autolysosomesof spin knockdown cells (Fig. S4A). In addition, electron mi-croscopy revealed that compared with control cells, spin knock-down cells accumulate what appear to be undigested material inautolysosomes after 10 h of starvation, and this is also observedin starved Drosophila spin mutants (Figs. S4C and S5). To testthe degradation capacity of lysosomes and autolysosomes, weloaded control or spin knockdown cells expressing Lamp1-cherryred with the lysosomal substrate DQ-BSA. If grown in normalgrowth medium, lysosomes of both control and spin knockdowncells exhibit normal degradation capacity. However, after 12 h ofstarvation, whereas the Lamp1-positive structures of control cellsstain positive for DQ- BSA, the majority of giant, distendedLamp1-positive structures in spin knockdown cells lack DQ-BSAstaining (Fig. S6). Thus, only after prolonged starvation does lossof spin lead to defects in the degradation capacity of autolyso-somes. This result indicates that the degradation capacity ofautolysosomes, but not lysosomes, is impaired upon spin knock-down, and suggests a possible role for ALR in maintaining thedegradation capacity of autolysosomes. To further test the degra-dation capacity of autolysosomes, we stained cells with BODIPYFL-conjugated Pepstatin A (BODIPY FL-Pep), a cathepsin Dsubstrate that stains only the active form of cathepsin D. Consis-tent with what we observed with DQ-BSA, we found that innutrient- rich conditions, spin knockdown only slightly decreasesBODIPY FL-Pep staining in lysosomes. However, after 12 h ofstarvation, most Lamp1-positive structures in spin knockdown cellsare BODIPY FL-Pep negative (Fig. S7), indicating that cathepsinD activity is impaired in starved spin knockdown cells.We next examined the mechanism of degradation impairment

in spin knockdown cells by staining cells with lysosensor greenDND-189, a pH indicator that exhibits a pH-dependent increasein fluorescence intensity upon acidification. We found that, after12 h of starvation, autolysosomes from spin knockdown cellshave much higher fluorescence intensity than nonspecific RNAi-

transfected cells (Fig. S8), suggesting that spin knockdown aug-ments lysosome acidification.

spin Is Required for mTOR Reactivation Following Starvation. Werecently reported that mTOR activity is inhibited during theinitiation of autophagy, but is reactivated after prolonged star-vation. We found that the degradation of autophagic cargo isrequired for mTOR reactivation after starvation, and thatmTOR reactivation attenuates autophagy and triggers ALR (10).

A

B

C

Fig. 4. Sugar transporter activity is essential for Spin’s role in the regulationof autophagic lysosome reformation. (A) Carbohydrates accumulate in spinknockdown cells under prolonged starvation. NRK (CFP-LC3) cells weretransfected with NS- or spin-RNAi. After 2 d, cells were starved for 18 h andsubjected to periodic acid-Schiff (PAS) staining. (Scale bars, 5 μm.) Arrowsindicate PAS-positive structures. Histogram shows the percentage of PAS-positive structures from multiple experiments counting a total of 100 cells,with mean and SD shown. (B) The alignment of the spin amino acid se-quence from representative species shown in standard single letter code.The conserved glutamic acid residue (E) which is essential for sugar trans-porter activity is highlighted in red. (C) NRK cells were transfected with RNAiagainst spin. Two days after transfection, cells were retransfected withRNAi and vector/Lamp1-YFP, hSpin-CFP-wt/Lamp1-YFP or hspin-CFP-mutant(spinE217K)/Lamp1-YFP. Twenty-four hours after the second transfection,cells were starved for 10 h and then observed by confocal microscopy. (Scalebar, 5 μm.)

Rong et al. PNAS | May 10, 2011 | vol. 108 | no. 19 | 7829

CELL

BIOLO

GY

Because mTOR directly regulates ALR, we next tested whethermTOR reactivation is impaired upon spin knockdown. We usedPhospho-p70 S6 Kinase (P-S6K) as a marker for mTOR activity.We found that, whereas in control cells P-S6K is lost after 2 h ofstarvation but recovers after 8 h of starvation, spin knockdownmarkedly inhibits the recovery of S6K phosphorylation (Fig. 5A).Thus, mTOR reactivation is impaired in spin knockdown cells.The addition of serum is sufficient to rapidly activate mTOR

(12). To test whether mTOR reactivation after prolonged star-vation suffices to trigger autophagic lysosome reformation in spinknockdown cells, we induced giant autolysosomes by starvingspin knockdown cells for 10 h, and then added 10% FCS. Wefound that phosphorylated S6K levels were rapidly boosted bythe addition of FCS (Fig. 5B). Interestingly, shortly after theaddition of FCS, tubules extended from giant autolysosomes inspin knockdown cells (Fig. 5C) similar to the tubules that havebeen associated with ALR and establishment of lysosome ho-meostasis (12). Within 2 h of the addition of FCS, most of thegiant autolysosomes had disassembled into small Lamp1-positivevesicles (Fig. 5C). These data demonstrate that reactivation ofmTOR signaling, which can be achieved by replacing nutrientsexogenously, is sufficient to induce ALR even in spin knockdowncells. This finding supports our conclusion that the stagnation ofautolysosome progression to ALR in starved spin-deficient cellsis due to a failure in mTOR activation, which likely results fromdefective sugar efflux, which has a deleterious effect on autoly-sosomal degradation of a variety of macromolecules, leading toa failure of mTOR activation in these cells.

DiscussionDrosophila spin mutants have been shown to exhibit progressiveneurodegeneration, and lysosomal abnormalities have been shown

to contribute to neurodegeneration in this context (4, 7).Nonetheless, how defects in spin lead to lysosomal abnormalitieshas been unclear. Here, we demonstrate that the structures thataccumulate in spin mutants are autolysosomes, and that spin isrequired for ALR following prolonged starvation.It is interesting that abnormalities in lysosome function and

morphology become apparent only under starvation conditionsin cultured mammalian cells. This might be explained by the factthat under nutrient-rich conditions, the influx of lysosome cargois limited, but when cells undergo autophagy, lysosome cargoinflux increases, magnifying the severity of the defect. If thishypothesis is correct, then one must wonder why spinmutant fliesexhibit an accumulation of slightly enlarged Lamp1-positivestructures even when fed (4). One possibility is that the metab-olism of flies in this respect is greater than the metabolism ofmammals. Another possible explanation is that the fluctuation ofnutrients in vivo is much greater than in cells grown in culturemedium. Thus, a nonstarved animal could have higher basallevels of autophagy compared with cells maintained in culture.This might be particularly important in the brain because selectivedeletion of Atg5 in neuronal cells leads to neurodegeneration evenin unstarved mice (13).One question is how Spin, a lysosomal efflux sugar transporter,

alters the protein degradation capacity of lysosomes. Lysosomeprotein degradation capacity is dependent on the lysosomal in-ternal environment. Lysosome pH is one of the major factorsregulating lysosomal degradation ability, as the optimal pH formany lysosomal proteases is around 4.5; in either lower or higherpH, lysosome protease activity is compromised. Interestingly, wefound that upon starvation, spin knock-down cells exhibit a dra-matically decreased lysosomal pH. However, it is unknown howspin knock-down leads to an increase in lysosomal acidity. Onepossibility is that Spin is a H+/sugar symporter. H+/amino acidsymporters have been indentified in lysosomes, for example,LYAAT1, a lysosome amino acid efflux transporter, is a H+/amino acid symporter, and the efflux transport of amino acids byLYAAT1 is driven by the H+ gradient (14). Similarly, cystinosin,a lysosomal cystine efflux transporter which also contains seventransmembrane domains, has been reported to be a H+/cystinesymporter which uses H+ to drive cystine efflux transport (14). Arecent structural study for FucP, a H+/Fucose symporter whichalso belongs to the major facilitator superfamily (MFS) oftransporters further supports this hypothesis. A structural studyshowed that a conserved E residue must be protonated forproper Fucose transport (15). Interestingly, in our rescue ex-periment, the E217K mutation failed to rescue ALR in spinknockdown cells. If this hypothesis is correct, we would expectthat spin knockdown would block both sugar and H+ effluxand lead to lysosomal acidification, and autolysosomal degra-dation defects. Further investigation will be required to explorethis possibility.There is a bidirectional regulation between autolysosomal

degradation and mTOR reactivation/ALR. On one hand,degradation of autolysosomal content is required for mTORreactivation and ALR; on the other hand, defective mTORreactivation/ALR also causes the impairment of degradation(10). For example, if mTOR reactivation/ALR is blocked duringstarvation by either adding the mTOR inhibitor rapamycin orknocking down mTOR, the degradation of autophagy substrateLC3 is impaired. Interestingly, we found that adding FCS rap-idly, albeit partially, restores the pH and degradation capacity ofspin knockdown cells (Fig. S9), indicating that mTOR may playa role in regulating the pH, and thus the degradation capacity, ofautolysosomes. These data raise the interesting possibility thatmTOR may regulate ALR by affecting the degradation capacityof autolysosomes.One important implication of our findings is that ALR may

play an important role in disease progression in LSDs. The

A

B

C

Fig. 5. spin is required for mTOR reactivation following starvation. (A) NRK-LC3 cells were transfected with nonspecific RNAi (NS) or RNAi against spin.Sixty hours after transfection, cells were starved 0, 2, 6, or 8 h, harvested, andanalyzed by Western blot using antibodies against Phospho-p70 S6 Kinase(P-S6K), Total S6 Kinase (S6K), and Actin. (B and C) NRK cells were trans-fected with spin-RNAi as in A, and starved for 10 h. FCS was then added backfor 0–120 min; cells were harvested and analyzed by Western blot usingantibodies against Phospho-p70 S6 Kinase (P-S6K) (B), or observed by con-focal microscopy for the Lamp1 intracellular pattern (C). (Scale bar, 5 μm.)(Inset) Enlargement of tubules extruding from lysosomes.

7830 | www.pnas.org/cgi/doi/10.1073/pnas.1013800108 Rong et al.

phenotype of spin mutant flies has been long considered to besimilar to certain LSDs (4). In this study, we demonstrate thata spin knockdown also leads to defects in autolysosomal degra-dation and ALR. These data clearly demonstrate that ALR isimportant to the maintenance of lysosome-based cellular deg-radation capacity. As we discuss above, the basal level ofautophagy may be higher in vivo than cultured cells due to thegreater fluctuation of nutrients during the feeding cycle. Thus, itis conceivable that a defect in ALR could cause LSD phenotypesin fed animals.A long-lasting question in LSDs is why a mutation in a single

lysosomal enzyme can cause overall lysosome degradation fail-ure. Because the regulation between lysosomal degradation andALR is bidirectional, our data suggest the interesting possibilitythat a minor defect in lysosomal degradation capacity couldcause a minor defect in ALR, which could in turn amplify thelysosomal/autolysosomal degradation defect. The positive feed-back nature of this regulation loop then may eventually cause theprogressive pathological features of LSDs.

Materials and MethodsReagents and Antibodies. Lysotracker-red, dextran-red (MW 10000), DQ-BSA-red, and DQ-ovalbumin-green were from Invitrogen (Carlsbad, CA). Anti-Lamp1 antibody was from Sigma-Aldrich (St. Louis, MO).

Cell Culture and Transfection. NRK cells were obtained from American TypeCulture Condition (ATCC) and cultured in DMEM (Life Technologies) mediumsupplemented with 10% FBS (5% CO2). Cells were starved by removal ofserum and glutamine. Cells were transfected via Amaxa nucleofection usingsolution T and program X-001, using 200 pmol of RNAi or total 2 μg of DNA.Cells were then cultured in growth medium for further analysis. For tworounds of transfection, cells were transfected with 200 pmol RNAi, and 72 hafter transfection, cells were transfected again with 100 pmol of RNAi andup to 2 μg of DNA.

Constructs. The human Spinster construct was kindly provided by D. Yamamoto(Waseda University). Lamp1 and LC3 constructs were provided by J. Lippincott-Schwartz (National Institutes of Health). GTP-Rab7 and Rab7 constructs wereprovided by J Bonifacino (National Institutes of Health). spinster mutant flieswere generously provided by G. Davis (University of California), and tub-Lamp1-GFP flies were provided by H. Kramer (University of Texas SouthwesternMedical Center).

Live Cell Imaging. Transfected cells were replated in Lab Tek Chamberedcoverglass (Nunc) the night before imaging, and cells were maintained at37 °C with 5% CO2 in a PeCon open chamber (PeCon, Erbach, Germany).Images were acquired by a Leica sp5 or Olympus FV1000 confocal micro-scope. Three-dimensional models were constructed by collecting images by

Z-stack scanning at 0.5-μm intervals, and images were collapsed to construct3D models by IMRIS. For fly experiments, Canton-S third instar larvaeexpressing tub-Lamp1-GFP (control), or transheterozygous spinster mutantlarvae expressing Lamp1-GFP (tub-Lamp1-GFP; spinster10403/spinsterK09905)were starved in moist Petri dishes at 25 °C. The fatbody was dissected andimaged immediately without fixation on a Zeiss AxiovisionZ.1 microscopewith fluorescence.

Staining. Cells were washedwith PBS (PBS), fixed in 2% paraformaldehyde for10 min, and permeabilized in 0.2% Triton X-100 for 5 min. Cells were blockedwith 10% FBS in PBS for 30 min, stained with 10 μg/mL of rabbit anti-Lamp1(Sigma) in blocking buffer for 1 h, and washed with PBS three times. Cellswere then stained with Fluorescein isothiocyanate conjugated (FITC)-antirabbit secondary antibody (BD, San Jose, CA) in PBS for 1 h and washed withPBS three times. For PAS staining, cells were fixed with 10% formalin andembedded in paraffin, and 5-μm sections were deparaffinized hydrated withwater. Sections were oxidized in 0.5% periodic acid solution for 5 min, rinsedin distilled water, and stained in Schiff reagent for 15 min. Sections werewashed in lukewarm tap water for 5 min before dehydration and mounting.For lysosensor staining, cells were labeled with 1 μM LysoSensor Green DND-153 for 30 min, and then the probe-containing medium was replaced withfresh medium. For BODIPY FL pepstatin A staining, cells were labeled with1 μM BODIPY FL pepstatin A for 30 min, and then the probe-containingmedium was replaced with fresh medium.

Electron Microscopy. Cells were fixed in 3% glutaraldehyde in 0.1 M Mopsbuffer (pH 7.0) for 8 h at room temperature, then 3% glutaraldehyde/1%paraformaldehyde in 0.1 M Mops buffer (pH 7.0) for 16 h at 4 °C. They werethen postfixed in 1% osmium tetroxide for 1 h, and embedded in Spurr’sresin, sectioned, doubly stained with uranyl acetate and lead citrate, andanalyzed using a Zeiss EM 10 transmission electron microscope. For fly EManalyses, spinster mutant larvae (spinster10403/spinsterK09905) were comparedwith control progeny of spinster mutant lines (either spinster10403/Cyo-TM6Bor spinsterK09905/Cyo-TM6B) crossed to wild-type Canton-S. third instar larvaewere starved for 12 h on moist plates and fat was dissected and fixed in 2%glutaraldehyde-4% paraformaldehyde in PBS. Fat was then fixed for EManalysis as described above.

ACKNOWLEDGMENTS. We are grateful to Olympus China and the Tsinghuacell biology core facility for providing technical support and O. Schwartz,Juraj Kabat, Lily Koo, Meggan Czapiga (Biological Imaging Facility, NationalInstitute for Allergy and Infectious Diseases, National Institutes of Health(NIH), Qi Dong, and Ying Li for assistance with confocal microscopy, TEM, andimaging processing. We thank J. Lippincott-Schwartz and J. Bonifacino forhelpful discussions and D. Yamamoto, G. Davis, and H. Kramer for constructsand fly strains. This research was supported by National Science FoundationGrants 31030043 and 30971484, 973 Program 2010CB833704, 2011CB910100,and Tsinghua University Grants 2010THZ0 and 2009THZ03071 (to L.Y.), NIHGrant GM079431 (to E.H.B.), and the Division of Intramural Research of theNational Institute of Allergy and Infectious Diseases, NIH, Department ofHealth and Human Services.

1. Mizushima N (2007) Autophagy: process and function. Genes Dev 21:2861–2873.2. Walkley SU (2009) Pathogenic cascades in lysosomal disease-Why so complex? J Inherit

Metab Dis 32:181–189.3. Eskelinen EL, Tanaka Y, Saftig P (2003) At the acidic edge: emerging functions for

lysosomal membrane proteins. Trends Cell Biol 13:137–145.4. Dermaut B, et al. (2005) Aberrant lysosomal carbohydrate storage accompanies endocytic

defects and neurodegeneration in Drosophila benchwarmer. J Cell Biol 170:127–139.5. Ruivo R, et al. (2008) Molecular pathogenesis of sialic acid storage diseases: insight

gained from four missense mutations and a putative polymorphism of human sialin.Biol Cell 100:551–559.

6. Vesa J, Peltonen L (2002) Mutated genes in juvenile and variant late infantileneuronal ceroid lipofuscinoses encode lysosomal proteins. Curr Mol Med 2:439–444.

7. Nakano Y, et al. (2001) Mutations in the novel membrane protein spinster interferewith programmed cell death and cause neural degeneration in Drosophila melanogaster.Mol Cell Biol 21:3775–3788.

8. Sweeney ST, Davis GW (2002) Unrestricted synaptic growth in spinster-a lateendosomal protein implicated in TGF-beta-mediated synaptic growth regulation.Neuron 36:403–416.

9. Young RM, et al. (2002) Zebrafish yolk-specific not really started (nrs) gene is a

vertebrate homolog of the Drosophila spinster gene and is essential for embryogenesis.

Dev Dyn 223:298–305.10. Yu L, et al. (2010) Termination of autophagy and reformation of lysosomes regulated

by mTOR. Nature 465:942–946.11. Yanagisawa H, Miyashita T, Nakano Y, Yamamoto D (2003) HSpin1, a transmembrane

protein interacting with Bcl-2/Bcl-xL, induces a caspase-independent autophagic cell

death. Cell Death Differ 10:798–807.12. Dennis PB, et al. (2001) Mammalian TOR: a homeostatic ATP sensor. Science 294:

1102–1105.13. Hara T, et al. (2006) Suppression of basal autophagy in neural cells causes

neurodegenerative disease in mice. Nature 441:885–889.14. Boll M, Daniel H, Gasnier B (2004) The SLC36 family: proton-coupled transporters

for the absorption of selected amino acids from extracellular and intracellular

proteolysis. Pflugers Arch 447:776–779.15. Dang S, et al. (2010) Structure of a fucose transporter in an outward-open con-

formation. Nature 467:734–738.

Rong et al. PNAS | May 10, 2011 | vol. 108 | no. 19 | 7831

CELL

BIOLO

GY