NATURE | VOL 405 | 15 JUNE 2000 | www.nature.com 797

letters to nature

.................................................................Excessive placental secretion ofneurokinin B during the thirdtrimester causes pre-eclampsiaN. M. Page*, R. J. Woods*, S. M. Gardiner†, K. Lomthaisong*,R. T. Gladwell*, D. J. Butlin*, I. T. Manyonda‡ & P. J. Lowry*

* School of Animal and Microbial Sciences, The University of Reading,Reading RG6 6AJ, UK† School of Biomedical Sciences, The University of Nottingham,Nottingham NG7 2UH, UK‡ Department of Obstetrics and Gynaecology, St George’s Hospital,London SW17 0QT, UK

.................................. ......................... ......................... ......................... ......................... ........

Pre-eclampsia is a principal cause of maternal morbidity andmortality, affecting 5–10% of first pregnancies worldwide. Mani-festations include increased blood pressure, proteinuria, coagulo-pathy and peripheral and cerebral oedema. Although the aetiologyand pathogenesis remain to be elucidated, the placenta isundoubtedly involved, as termination of pregnancy eradicatesthe disease. Here we have cloned a complementary DNA fromhuman placental messenger RNA encoding a precursor protein of121 amino acids which gives rise to a mature peptide identical tothe neuropeptide neurokinin B (NKB)1 of other mammalianspecies. In female rats, concentrations of NKB several-foldabove that of an animal 20 days into pregnancy caused substantialpressor activity. In human pregnancy, the expression of NKB wasconfined to the outer syncytiotrophoblast of the placenta,significant concentrations of NKB could be detected in plasmaas early as week 9, and plasma concentrations of NKB were grosslyelevated in pregnancy-induced hypertension and pre-eclampsia.We conclude that elevated levels of NKB in early pregnancy maybe an indicator of hypertension and pre-eclampsia, and thattreatment with certain neurokinin receptor antagonists may beuseful in alleviating the symptoms.

The mammalian tachykinins comprise substance P2 and neuro-kinins A3 and B1. Their biological actions include smooth musclecontraction, vasodilation, pain transmission, neurogenic inflam-mation and activation of the immune system4. They are normallyrestricted to nervous tissue and exert their effects peripherally bytheir release from nerve endings which activates the neurokininreceptors, NK1, NK2 and NK3. Each NK peptide shows preferencefor one receptor: substance P for NK1; NKA for NK2; and NKB forNK3. NK1 and NK2 induce hypotension, whereas NK3 receptorscause contraction of the portal vein5, venoconstriction of themesenteric bed6, and increased heart rate7—all potentially hyper-tensive. Although substance P and NKA can be found outside thebrain, NKB was undetectable in peripheral tissues examined in non-pregnant animals8, despite being the most potent at NK3 receptorsand exerting equally potent effects as substance P on the cerebralvasculature9.

During pregnancy, the activation of NK3 receptors by NKBwould reduce the large blood flow through the liver to satisfy theneed of the uterus and placenta. Pregnancy involves increases inmaternal cardiac output and blood flow to the uterus, and areduction in vascular resistance within the placental bed10. Tropho-blastic invasion of the intradecidual portion of the spiral arteriesrenders them dilated, flaccid and unresponsive to vasoconstrictoragents and is a key haemodynamic event. Defective trophoblasticinvasion is a consistent feature in pre-eclampsia, in which the spiralarteries retain their musculo-elastic properties and responsivenessto vasoactive substances11, resulting in trophoblastic ischaemia. Theendothelial cell damage that occurs in pre-eclampsia12 leads to areduction in the levels of the vasodilator and platelet-aggregationinhibitor, prostacyclin, with an increase in vasoconstrictorsubstances13.

Previously, we found that the increased concentrations of pla-cental corticotropin releasing factor that we had observed inpregnancy-induced hypertension14 did not appear to contribute tothe pathology of pre-eclampsia. We therefore sought other vaso-active placental neuropeptides using mRNA fingerprinting15 and thehuman data bases16,17 and found nine matches that showed highsimilarity to the bovine NKB precursor. Cloning the full-length gene

Figure 1 Localization of neurokinin B mRNA expression in vertical sections of theplacenta. a, Human at term (39 weeks) with human antisense probe. b, Human at term(39 weeks) with human sense probe. c, Rat 18-day placenta with rat antisense probe.

d, High magnification showing giant cells of the rat placenta expressing neurokinin B.Magnification: a, 10× original size; b, 10×; c, 16×; d, 40×.

© 2000 Macmillan Magazines Ltd

(GenBank accession number AF216586) revealed an open readingframe of 363 bases predicting the 121-residue human pro-neuro-kinin B (see Supplementary Information). The precursor proteingives rise to a mature peptide (DMHDFFVGLM-NH2) that isidentical to neurokinin B of other mammalian species. Expressionof placental NKB mRNA was detected in the first trimester andincreased fivefold at term. In situ hybridization studies on thehuman placenta at term located it to the outer syncytiotrophoblastsideally positioned to secrete its peptide into the maternal blood(Fig. 1a, b). Studies in rat placenta showed a similar expression ofNKB mRNA in the junctional layer and in the giant cells (Fig. 1c, d).

Significant amounts of immunoreactive human NKB were foundin early (21 pg g−1) and term (25 pg g−1) placentae; this NKB hadchromatographic properties as assessed by high pressure liquidchromatography (HPLC) that were identical to synthetic NKB.An identical elution pattern was observed on extraction fromterm plasma. Samples taken from non-pregnant laboratory person-nel all had low (near the sensitivity of the assay) or undetectablelevels of the peptide, as did most samples taken from normotensivewomen throughout gestation, a proportion of whom exhibitedslight increases at term (Fig. 2). Four individuals who had beenadmitted for elective abortions between weeks 9 and 14 hadconcentrations equivalent to the highest found at term (picomolarrange); suggesting that the placentae from these individuals havestarted to secrete supraphysiological concentrations of NKB early inpregnancy. Patients in the third trimester suffering from pregnancy-induced hypertension and pre-eclampsia had concentrations in thenanomolar range (Fig. 2). Simultaneous samples taken frommaternal and cord blood showed levels in cord blood that wereroughly one-third of that of the mother (data not shown), indicat-ing that hypersecretion of NKB could also affect NK receptors in thefeto-placental circulation. When NKB was undetectable in maternalblood, it was also undetectable in cord blood.

Samples taken from pregnant rats indicated that there is a rapidincrease to nanomolar concentrations in maternal blood NKB onday 19, peaking at 3.1 nM on day 20, and then declining at term to0.15 nM. In control female rats, concentrations were less than50 pM. Intravenous infusion of low doses of NKB resulted in nosignificant change in blood pressure, but, during infusion of thehighest dose (180 nmol h−1) when the concentration of NKBreached was 20-fold higher (80 nM) than that found in the bloodof a animal at 20-days pregnant, there was a significant, albeittransient, rise in mean arterial blood pressure (Fig. 3). Heartrate increase was not significant. The haematocrit value was37.7 6 0.6% in controls, and 39.0 6 1.1%, 36.0 6 0.7% and

36 6 1.2%, respectively, after each of the three infusion periods (seeMethods). In the animals receiving NKB there was a 37% gain inweight (0.624 6 0.03 g compared with 0.459 6 0.03 g, P , 0.018) ofthe uteri. Notably, infusion of NKB at the same doses in male ratshad no effects on arterial blood pressure (data not shown).

These studies in conscious rats demonstrate a pressor effect ofchronic high-dose infusion of NKB. We cannot determine theprecise mechanism underlying the rise in blood pressure, but it isunlikely to have involved a change in blood volume, as controlhaematocrit values were not different from those at the end of thethree infusion periods. There was, however, indirect evidence for anincrease in blood to the uterus, as shown by a 37% increase inweight.

Previous reports have shown in vivo transient depressor18 orpressor19 effects of bolus injections of NK3 receptor agonists inanaesthetized and conscious rats, respectively. In the former studythe effects were attributed to a Bezold–Jarisch reflex, and in thelatter the changes were not mediated by NK3 receptors. In consciousguinea pigs, intravenous bolus doses of an NK3 receptor agonist didcause a transient rise in blood pressure which was NK3 receptor-mediated20; however, to our knowledge we are the first to reportcardiovascular effects of continuous infusion of NKB in consciousfemale animals to simulate late pregnancy concentrations. Althoughthe effects were only significant at the highest dose, and then onlytransient, there are notable points. First, the effects occurred inconscious animals with fully competent cardiovascular reflexes.Second, endothelial release of nitric oxide acts to buffer thevenoconstricting effects of NKB, and, in pregnancy, there are anumber of changes in the nitric oxide system21, which mightinfluence the cardiovascular actions of the peptide. As the effectsof NKB are relatively selective for the venous side of the mesentericvascular bed, it is feasible that at low doses the only cardiovascularimpact would be a modest rise in venous return and hence cardiacoutput. In this way, release of NKB from the normal placenta couldcontribute to the hyperdynamic circulation of normal pregnancy inthe rat.

Could NKB be involved in the pathogenesis of pre-eclampsia?NKB could not cause the failed trophoblast invasion, but its potentand vasoactive properties might be involved in the clinical manifesta-tions of pre-eclampsia. In response to the ischaemia consequent

letters to nature

798 NATURE | VOL 405 | 15 JUNE 2000 | www.nature.com

Figure 2 Neurokinin B levels in the plasma of 30 normotensive women (filled circles) weredetermined in samples taken throughout pregnancy and at term in 16 different women(open circles). They are compared with measurements made in eight pre-eclampticsubjects (open squares) in the last trimester of pregnancy. Pre-eclampsia was defined asfollows: new hypertension (diastolic pressure consistently .90 mm Hg, with previouslower levels); and new proteinuria (.500 mg in 24 h) in the absence of urinary tractinfection; both remitting remotely after delivery. Figure 3 Changes in heart rate and blood pressure during infusion of saline (open circles,

n = 3) or incremental doses of NKB (filled circles, n = 6) in conscious unrestrained femalerats. a, Changes in heart rate; b, changes in blood pressure. NKB was infused at doses of1.8 nmol h−1 from t = 0; 18 nmol h−1 from t = 16 h; and 180 nmol h−1 from t = 20 h.Values are mean 6 s.e.m.; asterisk, difference from the original baseline and from thevalues at t = 20 h is significant (Friedman’s test).

© 2000 Macmillan Magazines Ltd

upon defective trophoblastic invasion, the feto-placental unit mustsend a signal to the mother, and the obvious signal would be tomaximize blood flow to the feto-placental unit by increasing thematernal blood pressure. Pre-eclampsia must be viewed as a con-tinuum, because in mild cases of defective invasion there might be asmall increase in blood pressure and none of the clinical features ofthe full-blown syndrome. When ischaemia within the feto-placentalunit is profound, a stronger, enhanced and sustained signal isrequired. As NKB concentrations increase above normal, stimula-tion of the NK3 receptors causes constriction and contraction of themesenteric and portal veins, increasing blood pressure, and poten-tially damaging the kidneys and liver. In severe cases, concentrationsof NKB may be sufficient to stimulate peripheral NK1 receptors, forexample, those found on platelets22 and neutrophils23,24, and tocontribute to other complications associated with activation ofthese cells. The NKB concentrations might also be responsible forthe cerebral complications, as high intravascular concentrations ofsubstance P have been shown to dilate cerebral blood vesselsthrough NK1 receptors9,25, which have been located to theendothelium26. We found significant levels of NKB in the plasmaof a small proportion of plasma samples from first trimesterpregnancies; however, most were low throughout normotensivepregnancies and were increased in only a proportion at term. It ispossible that increased secretion of NKB pre-dates the developmentof pre-eclampsia, and that increased levels of NKB in earlypregnancy may identify pregnancies destined to develop the disease.

Potent and selective antagonists for the NK receptors maypotentially be effective treatments for pre-eclampsia. NK3 receptorantagonists might alleviate the effects of high levels of NKB in theplasma of mothers suffering from hypertension, by reducing thevasoconstrictive effects of NKB on their mesenteric vascular bedsand hepatic portal vein. Indeed, the potent NK3 receptor antago-nist, GR138676, has been shown to be a competitive antagonist ofcontractions induced by the NK3 agonist, senktide, in the rat portalvein27. Where concentrations of NKB have increased sufficiently toexert effects on NK1 receptors, for example, NK1 receptor antago-nists might be the drugs of choice. Antagonists with a potency forNK receptors similar to that of NKB might offer better overallcontrol because specific NK3 receptor antagonists, while alleviatingthe hypertensive effects, may induce a compensatory increase inplacental NKB secretion and thus an increase in NK1 receptoragonist activity. M

MethodsPolymerase chain reaction with reverse transcriptionHuman placental tissue was obtained from pregnancy terminations at weeks 9 and 13, andat term, in compliance with and approval from the Local Research Ethics Committee. RNAwas extracted with Tri Reagent (Sigma). The full-length preproNKB precursor wasamplified from total RNA using SMART RACE cDNA amplification (Clontech, UK). 39RACE was performed using a 59 gene-specific primer (59-GGCACAGAGCTGCTCCACAGGCACCAT-39) derived from the human cDNA clone 138761 similar to bovine P08858NKB precursor. The PCR fragment was purified following gel electrophoresis, cloned intothe vector pGEM-T Easy (Promega), sequenced and compared with the GenBank databaseusing the BLAST program.

In situ hybridization for NKB in the placentaSense and antisense digoxygenin (DIG) riboprobes were synthesized by in vitro tran-scription from rat or human NKB cDNA. Cryostat sections (25-mm) of 18-day ratplacenta and 39-week human term placenta delivered by caesarian section were thaw-mounted on 3-aminopropyltriethoxysilane coated slides, fixed in parafomaldehyde andhybridized with the appropriate DIG-labelled riboprobe. Detection was performed usingan anti-DIG antibody (Roche Diagnostics) coupled with the Renaissance Tyramide SignalAmplification System (New England Nuclear). Nuclei were counterstained with TOTO-3,and sections visualized using confocal microscopy.

Extraction and measurement of NKB from placenta and plasmaNKB was extracted from placenta as described for sheep brain28 using similar weights oftissue to extraction medium. Extraction from plasma was performed as described for Met-enkephalin29, but using octadecasilyl-silica cartridges, Sep-Pak Vac/1CC, C18 (MilliporeCorporation). NKB eluted between 30 and 50% acetonitrile in aqueous 0.1% trifluor-

oacetic acid (v/v). Both extracts were reconstituted in 500 ml of NKB radioimmunoassaybuffer (kit RIK 7357; Peninsula Labs, Belmont, CA), and assays carried out in duplicateaccording to manufacturer’s protocols. Results were corrected by reference to extractedstandards.

Cardiovascular effects of NKB in conscious ratsFemale Sprague–Dawley rats (265–302 g; Charles River, UK) were anaesthetized withHypnorm (0.126 mg kg−1 fentanyl citrate, 4 mg kg−1 fluanisone intraperitonially (i.p.);Janssen, Belgium) and midazolam (5 mg kg−1 i.p.; Antigen Pharmaceuticals, Ireland) forthe implantation of intra-arterial (abdominal aorta via caudal artery) and intravenous(right jugular vein) catheters. After a 24-h recovery period, we measured resting arterialblood pressures and heart rates in the conscious unrestrained state for 2 h beforecontinuous intravenous infusion of incremental doses of NKB (1.8 nmol h−1 overnightfor 16 h, 18 nmol h−1 for 4 h, and 180 nmol h−1 for 2 h; n = 6; Bachem, UK) or saline(0.4 ml h−1; n = 3). Blood samples were measured before infusion of each dose forhaematocrit and for plasma levels of NKB. Uteri and fallopian tubes were removed andweighed at the end. The animals were kept under constant lighting conditions for 48 hbefore and during experiments to ensure that they were in a similar state of oestrus.

Received 8 December 1999; accepted 4 April 2000.

1. Kangawa, K., Minamino, N., Fukuda, A. & Matsuo, H. Neuromedin K: a novel mammalian tachykinin

identified in porcine spinal cord. Biochem. Biophys. Res. Commun. 114, 533–540 (1983).

2. von Euler, U. S. & Gaddum, J. H. An unidentified depressor substance in certain tissue extracts. J.

Physiol. 72, 74–87 (1931).

3. Kimura, S., Okada, K., Sugita, Y., Kanazawa, I. & Munekata, E. Novel peptides neuropeptides,

neurokinin a and b, isolated from porcine spinal cord. Proc. Japan Acad. B 59, 101–104 (1983).

4. Longmore, J., Hill, R. G. & Hargreaves, R. J. Neurokinin-receptor antagonists: pharmacological tools

and therapeutic drugs. Canadian J. Physio. & Pharmacol. 75, 612–621 (1997).

5. Mastrangelo, D. et al. The rat isolated portal vein: a preparation sensitive to neurokinins, particularly

neurokinin B. Eur. J. Pharmacol. 134, 321–326 (1987).

6. D’Orleans-Juste, P., Claing, A., Telemaque, S., Warner, T. D. & Regoli, D. Neurokinins produce

selective venoconstriction via NK-3 receptors in the rat mesenteric vascular bed. Eur. J. Pharmacol.

204, 329–334 (1991).

7. Thompson, G. W., Hoover, D. B., Ardell, J. L. & Armour, J. A. Canine intrinsic cardiac neurons

involved in cardiac regulation possess NK1, NK2 and NK3 receptors. Am. J. Physiol.-Regul. Integrative

Comp. Physiol. 44, 1683–1689 (1998).

8. Moussaoui, S. M. et al. Distribution of neurokinin B in rat spinal cord and peripheral tissues:

comparison with neurokinin A and substance P and effects of neonatal capsaican treatment.

Neuroscience 48, 969–978 (1992).

9. Jansen, I., Alafaci, C., McCulloch, J., Uddman, R. & Edvinsson, L. Tachykinins (substance P,

neurokinin A, neuropeptide K, and neurokinin B) in the cerebral circulation: vasomotor responses in

vitro and in situ. J. Cereb. Blood Flow Metab. 11, 567–575 (1991).

10. Baumann, H. & Huch, R. Maternal hemodynamics in pregnancy. Am. J. Obstet. Gynecol. 165, 237–238

(1991).

11. Lim, K. H. et al. Human cytotrophoblast differentiation/invasion is abnormal in pre-eclampsia. Am. J.

Pathol. 151, 1809–1818 (1997).

12. Lyall, F. & Greer, I. A. The vascular endothelium in normal pregnancy and pre-eclampsia. Rev. Reprod.

1, 107–116 (1996).

13. Hayman, R., Brockelsby, J., Kenny, L. & Baker, P. Preeclampsia: the endothelium, circulating factor(s)

and vascular endothelial growth factor. J. Soc. Gynecol. Investig. 6, 3–10 (1999).

14. Campbell, A. et al. Plasma corticotrophin-releasing hormone concentrations during pregnancy and

parturition. J. Clin. Endocrinol. Metab. 64, 1054–1059 (1987).

15. Page, N. M., Manyonda, I., Butlin, D. J., Bicknell, A. B. & Lowry, P. J. Identification of novel

differentially expressed secretory proteins from the human placenta. J. Endocr. 156 (Suppl.), P257

(1998).

16. Boguski, M. S., Lowe, T. M. & Tolstoshev, C. M. dbEST—database for ‘‘expressed sequence tags’’.

Nature Genet. 4, 332–333 (1993).

17. Adams, M. D. et al. Initial assessment of human gene diversity and expression patterns based upon 83-

million nucleotides of cDNA sequence. Nature 377, Genome Directory 3S (1995).

18. Couture, R., Laneuville, O., Guimond, C., Drapeau, G. & Regoli, D. Characterization of the peripheral

actions of neurokinins and neurokinin receptor selective agonists on the rat cardiovascular system.

Naunyn Schmied. Arch. Pharmacol. 340, 547–557 (1989).

19. Cellier, E., Barbot, L., Regoli, D. & Couture, R. Cardiovascular and behavioural effects of intracer-

ebroventricularly administered tachykinin NK3 receptor antagonists in the conscious rat. Br. J.

Pharmacol. 122, 643–654 (1997).

20. Roccon, A., Marchionni, D. & Nisato, D. Study of SR 142801, a new potent non-peptide NK3 receptor

antagonist on cardiovascular responses in conscious guinea-pig. Brit. J. Pharmacol. 118, 1095–1102

(1996).

21. Conrad, K. P., Kerchner, L. J. & Mosher, M. D. Plasma and 24h NOx and cGMP during normal

pregnancy and preeclampsia in women on a reduced NOx diet. Am. J. Physiol. 46, F48–F57 (1999).

22. Gecse, A., Kis, B., Mezei, Z. & Telegdy, G. The effect of bradykinin and substance P on the arachidonate

cascade of platelets. Immunopharmacology 33, 167–170 (1996).

23. Perianin, A., Snyderman, R. & Malfroy, B. Substance P primes human neutrophil activation: a

mechanism for neurological regulation of inflammation. Biochem. Biophys. Res. Commun. 161, 520–

524 (1989)

24. Perretti, M., Ahluwalia, A., Flower, R J. & Manzini, S. Endogenous tachykinins play a role in IL-1-

induced neutrophil accumulation: involvement of NK-1 receptors. Immunology 80, 73–77 (1993).

25. Kobari, M. et al. Intravascular substance P dilates cerebral parenchymal vessels through a specific

tachykinin NK1 receptor in cats. Eur. J. Pharmacol. 317, 269–274 (1996).

26. Shimizu, T. et al. Occurrence and distribution of substance P receptors in the cerebral blood vessels of

the rat. Brain Res. 830, 372–378 (1999).

27. Stables, J. M. et al. GR138676, a novel peptidic tachykinin antagonist which is potent at NK3 receptors.

Neuropeptides 27, 333–341 (1994).

letters to nature

NATURE | VOL 405 | 15 JUNE 2000 | www.nature.com 799© 2000 Macmillan Magazines Ltd

28. Baigent, S. M. & Lowry, P. J. Urocortin is the principal ligand for the corticotrophin-releasing factor

binding protein in the ovine brain with no evidence for a sauvagine-like peptide. J. Mol. Endocrinol.

24, 53–63 (2000).

29. Clement-Jones, V., Lowry, P. J., Rees, L. H. & Besser, G. M. Met-enkephalin circulates in human

plasma. Nature 283, 295–297 (1980).

Supplementary information is available on Nature’s World-Wide Web site(http://www.nature.com) or as paper copy from the London editorial office of Nature.

AcknowledgementsThis work was supported by the Medical Research Council.

Correspondence and requests for materials should be addressed to P.J.L.(e-mail: p.j.lowry@reading.ac.uk).

letters to nature

800 NATURE | VOL 405 | 15 JUNE 2000 | www.nature.com

.................................................................The g-subunit of the coatomercomplex binds Cdc42 tomediate transformationWen Jin Wu, Jon W. Erickson, Rui Lin & Richard A. Cerione

Departments of Molecular Medicine, and Chemistry and Chemical Biology, VMC,Cornell University, Ithaca, New York 14853-6401, USA

.................................. ......................... ......................... ......................... ......................... ........

The Ras-related GTP-binding protein Cdc42 is implicated in avariety of biological activities including the establishment of cellpolarity in yeast, the regulation of cell morphology, motility andcell-cycle progression in mammalian cells and the induction ofmalignant transformation1,2. We identified a Cdc42 mutant(Cdc42F28L) which binds GTP in the absence of a guaninenucleotide exchange factor, but still hydrolyses GTP with a turn-over number identical to that for wild-type Cdc42 (ref. 3).Expression of this mutant in NIH 3T3 fibroblasts causes cellulartransformation, mimicking many of the characteristics of cells

transformed by the Dbl oncoprotein, a known guanine nucleotideexchange factor for Cdc42 (ref. 4). Here we searched for newCdc42 targets in an effort to understand how Cdc42 mediatescellular transformation. We identified the g-subunit of the coat-omer complex (gCOP) as a specific binding partner for activatedCdc42. The binding of Cdc42 to gCOP is essential for a trans-forming signal distinct from those elicited by Ras.

A number of putative targets for Cdc42 have been identified,including the p21-activated kinases (PAKs), the Wiscott–Aldrichsyndrome proteins (WASPs) and IQGAP (refs 5–8). These targetsare involved in Cdc42-mediated changes in the actin cytoskeletonand cell morphology, but targets/effectors that are responsible forCdc42-regulated cell growth or malignant transformation haveremained elusive. Thus, we generated glutathione S-transferase(GST)–Cdc42 fusion proteins to use as affinity resins to screenNIH 3T3 cell lysates for novel Cdc42 targets. In addition topreviously identified binding partners of Cdc42, this screen revealeda protein with relative molecular mass of 160,000 (Mr 160K) whichbinds to activated, GTPase-defective GST–Cdc42Q61L, better thanwild-type GDP-bound GST–Cdc42 or GST alone (not shown). Wesequenced the Mr 160K protein band and found that it is the aCOPsubunit (Fig. 1), a component of the coatomer protein complex.This complex is important in intracellular trafficking because of itsability to assemble at Golgi membranes under the direction of theGTP-binding protein Arf (refs 9–11).

Figure 1 Coatomer subunits associate with activated Cdc42. Top, alignment of the nineamino acids sequenced from the Mr 160K putative Cdc42-target with residues 415–423of the aCOP subunit. Bottom, the first five lanes are the affinity precipitations of coatomersubunits (aCOP and gCOP) from NIH 3T3 cell lysates using different GST-coupled Cdc42fusion proteins. We loaded the GTPase-defective Cdc42 (Cdc42Q61L) with GTP and wild-type Cdc42 with GTPgS or GDP. In lane 5, we washed the glutathione agarose beads in1M NaCl (3×) to disassemble the coatomer subunits13. The last two lanes are HA-taggedCdc42 proteins that immunoprecipitated from COS-7 cells transiently transfected withpKH3 constructs encoding the indicated Cdc42 proteins. We detected the a- andg-coatomer subunits by western blotting with an antibody that recognizes both of thesesubunits.

Figure 2 A dilysine sequence within the carboxy-terminal end of Cdc42 is necessary forbinding coatomer. a, A synthetic peptide corresponding to the C-terminal tail sequence ofcargo receptors blocks the binding of activated Cdc42 to gCOP. We incubatedglutathione-agarose bound GST–Cdc42Q61L with NIH 3T3 cell lysates containing theindicated amounts of the synthetic peptide CYLRRFFKAKKLIE. We then precipitated thebeads and assayed for the association of gCOP using the anti-a/gCOP antibody, IQGAPusing an IQGAP antibody raised against the amino terminus29 or PAK using an anti-PAKpolyclonal antibody30. b, Substitution of lysine residues 183 and 184 of Cdc42 to serinesprevents the binding of Cdc42 to gCOP. We determined the association of coatomersubunits with GST–Cdc42, GTPase-defective Cdc42 (GST–Cdc42Q61L), and thecorresponding GST–Cdc42 proteins that contain serine rather than lysine at positions183 and 184 (GST–Cdc42ss and GST–Cdc42Q61Lss, respectively) as in Fig. 1. c, Lysineresidues 183 and 184 are not essential for the binding of Cdc42 to other targets andregulators. We incubated the indicated GST–Cdc42 proteins or GST alone with lysatesfrom NIH 3T3 cells that express endogenous IQGAP and PAK, lysates that ectopicallyexpressed HA-tagged ACK-2 or Myc-tagged WASP, or with purified recombinantCdc42GAP, and then precipitated and analysed for their binding by western blotting.IQGAP and PAK were detected using polyclonal antibodies29,30; HA–ACK2 and Myc–WASP were detected with anti-HA and anti-Myc antibodies, and the recombinant purifiedCdc42GAP was detected by Coomassie blue staining.

© 2000 Macmillan Magazines Ltd