A novel Dictyostelium RasGEF is required for normalendocytosis, cell motility and multicellular developmentAndrew Wilkins*, Jonathan R. Chubb*† and Robert H. Insall‡

Background: Dictyostelium possesses a surprisingly large number of Rasproteins and little is known about their activators, the guanine nucleotideexchange factors (GEFs). It is also unclear, in Dictyostelium or in highereukaryotes, whether Ras pathways are linear, with each Ras controlled by itsown GEF, or networked, with multiple GEFs acting on multiple Ras proteins.

Results: We have identified the Dictyostelium gene that encodes RasGEFB, aprotein with homology to known RasGEFs such as the Son-of-sevenless (Sos)protein. Dictyostelium cells in which the gene for RasGEFB was disruptedmoved unusually rapidly, but lost the ability to perform macropinocytosis andtherefore to grow in liquid medium. Crowns, the sites of macropinocytosis, werereplaced by polarised lamellipodia. Mutant cells were also profoundly defectivein early development, although they eventually formed tiny but normallyproportioned fruiting bodies. This defect correlated with loss of discoidin IγmRNA, a starvation-induced gene, although other genes required fordevelopment were expressed normally or even precociously. RasGEFB wasable to rescue a Saccharomyces CDC25 mutant, indicating that it is a genuineGEF for Ras proteins.

Conclusions: RasGEFB appears to be the principal activator of the RasSprotein, which regulates macropinocytosis and cell speed, but it also appears toregulate one or more other Ras proteins.

BackgroundThe ras subfamily of proto-oncogenes encode membrane-bound GTPases, which have been implicated in the controlof a wide variety of cellular processes [1]. Although initialstudies focused on the role of the Ras protein in growthcontrol in mammalian cells [2,3], Ras signalling pathwayshave now been found in almost all eukaryotes and shownto control processes as diverse as differentiation [4], apop-tosis [5], synaptic transmission [6], cell-cycle arrest andsenescence in mammalian cells [7], cytoskeletal rearrange-ments and pinocytosis [8], cytokinesis [9], chemotaxis [10]and yeast pheromone signalling [11]. Ras GTPases existin two conformations, an active GTP-bound state andan inactive GDP-bound form. The cyclic interconversionbetween these two states is controlled by GTPase-activat-ing proteins (GAPs), which inactivate Ras by stimulatingits weak intrinsic GTPase activity, and guanine nucleotideexchange factors (GEFs), which activate Ras by promot-ing the exchange of bound GDP for GTP. It is throughGAPs and GEFs that extracellular signals are thought tocontrol the activation state of Ras.

The first RasGEF to be identified was Cdc25p of Saccha-romyces cerevisiae, which is required for Ras-mediated acti-vation of adenylyl cyclase and is thus essential forproliferation and spore germination [12]. In Drosophila

melanogaster, the Son-of-sevenless (Sos) protein functionsupstream of Ras in R7 photoreceptor differentiation [13].In mammals, there are at least six RasGEFs: the closelyrelated Sos1 and Sos2 [14,15], RasGRP [16], the closelyrelated RasGRF1 and RasGRF2 proteins [17,18], andCNRasGEF [19]. The phenotypes produced by deletionor inhibition of RasGEFs have sometimes revealed previ-ously unidentified roles for Ras. For example, mice lackingRasGRF1 are impaired in memory consolidation, implicat-ing RasGRF1-mediated Ras signalling in synaptic eventsleading to formation of long-term memory [6]. Thus, anunderstanding of the physiological role of Ras may beaided by an investigation of RasGEF function.

In Dictyostelium, at least eight ras subfamily genes areexpressed at different times throughout the life cycle, fiveof which have been published: rasB, rasC, rasD, rasG andrasS. RasD and RasG are most similar to the prototypichuman H-Ras, both having only six conservative changesin the first 80 residues relative to H-Ras — the regionthought to contain the residues most important for effec-tor interaction. RasD and RasG are 82% identical, withonly two conservative changes in the first 80 residues.RasB, RasC and RasS are more distantly related to theprototypic Ras. Interestingly, the effector region of humanH-Ras (amino acids 32–40) is identical in all Dictyostelium

Addresses: *MRC Laboratory for Molecular CellBiology, University College London, Gower Street,London WC1E 6BT, UK. ‡School of Biosciences,Birmingham University, Edgbaston, BirminghamB15 2TT, UK.

†Present address: MRC Human Genetics Unit,Crewe Road, Edinburgh EH4 2XU, UK.

Correspondence: Robert InsallE-mail R.H.Insall@bham.ac.uk

Received: 13 June 2000Revised: 11 October 2000Accepted: 11 October 2000

Published: 1 November 2000

Current Biology 2000, 10:1427–1437

0960-9822/00/$ – see front matter © 2000 Elsevier Science Ltd. All rights reserved.

Research Paper 1427

Ras proteins except RasS and RasC. The rasB, rasCand rasS genes are expressed throughout the life cycle ofDictyostelium [20–22]; rasG is expressed maximally duringvegetative growth and is undetectable by the onset of aggre-gation [23] whereas rasD expression is restricted to the mul-ticellular stage of development being first detectable aroundthe tipped mound stage and maximal during the slugstage [23].

We have previously described the phenotypic conse-quences of single disruptions of the rasG, rasD and rasSgenes: rasG– cells have impaired cell movement and aredefective in cytokinesis [9]; rasD– cells exhibit impairedphototaxis during the slug stage of development, but nodefects in cell-fate specification or pattern formation areevident [24]; rasS– cells are impaired in solid- and fluid-phase endocytosis, and yet have an increased cell speed[25]. To date, only a single Dictyostelium RasGEF has beendescribed, encoded by the aimless gene [10]. The pheno-type caused by deletion of aimless is different from that ofthe ras null mutants generated so far — aimless– cells failto aggregate upon starvation and are unable to synthesiseor respond chemotactically to cAMP. Thus, otherRasGEFs must exist that regulate the Ras activity requiredfor cellular cytokinesis, endocytosis and for slug phototaxis.Here, we report the identification and characterisation of asecond Dictyostelium gene encoding a RasGEF.

ResultsIdentification and cloning of RasGEFBUsing the tBLASTn algorithm to search the Tsukuba Dic-tyostelium sequencing project databases [26], we identifieda partial cDNA (clone SSK260) with strong homology toknown GEF domains. This partial cDNA clone was usedto probe a random-primed λZAPII cDNA library. Multiple,overlapping, partial cDNA fragments were obtained which,

when assembled, contained a continuous open readingframe. The gene was designated rasgefB. The predictedopen reading frame encodes a protein of 1529 amino acidswith a mass of 174 kDa (Figure 1a). The sequence is char-acterised by stretches of repeated glutamine and asparagineresidues, a feature of unknown significance seen in manyDictyostelium proteins. The carboxy-terminal region of thisprotein contains the putative RasGEF catalytic domaincomprising the Cdc25 homology domain (Figure 1b; thisregion is found in the catalytic domains of known GEFsfor Ras and Ral subfamily proteins) and the REM box(Figure 1c; this region is found only in Ras-specificGEFs). The Cdc25 homology domain and the REM boxshare 32% and 40% amino acid identity, respectively, withthose of the human Sos1 GEF. Outside the conserved cat-alytic domain, the remainder of the RasGEFB proteincontains no common sequence motifs and has no obvioussequence homology with known proteins.

Disruption of rasgefBA rasGEFB gene disruption vector (see Supplementarymaterial) was transfected into AX3 cells, and transformantswere cloned following 7 days of selection in blasticidin S.Of 20 independent clones examined, five were found tocontain a simple disruption in rasgefB. All clones were phe-notypically similar under all assay conditions tested, andtherefore the data for only one rasgefB– clone are presentedhere. The wild-type rasgefB gene was retransfected backinto this clone, under the control of the constitutive act15promoter, as a control for non-specific effects of transfor-mation. This cell line, which will be referred to asrasgefB–rasgefB, behaved like AX3 in all experiments, con-firming that the rasgefB cDNA was fully functional.

The rasgefB cDNA was used to probe a northern blot ofDictyostelium total RNA. A single 5 kb mRNA was detected

1428 Current Biology Vol 10 No 22

Figure 1

The RasGEFB protein contains a carboxy-terminal RasGEF domain. (a) Predictedamino-acid sequence of RasGEFB (GenBankaccession number AF275723). The putativeRasGEF catalytic domain is shaded.(b,c) Alignment of amino-acid sequences ofthe (b) Cdc25 homology domain and (c) Ras-specific REM box from the RasGEFB andSos1 (hSos) catalytic domains. Plusesindicate conservative substitutions.(d) Northern blot. Total RNA (20 µg),extracted from parental AX3 and rasgefB–

cells at the indicated times of development (inhours), was separated on a 1% agarose geland transferred to a nylon membrane. Themembrane was cut horizontally and hybridisedsimultaneously to a radiolabelled fragment ofthe rasgefB cDNA and an IG7 probe used asa loading control.Current Biology

(a)MVVILWGWNENRFNTLDQVISFDFIEENKKLYLKILENEKQQKIYNISNLIGRRMDNYRDNNNNNGNGSNINNNGIIINSSMGSQYFNLNKGIIIVESDNETFLIEEIDYLGKFPILKQFIKCEEFIGISTLEIISNIEKQLSNIVNLKSNTTEQTDRKYKTNLIDFKESIIQLEKDCKNLLKQSNLINPSIKYSVVKSVKFLLKEIKFEFYHGDINIKRIELFQQLKKFLKIHFWNNPNLILNSGSNIDMTDSNDSISSEEINSPNDLNLSLSSFPSISSPSNSSYNNINTLTFSNNNNNNNNNSNNNNNNNNNNNNNNSIINSNNNNNSNNNSNISNSSNNNNNNSNNSNNISINVSKTGNNSVINSPTQSSSCYIDCDDTMSNYTDCTNTTNTNSTTTTTTTTTTYGNNLNVYNNNNNNNTNNNNNNVPYKKPALKMMKSPSFKIQLIQEVYLEIMDQCKKIQNYLYELLDTYLTDEEYINNNNNNNNKDISILFENNTSYCIAQEIFISLIKNSNDFLYQVQLIDLTDTRKKSKELSEVSNQLFYSTQQLFPSISDHLDFIEMDIFTNNNTTTTTTTTTTTTNTTTNPSNTIKSSHSSLLSSSYSSSKDDDIVLTINKVLSKLDTFSKEYTKLIGNNLSSNYYVDSGNNNESVSSPKCGLFLSHSLSTDSPMTYVKRNTSSGSSASTSSYTTPVSSSASLSFSSSTQTPSSAASQHHIQQILQQQQQQQQQQQQQQQQQQPQQIQSQSQQQHQQQQQQQQQQQQQQQQQQQQQQQQQQQQIQSQSQQQQQLNRKPHYSSISPTSMLVYNATGGKFGVVNQNNNSAESTPTFQLNTMDILNSLMGKEGQNSNTNNANNPNNNNNNNNNNNNNNNNNNNNNNNNENKNENKNETNKSGHNYNNSSTSTLSSSTTIGDLANVSSTNSPNSSTPNLLLPPHQFHNHNREYNHSHHHLLSSSSNNVMINNRSIGKSFSTGNFSQPQPIMPSNGIGGSGGGGLGTTLSVGNGGIVGGLQSRSSSFEENQLQYAIESFMEVYNHLTSKDKVDTSIWLEEEKEDINVYYINNTNIDGKFYRSIKYASLNQLILKLTCESNTELRFTKTFITTYRSFTTGDIFLMKLIQRYFTPNLKNIVPYQFVQKIQIPIQLRVLNVLRLWIEQYPSDFQQPPQITTQSTQPIQNSTTQPQPQPQPQQPQPQLQQSTNLQNSTIQQPSLFKMLTAFLNSTRRNGHGQYSDLILKKFKQQKQKDRLQLPIINNGQGIPKAKIFWMKYSTEFIFSQSSTDIAEQLTLLDFDSYKSIEEIELLNQAWSKPDQKTNTPNIASMVNRFNSFSSFVSWAILRENDIKQRSKMMLKMIKICYALYKLSNFNGLLAGLSGLMASGVYRLNHTKSLISKQYQKKFDFLCKFLDTKKSHKVYRDIIHSTCPPLIPYLGIYLTDLTFIEDGNQDEIKGLINFKKREMIFNTILEIQQYQQQGYTIKPKPSVLGFLCELPHMSDKKKFEDDTYEQSILLEPKNSTKLEVMNAKK

(b)RasGEfB: 1272 KYSTEFIFSQSSTDIAEQLTLLDFDSYKSIEEIELLNQAWSKPDQKTNTPNIASMVNRFN 1331 ++ T +++ +IA QLTLL+ D Y+ ++ EL+ W+K D++ N+PN+ M+ +hSos : 768 QFETFDLMTLDPIEIARQLTLLESDLYRKVQPSELVGSVWTKEDKEINSPNLLKMIRHTT 827

RasGEfB: 1332 SFSSFVSWAILRENDIKQRSKMMLKMIKICYALYKLSNFNGLLAGLSGLMASGVYRLNHT 1391 ++ + I+ +++++R ++ ++I+I + L+NFNG+L +S + + VYRL+HThSos : 828 NLTLWFEKCIVEAENFEERVAVLSRIIEILQVFQDLNNFNGVLEIVSAVNSVSVYRLDHT 887

RasGEfB: 1392 KSLISKQYQKKFDFLCKFLDTKKSH-KVYRDIIHSTCPPLIPYLGIYLTDLTFIEDGNQD 1451 + ++ +K +D + ++ + H K Y ++S PP +P++GIYLT++ E+GN+DhSos : 888 FEALQERKRKILD---EAVELSQDHFKKYLVKLKSINPPCVPFFGIYLTNILKTEEGNND 944

RasGEfB: 1452 EI----KGLINFKKRE 1463 + K LINF KRhSos : 945 FLKRKGKDLINFSKRR 960

(c)

(d)

RasGEfB: 1079 IKYASLNQLILKLTCESNTELRFTKTFITTYRSFTTGDIFLMKLIQRY 1127 IK ++ +LI +LT ++ F +TF+TTYRSF + +L LI+R+hSos : 562 IKGGTVVKLIERLTYHMYADPNFVRTFLTTYRSFCKPQELLSLLIERF 610

rasgefB–AX30 4 8 12 16 20 0 4 8

rasgefB

throughout growth and development in AX3 cells, withmaximal expression after 8 hours of starvation (Figure 1d).This mRNA species was absent in rasgefB– cells, confirm-ing the functional deletion of the rasgefB gene.

The rasgefB- cells are defective in axenic growthDictyostelium cells lacking the rasG or rasS genes are unableto proliferate in shaking axenic culture; rasG– cells have acytokinesis defect and become large and multinucleate [9]whereas rasS– cells are unable to perform fluid-phase endo-cytosis [25]. To determine whether the rasgefB mutantshared these phenotypes, the proliferation rate in shakingaxenic culture of vegetative rasgefB– cells relative to theAX3 parent was assessed (Figure 2a). AX3 cells proliferatedwith a doubling time of approximately 8 hours whereas,even after 100 hours, the rasgefB– cell lines failed to prolifer-ate significantly. When viewed under a haemocytometer,rasgefB– cells kept in shaking axenic culture for several daysdid not exhibit the huge increase in cell size seen inshaking, axenic rasG– cells (data not shown), suggesting thatthe proliferation defect is not due to a failure in cytokinesis.The failure of cells to grow in normal axenic mediumcould be due to reduced uptake of nutrients. Alternatively,rasgefB– cells could be defective in some form of nutrientsensing, for example, for glucose or amino acids, which isrequired for axenic growth. We therefore measured growthin MA medium, a richer medium designed for growth ofcells that lack axenic mutations. As shown in Figure 2b,rasgefB– cells proliferated in MA medium, albeit slowly.Under these conditions, we were unable to grow NC4, theparent of AX3 (Figure 2b). Similarly, rasgefB– cells couldproliferate when shaken in a dense bacterial suspension,albeit at a reduced rate relative to AX3 cells; rasgefB– cellshad a doubling time of 7 hours whereas AX3 cells doubledevery 4 hours under these conditions (Figure 2c).

The osmolarity of HL5 axenic medium is fivefold higherthan that of the KK2 buffer used for the bacterial suspen-sion experiments [27]. As rasgefB– cells were capable ofproliferation in a suspension of bacteria but not when cul-tured in axenic medium, it was therefore possible that theinability of rasgefB– cells to proliferate was a consequenceof the high osmolarity of the axenic medium. To addressthis, the ability of rasgefB– cells to proliferate in HL5 axenicmedium containing a dense suspension of heat-killed bac-teria was assessed. As in the previous set of experiments,the rasgefB– cells had a doubling time of 7 hours whereasAX3 cells doubled every 4 hours (Figure 2d). It thereforeseems likely that the proliferation defect of rasgefB– cellsin axenic medium was not caused by the change in osmo-larity experienced by the cells when transferred to axenicmedium. As a consequence of the proliferation defect ofrasgefB– cells in axenic culture, cells had to be maintainedon bacterial lawns or in a shaking bacterial suspension. Allthe results described below were obtained using cells frombacterial lawns.

The rasgefB– cells are impaired in both fluid-phaseendocytosis and phagocytosisOnly cells capable of efficient fluid-phase endocytosis canproliferate in liquid medium. A number of Dictyosteliummutants that have impaired axenic growth show a corre-sponding impairment in fluid-phase endocytosis [25,27].To investigate the fluid-phase endocytosis of rasgefB–

cells, their ability to take up the fluid-phase marker tetram-ethylrhodamine isothiocyanate (TRITC)–dextran in axenicculture was assessed (Figure 3a). AX3 cells took upTRITC–dextran rapidly, reaching a maximum after approx-imately 125 minutes, whereas rasgefB– cells showed almostno accumulation of the marker during this time. Thus, itseems that rasgefB– cells are incapable of normal fluid-phase endocytosis in axenic medium. This defect is con-sistent with the impaired proliferation of rasgefB– cells inaxenic medium.

As rasgefB– cells showed a reduced proliferation rate inbacterial suspension, it was possible that they were also

Research Paper Dictyostelium RasGEFB regulates endocytosis and cell speed Wilkins et al. 1429

Figure 2

Proliferation of wild-type and rasgefB– cells in axenic culture and bacterialsuspension. Cell proliferation in (a) shaken axenic culture, (b) richmedium, (c) shaken bacterial suspension, and (d) shaken bacterialsuspension in axenic medium. (a,c,d) AX3 and rasgefB– cells wereseeded at approximately 2 × 105 per ml into (a) flasks of HL5 medium, orinto a suspension of (c) live Escherichia coli B/r strain in KK2 medium or(d) heat-killed E. coli B/r in HL5 medium. (b) NC4 and rasgefB– cellswere seeded at approximately 2 × 104 per ml into flasks of MA medium.All cultures were shaken at 150 rpm and, at the indicated times, the cellnumber was determined using a haemacytometer. In all cases, the mean± SD of three independent cultures is plotted against time.

Current Biology

Time (h)Time (h)0 20 40 60 80 100

0 5 10 15 20Time (h)

rasgefB–AX3

(a)

(c) (d)

(b)

Time (h)0 5 10 15 20

rasgefB–AX3

rasgefB–AX3

rasgefB–NC4

0 50 100 150105

106

107

108

Num

ber o

f cel

ls p

er m

l105

106

107

Num

ber o

f cel

ls p

er m

l

105

106

Num

ber o

f cel

ls p

er m

l105

106

107

Num

ber o

f cel

ls p

er m

l

impaired in phagocytosis. The ability of rasgefB– cells tophagocytose live bacteria was assessed (Figure 3b). TherasgefB– cells reduced the optical density of a bacterial sus-pension at about a third of the rate of AX3 cells. TherasgefB– cells were also severely impaired in phagocyto-sis of Saccharomyces cells — AX3 cells took up the yeastcells rapidly, reaching a maximum after approximately150 minutes, whereas rasgefB– cells showed almost noaccumulation during this time (Figure 3c). Therefore,the reduced proliferation rate of rasgefB– cells in bacterialsuspension correlates with impairment of phagocytosis. Itshould be noted that these phagocytosis experiments usedbacteria at a tenfold lower density than that used for theproliferation assays described in Figure 2c. There is there-fore no inconsistency between the ability of cells to prolif-erate in bacterial suspension and the phagocytosis defectsobserved here.

Aberrant morphology of rasgefB– cellsDictyostelium cells lacking RasG or RasS have aberrant cel-lular morphology when attached to surfaces: rasG– cellsappear flattened and non-polar [9] whereas rasS– cells areelongate and highly polarised, with prominent pseudopods[25]. To examine the cellular morphology of the AX3 andrasgefB– cells from axenic culture, scanning electronmicroscopy (SEM) was performed (Figure 4). AX3 cellshad a rounded morphology with some membrane rufflesbut few filopods. Circular ruffles or crowns were promi-nent on the majority of the cells, usually on the dorsalsurface (214 crowns were present on 200 AX3 cells).These crowns are the formation sites of macropinosomes,the vesicles responsible for fluid-phase uptake by axeniccells [28]. In contrast, rasgefB– cells were strikingly moreflattened, elongate, polarised and dominated by a singlelarge lamella. Although there were an increased numberof membrane protrusions in the rasgefB– cells, none ofthe rasgefB– cells examined possessed the dorsal circular

ruffles typical of normal crowns. The absence of normalcrowns correlates well with the severe fluid-phase endocy-tosis defect of rasgefB– cells. The morphology of rasgefB–

cells was similar to that of rasS– cells, although the polarityof rasgefB– cells was if anything more exaggerated. It wasalso similar to that of normal Dictyostelium cells that havedeveloped to an aggregation competent state [29].

There were also striking differences between rasgefB– andAX3 cells in filamentous actin (F-actin) distribution(Figure 5). In axenic AX3 cells, the majority of F-actin wasrecruited into the macropinocytic crowns. In rasgefB– cells,which did not have crowns, F-actin was usually localisedin a single large lamella at the leading edge of the cell.These F-actin distributions were consistent with the cel-lular morphologies observed by SEM (Figure 4). Althoughthere were fewer differences between rasgefB– and AX3cells when they were grown on bacteria (data not shown),the F-actin distribution remained markedly different inthe two cell types. In bacterially grown AX3 cells, themajority of F-actin was recruited into several smallpseudopods whereas in rasgefB– cells the majority of F-actin was concentrated in a single large lamella (Figure 5).

The rasgefB– cells move unusually rapidlyCellular morphology and F-actin distribution were similarin rasgefB– and rasS- cells. In the case of rasS– cells, thiscorrelates with a threefold increase in the speed of move-ment of vegetative cells [25]. To determine whether thiswas also true for rasgefB–, cells were transferred to tissueculture plates and time-lapse images collected to allowanalysis of cell speed. Bacterially grown AX3 cells movedat a rate of 6 ± 1.6 µm/minute whereas rasgefB– cells movedat 13.3 ± 2.1 µm/minute. In the case of axenic cells, AX3cells moved at a rate of 1.8 ± 0.3 µm/minute whereasrasgefB– cells moved at 15.2 ± 4.1 µm/minute. Thereforeunder both sets of conditions, the speed of rasgefB– cells

1430 Current Biology Vol 10 No 22

Figure 3

(a) Fluid-phase endocytosis, and(b,c) phagocytosis of (b) bacteria and(c) S. cerevisiae by wild-type and rasgefB–

cells. AX3 and rasgefB– cells were seeded at(a) 5 × 106 per ml into flasks of HL5 mediumcontaining 2 mg/ml TRITC–dextran, or(b) 1 × 106 per ml into flasks containing asuspension of E. coli B/r in KK2 medium, or(c) 2 × 106 per ml into flasks of KK2containing a suspension of TRITC–dextran-labelled yeast (109 cells per ml). All cultureswere shaken at 150 rpm; 1 ml of each culturewas removed at various times and (a) theamount of fluid uptake, or (c) the relativeuptake of yeast, determined using afluorimeter, or (b) the optical density at600 nm determined using aspectrophotometer. In (a,c), the graphs show

the data points from three separateexperiments, the best-fit curve being fitted totheir mean value. In (b), the graph shows the

mean ± SD of three independent cultures; thedecrease in the density of the bacterial culturewas taken as a measure of phagocytosis.

Current Biology

Time (min)

Fluo

resc

ence

Time (min)

A60

0

0 25 50 75 100125

rasgefB–AX3

Time (min)

Fluo

resc

ence

0 30 60 90120150

rasgefB–AX3

0 100 200 300 400

rasgefB–AX3

(a) (b) (c)

0

10

20

30

40

50

60

0.2

0.3

0.4

0.5

0.6

0.7

05

10152025303540

was significantly greater than that of AX3 cells (two sidedt-test, p < 0.001). This increase in cell speed is consistentwith the observed F-actin distribution and cellular mor-phology of rasgefB– cells.

This unusually fast movement did not appear to compro-mise the chemotaxis of rasgefB– cells. When challengedwith a cAMP-containing micropipette , they moved up thecAMP gradient with great accuracy (Figure 6). In commonwith most highly polar cells, they most often moved byturning an established leading edge (Figure 6, first threepanels), but would also change cell polarity by advancingnew pseudopods up the gradient if stimulated appropri-ately (Figure 6, last six panels). Wild-type cells were lesschemotactic under these conditions (data not shown),either because of their slower movement or because theywere better at generating their own, competing, signals.To test the dose-response of chemotaxis, a two-drop assaywas performed in which approximately 0.1 µl of a cAMPsolution was spotted near a 0.1 µl drop of cells [30]. Underthese conditions, rasgefB– cells showed positive chemo-taxis to cAMP to the same extent and over the same con-centration range as AX3 cells. The weak aggregation ofrasgefB– cells (shown below) is therefore not due to asimple defect in chemotaxis.

Defects in multicellular developmentWhen spotted onto a lawn of bacteria, wild-type cells growby phagocytosis, clearing a plaque in the lawn, whichincreases in size as the cells proliferate. Roughly two

zones are visible in the plaque; a zone on the peripherywhere the bacterial lawn has thinned but is not completelycleared, called the ‘feeding front’, and a central zone inwhich the bacteria have been cleared and the Dictyosteliumcells are in various stages of multicellular development.When a similar amount of rasgefB– cells was spotted onto abacterial lawn, the plaque formed was strikingly differentfrom that of AX3 cells (Figure 7a). The plaques of rasgefB–

cells were two or three times larger than those of AX3cells. Although the bacterial lawn was thinned, rasgefB–

plaques were never fully cleared of bacteria. In addition,although some loose aggregates of cells appeared, therasgefB– plaques contained relatively few small fruitingbodies. The increased speed of vegetative rasgefB– cellscoupled with their reduced phagocytic rate providesan explanation for the large diffuse plaques formed byrasgefB– cells on bacterial lawns. The mutant cells are highlymotile, and thus can move faster across the bacterial lawn,causing an increase in colony diameter. Furthermore, therasgefB– cells only partially cleared the bacterial lawn asthey moved out, a phenotype consistent with the impairedphagocytosis observed in the mutant line. The size ofplaques formed by rasgefB– cells was strikingly similar tothat of plaques formed by rasS– cells [25]; however, themulticellular development of rasS– cells on bacterial lawnsis apparently normal.

Research Paper Dictyostelium RasGEFB regulates endocytosis and cell speed Wilkins et al. 1431

Figure 4

Morphology of wild-type and rasgefB– cells. Vegetative (a,b) AX3 and(c,d) rasgefB– cells were placed in HL5 medium for 24 h, thenprepared for SEM as described in the Materials and methods.Arrowheads indicate crowns, the dorsal, circular ruffles that are thesites of macropinosome formation. Although present on the majority ofAX3 cells, these structures were never seen on rasgefB– cells.

Current Biology

(a)

(b)

(c)

(d)

20µm 20µm

5µm 5µm

Figure 5

Distinct patterns of F-actin distribution in rasgefB– cells. (a) ParentalAX3 and (b) rasgefB– cells were cultured in HL5 medium for 24 hbefore fixation and staining with Texas red–phalloidin as described inthe Materials and methods. In the parent, F-actin accumulated mainlyin the crowns, which are the normal site of macropinosomeformation. This staining pattern was never seen on rasgefB– cells, inwhich F-actin accumulated in a single large pseudopod at theleading edge of each cell. Vegetative, bacterially grown (c) AX3 and(d) rasgefB– cells were similarly fixed and stained. There was aconcentration of F-actin in a single large pseudopod at the leadingedge of rasgefB– cells, which was not seen in AX3 cells. The scalebars represent 5 µm.

Current Biology

(a)

(d)(c)

(b)

Development of rasgefB– cells on nitrocellulose filters oron non-nutrient agar was also aberrant (Figure 7b). AX3cells completed multicellular development within 24 hours,forming spore-containing fruiting bodies. After 24 hours,rasgefB– cells had only begun to form very small aggre-gates. Over the next 12 hours, these structures proceededto undergo apparently normal morphogenesis, resulting inminute but morphologically normal fruiting bodies thatcontained mature, functional spores. Although belatedlyable to aggregate, rasgefB– cells never formed streams,which are characteristic of efficient aggregation, whendeveloped on non-nutrient agar (data not shown). Thedevelopmental defect of rasgefB– cells was unaffected by

synergistic development with AX3 cells and thereforeappears cell autonomous (see Supplementary material).

Gene expression during early developmentGene expression is tightly regulated throughout thedevelopment of Dictyostelium and certain genes serve asmarkers of the extent of development. One of the earliestgenes expressed after starvation is the gene encoding thelectin discoidin-Iγ; disIγ expression is first detectable after2 hours of starvation and expression is maximal afterapproximately 5 hours, thereafter its expression isrepressed by the pulses of extracellular cAMP that consti-tute the aggregation stimulus [31]. The next set of genesto be expressed includes those encoding the aggregation-stage cAMP receptor cAR1 and the adenylyl cyclase ACA,which are required to generate the extracellular cAMPthat constitutes the aggregation stimulus; cAMP signallingis also required for induction of a number of aggregationstage genes including the cell–cell adhesion molecule csA[32]. To analyse the physiological state of rasgefB– cellsduring growth and early development, we analysed theexpression of these early developmental genes in AX3and rasgefB– cells upon starvation (Figure 8a–c). In thecase of carA and acaA, mRNA levels were measured withor without treatment with exogenous cAMP pulses every6 minutes to mimic normal signalling. During the firstfew hours of development, AX3 cells exhibited typicalexpression patterns for disIγ, carA and acaA mRNA, andcsA protein (Figure 8a–c); rasgefB– cells showed similarexpression profiles for carA and acaA in the absence ofexogenous cAMP, and similar induction of the csAprotein. However, disIγ expression was almost entirelyabsent in rasgefB– cells (Figure 8a), and carA and acaA

1432 Current Biology Vol 10 No 22

Figure 6

Chemotaxis of rasgefB– cells. Bacterially grown rasgefB– cells weredeveloped in shaken suspension, in the presence of exogenouscAMP pulses, for 5 h; 5 × 106 cells were allowed to adhere to a24 mm glass coverslip, then stimulated by an adjacent micropipettecontaining 10 µM cAMP. The micropipette was moved at 220,290 and 350 sec.

Current Biology

0 60 sec 215 sec

225 sec 245 sec 285 sec

305 sec 330 sec 350 sec

355 sec 400 sec 420 sec

Figure 7

Development of wild-type and rasgefB– cells. (a) Vegetative AX3 andrasgefB– cells (103 of each) were spotted onto lawns of Klebsiellaaerogenes bacteria on SM agar plates. After 7 days’ incubation at22°C, the resultant plaques were photographed. (b) Vegetative AX3and rasgefB– cells were deposited on nitrocellulose filters at a densityof 3 × 106 cells/cm2, and photographed after 36 h of development. Thescale bars represent 1 mm.

Current Biology

rasgefB–AX3(a)

(b) AX3 rasgefB– rasgefB–

were expressed precociously in cAMP-treated cells. Theseresults have a number of important implications for under-standing of the phenotype of rasgefB– cells.

Firstly, rasgefB– cells had the properties of aggregatingDictyostelium cells: highly polarised morphology, a largenumber of filopods and a dramatic decrease in the rateof phagocytosis [29,33]; they were also highly motile,moving approximately three times faster than vegetativecells [34]. It therefore seemed possible that the pheno-type of rasgefB– cells might reflect premature develop-ment to an aggregation-competent state. However, rasgefB–

cells did not express any early developmental or aggrega-tion-stage genes during vegetative growth. Thus, theincreased cell speed, reduced endocytic rate andpolarised morphology of rasgefB– cells is not a simple con-sequence of the cells being inappropriately advancedinto development.

Secondly, although unable to aggregate, rasgefB– cells wereable to induce expression of acaA and carA normally uponstarvation. Moreover, the cells would appear to be capableof generating functional pulses of cAMP as the csA proteinwas induced normally in rasgefB– cells. Thus, the aggrega-tion defect of rasgefB– cells does not appear to be a conse-quence of the lack of a functional cAMP relay system or,as previously mentioned, an inability to perform normalchemotaxis to cAMP upon starvation. This may indicatethat RasGEFB controls some component of an as yetunidentified pathway that is required for aggregation.

Thirdly, although carA, acaA and csA expression werenormal in rasgefB– cells, there was clearly a defect in earlygene expression. The disIγ gene is, in part, induced byfactors secreted in response to nutrient depletion andincreasing cell density and is usually a good indicator ofthe onset of starvation [35]. It was therefore surprisingthat, although rasgefB– cells appeared to enter the devel-opmental programme, they failed to induce disIγ expres-sion. In addition, although able to respond to pulses ofexogenous cAMP, the rasgefB– cells seemed to be hyper-sensitive to cAMP, such that they maximally inducedacaA and carA expression several hours earlier than theAX3 parent cells. Taken together, these data suggest thatthe normal control of early developmental gene expres-sion is lost in rasgefB– cells, implicating RasGEFB incontrol of the early cellular response to starvation. Thecell-type specific genes ecmA and pspA were eachexpressed at apparently normal levels during multicellu-lar development (see Supplementary material), whichindicates that loss of the rasgefB gene does not greatlyaffect either the differentiation or the proportioning ofprestalk and prespore cells.

Complementation of a S. cerevisiae CDC25 mutant byrasgefBAlthough sequence homology predicts that RasGEFB willfunction as a GEF for Ras, an inability to produce soluble,recombinant RasGEFB protein precluded a direct demon-stration of this in vitro. The temperature-sensitive CDC25-5S. cerevisiae mutant has previously proved useful for assess-ing the catalytic activity of putative RasGEFs. Expressionof the catalytic domains of the murine RasGEFs, Sos andRasGRF, will complement the phenotype of this yeastmutant, allowing growth at the non-permissive tempera-ture (37°C) [36,37]. However, expression of the S. cerevisiaeBUD5 gene, which encodes a GEF for the Rap-relatedBud1 protein, is unable to complement the CDC25-5mutant [38]. To demonstrate that rasgefB encodes a func-tional RasGEF, DNA encoding the putative catalyticdomain (residues 1078–1463) was transfected into CDC25-5under the control of the strong ADH promoter (see Materi-als and methods). Expression of the putative catalyticdomain of RasGEFB allowed growth of CDC25-5 at thenon-permissive temperature in a similar manner to a

Research Paper Dictyostelium RasGEFB regulates endocytosis and cell speed Wilkins et al. 1433

Figure 8

Gene expression during early development of wild-type and rasgefB–

cells. (a) Vegetative AX3 and rasgefB– cells were plated at a density of3 × 106 cells/cm2 on nitrocellulose filters and allowed to develop. Cellswere harvested at the indicated times (in hours) and total RNAprepared; 20 µg RNA was separated on a 1% agarose gel, transferredto nylon membrane and hybridised with a radiolabelled disIγ cDNAprobe. (b) As in (a), except that harvested cells were dissolved inSDS–PAGE sample buffer, separated on a 10% polyacrylamide geland electroblotted onto a nitrocellulose membrane. The blot wasprobed with a monoclonal antibody to csA protein, and visualisedusing a Pierce Supersignal kit. (c) AX3 and rasgefB– cells werestarved in a suspension of KK2 at 5 × 107 cells per ml, with or withoutaddition of 100 nM cAMP every 6 min. Cells were harvested at theindicated hours of development and total RNA prepared; 20 µg RNAwas separated on a 1% agarose gel, transferred to a nylon membraneand hybridised with radiolabelled carA and acaA cDNA probes.

Current Biology

(a)0 62 4 8 10 0 62 4 8 10

rasgefB–AX3

(b)0 62 4 8 10

AX3 rasgefB–

0 62 4 8 10

(c)0 62 4 80 62 4 8

AX3 rasgefB–

+ cAMP

– cAMP

disIγ

csA

acaA

carA

acaA

carA

plasmid expressing the Cdc25 catalytic domain [12],whereas transformation with the pAD4 vector alone [39]did not (Figure 9). This strongly supports the idea thatRasGEFB functions as a genuine RasGEF in vivo.

DiscussionWe have identified a novel Dictyostelium RasGEF gene,rasgefB. Dictyostelium cells lacking rasgefB were unable toproliferate in axenic culture and exhibited a reduced rateof proliferation in a bacterial suspension. These prolifera-tion defects correlated with profound defects in fluid-phase endocytosis and phagocytosis. In axenic culture,rasgefB– cells had a highly polarised morphology, with anincreased number of membrane protrusions and a totalabsence of crowns, the sites of macropinosome formation.In addition, vegetative rasgefB– cells moved at about twicethe speed of the parental AX3 cells. Although capable ofcAMP chemotaxis, rasgefB– cells were unable to formnormal aggregation streams upon starvation, resulting inthe formation of small aggregates that developed intosmall but morphologically normal fruiting bodies.

Activation of RasGEFBAlthough the predicted protein sequence of RasGEFBcontains a putative RasGEF catalytic domain, the remain-der of the sequence has no obvious homology to knownproteins. In general, RasGEFs are thought to requirerecruitment to cellular membranes in order to come intocontact with their Ras protein targets [40]. In otherRasGEFs, this seems to be mediated via adaptor proteinssuch as Grb2 [41], which binds to a polyproline region onthe Sos RasGEF via its two Src homology 3 (SH3) domains,or by other domains in the molecule such as pleckstrin

homology (PH) domains, which can bind free βγ subunitsof heterotrimeric G proteins or to various membrane phos-pholipids. As RasGEFB contains no obvious means ofmembrane recruitment, it not clear how it is able to acti-vate its target Ras proteins. Preliminary data from expres-sion of a green fluorescent protein (GFP)–RasGEFB fusionprotein suggested that RasGEFB is cytoplasmicallylocalised and not constitutively membrane anchored. Itwill be interesting to determine whether RasGEFB proteinis recruited to the plasma membrane, and if so which cel-lular signals cause it to relocalize.

Signalling pathways downstream of RasGEFBCells disrupted in rasS– cells show a strikingly similar phe-notype to that of vegetative rasgefB– cells — they arestrongly impaired in solid- and fluid-phase endocytosisand have an increased cell speed [25]. The developmentaldefect of rasgefB– cells is, however, not apparent in any ofthe known Ras mutants. RasS may therefore be a directtarget of RasGEFB in a signalling pathway controllingthe balance between endocytosis and motility whereasanother, as yet uncharacterised Ras, perhaps RasB or RasC,may mediate the signals from RasGEFB required fornormal aggregation. Alternatively, RasS may mediate allsignalling from RasGEFB, but other Ras proteins couldtake over its roles during development. Single and multi-ple mutants for all the Dictyostelium ras genes will need tobe generated to resolve this question.

Of the accepted Ras effectors, only the class I phospho-inositide (PI) 3-kinases have been found in Dictyosteliumcells so far. Interestingly, cells lacking two class I PI 3-kinases (PIK1 and PIK2) show diminished fluid-phaseendocytosis and increased motility [42,43]. The endocyto-sis defect of pik1–/pik2– cells is far less severe than bothrasgefB– and rasS– mutants, but this could reflect partialcompensation by PIK3. Although a direct link betweenRasGEFB, RasS and PI 3-kinase has not yet been demon-strated, it seems likely that these three proteins are allmembers of the same signalling pathway controlling endo-cytosis and motility

Endocytosis and motility defectsThe loss of rasgefB resulted in a profound impairment offluid-phase endocytosis. This correlated with an inabilityto proliferate in axenic culture. The lack of nutrient avail-ability caused by diminished uptake of medium is a proba-ble reason for the proliferation defect of rasgefB– cells inaxenic culture, although it has not been possible todemonstrate this directly. The reduced rate of fluid-phaseendocytosis was accompanied by a reduction in the rate ofphagocytosis and an increase in cell speed. Together,these additional phenotypes of rasgefB– cells provide anexplanation for the enlarged and diffuse colony morphol-ogy seen when rasgefB– cells are grown on bacterial lawns.

1434 Current Biology Vol 10 No 22

Figure 9

Complementation of a S. cerevisiae CDC25 mutant by the putativecatalytic domain of RasGEFB. The temperature-sensitive CDC25S. cerevisiae strain LV25-5 was transformed with plasmids expressingeither the catalytic domain of Cdc25p or the putative catalytic domainof RasGEFB (residues 1078–1463). An empty pAD4 vectortransformation was used as a negative control. See the Materials andmethods for details of the rasgefB expression construct. Three clonesfrom each transformation were patched onto two separate Leu–

synthetic medium agar plates and incubated at either 30°C or 37°C.Colony growth was assessed after 5 days.

30ºC 37ºC

Current Biology

Vector only

CDC25

rasgefB

The cells were able to move outwards from the centre ofthe plaque more rapidly, creating a plaque of greater size,while the cells were unable to phagocytose the bacteria asrapidly as wild-type cells, resulting in partial clearance ofthe bacterial lawn and the characteristically diffuseappearance. This observation also serves to highlight thedanger of using plaque size as a measure of cell prolifera-tion rate; rasgefB– cells would appear to be growing fasterthan wild-type cells as the rasgefB– plaque size increased ata faster rate, but examination of the proliferation rate ofrasgefB– cells in shaking bacterial suspension showed thatthe cells proliferate more slowly.

A body of evidence suggests that the processes of endocy-tosis and cell migration are in constant competition withone another. Competition arises between pseudopods andthe membrane protrusions which mediate phagocytosisand macropinocytosis, as they are projected. It appearsthat the two structures can only coexist in the same cell fora short time [44]. Leading edges have been observed torecruit the actin-binding protein coronin from regressingphagocytic cups, and extending cups can sequester coroninfrom retracting pseudopods [44]. It therefore appears thatcompetition for cytoskeletal components limits the extentto which a cell can perform phagocytosis and migrate atthe same time. These data suggest an inverse correlationbetween cell speed and the intrinsic rate of endocytosisin the Dictyostelium cell. It may be that the function of theRasGEFB/RasS signalling pathway is to regulate thisbalance between endocytosis and motility, favouringendocytic feeding during growth and motility while cellsare aggregating.

Developmental defectsThe rasgefB– cells were both delayed in multicellulardevelopment and unable to aggregate normally. Althoughable to initiate expression of the aggregation-stage cAMPreceptor and adenylyl cyclase, rasgefB– cells failed toinduce expression of discoidin upon starvation. Surpris-ingly, rasgefB– cells were able to perform normal chemo-taxis to cAMP and seemed capable of producing pulses ofcAMP as the csA protein was induced normally. SomeDictyostelium mutants are unable to aggregate normally bychemotaxis and do so by random motion and adhesion[45,46]. In most cases, although the resultant aggregatesare smaller then normal, the time-scale of development issimilar to that of the parental cells. The delay in multi-cellular development in rasgefB– cells suggests that thereis a partial block in development, in addition to thedefect in aggregation. The lack of disIγ induction isunlikely to be the cause of the developmental delay orthe aggregation defect as cells lacking discoidin-I are ableto aggregate normally upon starvation [47]. However, theabsence of disIγ may indicate that the developmentaldelay is caused by a lack of induction of a subset of genesnecessary for aggregation. Alternatively, all the necessary

cellular components required for normal developmentmay be present but signalling through RasGEFB may berequired to attenuate a signal inhibitory to aggregation.Although required for normal aggregation, it does notappear that RasGEFB is required for cell-fate specifica-tion during later development as both prestalk and pres-pore cell mRNAs were expressed.

Physiological function of rasgefBIn several respects, vegetative rasgefB– cells behaved likeaggregation-competent amoebae [4]. They were highlypolarised, highly motile, and they had a reduced rate ofendocytosis. It was therefore surprising to find that themutant did not express early developmental or aggrega-tion-stage marker genes during growth. A possible expla-nation for this phenotype that fits the available data is thatthere are independent, parallel, rather than sequential,developmental signalling pathways required for the pro-gression of cells to an aggregation-competent state. Lossof one of these signalling pathways, perhaps involvingRasGEFB, might give rise to cells that manifest a specificbehaviour of starved amoebae, such as rapid migration,without any of the others. The physiological function ofRasGEFB may therefore be to control the balancebetween endocytosis and motility in the context of theearly developmental program, favouring endocytic feedingduring vegetative growth and motility during early devel-opment. In addition, signalling through RasGEFB maybe required to allow aggregation, either by inducing theexpression of genes required for aggregation or attenuat-ing an aggregation-blocking signal. The loss of RasGEFBwould therefore result in an inappropriate increase inmotility and concomitant decrease in endocytosis in theabsence of starvation and prevent aggregation.

Materials and methodsUnless otherwise stated, all chemicals were obtained from Sigma and allrestriction and DNA modifying enzymes were from New England Biolabs.

Cloning of rasgefB cDNA and vector constructionThe Tsukuba Dictyostelium cDNA sequencing project databases [26]were searched using the tBLASTn search engine for clones withhomology to the catalytic domain of the putative RasGEF encoded byaimless [10]. One clone, SSK260, was identified. This partial cDNAclone was obtained from T. Morio and the full DNA sequence deter-mined. To obtain a full-length rasgefB cDNA, a λZAPII cDNA library(Stratagene) was screened using the entire cDNA insert of cloneSSK260, radiolabelled by the random primer method using[α-32P]dATP (Amersham). Several partial cDNA clones were isolated,which when assembled, encompassed the entire rasgefB openreading frame. The rasgefB gene disruption vector was constructedby inserting the blasticidin resistance gene (bsr) from pBsr∆Bam[9,48] into the single BstZ117 site of a 1.6 kb, Spe I–Not I fragment of3′ rasgefB cDNA cloned into the EcoRI and Not I sites of pBluescriptKS+. From the resulting vector (pATW71), the 3.9 kb rasgefB/bsrfragment was excised using EcoRI and Not I and used to transformDictyostelium strain AX3 as described above. A rasgefB expressionvector was constructed by inserting the entire rasgefB cDNA into theHind III and XbaI sites of plasmid pDXA [49] to create plasmidpATW118. This vector expresses rasgefB from the strong constitutiveact15 promoter and was used to reintroduce the rasgefB cDNA into

Research Paper Dictyostelium RasGEFB regulates endocytosis and cell speed Wilkins et al. 1435

rasgefB– cells. The rasgefB gene disruption vector (pATW71) and therasgefB cDNA expression vector (pATW118) were transfected intocells by electroporation as previously described [9]. To assemble theyeast complementation vector, pATW99, containing the 3′ half of therasgefB cDNA, was digested with BstB I and Not I and a double-stranded oligonucleotide inserted (5′-GGCCTGCAGATGTT-3′ and5′-CGAACATCTGCA-3′) to create pATW123. This resulted in a Pst Iand in-frame ATG translational initiation codon being added before theputative catalytic domain (residues 1078–1463) of rasgefB. The mod-ified GEF domain was excised from pATW123 by Pst I/Sal I digestionand inserted into the Pst I/XhoI sites of pAD4 to create pATW124.This vector expresses the transgene using the strong constitutiveADH1 promoter.

Endocytosis assaysFluid-phase endocytosis was determined by a modification of previousmethods [27,44]. Briefly, bacterially grown cells were pelleted at2000 × g for 2 min, washed free of bacteria with KK2 and plated at5 × 106 cells per plate in HL5 on 90 mm tissue culture plates for 24 h.Cells were harvested by trituration, pelleted, then resuspended at5 × 106 cells/ml in 10 ml HL5 containing 2 mg/ml TRITC–dextran(MW 70000) and shaken at 150 rpm; 1 ml of culture was removed atregular intervals and 100 µl trypan blue added to quench extracellularfluorescence. Cells were then washed once, resuspended in 1 ml KK2and intracellular fluorescence measured by Fluorimeter (KontronSFM25). Excitation was at 544 nm and emission read at 574 nm. Foreach cell line, the assay was performed on three separate occasionsusing the same batch of TRITC–dextran. Results are presented interms of relative fluorescence with the data points from each separateexperiment and their mean value displayed on the same graph. Phago-cytosis of bacteria was performed as described previously [50]. Dic-tyostelium cells were inoculated at a density of 106 cells per ml into a20 ml suspension of E. coli strain B/r at 109 cells per ml in KK2; 1 mlculture was removed at regular intervals and the optical density at600 nm measured. Each assay was performed in triplicate, the meanand SD plotted against time and an estimate of phagocytic rate calcu-lated from the best-fit straight line. Phagocytosis of TRITC-labelledyeast was performed as described [44]. The assay was performedexactly as the fluid-phase endocytosis assay above except that vegeta-tive cells were assayed at 2 × 106 in 10 ml KK2 containing TRITC–dextran-labelled yeast at 109 cells/ml (a kind gift from M. Maniak).

MicroscopyVisualisation of F-actin in vegetative or axenic cells was performedexactly as described [9]. SEM was performed as follows. Cells werefixed on glass coverslips using 1% osmium tetroxide in KK2 for 5 secfollowed by 2% glutaraldehyde in KK2 for 2 h. The fix was removedand the samples were progressively dehydrated through an ethanolseries of 30–100% ethanol, then taken into hexamethyl disilane. Afterair-drying, the cells were sputter-coated in gold, and viewed on a JeolJSM5410 scanning electron microscope. The mean speed of bacteri-ally grown or axenic cells was estimated using time-lapse recordingsof digital phase-contrast images of the cells. Bacterially grown cellswere washed free of bacteria, plated in KK2 onto 90 mm tissueculture plates and allowed to adhere for 15 min. Axenic cells wereassayed in HL5 medium in 90 mm tissue culture plates; 40 frameswere captured at 20 sec intervals. Cell centroids were determined andthe mean centroid displacement between each frame was calculatedby the software.

Yeast strains, growth and transformationS. cerevisiae strain LV25-5 [12] (a strain carrying the temperature-sen-sitive CDC25-5 allele) was maintained at 30°C in YPD medium or onYPD agar plates. Yeast transformations were performed according tothe TRAFO protocol [51] and transformants selected on Leu– drop-outmedia agar plates (Bio101). Transformants were patched onto Leu–

plates and incubated at 30°C (permissive temperature) or 37°C (non-permissive temperature) to assess suppression of the temperature sen-sitive growth defect.

Supplementary materialSupplementary material including further methodological detail, twofigures showing the strategy used to disrupt the rasgefB gene, and theexpression of the ecmA and pspA genes during multicellular developmentin rasgefB– cells, is available at http://current-biology.com/supmat/supmatin.htm.

AcknowledgementsWe gratefully acknowledge Y. Tanaka and the D. discoideum cDNAproject in Tsukuba, Japan for the initial 3′ rasgefB clone. We thank AdrianHarwood for advice and assistance, Laura Machesky and two referees forconstructive comments on the manuscript, Markus Maniak and CaroleParent for help with experimental methods, and Jan Faix and GüntherGerisch for the monoclonal antibody against csA. R.H.I. was supported byWellcome Trust Career Development Fellowship 043754 and by an MRCSenior Fellowship.

References1. Valencia A, Sander C: The ras superfamily. In Guidebook to the

Small GTPases (eds Zenal M, Huber LA). Oxford: Oxford UniversityPress; 1995:12-19.

2. Stacey DW, Kung HF: Transformation of NIH 3T3 cells bymicroinjection of Ha-ras p21 protein. Nature 1984, 310:508-511.

3. Feig LA, Cooper GM: Inhibition of NIH 3T3 cell proliferation by amutant ras protein with preferential affinity for GDP. Mol Cell Biol1988, 8:3235-3243.

4. Kayne PS, Sternberg PW: Ras pathways in Caenorhabditiselegans. Curr Opin Genet Dev 1995, 5:38-43.

5. Kauffmann-Zeh A, Rodriguez-Viciana P, Ulrich E, Gilbert C, Coffer P,Downward J, Evan G: Suppression of c-Myc-induced apoptosis byRas signalling through PI(3)K and PKB. Nature 1997, 385:544-548.

6. Brambilla R, Gnesutta N, Minichiello L, White G, Roylance AJ,Herron CE, et al.: A role for the Ras signalling pathway in synaptictransmission and long-term memory. Nature 1997, 390:281-286.

7. Lloyd AC: Ras versus cyclin-dependent kinase inhibitors. CurrOpin Genet Dev 1998, 8:43-48.

8. Bar-Sagi D, Feramisco JR: Induction of membrane ruffling andfluid-phase pinocytosis in quiescent fibroblasts by ras proteins.Science 1986, 233:1061-1068.

9. Tuxworth RI, Cheetham JL, Machesky LM, Spiegelmann GB, Weeks G,Insall RH: Dictyostelium RasG is required for normal motility andcytokinesis, but not growth. J Cell Biol 1997, 138:605-614.

10. Insall RH, Borleis J, Devreotes PN: The aimless RasGEF is requiredfor processing of chemotactic signals through G-protein-coupledreceptors in Dictyostelium. Curr Biol 1996, 6:719-729.

11. Fukui Y, Kozasa T, Kaziro Y, Takeda T, Yamamoto M: Role of a rashomolog in the life cycle of Schizosaccharomyces pombe. Cell1986, 44:329-336.

12. Broek D, Toda T, Michaeli T, Levin L, Birchmeier C, Zoller M, et al.:The S. cerevisiae CDC25 gene product regulates theRAS/adenylate cyclase pathway. Cell 1987, 48:789-799.

13. Simon MA, Bowtell DD, Dodson GS, Laverty TR, Rubin GM: Ras1and a putative guanine nucleotide exchange factor performcrucial steps in signaling by the sevenless protein tyrosinekinase. Cell 1991, 67:701-716.

14. Chardin P, Camonis JH, Gale NW, van Aelst L, Schlessinger J,Wigler MH, Bar-Sagi D: Human Sos1: a guanine nucleotideexchange factor for Ras that binds to GRB2. Science 1993,260:1338-1343.

15. Bowtell D, Fu P, Simon M, Senior P: Identification of murinehomologues of the Drosophila son of sevenless gene: potentialactivators of ras. Proc Natl Acad Sci USA 1992, 89:6511-6515.

16. Ebinu JO, Bottorff DA, Chan EY, Stang SL, Dunn RJ, Stone JC:RasGRP, a Ras guanyl nucleotide-releasing protein with calcium-and diacylglycerol-binding motifs. Science 1998, 280:1082-1086.

17. Farnsworth CL, Freshney NW, Rosen LB, Ghosh A, Greenberg ME,Feig LA: Calcium activation of Ras mediated by neuronalexchange factor Ras-GRF. Nature 1995, 376:524-527.

18. Fam NP, Fan WT, Wang Z, Zhang LJ, Chen H, Moran MF: Cloningand characterization of Ras-GRF2, a novel guanine nucleotideexchange factor for Ras. Mol Cell Biol 1997, 17:1396-1406.

19. Pham N, Cheglakov I, Koch CA, de Hoog CL, Moran MF, Rotin D:The guanine nucleotide exchange factor CNrasGEF activates rasin response to cAMP and cGMP. Curr Biol 2000, 10:555-558.

1436 Current Biology Vol 10 No 22

20. Daniel J, Spiegelman GB, Weeks G: Characterization of a third rasgene, rasB, that is expressed throughout the growth anddevelopment of Dictyostelium discoideum. Oncogene 1993,8:1041-1047.

21. Daniel J, Bush J, Cardelli J, Spiegelman GB, Weeks G: Isolation oftwo novel ras genes in Dictyostelium discoideum; evidence for acomplex, developmentally regulated ras gene subfamily.Oncogene 1994, 9:501-508.

22. Robbins SM, Suttorp VV, Weeks G, Spiegelman GB: A ras-relatedgene from the lower eukaryote Dictyostelium that is highlyconserved relative to the human rap genes. Nucleic Acids Res1990, 18:5265-5269.

23. Robbins SM, Williams JG, Jermyn KA, Spiegelman GB, Weeks G:Growing and developing Dictyostelium cells express different rasgenes. Proc Natl Acad Sci USA 1989, 86:938-942.

24. Wilkins A, Khosla M, Fraser DJ, Spiegelman GB, Fisher PR, Weeks G,Insall RH: Dictyostelium RasD is required for normal phototaxis,but not differentiation. Genes Dev 2000, 14:1407-1413.

25. Chubb JR, Thomas GM, Insall RH: The Dictyostelium RasS proteinis required for macropinocytosis, phagocytosis and the control ofcell movement. J Cell Sci 2000, 113:709-719.

26. Morio T, Urushihara H, Saito T, Ugawa Y, Mizuno H, Yoshida M, et al.:The Dictyostelium developmental cDNA project: generation andanalysis of expressed sequence tags from the first-finger stage ofdevelopment. DNA Res 1998, 5:335-340.

27. Hacker U, Albrecht R, Maniak M: Fluid-phase uptake bymacropinocytosis in Dictyostelium. J Cell Sci 1997, 110:105-112.

28. de Hostos EL, Bradtke B, Lottspeich F, Guggenheim R, Gerisch G:Coronin, an actin binding protein of Dictyostelium discoideumlocalized to cell surface projections, has sequence similarities toG-protein ββ-subunits. EMBO J 1991, 10:4097-4104.

29. De Chastellier C, Ryter A: Changes on the cell surface and of thedigestive apparatus of Dictyostelium discoideum during thestarvation period triggering aggregation. J Cell Biol 1977,75:218-236.

30. Konijn TM: Microbiological assay for cyclic 3′′,5′′-AMP. Experientia1970, 26:367-369.

31. Vauti F, Morandini P, Blusch J, Sachse A, Nellen W: Regulation ofthe discoidinIγγ-gene in Dictyostelium discoideum — identificationof individual promoter elements mediating induction oftranscription and repression by cyclic AMP. Mol Cell Biol 1990,10:4080-4088.

32. Noegel A, Gerisch G, Stadler J, Westphal M: Complete sequenceand transcript regulation of a cell adhesion protein fromaggregating Dictyostelium cells. EMBO J 1986, 5:1473-1476.

33. Takeuchi I, Ishida S, Amagai A: Phagocytosis and exocytosis bydifferentiating Dictyostelium discoideum cells. Plant Cell Physiol1983, 24:395-402.

34. Varnum B, Edwards KB, Soll DR: The developmental regulation ofsingle-cell motility in Dictyostelium discoideum. Dev Biol 1986,113:218-227.

35. Clarke M, Gomer RH: PSF and CMF, autocrine factors that regulategene expression during growth and early development ofDictyostelium. Experientia 1995, 51:1124-1134.

36. Liu BX, Wei W, Broek D: The catalytic domain of the mouse sos1gene product activates Ras proteins in vivo and in vitro. Oncogene1993, 8:3081-3084.

37. Martegani E, Vanoni M, Zippel R, Coccetti P, Brambilla R, Ferrari C,et al.: Cloning by functional complementation of a mouse cDNAencoding a homologue of CDC25, a Saccharomyces cerevisiaeRAS activator. EMBO J 1992, 11:2151-2157.

38. Powers S, Gonzales E, Christensen T, Cubert J, Broek D: Functionalcloning of BUD5, a CDC25-related gene from S. cerevisiae thatcan suppress a dominant-negative RAS2 mutant. Cell 1991,65:1225-1231.

39. Park W, Mosteller RD, Broek D: Amino acid residues in the CDC25guanine nucleotide exchange factor critical for interaction withRas. Mol Cell Biol 1994, 14:8117-8122.

40. Boguski MS, McCormick F: Proteins regulating Ras and itsrelatives. Nature 1993, 366:643-664.

41. Rozakis-Adcock M, Fernley R, Wade J, Pawson T, Bowtell D: TheSH2 and SH3 domains of mammalian Grb2 couple the EGFreceptor to the Ras activator mSos1. Nature 1993, 363:83-85.

42. Zhou K, Pandol S, Bokoch G, Traynor-Kaplan AE: Disruption ofDictyostelium PI3K genes reduces [32P]phosphatidylinositol 3,4bisphosphate and [32P]phosphatidylinositol trisphosphate levels,alters F-actin distribution and impairs pinocytosis. J Cell Sci 1998,111:283-294.

43. Buczynski G, Grove B, Nomura A, Kleve M, Bush J, Firtel RA,Cardelli J: Inactivation of two Dictyostelium discoideum genes,DdPIK1 and DdPIK2, encoding proteins related to mammalianphosphatidylinositide 3-kinases, results in defects in endocytosis,lysosome to postlysosome transport, and actin cytoskeletonorganization. J Cell Biol 1997, 136:1271-1286.

44. Maniak M, Rauchenberger R, Albrecht R, Murphy J, Gerisch G:Coronin involved in phagocytosis: dynamics of particle-inducedrelocalization visualized by a green fluorescent protein tag. Cell1995, 83:915-924.

45. Ma H, Gamper M, Parent C, Firtel RA: The Dictyostelium MAPkinase kinase DdMEK1 regulates chemotaxis and is essential forchemoattractant-mediated activation of guanylyl cyclase. EMBO J1997, 16:4317-4332.

46. Vithalani KK, Parent CA, Thorn EM, Penn M, Larochelle DA,Devreotes PN, De Lozanne A: Identification of darlin, aDictyostelium protein with armadillo-like repeats that binds tosmall GTPases and is important for the proper aggregation ofdeveloping cells. Mol Biol Cell 1998, 9:3095-3106.

47. Alexander S, Sydow,LM, Wessels D, Soll DR: Discoidin proteins ofDictyostelium are necessary for normal cytoskeletal organizationand cellular morphology during aggregation. Differentiation 1992,51:149-161.

48. Sutoh K: A transformation vector for Dictyostelium discoideumwith a new selectable marker Bsr. Plasmid 1993, 30:150-154.

49. Manstein DJ, Schuster HP, Morandini P, Hunt DM: Cloning vectorsfor the production of proteins in Dictyostelium discoideum. Gene1995, 162:129-134.

50. Witke W, Schleicher M, Noegel AA: Redundancy in themicrofilament system - abnormal development of Dictyosteliumcells lacking two F-actin cross-linking proteins. Cell 1992,68:53-62.

51. Agatep R, Kirkpatrick RD, Parchaliuk DL, Woods RA, Geitz RD:Transformation of S. cerevisiae by the lithium acetate/singlestranded carrier DNA/PEG (LiAc/ss-DNA/PEG) protocol.Technical Tips Online (http://tto.trends.com) 1998.

Research Paper Dictyostelium RasGEFB regulates endocytosis and cell speed Wilkins et al. 1437