7/28/2019 Guo Sk1995

1/10

Cell, Vol. 61, 61 l-620, May 19, 1995, Copyright 0 199 5 by Cell Press

par-$ a Gene Required for Establishing Polarityin C. elegans Embryos, Encodes a Putative Ser/ThrKinase That Is Asymmetrically DistributedSu Guo and Kenneth J. KemphuesSection of Genetics and DevelopmentCornel l Universi tyIthaca, New York 14853

SummaryThe f i rst cleavage of C. elegans is asym metr ic, gener-ating daughter cel ls wi th di fferent sizes, cytoplasmiccompon ents, and fates. Mutations in the par-7 genedisrupt this asym metry. We report here that par-7 en-codes a putative Ser/Thr kinase w ith simi lar i ty to ki-nases from yeasts and mam mals. Two strong al leleshave mutations in the kinase domain, suggesting thatkinase activi ty is essential for par-l fun ction. PAR-1protein is localized to the posterior periphery of thezygo te and is distributed in a polar fashion precedingthe asym metr ic divisions of the germl ine l ineage. Be-cause PA R-1 distr ibution in the germl ine correlateswith the distribution of germl ine-speci fic P granules,i t is possible that PAR-1 functions in germl ine develop-ment as wel l as in establ ishing embryonic polar i ty.IntroductionAsym metr ic cel l division, by which the two daughter cel lsadopt di fferent fates, plays an important role in the genera-t ion of cel lular d iversi ty dur ing developm ent. Daughtercel ls wi th di fferent fates can ar ise via extr insic cel l -cel ls ignaling or by intr insical ly asym metr ic divisions (Horvi tzand Herskow itz, 1992). Intr insical ly determined asym met-r ic divisions occur in a number of invertebrates and appearto play crucial roles in establ ishing embryonic axes orspeci fying part icular di fferentiated cel l fates (Davidson,1986; Slack, 1991). Mos t evidence supports the model ofcytoplasm ic local ization as the basis for such intr insicallypatterned asym metr ic divisions. In cytoplasm ic local iza-t ion, maternal ly provided determinants are unequal lypartitioned to particular cells through ase riesof reproduci-bly or iented cleavage divisions (Davidson, 1986).

The nematode Caenorhabdit is elegans provides an op-portuni ty to study the mecha nisms responsible for intrinsi-cal ly asym metr ic divisions. C . elegans em bryos have afixed cel l l ineage, and a precise cel l fate map has beenestabl ished (Sulston et al ., 1983). The asymme tr ic divi -sions leading to the formation of the germl ine play an im-portant role in establ ishing embryonic organization. Eachof these divisions is asym metr ic, producing daughter cel lswith di fferences in size, cel l division t iming, cleavage pat-tern, and ult imate fate (see Figure 1A for a summ ary ofthe cell l ineage). In addition, germline -specific P granulesare localized to one pole of the cell prior to each divisionand partitioned into the sma ller da ughter cell (Strome andWood, 1982, 1983).Al though the mecha nisms under lying this ser ies ofasym metr ic divisions are largely unknown, microfi laments

appear to play a role in at least the firs t division. Briefpulses of the microfi lament-disrupting drug cytochalasinduring a critical period of the first cell cycle preven t theposter ior local ization of the P granules (Strome and Wood,1983; Hi ll and Strome, 1988). Cytochalasin pulses duringthis same per iod som etimes also lead to symm etr ic divi -sions producing daughter cel ls of equal sizes, simi lar cel lcycle rates, and variable spindle orientations (Hill andStrome, 1988, 1990).

Maternal-effect lethal mutations in the par genes pro-duce phenotypes simi lar to the effects of cytochalasintreatments (Kemphues et al ., 1988). In par mutant e m-bryos, normal ly unequal divisions are equal, cleavagespindlesare misor iented, Pgranules are mislocal ized, andthe blastomere fates are al tered. This impl ies that the pargenes function in both spindle placement and cytoplasmiclocal ization (Kemphues et al ., 1988; Kirby et al ., 1990;Morton et al ., 1992; Cheng et al ., 1995).

The par-7 gene is required for several aspects of ear lyembryonic polar i ty (Kemphues et al ., 1988). As schema ti-cal ly diagrammed in Figure 1 B, par-7 m utant embryos ex-hibit equal first cleavage and fail to localize P granules.The 2-cel l stage blastomeres divide synchronously, andthe cleavage patterns are aberrant for al l subsequentstages. The par-7 embryos arrest as amorphous mass eswith large numbers of di fferentiated cel ls. Mos t embryoscontain e xcess b ody wal l and pharyngeal muscle ce l ls(normal ly produced by the MS lineage) and are completelylacking intestine. The putative MS cel l fate determinant,SKN-1, normal ly distr ibuted asymm etr ical ly among ear lyblastomeres, is found in equal amoun ts in al l the cel ls atthe 2- and 4-cel l stages in par-7 mutant embryos (see Fig-ure 1 B). The excessive pharyngeal and body wal l musclecel ls observed in par-7 mutant embryos may ar ise as aconsequence of the abnormal distribution of SKN-1 (Bow-erman et al., 1993).We demonstrate here that par-7 encodes a conservedputative Ser/Thr kinase that is asymm etr ical ly local ized tothe posterior periphery of the l-cell emb ryo and differen-t ial ly distr ibuted, preceding al l the asym metr ic divisionsof the germl ine l ineage. Mutations in the kinase domainel iminate PAR-1 function, suggesting PAR-1 kinase activ-i ty is essential for i ts role in establ ishing asym metry.ResultsMolecular Cloning of the par- f GeneIdentificatio n of the par-l RegionAs summ arized in Figure 2A , par-7 m aps to chromosom eV between genetic markers rol-4 and uric-76, an intervalof -3 m.u . Tel fer (1991) had placed the par-7 gene onthe C. elegans physical map within or near the region cov-ered by the yeast ar t i fic ial chrom osome (YAC) Y39Hl. Wefurther mapped par-7 relative to two closely l inked genes,ogr-7 and him-5 The mappingdataputpar-7 -0.075 m.u.(- 60 kb) to the right of him-5 in a region only partiallycovered by cosmid clones (descr ibed in Exper imental Pro-

7/28/2019 Guo Sk1995

2/10

Cell61 2

0 0 zygote

AB($h & heI1+AB o

EM SFigure 1. Cell Lineages and Early Div is ions in C. elegans(A) Diagram showing a s implif ied cell l ineag e and majo r cell typesproduced by the founder cells (Sulston et al., 1963). PO is the zygote.Posterior is to the right. The blastomeres of the P lineage , namely, POto 22 and 23, are outl ined . Those P lineag e blastomeres that undergoasymmetric divisions are marked with asterisks. Note that asymmetricdiv is ion of EMS segregates the abil ity to produce intestine to the Edaughter and the abil ity to produce pharyngeal and body wall musclesto the M S daughter.(B) Schematic drawing of I-, 2-, and 4-cell wild-type ( left) and par-lmutant (r ight) embryos, showing the div is ions and the distr ibution ofP granules (represented by dots in the cytoplasm) and transcriptionfactor S KN-1 (represented by wavy l ine s in the nucleus; the extentof shading represents the relative concentration of SKN-1 protein).Posterior is to the r ight. In wild-type embryos, f irst and second div is ionsproduce daughter cells differing in s ize, div is ion t iming, c leavage pat-tern, and cell fates. P granules are localized to the posterior end beforethe first div is ion; they are segregated into the Pl blastomere by f irstasymmetric c leavage and further loca lized to the P2 blastomere(Strome and Wood, 1963). SKN-1 protein is highly concentrated in thePl nucleus a nd remains concentrated in the s ister blastomeres EMSand P2 derived from Pi (Bowerman et al., 1993). In par-l mutantembryos (from homozy gous par-l mothers), asymmetry in size, divi-s ion t iming, a nd c leavage pattern is disrupted, and cell fates are dra-matically altered. P granules fail to localize. Since the P granules areundetectable in the late l-cell par-7 embryos and in the 2-cell par-lembryos, they are only shown in the 4-cell embryo. SKN-1 protein isfound in equa l concentration in all blastomeres at 2-, 4-, and &cellstages.

cedures). We were able partially to rescue par-7 mutantsby transformation with the YAC, Y51 F3 (see Exper imentalProcedures). Rescued par-7 homozygous animals pro-duced about f ive hatching progeny per worm, but all ofthese progeny were agametic. This agametic phenotypeis a maternal effect, s ince par-7 homozygous hermaphro-dites derived from par-l/+ mothers produced gametes.We interpret this result to mean that the expression of

ALGVgeneticp:,- - I, /,fl, ofFI im-5 py-I\ ,,2!lc -camids Md YACS:

E2sEu \D.1t

0 \00 Y-0 khl m \ \, YyF3 (-2Sa kb)YFA1.4 Ad YFAlsa-

cDNA clone: ZC22(4 W

I3 100 kb probeN2 bn2c ZC22 probe

Bcl IAlw39 N2~.,

Figure 2. Identif ication of par-7(A) Placemen t of par-7 on the physical map. The diagram shows thegenetic map and physical map in the par-l region. Genetic ma ppingputs par-7 -0.075 mu. (estimated - 60 kb) to the r ight of him-5 (seeExperimenta l Procedures). The YAC Y5lF3 rescued par-7 mutant em-bryos in a germline transformation assay. Y51F3 was cut with therestr ic tion enzyme Ascl, and two fragments, YFAlOO and YFA15 0,were generated.(B) Northern blot analysis of maternally enriched transcripts presentin the 100 kb fragment of Y5lF3. Poly(A)t RNA was isolated fromwild-type (N2) and the temperature-sensitive germline-defic ient straing/p-4(bn2), resolved by electrophoresis, blotted, and probed with the100 kb fragments. A s ingle maternally enriched transcript is detectedand indicated with an arrowhead. It is -4.4 kb in length.(C) Genomic Southern blot analysis of par-l(lw39) and wild type (N2).Genomic DNA was isolated from wild-type (N2) and par-l( lw39) hetero-zygous animals, digested with Bell, resolved by electrophoresis, blot-ted onto n ylon fi l ter, and probed with the cDNA clone ZC22. Threebands, 1.6 kb, 7.5 kb, and 10 kb, were detected in N2; in par-l(lw39)heterozygous animals, 7.5 kb and 10 kb bands were found to be atdecreased intensity; in addit ion , a novel ban d of - 12 kb was detected.

par-l in the extrachromosoma l array is suff ic ient to res-cue some embryos from maternal-effect lethal ity bu t notsuff ic ient to prevent a maternal-effect ster i l ity. Consistentwith this interpretation, escapers from the maternal-effectlethal i ty of the leaky par-7 allele lw7 are also agametic(unpublished data).

Our attemp ts to identi fy sm al ler segments of chromo-somal DNA that could rescuepar- fai led because we wereunable to recover clones of DNA that contiguously covered

7/28/2019 Guo Sk1995

3/10

Kinase Involvement in Asymmetric Div is ions61 3

Table 1. Scoring for Embryonic Lethality in RNA Injections

Molecules InjectedNumber o f Worms PercentInjected Embryonic Lethality

X.22 @r-I) antisense 16 52ZC22 @af-7) sense 12 54Clone 7S ant isense 8 0Clone pZ1 antisense 8 0DEPC-dHz0 4 0ZC22is the nearly full- length cDNAclone (4 kb)forpar-7; TSis twinstar,a Drosophila gene encoding cofi l in (K. Gunsalus and M. Goldberg,personal communication); c lone pZ1 is a zygotic cDNA clone identif iedby J. Wans (personal communication). The number of hatched versusunhatched embryos was counted for the f irst 2 days following theinjections (a minim um of 100 embryos from each injected worm werecounted).

this region. Therefore, we took an al ternative approach toclone the par-7 gene.Isolation of a Candidate cDNA and Detectionof an Al lele-Associated PolymorphismWe cut Y5lF3 into 100 kb and 150 kb fragments withthe restr ict ion endonuclease Ascl . Southern blot analysisusing the end of the cosmid D1086 as a probe showedthat the 100 kb fragment is proximal to him-5 (data notshown) and thus l ikely tocontainpar-7. Aspar- mutationsare str ict ly maternal (Kemphues et al ., 1988) we searchedthis 100 kb region for the presence of maternal ly enr ichedmR NAs that would be candidates for the par-7 gene. Wedetected one message of -4.4 kb that was greatly en-riched in wild type relative to the germline-de ficient strain,g/p-4(bn2) (Figure 28). We isolated cDNA s correspondingto this transcr ipt (see Exper imental Procedures). By usingthe longest cDNA (X22) as a probe, w e detected an al lele-associated restr ict ion fragme nt length polymorphism (Fig-ure 2C) in the y-irradiation-induced par-7 allele, lw39 (agi ft from J. Shaw). Further Southern blot analyses usingfragments of the cDNA clone as probes suggest that thelesion in the lw39 allele is a deletion of - 5 kb in the tran-scr ipt ion uni t (data not shown). This resul t indicates thatthe ZC22 cDNA may be der ived from par-7.Phenocopying par- l in Wi ld Type via AntisenseRNA InjectionGerml ine transformation rescue with agenomic clone con-taining only the gene of interest has been routinely usedto confi rm cloning of genes in C. elegans. Howev er, whenwe searched for the corresponding genomic clone in agenomic DNA l ibrary (provided by H. Browning) or in theYAC Y51 F3 subclone l ibrar ies we previously constructed,we fai led to identi fy clones that would represent the ZC22genomic region. Therefore, we decided to determinewhether the cDNA we identif ied corresponds to the par-7gene using antisens e inhibition. An tisense inhibition hasbeen previously accompl ished in transgenic worms car-rying antisense DNA con structs (Fire et al ., 1991). Be-causepar- wasexpected to beexpressed in thegerml ine,we carr ied out a more direct inhibi t ion assay by injectingantisense RNA into gonads of wi ld- type worms (see Exper-imental Procedures).

The resul ts of these injections are summ arized in Table

1. An average of 50% of the emb ryos from the ZC22 anti -sense RNA-injected wi ld- type animals arrested develop-ment with par-7 terminal phenotypes: no morphogenesis,no intestinal cel ls, and excess of pharynx. Furthermore,mos t of the 50 or so ear ly embryos we dissected frominjected worms exhibi ted cel l division patterns character is-t ic of par-7 loss-of- function mutations. For example, manyembryosexhibi tedpar- l-speci f ic pseudocleavage defects(Kirby et al ., 1990) equal f i rst cleavage, and synchronousear ly divisions (data not shown). Also, consistent wi th ourresul ts from germl ine transformation, some of the surviv-ing progeny from injected worms were agame tic. In controlexper iments, in which Hz0 and antisense RNA made fromtwo unrelated cDNA clones w ere injected, no effects wereobserved.

Surpr isingly, injection of in vi tro synthesized sense RNAfrom the cDNA ZC22 also induced par-7 phe notypes at ahigh frequency among the progeny of injected w orms. Itis not clear what accounts for this effect. Moreover, thesense effec t appears to be restr icted to the putative trans-lated region of the RNA whi le the antisense effect is not.Injection of both sense and antisense RNA from the 5region (lacking the 3 untranslated region) also gave par-7phenocopies, whi le only the antisense RNA from the 3untranslated region gave an effect (data not shown). Thus,the antisense and sense effects appear to be separableand probably involve di fferent m echanism s. The basis forthe sense effect is under investigation and wi l l not be dis-cussed fur ther. Overal l , the speci f ic i ty of the antisenseand sense phenocopies provides strong evidence that theZC22 cDNA represents thepar- transcr ipt. Addit ional evi-dence is provided below.The par-l Gene Encodes a Conserved PutativeSerlThr KinaseWe sequenced the part ial cDNA Z C22 and used the rapidampl i f ication of cDNA ends-polymerase chain reaction(RACE-PCR ) to identi fy the 5 end of the transcr ipt (seeExper imental Procedures). Four di fferent 5ends were de-tected, al l of which w ere frans-spl iced to the SLl leader(Krause and Hirsh, 1987; Bektesh et al ., 1988). The nucle-otide sequence of the longest mR NA der ived from theRACE -PCR resul ts has been submitted to GenBank andis not reproduced here. A single long open reading frameis detected in the messa ge; the putative star t codon ispreceded by stop codons, and a C. elegans translationini tiat ion consensus sequence, AACA (M. Perry, personalcomm unication), is found just upstream of this ATG. Thepredicted PAR-l protein consists of 1192 amino acids witha molecular mas s of - 126 kDa. Protein data basesearches reveal that PAR-1 is a putative Serf lhr kinasehighly conserved among di fferent organisms (Figure 3A).PAR-l shows str ik ing simi lar i ty to i ts mamm alian homo-logs. In the kinase domain, PAR-1 shares -80% identi tywiththe human Kp78(217of260)(Parsa, 1988)and mouseEmk (220 of 260) ( Ingl is et al ., 1993) and - 46% identi ty(70% simi lar i ty) wi th yeast SN Fl (128 of 260) (Celenzaand Car lson, 1986) the rat AMP-dependent kinase (122of 260) (Car ling et al ., 1994) and yeast KINlIKIN2 (113of 260) (Levin et al., 1987; Levin and Bishop, 1990). Intrigu-ingly, besides strong homologies in the kinase domain,

7/28/2019 Guo Sk1995

4/10

Cell614

Figure 3. Sequence A l ignment of PAR- I w i th Members of the SNFlProte in K inase Subfami ly(A) Amin o acid sequence comparison of the k inase domain from PAR-1with homolog ous k inases identif ied in mamm als and yeast: huma n

N2 lw39

- 205PAR-1 w - 123

896750

Figure 4. Detection of Endoge nous PAR-1 Protein on Western BlotsExtracts from wild-type (N2) and par-l( lw39) homozygotes were re-solved by SDS-PA GE, blotted onto nitrocellulose fi l ters, and probedwith the aff inity-purif ied anti-PAR-l antibody (see Experimental Proce-dures). A major b and o f - 126 kDa (the expected s ize of the PAR-1protein) was detected in N2, but not in paf-7(/w39) homozygotes. Thesame blot was reprobed with a monoclo nal anti-a-tubulin antibody asa loadin g control. Posit ions of molecular mass markers are indicatedon the r ight.

an addit ional conserved domain at the C-termini of theproteins is present in PAR-l , Kp78, E mk, and KINl/KIN2(Figure 3B). This domain is highly conserved amongPAR-1 and the mam malian kinases, but much divergedin the yeast kinases.To ver i fy that this gene is indeed par- l, we sequencedthe kinase domain in 10 mutan t al leles and found al ter-ations of invar iant kinase residues (Hanks and Quinn,1991) in twop af-7 mu tants . In allele it90, an invariant resi-due Gly of the putative ATP-binding si te in the kinase sub-doma in I is changed to Glu (Figure 3A). In allele ir57, aninvariant residue Arg in subdom ain Xl is changed to Lys(Figure 3A). Although Arg and Lys are both positivelycharged , the invariance of this position in all protein ki-nases that have been sequenced to date suggests that thisArg residue can not be substi tuted. A simi lar conservedsubsti tut ion (Asp to Glu) in an invar iant kinase residue hasalso been reported in a mu tant allele of the cfrl gen e inArabidops is (Kieber et al., 1993).PAR-l Protein Is First Detected in NewlyForming OocytesTo examine the distr ibution of PAR-1 in adul t hermaphro-di tes and ear ly embryos , we raised polyclonal antibodiesagainst a port ion of the PAR-1 protein (see Exper imentalProcedures). The antibody speci f ic i ty was examined on

Kp76 (sequence unpublishe d but found in GenBank), mouse Em k (ln-gliset al., 1993) rat AMP-activated protein k inase(Carling etal., 1994)SNFl (Celenza and Carlson, 1966) and KIN1 in S. pomb e (Levinand B ishop, 1990) . The a l ignment was generated us ing the CLUSTALmethod (Higgins and Sharp, 1969) with DNAST AR software. Identicalamino acids are boxed. The k inase su bdomains as defined by Hankset al. (1966) are shown. The invariant k inase residues are indicatedby asterisks. The percentage of amino acid identity between PAR-Iand other k inases is indicated. The amino acid alterations of invariantresidues observed in if57 and it90 are indicated by arrowheads.(6) Alignm ent of the C-terminal domains in PAR-l, Kp76, Emk, andKINIIKINP. Identical residues are boxed, and the percentage of iden-tity is indicated.

7/28/2019 Guo Sk1995

5/10

Kinase Involvement in Asymmetric Div is ions61 5

Figure 5. Distr ibution of PAR-1 in the Gonad and Early Embryolmmunofluorescence imageo f gonads and embryos showing the distr i-bution of the anti-PAR-l antibody (green) an d nuclei (blue).(A) lmmunofluorescence image of a wild-type gonad showing strongfluorescence at the periphery of newly forming oocytes. Althoug h wedetect s ignal aroun d the periphery of these young oocytes, the outermembran e signal is faint and is not apparent in this reproduction.Abbreviations: MN, meiotic nuclei; NFO, newly forming oocytes; MO,mature oocytes.(6) Double exposure of the same gona d shown in (A), showing thelocalization of anti-PAR-l relative to nuclei.(C) Double exposure showing absence of peripheral staining in par-l( i t86) mutant gonad.(D-E) Wild-type l- and 2-cell embryos, respectively, showing strongperipheral f luorescence in the posterior.(F-G)Mutant l-and2-cellpar- l( lw39)embryosshowingtheabsenceofposterior peripheral staining and the reduction of cytoplasmic staining.The scale bar in (A)-(C) is 32 pm; in (D)-(G), 6 pm.

Western blots (Figure 4). Using this antibody, we havestained w orms and isolated gonads and ear ly embryos ofwi ld- type and par-7 mutants (see Exper imental Proce-dures). The patterns we descr ibe below were seen in al lsamples of a given genotype, were observed to exhibi tl ittle variability, and were eliminated if PAR-l fusion proteinwas added to the antibody binding reaction.

Antibody staining is f i rst detected at the bend of thereflexed h ermaphrodi te gonad. The section of the gonaddistal to the bend contains mitotic and meiotic nuclei in asyncytium . Individual meiotic nuclei at the bend are en-closed by cel l memb ranes forming oo cytes. Staining isseen associated with the cel l memb ranes of the newlyforming oocytes (Figures 5A and 58). Since we have notdetermined whether this staining is at the cortex or in thememb rane, we wi l l refer to this pattern as per ipheral stain-ing. Although we see staining around the entire ce ll periph-

ery, only the strong staining where oocytes abut is visiblein Figure 5. As the oocytes mature, the staining at theper iphery become s fainter . Per ipheral staining of oocytesis not detected in par-7 mutan ts (Figure 5C), and thus i tis due to the PAR-1 protein.PAR-1 Protein Is Distr ibuted Unequal lybefore the First Asym metr ic CleavageLocalized antibody staining is not detecta ble in unferti l izedmature oocytes , nor in newly fert i l ized zygotes that are st i l lundergoing meiotic divisions. We start to detect staining atthe poster ior per iphery of the zygote, when both pronucleiare decondensed and the female pronucleus is just star t-ing to migrate toward the poster ior . Thisvery ear ly stainingis faint and extends from the poster ior pole to - 50% egglength (data not shown). As the cel l cycle proceeds towardmetapha se, the poster ior per ipheral staining gets strongerand becomes restr icted to the poster ior third of the embryo(Figure 5D). The staining becom es fainter when the f i rstcel l cycle passes metaph ase and approaches telophase(data not shown).

In addition to the posterior peripheral signal, staining isalso found in the internal cytoplasm of the blastomeres(Figures 5D and 5E). Using the same cond it ions as forwi ld type, we stained par-7 m utant embryos. In nine par-7alleles exam ined, the posterior peripheral staining is ab-sent (Figures 5F and 5G). Thus, this posterior per ipheralstaining is due to the PAR-l protein. In allele lw3 9, whichcontains a deletion of -5 kb in the par-7 gene and lacksdetectable PAR-l protein (see Figure 4) the cytoplasm icstaining is reduced, suggesting some staining in wi ld- typecytoplasm is due to unlocal ized PAR-1 protein.PAR-1 Protein D istr ibution Is Asym metr ic in Cel lsof the Germline LineageAs descr ibed above, PAR-1 is asymm etr ical ly local izedto the posterior per iphery of the zygote. In addi t ion, weobserve PAR-l staining in the cel ls of the germl ine l ineage.PAR-1 p rotein is faintly detectable around the peripheryof the newly forme d Pl cel l (data not shown). As the Plcel l cycle progresses, PAR-1 protein become s local ized tothe posterior periphery of Pl (Figure 5E), and this localizedPAR-1 is segregated into P2 at the division. Because thispattern of localization is similar to the localization of germ-l ine-specif ic P granules (Strome and Wood , 1983) weexamined PAR-l and P granule local ization in the sameembryos . As shown in Figure 6, we observe a str ik ingcorrelation betwe en PAR-1 and P granule localization. Inlate stage l -cell and 2-cel l embryos, both PAR-1 and Pgranules are localized to the posterior, and the anteriorexte nt of the peripheral PAR-1 distribution is correlatedwith the anterior exten t of the P granule distribution (Fig-ures 6A and 6B).

The asymm etr ic PAR-1 local ization at the per iphery isagain detected in the P2 and P3 cel ls. In P2 and P3, theconstraints of the eggshel l combine with a polar i ty reversal(Schierenberg, 1987) to position the two divisions alongthe dorsal-ventral rather than the anterior-posterior axis,with the P granules being segregated to the ventral side.In these two cel ls, PAR-1 become s local ized to the ventral

7/28/2019 Guo Sk1995

6/10

Cel l61 6

/A

-*B P-EM S

Figure 6. Correlation between PAR-l P eripheral Distr ibution and PGranule Localization(Left column) C onfocal images showing PAR-l (green) and P granule(red) distr ibution in wild-type embryos at different stages. Posterior isto the r ight.(A-B) PAR-1 is localized at the posterior periphery in l- and 2-cellembryos. The peripheral PAR-1 staining is correlated with the extentof P granule distr ibution.(C) An early 4-cell embryo (ABa and ABp at prophase; EMS and P2at interphase) showing the distr ibution of PAR-l not only at the periph-ery of P2 (the posterior cell) , but also at the boundaries between allfour cells . E MS is ventral.

per iphery (Figures 6D and 6E). Again, at the late stagesof the P2 and P3 cel l cycle, the P granule distributioncorrelates with the PAR-l local ization. Interestingly, in P4,which d ivides symm etr ical ly, PAR-1 is always found to bearound the enti re per iphery(Figure 6F). As a resul t, PA R-l ,l ike P granules, is found in both daughter cells, 2 2 and23. PAR-1 is detected around the enti re per iphery of bothcel ls (data not shown). This PAR-1 staining gradual ly be-come s fainter and is no longer detectable by the onset ofmorphogenesis.

Some PAR-l staining is seen in cel ls outside the germ-l ine lineage. In the newly formed 4cel l embryo, PAR-l isfaintly dete ctable only around the periphery of P2 (datanot shown). As the cell cycle progresses, PAR-1 is notonly found at the periphery of P2, but is also found alongthe boundar ies between the three soma tic blastomeres(Figure 6C). By the late 4-cell stage, PAR-l is barelydetect-able in the soma tic cel ls, and the protein is mainly concen-trated at the per iphery on the ventral hal f of the P2 cel l(Figure 6D). Al though al l of the Par- l phenotypes can beexplained by defe cts in intracellular localization, the tran-sient appearance of strong PAR-l staining along theboundar ies between cel ls at the 4-cel l stage could reflectan additional role for PAR-1 in the cell-cell interaction sthat are bel ieved to be occurr ing at this t ime (Pr iess andThoms on, 1987; Goldstein, 1992, 1993; Schnabel, 1994;Mel lo et al ., 1994). Some weak and var iable PAR-1 stain-ing is also seen along the boundar ies between some so-matic cel ls at later embryonic stages (Figures 6E and 6F).PAR-1 Klnase Activi ty Is Required for theLocal ization of P Granules and SKN-1Two par-7 al leles, i t57 and i t90, carry mutations in thekinase domain. We examined PAR-l , P granule, andSKN-1 local ization in these mutan ts. As shown in Figure7A, the peripheral localization of PAR-1 in the it57 zygo teis indistinguishable from that of wi ld type (compare withFigures 5D and 6A). In 2- and 4cel l embryos, per ipheralstaining at the ear ly stages of the cel l cycle is simi lar to wi ldtype (compare Figures 7C and 6C), but the asymme tr iclocalization to the posterior in Pl and ventral periphery inP2 is incomplete (data not shown). In later stage em bryos,the mutant PAR-1 protein is present at the per iphery of asubset of blastomeres, but is not asymm etr ical ly local izedin those cel ls.The P granule distribution in if57 and if90 emb ryos isabnorm al and is indistinguishable from the distributionsof P granules in other stron g par-7 alleles. P granules arepresent in the wi ld- type pattern in gonads and early zy-

(D) In the late 4-cell embryo (ABa and AB p at telophase; EMS at ana-phase; P2 at prophase), PAR-l is mainly concentrated at the peripheryon the ventral half of P2. EMS is ventral.(E) PAR-l and P granules are asymmetrically localized to the ventralperiphery.of P3 (- 15-cell s tage). Scale bar, 6 urn.(F) PAR-1 is detected around the periphery of P4 (-50-cell embryo),which no longer div ides asymmetrically .(Right column) Diagramsof embryos shown in the left column (reducedin s ize). The blastomeres of the germline l ineage , Pi -P4, a re hatched.

7/28/2019 Guo Sk1995

7/10

Kinase Involvement in Asymmetric Div is ions61 7

Figure 7. PAR-l, SKN-1, and P Granule Distr ibution in par-l( it57)(A) A l-cell i t5 7 embryo showing normal PAR-1 distr ibution.(B) A Z-cell i t57 embryo showing SKN-1 present in equa l concentrationin both blastomeres, in contrast with its asymmetric localization in wildtype (see Figure 16).(C) A 4-cell i t57 embryo showing normal PA R-1 distr ibution. Nucleiare sta ined w i th DAPI .(D) The same embryo as in (C), showing abnorm al P granule distr ibu-tion. Scale bar, 10 pm.

gotes before pronuclear migration (data not shown), butare undetectable in late zygotes (such as the embryoshown in Figure 7A) and in 2-cel l em bryos. This absenceof detectable P granules in l - and 2-cel l embryos is notspeci f ic to i t57 or i fgO, but is character ist ic of par-7 muta-t ions (unpublished data). At the 4-cel l stage, P granules areeasi ly detected, but are not local ized (Figure 7D). SKN-1protein distributions in if57 and it90 are also abnormal(Figure 7 8). By Western blot analysis, the amount ofPAR-l protein in i t57 and it90 is the same as wi ld type(data not shown).

These resul ts suggest that the kinase activi ty of PAR-1is essential for i ts normal func tion. The asym metr ic local-ization o f PAR-1 i tsel f, at least in the zygote, does no trequire i ts own kinase activi ty.

We have described the molecular cloning of par- l, a genenecessary for establ ishing asym metry in ear ly C. elegansembryo s. We have shown that par-7 encodes a conservedputative Ser/Thr kinase. The PAR-1 protein is found to beasymm etr ical ly local ized to the poster ior per iphery of thezygote before i ts mitotic division. It is then di fferential lydistributed in the germline precursor cells, and its periph-eral distribution correlates with P granule localization.The Role of Kinase Activi ty in par-l FunctionThe f inding of sequence simi lar ity to Ser/Thr protein ki-nases suggests that PAR-1 is a protein kinase. The identi f i -cation o f mutations in the kinase domain in two par-7 al-leles argues strongly that the kinase activi ty is importantfor par-7 function. Because the phenotypes caused byboth alleles are indistinguishable from other strong par-7

mutations, we conclude that the kinase domain is essentialforpar- function. In addi t ion, the fact that the PAR-1 pro-tein in these two alleles is localized properly in the zygo tesuggests that the kinase function is not required for PAR-1localization.PAR -l and Anterior-Posterior PolarityThe initial cue that set s the anterior-posterior polarity in C.elegans embryos is not wel l understood. Mature oocytesshow polar i ty in the acentr ic placement of the nucleus,but have no other obvious polar ity wi th respect to cytoskel-eton, distr ibution of cytoplasmicorganel les, orvar ious sur-face markers (Strome, 1986a). In mo st embryo s, the polarbodies mark the future anter ior and the sperm marks thefuture poster ior . I t is not clear whether the sperm entrypoint is restr icted by preexist ing oocyte polar i ty or whetherthe sperm entry point provides the cue for embryonic po-larity.

Whatever the ini tial cue, addi t ional events that occurdur ing the f i rst embryonic cel l cycle appear to transducethat cue into a functional polar ity in the embryo. By lateprophase of the f i rst cell cycle, the zygote is discernablypolar. Cort ical actin is concentrated at the anter ior (Strome,1986b), the poster ior cytoplasm has acquired the abi l ityto direct unequal cleavages (Schierenberg, 1985, 1988)and P granules are localized to the posterior (Strome andWood, 1982, 1983). Furthermore, transient treatment ofembryos with cytochalasin dur ing the f i rst cel l cycle affectsthese visible polar it ies and al ters the fates of the 2-cellstage blastomeres (Hi ll and Strome, 1988, 1990).

Previous work indicates that par-7 play s a crucial rolein establ ishing asym metry dur ing the f i rst cel l cycle. Thepar-7 function is required for unequal f i rst cleavage, asym -me tric P granule localization, and generation of daughtercel ls wi th di fferent fates (Kemphues et al ., 1988). par-7is also required for the asymm etr ic distr ibution of SKN-1(Bowerman et al ., 1993).

How does the par-7 kinase a ctivi ty contr ibute to asym-metry in the l -cell embryo? Although many possibi l i t iesare consistent wi th our resul ts, the local accumulation ofPAR-1 protein at the posterior per iphery leads us to sug-gest that the protein acts by modifying the cort ical cy-toskeleton in the poster ior . As a consequence of this modi-f ication, P granules and other cytoplasmic compone ntsare local ized to the poster ior . Modif ication of the poster iorcort ical cytoskeleton could also be required for the asym-metr ic placem ent of the spindle that leads to unequal f i rstcleavage. In this model, many of the par-7 phenotypeswould be indirect consequences of fai lure in local ization.

Interestingly, there may be a role for a Serf lhr kinasein asymm etr ic cel l division in yeast. Studies of buddingyeast identif ied a smal l GTP-binding protein, C dc42,whose activi ty is required for the asym metr ic assem bly ofactin cytoskeleton at the bud si te ( for review see Chant,1994). The direct downstream target of Cdc42 in yeast isunknown. Howeve r, a Ser/Thr kinase has been shown tobe the probable target of the human Cdc42 homolog(Manser et al ., 1994), suggesting that kinases may alsobe activated by yeast Cdc4 2.It is possible that the role of PAR-1 protein in f i rst cel l

7/28/2019 Guo Sk1995

8/10

Cd l61 8

cycle asymme try is not dependent upon i ts local ization tothe poster ior . The protein f i rst accumu lates symm etr ical lyat the per iphery of the forming oocytes and appears to bepresent throughout the cytoplasm of the zygote. I t couldbe that PAR-l acts in an unlocal ized manner, but thatasym metr ic local ization of the protein is a mecha nism toensure a high concentration of PAR-1 in the germl inelineage.PAR-1 and Germl ine Cel l FateAlthough it is possible that PAR-l is only required for estab -l ishing asym metry in the zygote, the local ization of PAR-lto the cel ls of the germl ine l ineage suggests a continuedrole of PAR-l in the germl ine. Consistent wi th this hypothe-sis, we have noted agametic phenotypes in associationwith insuff ic ient activ i ty of the maternal par-7 gene in prog-enyof the fol lowing: thepar-l homozygotescarryingpar-1+transforming DNA (YAC Y51 F3), wi ld- type worms injectedwith the par-l antisense or sense RN A, and homozygousherma phrodites bearing a wea k allele of par-l(lw7). Thes eobservations suggest tha t PAR-1 function is essential forgerml ine spe ci f ication, di fferentiat ion, or maintenance.Owing to the correlation between PAR-1 distr ibution andP granule localization in wild-type em bryos , it is possiblethat PAR-l is required for the asym metr ic distribution ofthe P granules, which, in turn, may be involved in speci-fying germl ine cel l fates.A Conserved Mecha nism for GeneratingIntracel lular Asymmetry?Our work presented here shows tha t par-7, whose functionis required fo r generating intracellular asy mm etry in earlyC. elegans embryos, belongs to a conserved kinase sub-fami ly. Interestingly, two mem bers of this KlNlSNFl ki -nase subfam ily may also play roles in generating cel lularpolar i ty. The Schizosaccha romyces pombe K/N7 genehas been shown to be required for cell growth polarity(Levin and Bishop, 1990). The human Kp78 is asymm etr i -cally localized to the apical surface of highly p olarizedepithelial cel ls forming the lumen of the fetal pancreas,and the loss of Kp78 is associated with carcinogenesis(Parsa, 1988). The str ik ing simi lar i ty between PAR-1 andKp78 strongly suggests that they may have simi lar func-t ions. Therefore, the study of PAR-1 function may help usunderstand the role of Kp78 and other kinases in establ ish-ing intracel lular asym metry.Experimental ProceduresStrains and AllelesThe Bristol strain N2 was used as the standard wild-type strain. Thegenetic markers and balancer chromosomes used are the following :LG I, g/p4 (bn2); LG V, rol-4 (scE), h im-5 (e1490), par-l (b274, it32,it57, it60, it78, it86, it89, itgO, it702, lw 7, lw39), uric-76 (e977), ogr-7,nT7 uric (n754) IV; V. Nemato de strains were cultured by standardtechniques (Brenner, 1974).Genetic AnalysisAlleles of par-7 were isolated in screens for recessive, maternal-effect,embryonic lethal mutants, and the par-7 loctis was mappe d 1.5 m.u.to the r ight of ro/-4 and 1 m.u. to the left of uric-760n LG V by three-factorrecombination analysis (Kemph ues et al., 1988). To facil i tate molecularc loning, par-7 was further mappe d with respect to c loned markers

ogr-l and him-5 The ogr-7 locus was tagged with a transposable ele-ment Tel but with no observable phenotype (Telfer, 1991). A strain,rol-4 ogr-7 uric-76/par-7, was constructed, and standard multifactorrecombination analysis was carried out. except that the presence ofogr-7 (Tel) was identif ied using the s ingle worm P CR technique(Barstead et al., 1991). The recombinants obtained (87 between rol-4and ogr-7, 5 between ogr-7 and par-l, 57 between par-7 and uric-76)placed par-7 -0.1 m.u. to the r ight of ogr-7. A nother gene, him-5, wasalso placed in this region of the physical map (S. Penn ington and P.Meneely, personal commun ication). Similarly, rol-4 ogr-7 par-l/him-5was constructed, and the recombinants obtained (219 between rol-4and ogr-7, 4 between ogr-I a nd him-5, and 10 between him-5 andpar-7) indicate that the genetic distance between him-5 and par-7 isabout three times the distance between ogr-7 and him-5, placing par-7-0.075 m.u. to the r ight of him-5.

Germline TransformationSeveral overlapping cosmids, covering -40 kb of genomic DNA tothe r ight of him-5, w ere injected into the gonad of KK29 0 heterozygotes(par-7 (b274)unc-76/nT7 unc(n754)IV; V) indiv idually or in combinationusing the procedures of Mello et al. (1991). N one o f the cosmids res-cued the par-7 phenotype. Total genomic DNA from a yeast strainconta in ing the YAC Y5 1 F3 was embedded in agarose p lugs and sepa-rated by pulsed field gel electrophoresis (PFGE). The posit ion of theYAG was determined by Southern blot analysis. The YAC DNA bandwas excised from the gel and treated with gelase (Epicentre Technolo-gies). The treated DNA sample was dialyzed overnight to remove freeoligosaccharides and concentrated using centr icon (Amicon). The con-centrated DNA was mixed with plasmid DNA carrying the domina ntrol-6 marker an d injected into gonads o f KK29 0 heterozygotes. Rolnon-Uric-78 progeny were picked as indiv iduals to establish transgeniclines. One line segregated Rol Uric-76 herm aphrodites that gave v iablebut agametic progeny (19 of 189 Rol Uric-76 mothers produced94 viable progeny). Non-Rol Uric-76 s iblings were Par (no v iablewwv).Molecular Analyses and cDNA Clonin gStandard methods (Sambrook et al., 1989) were used for all molecularanalyses except where indicated. The band corresponding to Y51F3was excised from the gel after PFGE as described above and usedin agarose-embedde d restr ic tion digestion with Ascl (New Engla ndBiolabs) overnight at 37%. The digested DNA was then separatedby PFGE, and two fragments, 100 kb and 150 kb, were obtaine d. The100 kb fragment was found by Southern blot analysis to contain him-5but not ogr-7. The 100 kb fragment was purif ied by GeneClea n (810101) and used as a probe in standard Northern blot analysis. A mater-nally enriched transcript of - 4.4 kb was detected. cDNAs correspond-ing to this transcript were obtained by screening a C. elegans hZAPcDNA library (Barstead and Waterston, 1989) using the 100 kb frag-ment as the probe. Since the YAC vector contained within the 100 kbfragm ent cross-hybridizes to the lZA P vector, extensive wa shes werecarried out. Five strongly hybridiz ing c lones were isolated, four ofwhich hybridized to a 4.4 kb transcript by Northern blot analysis. AcDNA clone w ith insert s ize of 2.5 kb was used as the probe to furtherisolate a 4 kb cDNA clone, named iC22, which also hybridized to the4.4 kb transcript.

ZC22 was u sed as the probe in Southern blot analysis to detectallele-associated restr ic tion fragment length polymorphisms. GenomicDNA was prepared from wild-type (N2) and par-7 strains, and standardSouthern blot analysis was performed.Antisense Phenocopy AnalysisThe ZC22 clone was used as a template to synthesize in v itro tran-scribed, capped antisense RNA. The plasmid DNA was linearized withappropriate restr ic tion endonucleases and then treated with protein-ase K, followed by phenol-chloroform extractions and ethano l precipi-tation. Reagents used were treated with diethyl pyrocarbonate (DEPC)to inactivate RNases. RNAs were synthesized using Stratagene invitro transciiption k it, except th at a cap analog (Stratagene) and thehuman placental RNase inhibitor (Bethesda Research Laboratories)were added. The concentration of injected RNA ranged from 1 mglml up to 5 mglm l; the higher concentrations appeared to give betterresults. Wild-type animals were injected with RNAs in the distal arm

7/28/2019 Guo Sk1995

9/10

Kinase Involvement in Asymmetric Div is ions61 9

of one or both gonads. Injected worms were plated out indiv idually at20C to allow e gg laying a nd were transferred to a new plate everyday. Embryonic lethality (failure to hatch) was checked a minim umof 24 hr after each transfer. The early embryonic phenotypes wereobserved by cutting worms open in disti l led HZ0 and transferring thelive early embryos to agar pads for microscopy.Sequenc ing and ldent l f lcat ion of the 5 EndsDNA sequence was determined using the dideoxy chain terminatormethod (Sanger et al., 1977). Seque ncing was performed by usingSequen ase (United States Biochemicals) and double-stranded plas-mid DNA; some sequences were determined by the Cornell UniversityBiotechnology Program Automatic S equencing Facil ity . Both strandsof the cDNA were sequenced. The Sen d o f the par-1 gene was deter-mined u sing RACE-PC R, according to the instructions of the manufac-turer (Clonetech). Sequen ce analysis was done using the FASTA(Pearson and Lipma n, 1988) and BLAS T programs (Altschul et al.,1990) within the Genetics Computer Group package (Madison, Wis-consin).

For sequencing of the k inase d omain in the par-7 m utant alle les,synthetic oligonuc leotide primers were made that would enab le theentire k inase doma in to be amplif ie d from the mutants by reversetranscription-PCR (RT-PCR). Total RNA was extracted from - 50 ho-mozygous par-l hermaphrodites (N2 was used as a control) and usedfor reverse transcription (first-strand cDNA synthesis kit; Pharm acia)and PCR. Alterations reported in the mutant alle les (it51 nd if90) weresequenced from at least f ive indepe ndent PCR products.Ant ibodies and lmmunos ta in ingTo generate polyclonal antibodies against PAR-I, we c loned a 0.65 kbfragment 3to the k inase domain cod ing region into pGEX-3X (AmradCorporation), and a fusion protein of the expected s ize (- 45 kDa) wasobtained . This protein was purif ied by aff inity bin ding to agarose-immob ilized glutathio ne and injected into two rabbits.

For aff inity purif ication, the crude antisera were first passed overa column containing the GST protein coupled to Affigel 102 (Bio-Rad)to elimina te any anti-GST antibodies and then were aff inity purif iedby bindin g to the GST-PAR-1 fusion p rotein co upled to the same ma-tr ix . After washing, the bound a ntibodies were eluted w ith low pHglycine and dialyzed against phosphate-buffered saline (PBS).

The abil ity of the purif ied antibody to recognize the PAR-l fusionprotein and endogeno us PAR-1 protein was verif ied on Western blots.To detect endoge nous PAR-l, we hand picked - 100 gravid hermaph-rodites of each genotype. After washes with M9 buffer, the wormswere boile d and ground in SDS sample buffer. The whole worm ex-tracts were resolved by SDS-PA GE and blotted onto nitrocellulosemembranes. The Western blots were performed with purif ied PAR-1antibody and horseradish peroxidase-linked donkey anti-rabbit IgG.The ECL detection system (Amersham) was used.

Embryos were stained by a modification of the protocol describedpreviously (Albertson, 1984). For each genotype examined, severalsamples of lo-20 hermaphrodites each were transferred to disti l ledHz0 (- 8 ~1) on poly lysine s lides. A coverslip (18 mm x 18 mm) wasadded, pressure was applie d to extrude embryos, and s lides wereplaced in l iquid nitrogen for 10 min. After removal of the coverslip,s lides were incubated for 15 min in -20% absolute metha nol andthen transferred to room-temperature PBS . The samples were blockedin goat serum for 1 hr and incubated with PAR-l antibody (diluted in1% bovine serum album in, 10% goat serum in PBS) for 1 hr at roomtemperature or were coincubated with anti-P granule an tibody(OICiD4 or K76 or both) at 16% for 4 hr. Slides were washed twicefor 10 m in in PBS and once for 10 m in in PBST (0.5% Tween inPBS) in between the two PBS washes. Slides were then incubated insecondary antibodies(for PAR-I, g oat anti-rabbit FITC; for P granules,goat anti-mouse rhodamine) for 1 hr at room temperature, followedby washes as described above, except that DAPI was included in thefirst wash. After the f inal wash, s lides were mou nted with moun tingmedia and v iewed und er f luorescence microscopes or confocal m icro-scopes (Cornell University Biotechnology Program Confoca l Micros-copy Facil ity). Typically, each s lide had between 10 and 40 embryosof each early stage (from zygote to 32-cell s tage, low er numbers oflater stages); a minim um of 50 s lides were e xamined for each of N2,it50, itSO, and lw3 9 genotypes.

For immunofluorescence detection of PAR-1 in the gonad, a proto-col adapted from the laboratory of T. Schedl (Wa shington UniversitySchool o f Medic ine, St. Louis, Missouri) was used: heads of hermaph-rodites were cut off (using a syringe needle) in PB S containing 0.2mM levamisole to allow the extrusion of gonads; the worm carcasseswere transferred to poly lysine s lides after f ixation in -20% methan olfor 5 min. The same staining procedures used for embryos were ap-plied thereafter. For SKN-1 staining, the monoclo nal antibody FA2 andthe procedure described by Bowerma n et al. (1993) were used.

To test for specif icity in the immunfluorescence assay, we incubatedsamples with aff inity-purif ied PAR-l antibody as described above ex-cept that the dilution buffer con tained 4 mglm l of the GST-PAR-1fusion protein.

AcknowledgmentsCorrespondence should be addressed to K. J. K. The authors thankthe C. elegans genetics stock center for providing strains; Alan Coul-son and John Sulston for providing cosmids and YACs that enab ledus to map and c lone par-l; Abby Telfer for providing the Tel-taggedogr-7 strain and for communicating unpublishe d data on the physicalmap location of par-l; Jocelyn Shaw for par-l( lw39) and par-l( lw7)alleles; Susan Strome for anti-P gra nule antibody; Bruce Draper andJim Priess for anti-SKN-1 antibody. We would l ike to thank Bingw ei Lu,Lesilee Rose, and Bijan Etemad -Mogha dam for helpfu l discussions;members of the Kemphu es lab for valuable comments on the manu-script; Drs. Anthony Bretscher, Mariana W olfner, Thomas Fox, MichaelGoldberg, and three anonymous reviewers for useful suggestions andcrit ical readings of the manuscript; J im Slattery an d Carol Bayles forassistance on use of the confoca l microscope. This work was sup-ported by Nationa l Institutes of Health grant HD2768 9.Received January 19, 1995; revised March 1 0, 1995.ReferencesAlbertson. D. (1984). Formation of the f irst c leavage spindle in nema-tode embryos. Dev. Biol. 707, 61-72.Altschul, S. F., Mil ler, W., Myers, E. W., and Lipma n, D. J. (1990).Basic local alignm ent search tool. J. Mol. Biol . 275, 403-410.Barstead, R. J., and Waterston, R. H. (1989). The basal componen tof the nematod e dense-body is v inculin. J. Biol. Chem. 264, 10177-10185.Barstead, R. J., Kleim an, L., and Waterston, R. H. (1991). C loning,sequenc ing and mapping of an a-ac tin gene from the nematode Caeno-fhabdit is elegans. Cell. Motil. Cytoskel. 20, 69-78.Bektesh, S., Van Doren, K., and Hirsh, D. (1988). Presence of theCaenorhabd ir is elegans spliced leader on different mRNAs and in dif-ferent genera of nematodes. Genes Dev. 2, 1277-1283 .Bowerma n, B., Draper, B. W., Mello, C. C., and Priess, J. R. (1993). Thematernal gene &n-l encodes a protein that is distr ibuted unequa lly inearly C. elegans embryos. Cell 74, 443-452.Brenner, S. (1974). The genetics of Caenorba bdit is elegans. Genetics77, 71-94.Carling, D., Aguan , K., Woods, A., Verhoeven, A., Beri, R., Brennan,C. H., Sidebo ttom, C., Davison, M. D., and Scott, J . (1994). Mam malianAMP-activated protein k inase is homolog ous to yeast and plant proteinkinase involved in the regulation of carbon metabo lism. J. Biol. Chem.269, 11432- l 1448.Celenza, J. L., and Carlson, M. (1986). A yeast g ene that is essentialfor release from glucose repression encodes a protein k inase. Science233, 1175- I 180.Chant, J. (1994). Cell polarity in yeast. Trends Genet. 10, 328-333.Cheng, N. N., Kirby, C. M., and Kemphues, K. J. (1995). Control ofc leavage spindle organization in Caenorhab dit is elegans: the role ofthe genes par-2 and par-% Genetics 739, 549-559.Davidson, E. (1986). Gene Activ ity in Early Developmen t, Third Edit ion(New York: Academic Press).Fire, A., Albertson, D., Harrison, S. W., and Moerman, D. G. (1991).Production of antisense RNA leads to effective and specif ic inhibit io n

7/28/2019 Guo Sk1995

10/10

Cel l62 0

of gene expression in Caenorhab dit is elegans muscle. De velopment7 73, 503-514.Goldstein, 8. (1992). lnductionofgut inCaenorha bdii ise/egans. Nature357, 254-257.Goldstein, B. (1993). Establishment of gut fate in the E lineag e of C.elegans: the roles of l ineage-de penden t mechanisms and cell interac-tions Developmen t 778, 1267-1277 .Hanks, S. K., and Quinn, A. M. (1991). Protein k inase catalytic doma insequence database: identif ication of conserved features of primarystructure and c lassif ication of family members. Meth. Enzymol. 200.38-62.Hanks, S. K., Quinn, A. M., and Hunter, T. (1988). The protein k inasefamily: conserved features a nd deduced phylogeny of the catalyticdomains. Science 247, 42-52.Higgins, D. G., and Sharp, P. M. (1989). Fast and sensit ive multip lesequence alignmen ts on a microcomputer. Comput. Appl. Biosci. 5,151-153.Hil l, D. P., and Strome, S. (1988). An analysis of the role of microfi la-ments in the establishment and maintenan ce of asymmetry in Caeno-rhabdit is elegans zygotes. Dev. Biol. 725, 75-84.Hil l, D. P., and Strome, S. (1990). Briefcytochalasin-induced disruptionof microfi laments during a crit ical interval in l-cell C. elegans embryosalters the partit ioning of developm ental instructions to the 2-cell em-bryo. Developmen t 708, 159-172.Horvitz, H. R.. and Herskowitz, I. (1992). Mecha nisms of asymmetriccell div is ion: two Bs or not two Bs, that is the question. Cell 68, 237-255.Inglis , J. D., Lee, M., and Hil l, R. E. (1993). Emk, a protein k inase withhomologs in yeast maps to mouse chromosome 19. Mam mal. Genome4, 401-403.Kemphu es, K. J., Priess, J. R., Morton, D. G., and Cheng, N. (1988).Identif ication of genes required for cytoplasmic localization in earlyembryos of C. elegans. Cell 52, 31 l-320.Kieber, J. J, Rothenberg, M., Roman, G., Feldman n, K. A., and Ecker,J. R. (1993). CTRI, a negative regulator of the ethylene responsepathway in Arabidopsis, encodes a member of the raf family of proteinkinases. Cell 72, 427-441.Kirby, C., Kusch, M., and Kemphues, K. (1990). Mutations in the pargenes of Caenorhab dit is elegans affect cytoplasmic reorganizationduring the first cell cycle. Dev. Biol. 742, 203-215.Krause, M., and Hirsh, D. (1987). A trans-spliced leader sequence onac t in mRNA in C. e legans . Cel l 49, 753-761.Levin, D. E., and Bishop, J. M. (1990). A putative protein k inase gene(kinl) is important for growth polarity in Schizosaccharomycespombe.Proc. Natl. Acad. Sci. USA 87, 8272-8276 .Levin, D. E., Hammo nd, C. I., Ralston, R. O., and Bishop, J. M. (1987).Two yeast genes that encode unusual protein k inases. Proc. Natl.Acad. Sc i . USA 84, 6035-6039.Manser, E., Leung, T., Salihu ddin, H., Zhao, Z., and Lim, L. (1994).A brain serinelthreonine protein k inase activated by Cdc42 and Racl.Nature 367, 40-46.Mello, C. C., Kramer, J. M., Stinchcomb, D., and Ambros, V. (1991).Effic ient gene transfer in C. elegans: extrachromosomal maintenan ceand integration of transforming sequences. EMBO J. 70, 3959-3970 .Mello, C. C., Draper, B. W., and Priess, J. R. (1994). The matern algenes apx-7 and g/p-l and establishment of dorsal-ventral polarity inthe early C. elegans embryo. Cell 77, 95-106.Morton, D. G., Roos, J. M., and Kemphue s, K. J. (1992). par-4, a generequired for cytoplasmic localization and determina tion of specif ic ce lltypes in Caenorha bdit is elegans embryogenesis. Genetics 730, 771-790.Parsa. I. (1988). Loss of a M, 78,000 marker in chemically inducedtransplantable carcinomas and primary carcinoma of huma n pan-creas Cancer Res. 48, 2265-2272.Pearson, W. R., and Lipm an, D. J. (1988). Improved tools for biologica lsequence analysis. Proc. Natl. Acad. S ci. USA 85, 2444-2448 .Priess, J. R., and Thomson, J. N. (1987). C ellular interactions in earlyC. elegans embryos. Cell 48, 241-250.

Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989). Molecular Clon-ing: A Laboratory Manu al, Secon d Edit ion (Cold Spring Harbor, NewYork: Cold Spring Harbor Laboratory).Sanger, F., Nicklen, S., and Coulson, A. (1977). DNAsequencing withchain-terminating inhibitors. Proc. Natl. Acad. Sci. USA 74, 5463-5467.Schierenberg, E. (1985). Cell determina tion during earlyembryogene-sis of the nematod e Caenorhab dit is elegans. Cold Spring HarborSymp. Quant. Biol. 50, 59-68.Schierenberg, E. (1987). Reversal of cellular polarity a nd early cell-cell interaction in the embryo of Caenorhab dit is elegans. Dev. Biol.722, 452-463.Schierenberg, E. (1988). Localization and segregation of l ineage-specif ic c leavage potentia l in embryos of Caenorhabd it is elegans.Rouxs Arch. Dev. Biol. 797 , 282-293.Schnabe l, R. (1994). Autonomy and nonautono my in cell fatespecif ica-tion of muscle in the Caenorhab dit is elegans embryo: a reciprocal in-duction. Science 263, 1449-1452 .Slack, J. (1991). From Egg to Embryos, Second Edit ion , P. W. Barlow,D. Bray, P. 8. Green, and J. M. W. Slack, eds. (Cambridge, Engla nd:Cambridge University Press).Strome, S . (1986a). Asymmetric movements of cytoplasmic compo-nents in Caenorhab dif is elegans zygotes. J. Embryol. Exp. Morphol.97, 15-29.Strome, S. (198613). Fluorescence v isualization of the distr ibution ofmicrofi laments in gonads and early embryos of the nemato de Caeno-rhabdif is elegans. J. Cell Biol. 703, 2241-2252 .Strome. S., and Wood, W. B. (1982). lmmunofluorescence v isualiza-tion of germ-line-specific cytoplasmicgranules in embryos, larvae, andadultsof Ca enorhabdit iselegans. Proc. Natl. Acad. Sci. USA 79,1558-1562.Strome, S., and Wood, W. 8. (1983). Generation of asymmetry andsegregation of germ line granules in early C. elegans embryos. Cell35, 15-25.Sulston, J., Schierenberg, E., White, J.. and Thomson, N. (1983). Theembryonic cell l ineag e of the nemato de Caenorhabd it is elegans. Dev.B io l . 700, 67- l 19.Telfer, A. (1991). Identif ication and molecular analysis of ogr-7, a ma-ternally expressed gene of Caenorhab dit is elegans. PhD thesis, Har-vard University, Cam bridg e, Massachusetts.GenBank Accession NumberThe accession number for the par-7 sequence reported in this paperis U22183.