document

© 2000 Macmillan Magazines Ltd

brief communications

NATURE CELL BIOLOGY VOL 2 NOVEMBER 2000 http://cellbio.nature.com 855

Type II myosin regulatory light chain relieves auto-inhibition ofmyosin-heavy-chain function

Naweed I. Naqvi*, Kelvin C. Y. Wong*, Xie Tang* and Mohan K. Balasubramanian*†*Cell Division Laboratory, Institute of Molecular Agrobiology, The National University of Singapore, 1 Research Link, Singapore 117604

†e-mail: [email protected]

The F-actin based motor protein myosin II has a key role incytokinesis. Here we show that the Schizosaccharomycespombe regulatory light chain (RLC) protein Rlc1p binds toMyo2p in manner that is dependent on the IQ sequencemotif (the RLC-binding site), and that Rlc1p is a componentof the actomyosin ring. Rlc1p is important for cytokinesisat all growth temperatures and is essential for this processat lower temperatures. Interestingly, all deleterious pheno-types associated with the loss of Rlc1p function are sup-pressed by deletion of the RLC binding site on Myo2p. Weconclude that the sole essential function of RLCs in fissionyeast is to relieve the auto-inhibition of myosin II function,which is mediated by the RLC-binding site, on the myosinheavy chain (MHC).

The importance of myosin II in cytokinesis has been establishedin several organisms1. Type II myosin dimers associate with twoessential light chains (ELC) and two regulatory light chains2.

RLCs have received considerable attention, given the potential roleof M-phase-promoting factors in regulating the timing of cytokine-sis through phosphorylation of RLC1. However, the functions ofRLCs remain largely unresolved. Interestingly, although RLCs havebeen shown to be essential for cytokinesis in Drosophila3 and in sus-pended Dictyostelium cells4, Dictyostelium mutants expressing typeII myosin lacking the RLC-binding site do not exhibit any obviousdefects5. Our analysis of S. pombe Rlc1p, an RLC of type II MHCsthat is essential for actomyosin-ring function6–9, might provide anexplanation for this paradox.

The S. pombe myo2+ gene encodes a type II MHC that is essen-tial for assembly of the actomyosin ring6,7,10,11 at all growth temper-atures from 19–36 °C (K.C.Y.W. and N.I.N., unpublished observa-tions). A second MHC gene, myp2+ (refs 8, 9, 11), which is essentialfor cytokinesis under conditions of stress, has also been identifiedin S. pombe. The S. pombe genome-sequencing project predicted anopen reading frame (ORF, SPAC926.03) encoding a 184-amino-acid peptide that is closely related to myosin II RLCs from Japanesesquid, rat and Drosophila (Fig. 1a; 36% identity over the entirelength of these proteins). We hereafter refer to this protein as Rlc1p,and to the gene that encodes it as rlc1+. Rlc1p, an EF-hand protein,is less closely related (23% identity) to Cdc4p12, a previously char-acterized light chain that associates with Myo2p13,14 and Myp2p14.To determine whether Rlc1p is physically associated with Myo2p,we constructed a strain in which the endogenous copy of rlc1+ wastagged with the gene encoding green fluorescent protein (GFP). Weimmunoprecipitated lysates prepared from wild-type cells express-ing GFP-tagged Rlc1p and from untagged wild-type cells with anti-bodies against GFP or Myo2p, and analysed the resulting immunecomplexes by immunoblotting with Myo2p-specific antibodies. Asshown in Fig. 1b, GFP–Rlc1p specifically associated with Myo2p. Todeteremine whether Cdc4p, which has previously been shown to

associate with Myo2p13,14, was present in the Rlc1p–Myo2p com-plex, we analysed immune complexes, generated with antibodiesagainst GFP from wild-type cells expressing GFP-tagged Rlc1p, byimmunoblotting using antibodies against Myo2p, GFP or Cdc4p(Fig. 1c). Anti-GFP antibodies precipitated proteins of relativemolecular mass 190,000 (Mr 190K; recognized by anti-Myo2p anti-bodies), 47K (GFP–Rlc1p, recognized by anti-GFP antibodies), and15K (recognized by anti-Cdc4p antibodies). In the untagged wild-type strain, none of these proteins were immunoprecipitated byanti-GFP antibodies (Fig. 1c). The immunoprecipitation experi-ment established that Myo2p, Rlc1p and Cdc4p are present in acomplex in vivo. We used GST–Rlc1p pull-down assays to confirmindependently the association between Rlc1p and Myo2p (Fig. 1d).To ascertain whether this interaction is mediated by the previouslypredicted RLC-binding site on Myo2p7 (the IQ2 motif, betweenamino acids 791 and 813 of Myo2p), we created a strain in whichthis region was precisely deleted. Surprisingly, growth rates and thegeneral phenotype of cells expressing Myo2p(∆IQ2) as their onlyMyo2p source were comparable to those of wild-type cells (Fig.3e–f and data not shown). The interaction between Myo2p andRlc1p was completely abolished by this mutation (Fig. 1d, rightpanel), establishing that Rlc1p binds to Myo2p through the secondIQ sequence motif. In further analyses, Rlc1p was capable of asso-ciating with GFP–Myo2p(R770A) (ref. 13), a derivative of Myo2pthat is unable to bind to Cdc4p (Fig. 1e); and Cdc4p was able tobind to Myo2p(∆IQ2) in co-immunoprecipitation experiments(Fig. 1f). On the basis of its sequence characteristics and its exclu-sive binding to the IQ2 motif of Myo2p, we conclude that Rlc1pmay serve as a bona fide RLC that associates with Myo2p.

Rlc1p is a component of the actomyosin ring and was detectedin these rings in early and late mitotic cells, as well as in cells inwhich the rings were undergoing constriction (Fig. 2a). We there-fore tested whether the localization of Rlc1p to the actomyosin ringis dependent on Myo2p. As described earlier, Rlc1p was detected atthe medial ring in early as well as late mitotic wild-type cells.Surprisingly, Rlc1p was still detected in medial rings in cells express-ing Myo2p(∆IQ2) (Figure 2b). However, in cells expressingMyo2p(∆IQ2), Rlc1p–GFP rings were detected in late, but not early,mitotic cells. Previous studies have shown that S. pombe Myp2p is anon-essential MHC8,9 that localizes to the actomyosin ring distinct-ly later during anaphase than Myo2p15. We reasoned that the abilityof Rlc1p to localize to the actomyosin ring in cells expressingMyo2p(∆IQ2) (Fig. 2b, arrowhead) might be a result of theRlc1p–Myp2p interaction. Consistent with this possibility is thefinding that Rlc1p–GFP localizes to the medial ring only in latemitotic cells expressing Myo2p(∆IQ2) as their sole Myo2p source.We therefore investigated whether localization of Rlc1p to the medi-al ring was abolished in myp2-null mutant cells expressingMyo2p(∆IQ2) as their sole Myo2p source. Indeed, Rlc1p was notdetected at the division site in myp2∆ cells expressing Myo2p(∆IQ2)



(Fig. 2b), indicating that Rlc1p may also associate with Myp2p.Thus, the assembly of Rlc1p at the division site depends on inter-actions with the two type II MHC proteins, Myo2p and Myp2p.

To study the cellular function of Rlc1p, we constructed a strainin which the rlc1 gene was deleted (rlc1∆). rlc1∆ cells were capableof division and colony formation at temperatures ranging from24–36 °C, but were unable to do so at 19 °C (Fig. 3a). However,even at the termperatures at which these cells were able to formcolonies, abnormalities related to cytokinesis were generallyobserved. For example, growing populations of rlc1∆ cells con-tained a higher proportion of binucleate (~50%) and tetranucleate(~14%) cells relative to wild-type populations (15% binucleate and0% tetranucleate; data not shown). At 19 °C, rlc1∆ cells were unableto assemble proper actomyosin rings and, as a result, assembledimproper septa or apparently normal-looking septa that failed tocleave, leading to a failure in cytokinesis (Fig. 3b–d). Both F-actinand Myo2p accumulated at the division site and formed non-func-tional actomyosin rings. Thus, Rlc1p is not required for the trans-port and accumulation of Myo2p at the division site; rather itseems essential for optimal Myo2p function. It is also noteworthythat although rlc1∆ cells exhibited cytokinesis abnormalities at alltemperatures, Rlc1p is essential only at lower temperatures. Onepossibility is that the function of myosin II is itself cold-labile,thereby amplifying the effects of lack of Rlc1p at lower tempera-tures. Alternatively, other proteins related to Rlc1p might substitutefor Rlc1p at higher temperatures.

The finding that cells expressing Myo2p(∆IQ2) as their soleMyo2p source did not exhibit any obvious phenotypic changes wasin stark contrast to the defective cytokinesis observed in rlc1∆ cells.These data are similar to the results of genetic studies involvingDictyostelium, which have shown that whereas RLC4 and MHC16,17,18

NATURE CELL BIOLOGY VOL 2 NOVEMBER 2000 http://cellbio.nature.com856

Anti-Cdc4pAnti-Myo2p

IPf

d Wild typeLYS GST RLC LYS GST RLC

myo2::∆IQ2

myo2::∆IQ2LYS GST RLCGFPmyo2::R770A

e

Wild type

Pre PreAnti-GFP Anti-GFPAnti-Myo2p Anti-Myo2pLys Lys

Myo2p

rlc1::GFPb

cIP: anti-GFP

rlc1::GFP Wild type

Myo2p

Rlc1p

Cdc4p

IgG

14.3

46

220

Mr(K)

a 1 MFSSKENSLGAKRAPFSSNTTSSQRVAAQAAKRASSGAFAQLTSSQIQEL 1 --------------------------------AEEAPRRVKLSQRQMQEL 1 ----------------MSSKRAKAKTTKKRPQSATSNVFAMFDQSQIQEF 1 --------------MSSRKTAGRRATTKKRAQRATSNVFAMFDQAQIAEF

51 KEAFALLDKDGDGNIGREDVKTMLTSLNQDASEDSINHMFESINPPINLA 19 KEAFTMIDQDRDGFIGMEDLKDMFSSLGRVPPDDELNAMLKECPGQLNFT 35 KEAFNMIDQNRDGFIDKEDLHDMLASLGKNPTDEXLEGMMNEAPGPINFT 37 KEAFNMIDQNRDGFVEKEDLHDMLASLGKNPTDDYLDGMMNEAPGPINFT

101 AFLTAMGSMLCRISPRNDLLEAFSTFDDTQSGKIPISTMRDALSSMGDRM 69 AFLTLFGEKVSGTDPEDALRNAFSMFDEDGQGFIPEDYLKDLLENMGDNF 85 MFLTMFGEKLNGTDPEDVIRNAFACFDEEASGFIHEDHLRELLTTMGDRF 87 MFLTLFGERLQGTDPEDVIKNAFGCFDEENMGVLPEDRLRELLTTMGDRF

151 DPQEVESILRSYTS-HGVFYYEKFVDAIAGSKDSN--- 119 SKEEIKNVWKDAPLKNKQFNYNKMVDIKGKAEDED--- 135 TDEEVDEMYRERIDKKGNFNYVEFTRILKHGAKDKDD- 137 TDEDVDEMYREAPIKNGLFDYLEFTRILKHGAK EQ

S. pombeT. pacificus

R. norvegiusD. melanogaster

S. pombeT. pacificusR. norvegiusD. melanogaster



Figure 1 Rlc1p is a regulatory light chain that associates with Myo2p. a,Comparison of the amino-acid sequence of S. pombe Rlc1p with those of the RLCfrom Todpa pacificus (LC-2, P08052); Rattus norvegicus (MLRN_Rat, Q64122) andDrosophila melanogaster (MRLC-C, P40423). Identical amino acids are shown inwhite against black; conservative changes are shaded. b, Rlc1p associates withMyo2p in vivo. Non-denaturing extracts prepared from wild-type and rlc1::GFPcells were immunoprecipitated with anti-GFP and anti-Myo2p antibodies and withpre-immune sera (Pre), and were processed for detection of Myo2p byimmunoblotting with anti-Myo2p antibodies. Lysates prepared under denaturingconditions from each strain (Lys) were used as positive controls for Myo2p detec-tion. c, The Rlc1p–Myo2p complex contains Cdc4p. Non-denaturing lysates wereprepared from wild-type cells expressing GFP-tagged Rlc1p and from untaggedwild-type cells, and immunoprecipitated (IP) with anti-GFP antibodies. The resultingprotein complexes were immunoblotted with antibodies against Myo2p, GFP andCdc4p. IgG, immunoglobulin G. d, Rlc1p associates with the IQ2 motif of Myo2p.Non-denatured protein extracts prepared from wild-type S. pombe cells and fromthe myo2::∆IQ2 strain were allowed to bind either to GST beads or to beads con-taining bound GST–Rlc1p (RLC). The respective eluates were analysed byimmunoblotting with anti-Myo2p antisera. Cell lysates (Lys) were also included inthe immunoblots to positively identify Myo2p and Myo2p(∆IQ2). e–f, Specificity ofthe binding of RLC and Cdc4p to the IQ2 and IQ1 motifs in Myo2p, respectively. e,Lysates obtained from cells expressing GFP–Myo2p(R770A) as their sole Myo2psource were allowed to bind to GST or GST–Rlc1p beads. Bound proteins wereanalysed by immunoblotting with anti-Myo2p antibodies. Cell lysates (Lys) werealso included in the immunoblots to positively identify GFP–Myo2p(R770A). f,Lysates were prepared from the myo2::∆IQ2 strain and immunoprecipitated withantibodies against Myo2p or Cdc4p. Immune complexes were analysed byimmunoblotting using antibodies against Myo2p. Arrowheads in b and d–f showthe position of the Mr 200K marker protein.

Wild type myo2::∆IQ2 myo2::∆IQ2 myp2∆

DNA

Rlc1–GFP

b

F-actinRlc1–GFPDNAa

Figure 2 Rcl1p localization and dependencies. a, Rlc1p is a component ofmedial actomyosin rings. Cells of the rlc1::GFP strain were fixed with formaldehydeand stained with rhodamine-conjugated phalloidin and DAPI to visualize F-actin andnuclei, respectively. Arrowhead indicates a Rlc1p–GFP ring undergoing constriction.b, Dependency of Rlc1p localization on Myo2p and Myp2p. Localization ofRlc1p–GFP was analysed in early and late mitotic cells of strains of the relevantgenotypes: rlc1::GFP; rlc1::GFPmyo2::∆IQ2; and rlc1::GFPmyo2::∆IQ2 myp2∆.Nuclei were stained with DAPI. Arrowhead indicates the lack of rlc1::GFP localiza-tion in an early mitotic cell of the rlc1::GFPmyo2::∆IQ2 strain.



NATURE CELL BIOLOGY VOL 2 NOVEMBER 2000 http://cellbio.nature.com 857

are essential for cytokinesis in suspension, Dictyostelium cellsexpressing mutant MHC lacking the RLC-binding site have no obvi-ous defects5. On the basis of crystallographic studies of type II MHC,it has been proposed that the region containing the RLC-binding siteserves as a cis-inhibitory domain of MHC function19. Thus, it is pos-sible that the function of Rlc1p is to relieve MHC of inhibition medi-ated by its RLC-binding site. Alternatively, RLC might associate withother proteins that are essential for cytokinesis. To distinguishbetween these possibilities, we crossed a strain expressingMyo2p(∆IQ2) as the sole Myo2p source to an rlc1∆ strain. If the for-mer hypothesis is correct, cells expressing Myo2p(∆IQ2) should becapable of rescuing the cytokinesis defect of rlc1∆ cells. We foundthat from among the four meiotic-spore products in tetrads of thetetratype configuration, three were capable of division and colonyformation at 19 °C. We established the genotypes of the four strainsby polymerase chain reaction (PCR, Fig. 3g). Phenotypic characteri-zation established that rlc1∆ cells expressing Myo2p(∆IQ2) as theirsole Myo2p source resembled wild-type cells in morphology andgrowth rate, and were capable of proper cytokinesis (84% uninucle-ate; 15.6 % binucleate; 0.4% tetranucleate) and colony formation at19 °C, unlike rlc1∆ cells (Fig. 3e, f). Thus, deletion of the RLC-bind-ing site in Myo2p creates a myosin molecule that does not requireregulation mediated by RLC, and therefore is able to suppress thedeleterious effects resulting from loss of Rlc1p. These findings alsorule out the latter possibility, that Rlc1p might interact with otherpartners that are essential for cytokinesis.

We have provided evidence that the only essential function ofRlc1p in S. pombe is to relieve the auto-inhibitory effects of Myo2p,which are mediated by its RLC-binding domain (IQ2 sequencemotif). We have also established that regulation by RLC is not aprerequisite for type II myosin function. Future studies are requiredto address the question of how Rlc1p is regulated.

MethodsConstruction of disruption strain.We disrupted the complete rlc1+ ORF with ura4+ using a one-step gene-replacement strategy. The final

disruption construct consisted of the 1.8-kilobase (kb) ura4+ gene flanked by 5′ and 3′ non-coding

regions (600 base pairs (bp) each) of the rlc1 locus. This deletion construct was introduced as a lin-

earized KpnI–NotI fragment into a uracil auxotrophic diploid strain. Ura+ diploid transformants were

selected and screened by PCR to identify the diploid strains in which the desired gene-replacement

event had occurred.

Expression of GST–Rlc1pA 0.55-kb XhoI–HindIII fragment encoding complete Rlc1p was expressed as a fusion protein with

glutathione-S-transferase in Escherichia coli20 and purified using glutathione–agarose beads (Sigma).

Generation of rlc1::gfp strain.A strain expressing Rlc1p fused to GFP at its carboxy terminus was constructed by first amplifying

rlc1+ together with its 5′ UTR as a 1.2-kb PCR fragment. The PCR product was cloned into the inte-

grating vector pJK21021, using the KpnI–BamHI sites from the primers. The GFP(F64L S65T) DNA

fragment was cloned in-frame at the 3′ end of the rlc1+ gene as an NdeI–BamHI fragment, to yield

plasmid pCDL527. The final construct was linearized with EcoRI and introduced into the uracil aux-

otrophic strain MBY192; correct integration was confirmed by PCR.

Deletion of the IQ2 motif in myo2+.The region encoding amino acids 791–813 of myo2+ (IQ2 motif) was deleted using mutagenic PCR

primers MOH446 (5′-AAACAAGGATCCGTGGCTTTCTTCAAAGAAAAATTTTTCAAAAGC-

GATTGAAGGACATACAAGCCATCAAATTAAGGCCCTTATTAAGTAGTACTC) and MOH127 (5′-CGTCAAACTCGATACCTT). The resulting 1.2-kb PCR fragment was digested with BamHI–BglII and

used to replace the 1.2-kb BamHI–BglII region in pRep81GFP–Myo2 (ref. 13), to generate plasmid

pRep81GFP–Myo2(∆IQ2) (pCDL534). A 4.5-kb NheI–SmaI fragment of myo2+ encompassing the

∆IQ2 mutation in pCDL534 was cloned into the XbaI–SmaI sites of pJK210 (ref. 21). This construct

was digested with NdeI (460 bp downstream of the NheI site); the resulting cohesive ends were filled in

using Klenow polymerase and then re-ligated. This final construct was linearized with BamHI and

introduced into a haploid wild-type strain MBY192. Stable Ura+ transformants were analysed by PCR

and by nucleotide-sequence determination to identify strains that had the myo2+ locus replaced with

myo2::∆IQ2ura4+. This integration event leaves a partial and defunct copy of myo2+ (lacking the pro-

moter and the proximal 190 amino acids, and frame-shifted as a result of the NdeI fill-in and re-liga-

tion described above), just downstream of the mutant myo2::∆IQ2ura4+ locus. To confirm the

myo2::∆IQ2 mutation, a 1-kb fragment (spanning bases 1,503–2,448 for the myo2::∆IQ2 mutation, or

representing bases 1,503–2,517 in myo2+) was amplified by PCR using primers MOH348 (5′-ACAGC-

CGACTATTGATTT) and MOH351 (5′-CTCGTACTTCAGCTCTAT). The PCR product was digested

with Kpn1 (at 625 bp) and the products were resolved on an agarose gel. Digestion of the PCR product

with KpnI generated a 625-bp and a 389-bp fragment for the wild-type myo2+, and in the case of

myo2::∆IQ2 mutation resulted in a 625-bp and a 320-bp fragment, along with the 625-bp and 389-bp

fragments contributed by the defunct and frame-shifted copy of myo2 (see above).

Immunoprecipitation and immunoblotting.Primary antisera (anti-Myo2p or anti-GFP) were used at a 1:1000 dilution. Non-denaturing extracts

were produced in TritonX-100 buffer; immunoprecipitation was carried out as described13.

Immunoblot detection was carried out using the chemiluminescent method with peroxidase-conjugat-

ed secondary antibodies (Sigma).

Fluorescence microscopy.F-actin staining was carried out as described22. Staining with anti-Myo2p antibodies was carried out on

methanol-fixed cells as described22. Alexa594-conjugated secondary antibodies were from Molecular

Probes. Cells were viewed using a Leica DMLB microscope and images were obtained with an

Optronics DEI-750T cooled CCD (charge-coupled device) camera and Leica Qwin software. Images

were processed with Adobe Photoshop 5.5 and assembled using Canvas 5.

RECEIVED 11 MAY 2000; REVISED 17 JULY 2000; ACCEPTED 1 AUGUST 2000;PUBLISHED 17 OCTOBER 2000.

1. Satterwhite, L. L. & Pollard, T. D. Curr. Opin. Cell Biol. 4, 43–52 (1992).

2. Harrington, W. F. & Rodgers, M. E. Annu. Rev. Biochem. 53, 35–73 (1984).

a

Septa

b

DNA

Myo2p

d

F-actin

DNA

c

e

WT myo2∆IQ2

rlc1∆

1 2

3 4

f

1 2 3 4M

750Size (bp)

600

625389320

g

rlc1∆ myo2∆IQ2

30 ̊ C 19 ̊ CWT WTrlc1∆ rlc1∆

Figure 3 Suppression of rlc1∆ by deletion of the Rlc1p-binding site onMyo2p. a, rlc1+ is essential for cell viability at 19 °C. The abilities of wild-type (WT)and rlc1∆ strains to form colonies at 30 °C and at 19 °C were tested. b–d,Rearrangements of cell-wall material (b), F-actin (c) and Myo2p (d) in rlc1∆ cells.The rlc1∆ mutant was grown at 30 °C. The culture was then split two ways; onewas retained at 30 °C and the other was shifted to 19 °C. Cells were fixed andstained to visualize cell walls (b), DNA (c, d), F-actin (c) and Myo2p (d). e, Themyo2::∆IQ2 allele totally suppresses the growth defect associated with rlc1+ dele-tion. Four strains (wild-type, rlc1∆, myo2::∆IQ2 and rlc1∆ myo2::∆IQ2), derivedfrom a single tetrad in a cross between rlc1∆ and myo2::∆IQ2 strains, were testedfor colony formation at 19 °C on YES agar plates. f, Morphology of cells from wild-type (1), rlc1∆ (2), myo2::∆IQ2 (3) and rlc1∆ myo2::∆IQ2 (4) strains at 19 °C.Note that whereas cells of strains 3 and 4 resemble those of strain 1, cells ofstrain 2 are elongated and fail to form visible colonies. g, Ethidium-bromide-stainedagarose gels confirming the genotypes of the strains described above. Upper andmiddle panels, amplification products associated with the 5′ and 3′ region, respec-tively, of the rlc1::ura4+ (rlc1∆) allele. Lower panel, identification of the myo2 andmyo2∆IQ2 loci by PCR and restriction analysis. Arrowhead indicates the 320-base-pair fragment that is characteristic of the functional myo2∆IQ2 locus (seeMethods). The leftmost lane (M) was loaded with molecular-mass standards.



3. Karess, R. E. et al. Cell 65, 1177–1189 (1991).

4. Chen, P., Ostrow, B. D., Tafuri, S. R. & Chisholm, R. L. J. Cell Biol. 127, 1933–1944 (1994).

5. Uyeda, T. Q. & Spudich, J. A. Science 262, 1867–1870 (1993).

6. Kitayama, C., Sugimoto, A. & Yamamoto, M. J. Cell Biol. 137, 1309–1319 (1997).

7. May, K. M., Win, T. Z. & Hyams, J. S. Cell. Motil. Cytoskeleton 38, 385–396 (1997).

8. Bezanilla, M., Forsburg, S. L. & Pollard, T. D. Mol. Biol. Cell 8, 2693–2705 (1997).

9. Motegi, F., Nakano, K., Kitayama, C., Yamamoto, M. & Mabuchi, I. FEBS Lett. 420, 161–166 (1997).

10. Balasubramanian, M. K. et al. Genetics 149, 1265–1275 (1998).

11. Mulvihill, D. P., Win, T. Z., Pack, T. P., & Hyams, J. S. Microsc. Res. Tech. 49, 152–160 (2000).

12. McCollum, D. et al. J. Cell Biol. 130, 651–660 (1995).

13. Naqvi, N., Eng, K., Gould, K. L. & Balasubramanian, M. K. EMBO J. 18, 854–862 (1999).

14. Motegi, F., Nakano, K., & Mabuchi, I. J. Cell Sci. 113, 1813–1825 (2000).

15. Bezanilla, M., Wilson, J. M., & Pollard, T. D. Curr. Biol. 10, 397–400 (2000).

16. De Lozanne, A. & Spudich, J. A. Science 236, 1086–1091 (1987).

17. Knecht, D. A., & Loomis, W. F. Science. 236, 1081–1085 (1987).

18. Neujahr, R., Heizer, C. & Gerisch, G. J. Cell Sci. 110, 123–137 (1997).

19. Rayment, I. et al. Science 261, 50–58 (1993).

20. Smith, D. B. & Johnson, K. S. Gene 67, 31–40 (1988).

21. Keeney, J. B. & Boeke, J. D. Genetics 136, 849–856 (1994).

22. Balasubramanian, M. K., McCollum, D. & Gould, K. L. Methods Enzymol. 283, 494–506 (1997).

ACKNOWLEDGEMENTS

We thank T. Pollard for the myp2-null strain, and P. Silver and K. Sawin for anti-GFP antibodies. We

also thank A. Munn, S. Oliferenko, K. Sampath and all members of the IMA yeast laboratories (espe-

cially S. Naqvi and V. Rajagopalan) for discussion and critical reading of the manuscript. Rlc1p was

independently isolated by V. Simanis and colleagues, whom we thank for sharing unpublished infor-

mation. This work was supported by research funds from the National Science and Technology Board,

Singapore.

Correspondence and requests for materials should be addressed to M.K.B.

NATURE CELL BIOLOGY VOL 2 NOVEMBER 2000 http://cellbio.nature.com858

Documents

document