NATURE|VOL 408|16 NOVEMBER 2000|www.nature.com 331articlesFunctional genomic analysis of celldivision in C. elegans using RNAi ofgenes on chromosome IIIPierre GoÈnczy*²³, Christophe Echeverri*²³, Karen Oegema*², Alan Coulson§, Steven J. M. Jonesk, Richard R. Copley², John Duperon*,Jeff Oegema*, Michael Brehm*³, Etienne Cassin*, Eva Hannak*, Matthew Kirkham*, Silke Pichler*², Kathrin Flohrs*, Anoesjka Goessen*,Sebastian Leidel*, Anne-Marie Alleaume*³,CeÂcilie Martin*³, Nurhan OÈzluÈ*, Peer Bork²& Anthony A. Hyman*²* Max-Planck-Institute for Cell Biology and Genetics (MPI-CBG), Pfotenhauerstrasse 108, D-01307 Dresden, Germany²European Molecular Biology Laboratory (EMBL), Meyerhofstrasse 1, D-69117 Heidelberg, Germany§ The Sanger Centre, Wellcome Trust Genome Campus, Hinxton, Cambridge, CB10 1SA, UKk Genome Sequence Centre, British Columbia Cancer Research Centre, 600 West 10th Avenue, Vancouver, British Columbia, V5Z-4E6, Canada³Present addresses: Swiss Institute for Experimental Cancer Research (ISREC), CH-1066 Epalinges/Lausanne, Switzerland (P.G); Cenix BioScience GmbH,Pfotenhauerstrasse 108, D-01307 Dresden, Germany (C.E., M.B., A.-M.A., C.M.).............................................................................................................................................................................................................................................................................Genome sequencing projects generate a wealth of information; however, the ultimate goal of such projects is to accelerate theidenti®cation of the biological function of genes. This creates a need for comprehensive studies to ®ll the gap between sequenceand function. Here we report the results of a functional genomic screen to identify genes required for cell division in Caenorhabditiselegans. We inhibited the expression of ,96% of the ,2,300 predicted open reading frames on chromosome III using RNA-mediated interference (RNAi). By using an in vivo time-lapse differential interference contrast microscopy assay, we identi®ed 133genes (,6%) necessary for distinct cellular processes in early embryos. Our results indicate that these genes represent most of thegenes on chromosome III that are required for proper cell division in C. elegans embryos. The complete data set, including sampletime-lapse recordings, has been deposited in an open access database. We found that ,47% of the genes associated with adifferential interference contrast phenotype have clear orthologues in other eukaryotes, indicating that this screen providesputative gene functions for other species as well.Genome sequencing projects identify large numbers of genes in thesearch for biological function. Function can be tested directly in theorganism of interest by using an appropriate assay or inferred onthe basis of knowledge gained from the study of related genes inother organisms. However, a signi®cant fraction of genes uncoveredby sequencing projects are `new' and cannot be rapidly assignedfunctions by either method. Thus, there is a need for large-scalestudies to test directly the function of new genes, and validate genefunctions inferred from studies of homologous proteins in otherorganisms. Large-scale functional genomic studies have alreadybeen successful in S. cerevisiae (ref. 1; http://www.mips.biochem.mpg.de/proj/eurofan/index.html; http:://genome-www.stanford.edu/cgi-bin/sgd/search), but comparable analysis is still lacking inmetazoans. Therefore, one of the challenges facing biologists todayis to design high-throughput assays to assign cellular functions togenes emerging from metazoan sequencing projects.Designing a large-scale screen for cell division genesWe sought to use a functional genomic approach in the earlyC. elegans embryo to identify a large set of genes necessary forcell-division processes, for the following reasons. First, the genomeis sequenced, offering the possibility to conduct a comprehensiveanalysis2. Second, cell-division processes can be examined with highspatial and temporal resolution using time-lapse differential inter-ference contrast (DIC) microscopy, allowing detection of evensubtle deviations from the wild-type sequence of events3. Third,using RNA-mediated interference (RNAi), expression of a givengene in the embryo can be abolished in a sequence-speci®c mannerby subjecting parental germ cells to corresponding double-strandedRNA (dsRNA)4. Notably, RNAi can be also applied to genes essentialfor cell division, as no mitoses occur between the time germ cells aresubjected to dsRNA and fertilization. In contrast, although thetargeted disruption of the ,6,200 open reading frames (ORFs) ofTable 1 Positive controlsPhenotypeUndiluted 13/13 (100%)1:1 dilution 12/13 (,92%)1:3 dilution 6/13 (,46%)1:7 dilution 4/13 (,31%).............................................................................................................................................................................Double-stranded RNA corresponding to 13 genes whose phenotype in the early embryo is knownfrom mutational analysis were generated and injected, either undiluted or diluted (1:1, 1:3, 1:7) witha mixture of 3±7 different dsRNA species corresponding to genes not associated with phenotypes.The resulting embryos were analysed by time-lapse DIC microscopy; phenotypes were scored asdescribed in Methods. The fraction of dsRNAs giving rise to the expected phenotype is given foreach dilution.The following 13 genes were tested cyk-1 (ref. 14), let-99 (ref. 35), mei-1 (ref. 36), nmy-2 (ref. 37),ooc-3 (ref. 38), par-1 (ref. 39), par-2 (ref. 18), par-3 (ref. 17), par-5 (K. Kemphues, personalcommunication), par-6 (ref. 40), zen-4 (refs 41, 42), zyg-9 (ref. 43), zyg-11 (ref. 44).Table 2 Overall outcome of screenDistinct genes tested 2,174Total with phenotype 281 (,12.9%)DIC phenotypes 133 (,6.1%)DIC + EL 125*DIC + L/A 5DIC only 3²Progeny phenotypes (only) 139 (,6.4%)EL only 78L/A only 61L1/L2 33L3/L4 15³Adult 13F0 sterile 9 (,0.4%).............................................................................................................................................................................EL, embryonic lethal; L/A, larval/adult phenotype; L1/L2, early larval defect; L3/L4, late larval defect;Adult, adult defect. See Methods, Table 3 and database (http://mpi-web.embl-heidelberg.de/dbScreen/) for more information.* In some cases, embryonic lethality was partial only; progeny