Skip to main content

Functional genomics of cell morphology using RNA interference: pick your style, broad or deep

Abstract

Several new studies have used RNA interference to screen for protein functions affecting cell shape, mitosis and cytokinesis of Drosophila cells in culture. One broad survey of nearly 1,000 proteins and three studies focused on cytoskeletal and motor proteins have identified key proteins essential for these processes in animal cells.

Given complete genome sequences from a growing number of organisms, investigators are confronted with how, most efficiently, to complete the inventory of proteins that participate in complex cellular processes, such as cytokinesis, cellular motility or the establishment of asymmetric cell shapes. Such an inventory is the essential first step in beginning to think mechanistically about how any such system of interdependent parts functions as a whole. In the pre-genomic era, classical forward genetics (screening of mutants), biochemical isolation with reconstitution, and pharmacology provided laborious but definitive methods to connect genes with molecular functions, working on one protein at a time. Now, strategies can aim for broad or even complete coverage. Experiments with budding yeast led the way with four sorts of experiments: first, a complete set of deletion mutations, which showed that only 19% of the 6,200 genes are required for viability in the laboratory [1]; second, crosses between viable deletion strains, revealing synthetic interactions [2]; third, large-scale two-hybrid assays mapping out networks of protein-protein interactions [3]; and fourth, comprehensive tagging with green fluorescent protein (GFP) to visualize subcellular localization [4]. Such experiments are complicated in higher eukaryotes, where homologous recombination is more difficult to achieve.

Fortunately, RNA interference (RNAi) came to the rescue, providing a simple method for depleting specific mRNAs, and making it possible to assay for the effects of loss of protein function on a grand scale. Pioneering experiments with Caenorhabditis elegans examined the effects of systematically deleting mRNAs for most of the 2,769 genes on chromosome I [5] and 2,300 genes on chromosome III [6]. Both of these studies assayed for gross developmental and functional defects – supplemented in the chromosome III study with high-resolution time-lapse movies of the first two embryonic cell divisions. Both studies yielded phenotypes for about 13% of the genes screened, including most, but not all, of the genes on these chromosomes that were already known to give developmental phenotypes.

The Drosophila community has followed suit, with several new studies using RNAi to deplete mRNAs and proteins from tissue-culture cells [7–10]. The method is simple and efficient. Where antibodies were available for testing, the levels of target proteins were found to be reduced by more than 90%, and sometimes more than 99%. These studies represent a range of styles, from a highly focused examination of the role of 26 microtubule motors in mitosis [7] to a much more ambitious screen of nearly 1,000 mRNAs for possible effects on cellular morphology and division [8]. Where they overlap in the genes studied, these studies generally agree on which depletions produce defects in cytokinesis or in the actin cytoskeleton in the lamellae of spread cells (Table 1).

Table 1 Screens for lamellar and cytokinesis phenotypes in cultured Drosophila cells

When initiating such a screen, a decision must be made regarding breadth or depth, because for practical reasons the assays can be more detailed in a small screen than in a very broad one. For example, the relatively small size of the screen of microtubule motors enabled Goshima and Vale [7] to use time-lapse microscopy of cells expressing GFP-tubulin to identify subtle phenotypes. This kinetic information allowed them to detect, for example, that depletion of dynein caused the cells to delay at the metaphase-to-anaphase transition, before completing an otherwise normal mitosis. The other studies relied on fluorescence microscopy of fixed cells to reveal effects of protein depletion on cellular shape, the actin cytoskeleton, microtubules or the process of cytokinesis. Somma et al. [9] focused on eight proteins suspected to participate in cytokinesis. Kiger et al. [8] hand-picked a selection of 994 Drosophila genes, many known to be functional from mutagenesis studies and others suspected to be functional from work in other systems; their targets included not only cytoskeletal and motor proteins, but also a large number of cell-cycle, signal-transduction and receptor proteins. Rogers et al. [10] screened 90 targets known to be components of the actin cytoskeleton for defects in cytokinesis and/or extension of leading lamellae. In addition, Kiger et al. [8] examined two Drosophila cell lines with distinct morphologies: large, flat S2R+cells and small, round Kc167 cells. Some RNA depletions produced phenotypes in both lines, but others produced different phenotypes in the two lines or a phenotype in only one of the two lines. The other groups [7, 9, 10] tested S2 cell lines; and Rogers et al. [10] plated their S2 cells on concanavalin A, to promote spreading on the microscope slide.

These studies all agree on target RNAs that give cytokinesis phenotypes: those encoding anillin, citron kinase, cofilin, Diaphanous, myosin-II heavy chain, myosin-II regulatory light chain, Pavarotti/kinesin, Pebble Rho GTPase exchange factor (GEF) and Rho1 GTPase. Of the three groups that performed the studies, only Rogers et al. [10] found a cytokinesis phenotype with RNAi targeting profilin and cyclase-associated protein. The Rogers et al. [10] and Kiger et al. [8] screens also agree on the RNA depletions that cause defects in the actin-based lamella and alter cellular shape: Abl tyrosine kinase, Arp2/3 complex, capping protein, Cdc42, cofilin, cofilin phosphatase, cyclase-associated protein, profilin and Scar. The screens did not agree on whether lamellar or cell-shape defects were found with the depletion of the following 15 RNAs: Aip1, formin, anillin, Rho1 GTPase, Aurora kinase, cortactin, Nck, Rac2, Abelson-interacting protein, Enabled kinase, Rho kinase, PAK, Rho GEF (vav), phosphatidylinositol trisphosphate (PIP3) phosphatase (Pten), and myosin essential light chain (unnamed open reading frame CG15780). In four of these cases, one of the labs did not test the RNA, but with depletion of the remaining 11 RNAs one lab found a lamellar defect whereas the other lab did not (Table 1).

Given the functional bias in the selection of the target RNAs for all of these Drosophila RNAi screens, it is striking how many individual proteins can be depleted with little or no effect on tissue-culture cells. The microscopic assays by Kiger et al. [8] revealed phenotypes such as changes in cell shape or the accumulation of multinucleated cells upon depletion of only 160 of the RNA targets (one third of which correspond to a gene lacking a previously characterized mutant allele). In the Rogers et al. screen [10] depletion of 66 of 90 known cytoskeletal targets had no effects on lamellar morphology or cytokinesis; these 66 include a number of overlaps with negative results from the Kiger et al. screen: Ciboulot, filamin, gelsolin, several myosins, tropomodulin, tropomyosin and Wiskott-Aldrich syndrome protein (WASP).

What is the explanation for the RNA depletions with no phenotype in these assays? Are these true negatives or false negatives? A trivial explanation is a failure of depletion, but the efficiency of depletion is impressive in the examples tested. Somma et al. [9], Rogers et al. [10] and Goshima and Vale [7] confirmed that S2 cells express most of the RNAs that they tested, ruling out another trivial explanation for a negative result, namely absence of the target. Kiger et al. [8] included many mRNAs encoding proteins with specialized functions in differentiated tissues such as muscle, nerve or germ cells, so some of their negative results are expected for S2 cells as they lack these functions.

Geneticists might tell us not to worry about the false negatives, since they are inevitable in any screen. But if one wants to reach the goal stated at the outset – a complete inventory of the proteins that participate in a complex cellular process – one cannot ignore the negatives. The highly focused test of microtubule motors in mitosis [7] provides some clues about the negative results. Goshima and Vale [7] targeted all 25 kinesins and the single dynein in the Drosophila genome. Their assay on live cells detected subtle phenotypes, such as the delay in metaphase caused by depletion of the dynein heavy chain, which might have been reported as a false negative in a study with fixed cells. Depletion of any of eight single kinesins compromised formation of a bipolar spindle or the alignment of chromosomes on the metaphase plate, but depletion of multiple kinesins with related functions was required for severe misalignment of metaphase chromosomes. Loss of one kinesin (Pavarotti) compromised cytokinesis owing to disruption of the central spindle, but depletion of two other kinesins had no effect on cytokinesis in spite of the ability of mutations in the respective genes to have an effect. One-off depletion of 17 other kinesins had no detectable effect on cell division, and no single kinesin depletion slowed anaphase chromosome movements. Only 3 of the 26 depletions caused defects so substantial that the cells failed to find an alternative pathway to complete mitosis.

Thus, redundant function, alternative pathways and antagonistic actions are the rule rather than the exception in cell division and probably in many other vital processes. The frequent lack of phenotypes upon depletion of single proteins simply reflects the complexity of redundant molecular machines with multiple safeguards to protect function. Most negatives in the RNAi screens are therefore likely to be real negatives rather than false negatives, but this does not indicate lack of function.

The occurrence of proteins lacking RNAi phenotypes opens the door to combinatorial RNAi screens for synthetic interactions and the construction of protein interaction networks (as in yeast [3]). Double RNAi depletions have already revealed synthetic interactions between some kinesins [7] and between the signaling molecules Nck and Rac [10]. Other modifier screens are feasible; for example, Kiger et al. [8] show RNAi suppressor interactions between the Pten phosphatase and 20 kinases. The targets that do have positive RNAi phenotypes need to be followed up with targeted gene disruption and/or gene replacement, to investigate mechanisms. This has been the Achilles' heel for Drosophila, but fortunately, gene targeting is becoming more efficient [11].

Both focused and broad approaches to RNAi have strengths, so that the way that one screens is largely a matter of taste. As someone interested in the details of molecular mechanisms, I am inclined toward slow, deliberate screens with high-resolution assays in an effort to maximize the information extracted on the first try. Such 'second generation' screens should include verification of the expression of each target in controls and its depletion in test cells. On the other hand, one cannot deny the importance of broad depletion screens designed to give a comprehensive account of genes that contribute to cellular function, such as the shape and morphology assay of Kiger et al. [8]. Fortunately, the groups of Norbert Perrimon at Harvard Medical School and Renato Paro in Heidelberg, Germany have set up the 'Drosophila RNAi Screening Center' (DRSC) [12], with a library of RNAi reagents for all of the known open reading frames in the Drosophila genome and facilities for high-throughout screens. Support from the National Institute of General Medical Sciences makes this resource available to other investigators, and the DRSC is eager to collaborate with experts with sophisticated assays for various cellular activities.

Having struggled to compare the recent RNAi studies in Drosophila [7–10], I feel compelled to finish with an observation about gene names. The clever and amusing names that geneticists have assigned to mutations and/or genes are genuinely hampering communication. Even for someone familiar with a particular field, matching dozens of meaningless gene names with generic protein names has become impossible. FlyBase [13] is an indispensable resource that allowed me to sort out most of the nomenclature, but a trip to FlyBase or PubMed [14] should not be required to match pavarotti with a kinesin, sanpodo with tropomodulin or zipper with cytoplasmic myosin-II heavy chain. The three letter and number nomenclature used by the yeast community to name yeast genes is just as opaque. Although it is strenuously resisted ("our history will be lost"), it is past time for gene names to be converted as soon as practical to common usage in all of the genetic model organisms. Nomenclature also needs to be consolidated across the phylogenetic tree, as has been done for myosins [15] but not yet for kinesins, where some family members have five or more names. The insights promised by new technologies such as RNAi may be wasted if jargon impedes communication with many interested readers.

References

  1. Giaever G, Chu AM, Ni L, Connelly C, Riles L, Veronneau S, Dow S, Lucau-Danila A, Anderson K, Andre B, et al: Functional profiling of the Saccharomyces cerevisiae genome. Nature. 2002, 418: 387-391. 10.1038/nature00935.

    Article  CAS  PubMed  Google Scholar 

  2. Tong AH, Evangelista M, Parsons AB, Xu H, Bader GD, Page N, Robinson M, Raghibizadeh S, Hogue CW, Bussey H, et al: Systematic genetic analysis with ordered arrays of yeast deletion mutants. Science. 2001, 294: 2364-2368. 10.1126/science.1065810.

    Article  CAS  PubMed  Google Scholar 

  3. Tong AH, Drees B, Nardelli G, Bader GD, Brannetti B, Castagnoli L, Evangelista M, Ferracuti S, Nelson B, Paoluzi S, et al: A combined experimental and computational strategy to define protein interaction networks for peptide recognition modules. Science. 2002, 295: 321-324. 10.1126/science.1064987.

    Article  CAS  PubMed  Google Scholar 

  4. Kumar A, Agarwal S, Heyman JA, Matson S, Heidtman M, Piccirillo S, Umansky L, Drawid A, Jansen R, Liu Y, et al: Subcellular localization of the yeast proteome. Genes Dev. 2002, 16: 707-719. 10.1101/gad.970902.

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  5. Fraser AG, Kamath RS, Zipperlen P, Martinez-Campos M, Sohrmann M, Ahringer J: Functional genomic analysis of C. elegans chromosome I by systematic RNA interference. Nature. 2000, 408: 325-330. 10.1038/35042517.

    Article  CAS  PubMed  Google Scholar 

  6. Gonczy P, Echeverri C, Oegema K, Coulson A, Jones SJ, Copley RR, Duperon J, Oegema J, Brehm M, Cassin E, et al: Functional genomic analysis of cell division in C. elegans using RNAi of genes on chromosome III. Nature. 2000, 408: 331-336. 10.1038/35042526.

    Article  CAS  PubMed  Google Scholar 

  7. Goshima G, Vale RD: The roles of microtubule-based motor proteins in mitosis: comprehensive RNAi analysis in the Drosophila S2 cell line. J Cell Biol. 2003, 162: 1003-1016. 10.1083/jcb.200303022.

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  8. Kiger AA, Baum B, Jones S, Jones MR, Coulson A, Echeverri C, Perrimon N: A functional genomic analysis of cell morphology using RNA-interference. J Biol. 2003, 2: 27-10.1186/1475-4924-2-27.

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  9. Somma MP, Fasulo B, Cenci G, Cundari E, Gatti M: Molecular dissection of cytokinesis by RNA interference in Drosophila cultured cells. Mol Biol Cell. 2002, 13: 2448-2460. 10.1091/mbc.01-12-0589.

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  10. Rogers SL, Wiedemann U, Stuurman N, Vale RD: Molecular requirements for actin-based lamella formation in Drosophila S2 cells. J Cell Biol. 2003, 162: 1079-1088. 10.1083/jcb.200303023.

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  11. Gong WJ, Gorlic KG: Ends-out, or replacement, gene targeting in Drosophila. Proc Nat Acad Sci USA. 2003, 100: 2556-2561. 10.1073/pnas.0535280100.

    Article  PubMed Central  CAS  PubMed  Google Scholar 

  12. Drosophila RNAi Screening Center. [http://www.flyrnai.org]

  13. FlyBase. [http://flybase.bio.indiana.edu]

  14. PubMed. [http://www.ncbi.nih.gov/entrez/query.fcgi]

  15. Berg JS, Powell BC, Cheney RE: A millennial myosin census. Mol Biol Cell. 2001, 12: 780-794.

    Article  PubMed Central  CAS  PubMed  Google Scholar 

Download references

Acknowledgements

I thank Amy Kiger for suggestions on the text and help with gene names.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Thomas D Pollard.

Rights and permissions

Reprints and permissions

About this article

Cite this article

Pollard, T.D. Functional genomics of cell morphology using RNA interference: pick your style, broad or deep. J Biol 2, 25 (2003). https://doi.org/10.1186/1475-4924-2-25

Download citation

  • Published:

  • DOI: https://doi.org/10.1186/1475-4924-2-25

Keywords