Oncogene (2004) 23, 8353–8358

& 2004 Nature Publishing Group All rights reserved 0950-9232/04 $30.00 www.nature.com/onc

High-throughput gene silencing using cell arrays Dominique Vanhecke1 and Michal Janitz*,1 1

Max Planck Institute for Molecular Genetics, Department Vertebrate Genomics, Fabeckstr. 60-62, Berlin 14195, Germany

A recently established transfected cell array (TCA) technology has opened new experimental dimensions in the field of functional genomics. Cell arrays allow for transfection of several thousands different DNA molecules in microarray format. The effects of overexpression of hundreds of proteins on cellular physiology can be observed in a single experiment. The TCA technique has also found its application in RNA interference (RNAi) research. Small interfering RNAs (siRNA) as well as plasmid expressing short hairpin RNAs can be transferred into the cells through the process of reverse transfection. The silencing of numerous genes in spatially separated manner can be thus monitored. This review will provide an overview on current concepts concerning combination of cell array and RNAi for high-throughput loss-of-function studies. Oncogene (2004) 23, 8353–8358. doi:10.1038/sj.onc.1208027 Keywords: RNA interference; microarray; reverse transfection; cell array

(TCA) (Ziauddin and Sabatini, 2001; Chang et al., 2004; Redmond et al., 2004) paved a way for the highthroughput gene analysis in the field of functional genomics. The principle of the TCA technique is based on the transfection of DNA or RNA molecules immobilized on a solid surface into mammalian cells with subsequent detection of the physiological effects caused by the introduction of the foreign nucleic acid on these cells. Full-length open-reading frames of genes inserted in expression vectors or RNA/DNA oligonucleotides are printed at a high density on a glass slide along with a lipid transfection reagent using a robotic arrayer (Figure 1). Densities of up to 8000 cell clusters per standard slide can be achieved. When the microarray of DNA constructs is covered with a monolayer of adherent cells only the cells growing on top of the DNA spots become transfected, resulting in the expression or silencing of specific proteins in spatially distinctive groups of cells. The phenotypic effects of this reverse transfection of hundreds or thousands of genes can be detected using specific cell-based bioassays, with readouts based on autofluorescence or immunofluorescence.

Introduction

Cell array as a functional genomics tool

Decoding of the human and mouse genomes provided a huge amount of sequence information giving the scientific community an opportunity to analyse genetic information on a global scale. The combination of gene expression profiling studies using genome-wide DNA microarrays and the annotation of thousands of ESTs to concrete genes resulted in unprecedental acceleration of expression data acquisition and interpretation. Yet, in vivo functional analysis of thousands of novel genes still presents a challenge toward our understanding of the cellular processes on a global scale. The functional analysis towards the putative causal role of newly discovered candidate genes and proteins in cellular processes and its potential clinical impact or suitability as a drug target revealed to be slow and rate limiting. Functional validation is usually accomplished in molecular- and cell-based assays on a gene-by-gene basis, creating a bottle-neck effect for the characterization of the huge numbers of targets arising from genomics and proteomics surveys. A recently described, cell-based microarray system, called the transfected cell array

The TCA (shortly cell array) combines microarray technology with cellular biology methods such as transfection and cell culture. A process of assay preparation and performance can be divided into three distinct steps as follows: (i) microarray preparation, (ii) reverse transfection and (iii) detection assay (Figure 1a). In contrast to microarrays that are used for gene expression profiling studies the DNA that has to be transfected in the cell arrays cannot be immobilized permanently on the glass surface. The reason is that in the following steps of the procedure DNA should enter cells growing on top of the DNA spot in the process of reverse transfection (Figure 1b). Therefore, the DNA to be arrayed is first dissolved in a gelatin solution, which facilitates transfer of the DNA from a solid support into the cells. The generation of arrays from DNA/gelatine solutions is performed using either stealth or solid pins and any standard robotic arrayer or alternatively using a noncontact piezoelectric dispensing printer. The diameter of the single DNA spot can range from 100 to 300 mm, whereas spots are separated from each other by 400–1000 mm, depending on the detection assay used. Design of the cell array, in terms of spot arrangement

*Correspondence: M Janitz; E-mail: [email protected]

RNAi using cell arrays D Vanhecke and M Janitz

8354

Figure 1 TCA technology combines the microarray technique with living cells (a). Reverse transfection allows for multiple, spatially separated transfections of nucleic acids in a single experiment (b)

throughout the slide, does not differ from the classical DNA array and is determined by the number of pins attached to the printing head of the roboter and individual settings of the robotic arrayer software (Murphy, 2002). Before covering the array with a monolayer of cells, the spots have to be incubated with transfection reagent so that the DNA becomes complexed with liposome-like structures and consequently can be taken up by the overlying cells. The time of incubation with the transfection reagent is crucial for the successful transfection and has to be optimized for each cell line tested. Subsequently, the prepared array is placed in a culture dish and covered with cells. The cells are let to grow on the slide surface for up to 3 days. Longer incubation times lead to a decline of the signals detected following transfection because the viability of cells cultured at such a high density decreases gradually in time. Therefore, the TCA is limited to those cellular events, which are induced up to 72 h after reverse transfection. This time frame is sufficient for most gene expression and silencing experiments. Depending on the nature of the molecular event to be monitored, signal detection can be performed either on living (e.g. calcium influx detection) or fixed cells (e.g. immunohistochemical protein detection). A common Oncogene

detection method uses signals emitted by fluorescent dyes (e.g. Cy3 or Cy5) coupled to primary or secondary monoclonal antibodies. Alternatively, autofluorescent reporter proteins such as green fluorescence protein (GFP) can be readily detected. Preference in fluorescence detection is defined by the possibility to collect signal intensities from the whole slide using standard microarray scanners. Alternatively, fluorescence microscopy allows for the detection of cell clusters by collecting the signal intensities from single cells and application of statistical analysis of signal to background ratios. Analysis of the fluorescence signals and statistical analysis of the data acquired by microscope or scanner can be performed using software commonly used for DNA arrays. Alternatively, in the case of studies where apoptosis or cell adhesion is of interest abnormal morphology of the transfected cells, as observed using bright field or phase-contrast microscopy, can be used as a functional read out. For studies involving apoptosis induction, particular cell clusters on the array can be further examined for DNA fragmentation (TUNEL assay) or for expression of proteins serving as early markers of apoptosis such as Annexin V binding or cleavage of poly(ADP-ribose) polymerase (PARP). Recent advances in high-throughput microscope technology provide a number of automated imaging systems, which make the genome-wide screening for gain-of-function or loss-of-function phenotypes feasible (Price et al., 2002; Wolf, 2003; Zemanov et al., 2003). TCAs offer a number of advantages when used to study gene function. One of the most important features is that this assay provides a unique opportunity to study protein function in the context of the living cell. For obvious reasons studying mammalian cells using TCA gives several advantages over analysing gene expression in yeast and Drosophila melanogaster. Not the least important is that proteins expressed in the cell array can undergo post-translational processing such as glycosylation and folding, which can be very distinct in different organisms and different cell types (Colosimo et al., 2000). Second, functional analysis of proteins in their natural environment guarantees that additional molecules, involved in the functional process, are present during the assay (e.g. cofactors, proteins involved in large complex formation, etc.). There are a number of advantages in using TCA as compared to functional assays performed on mammalian cells in microwell plate format. The array approach requires less cells per number of genes tested. This is especially relevant for human primary cells (cells recently isolated from organs or tissues), since only small numbers of these can be isolated out of a tissue and the in vitro expansion of these cells is rather limited. Although primary cells are notorious for their capacity to be transfected, reverse transfection of these cells is possible, but requires careful optimization of the experimental conditions for each cell type tested (Yoshikawa et al., 2004; Vanhecke et al., 2004). TCAs also require far less DNA/RNA as well as transfection

RNAi using cell arrays D Vanhecke and M Janitz

8355

and signal development reagents as compared to assays performed in microwell plate format. Since in the case of high-throughput technologies the cost of the single sample analysis must be reduced to absolute minimum, cell arrays are at the moment the most cost-effective functional genomic tool available.

RNA interference (RNAi) In the case of loss-of-function studies, the cell array has been combined with the technique of RNAi (Kumar et al., 2003; Mousses et al., 2003; Silva et al., 2004). RNAi has been first described in nematodes (Fire et al., 1998) and quickly became the fastest expanding field in molecular biology (Hannon, 2002). RNAi is a posttranscriptional gene silencing process that is based on sequence-specific interactions of small interfering RNAs (siRNA) with targeted mRNA molecules. Doublestranded RNA (dsRNA) has been observed in most eucaryotic cells (Cerutti, 2003; Denli and Hannon, 2003; Dykxhoorn et al., 2003). For gene silencing to occur dsRNA has to be processed by Dicer into short siRNA duplexes of 21–23 nucleotides in length (Bernstein et al., 2001). Subsequently, one strand of the siRNA duplex interacts with the RNA-induced silencing complex, which triggers a number of enzymatic reactions resulting in the degradation of the targeted mRNA molecules (Liu et al., 2003; Schwarz et al., 2003; Martinez and Tuschl, 2004). Although discovered as a natural process RNAi quickly proved to be excellent for experimental gene silencing not only in worms but also in mammalian cells. Application of short interfering RNAs (Elbashir et al., 2001a) paved the way for in vitro and in vivo loss-of-function studies in mammalian cells. Utilization of siRNA helped to overcome triggering the interferon pathway that would otherwise result in dsRNA-activated cell death (Stark et al., 1998; Williams, 1999). Genes of interest can efficiently be turned off in mammalian cell cultures both through transfer of synthetic siRNA into the cells (Elbashir et al., 2001b; Martinez et al., 2002) or transfection of plasmids expressing short hairpin RNA (shRNA) (Brummelkamp et al., 2002; Lee et al., 2002; Paddison et al., 2002). RNAi has been employed in genome-wide phenotype screens in Caenorhabditis elegans (Ashrafi et al., 2003; Kamath et al., 2003; Lee et al., 2003; Pothof et al., 2003) and D. melanogaster (Somma et al., 2002; Kiger et al., 2003; Lum et al., 2003). Although siRNAs synthesized chemically (McManus et al., 2002) or enzymatically (Yang et al., 2002; Shirane et al., 2004) have mainly been used to study function of individual mammalian genes, RNAi-based genome-wide functional analysis for mammalian cells is also well underway. Indeed, several laboratories have already created and successfully applied siRNA-based libraries targeting large groups of human genes (Aza-Blanc et al., 2003; Brummelkamp et al., 2003; Berns et al., 2004; Zheng et al., 2004).

Cell arrays in RNAi research Cell arrays soon revealed to be a suitable platform for high-throughput cellular assays using RNAi. Several groups, including author’s laboratory, focused on adaptation of the TCA platform for the purpose of large-scale loss-of-function studies (Kumar et al., 2003; Mousses et al., 2003; Silva et al., 2004). The basic protocol for preparation of RNAi cell arrays includes spotting of siRNAs in gelatin solution onto a modified glass surface using a standard robotic arrayer. The RNA array is subsequently treated with transfection reagent and covered with a monolayer of recipient cells. Transfection efficiency can be monitored by using fluorochrome (for example, rhodamine) conjugated siRNA (Mousses et al., 2003). Alternatively, the siRNA are cotransfected with a reporter plasmid, and gene silencing can be monitored in cells expressing the reporter gene (Silva et al., 2004). As it was observed for normal transfection of siRNA (McManus and Sharp, 2002), transfection efficiency of siRNAs on cell arrays is significantly higher then plasmid DNA (author’s own observations). It has been described, however, that some cells remain refractory to RNAi (Tavernarakis et al., 2000). The availability of vector based RNAi systems (Brummelkamp et al., 2002; Miyagishi and Taira, 2002; Paul et al., 2002; Paddison et al., 2004) offers a possibility to apply cell array using DNA-based reverse transfection protocols. The potential drawback of this approach is substantial delay in exertion of the silencing effect due to the de novo siRNA synthesis within the cell. Nevertheless, Silva and co-workers have recently shown effective silencing of the reporter protein using cotransfection of shRNA-expressing plasmid (Silva et al., 2004). Silencing of targeted genes is monitored using, for example, fluorescent-labeled monoclonal antibodies or autofluorescent reporter proteins. Targeted genes are either exogenous genes, introduced by transient transfection or stable integration of gene expression plasmids, or endogenous genes. The former approach allows for the evaluation of silencing capacity of different siRNAs specific for the same target gene and can be used as a robust validation platform for functional siRNA selection (Kumar et al., 2003; Silva et al., 2004). Functional validation of potential siRNA molecules is still required despite a number of algorithms that exist for sequence design (Paddison and Hannon, 2003). Indeed, for most algorithms it is estimated that only one out of two or even one out of five designed siRNA will efficiently knockdown a target gene. A more challenging application is the targeting of endogenous genes by reverse transfection of siRNA molecules. Endogenous gene silencing on cell arrays can be monitored by means of specific fluorescent-labeled monoclonal antibodies. This results in clusters of cells that are negative for the protein, thus creating arrays of black holes in monolayers of cells that are expressing the protein (Mousses et al., 2003; Vanhecke et al., 2004). However, detection with gene-specific antibodies puts a limit to the number of genes that can be simultaneously Oncogene

RNAi using cell arrays D Vanhecke and M Janitz

8356

assayed in the cell array. For the functional evaluation of a genome-wide collection of siRNA molecules using cell arrays, other detection methods have to be used. For example, downstream cellular processes that are affected by silencing endogenous genes can be monitored by specific antibodies that can detect changes in the phosphorylation state of cell membrane-bound receptors or transcription factors. This would allow for the identification of essential or even novel cell signaling molecules using genome-wide siRNA libraries (Figure 2a). This is of particular importance in view of the increasing interest in protein tyrosine phosphatases (van Huijsduijnen et al., 2002) and kinases (Gschwind et al., 2004) as drug targets for several human diseases such as diabetes, cancer and autoimmunity. Alternatively, the effects of silencing genes can result in cellular

processes such as apoptosis or changes in intercellular adhesion, which can be monitored by specific reagents (e.g. Annexin V, cleaved PARP, TUNNEL assay) or by simply analysing morphological changes of the transfected cells (Ziauddin and Sabatini, 2001). For the identification of genes that confer resistance to apoptosis in tumor cells, reverse transfection of siRNAs can be combined by low doses of prodrug (e.g. doxorubicin, TNF-a) treatments (Figure 2b). In a number of cases, gene-specific siRNAs were shown to sensitize efficiently cancer cells towards proapoptotic stimuli (CrnkovicMertens et al., 2003; Trougakos et al., 2004). Testing siRNA libraries on drug pretreated cancer cells would increase the likelihood of finding genes that play a role in apoptosis, thus bearing a great potential to serve as a basis for the development of novel therapeutic strate-

Figure 2 Genome-wide application of the RNAi in combination with the cell array technology could accelerate elucidation of cell signaling pathways (a) or mechanisms of programmed cell death (b) Oncogene

RNAi using cell arrays D Vanhecke and M Janitz

8357

gies. As an alternative application, genome-wide siRNA libraries could be analysed in cell arrays to identify molecules that affect promoter activity. Predicted promoter regions (Dieterich et al., 2003) can be cloned upstream of reporter genes (e.g. EGFP) and stable transfected cell lines expressing the promoter–reporter construct can be subjected to genome-wide siRNA arrays. Changes in promoter activity can be monitored by quantitative fluorescence microscopy and could lead to the identification of upstream regulatory factors involved in gene-specific expression activity.

Concluding remarks and perspectives Since TCAs allow the analysis of many genes in parallel and, furthermore, can be adopted to many existing cellbased assays, they could become an efficient molecular profiling tool for disease-related proteins or to study protein–protein and protein–drug interactions. Interesting perspectives arise when double knockdowns are considered. Two different siRNA or vectors expressing shRNAs could be cotransfected together or single siRNA could be transfected to the cell line stably transfected with siRNA vector (van de Wetering et al., 2003). Cell arrays would have enough spotting capacity to test hundreds of different siRNA combinations in a single experiment. This would open a possibility for screening synthetic lethal interactions, which occur by the appearance of two nonlethal mutations that act together resulting in cell death. Such a functional interaction of two corresponding gene products greatly

helps in the identification of novel drug targets, for example, against cancer. Most of the synthetic lethal interactions have been determined in lower eucaryotes such as C. elegans (Ooi et al., 2003; Wang et al., 2004). In contrast, in mammalian systems synthetic lethality has been detected only in few studies, mainly because efficient molecular tools for the identification of such reactions has been lacking so far. Application of highthroughput RNAi-based cell arrays could substantially accelerate progress in this field. The combination of cell arrays and cell lines stably transfected with inducible shRNA plasmid constructs is also well suited for screening of small molecules for their ability to interact specifically with protein targets such as G-protein-coupled receptors (Mishina et al., 2004). Also in this case, the cell microarrays present a more advanced alternative to standard protein arrays, since in TCA the proteins are synthesized in situ in a physiological environment of the cells, thus enabling proper post-translational protein folding and glycosylation. Taken together, cell arrays represent an emerging high-throughput platform in RNAi research, especially in applications where cell numbers are a limiting factor. Further optimization of the reverse transfection protocol for different cell types should open the possibility to perform loss-of-function studies in a tissue- or organspecific environment. Acknowledgements Dominique Vanhecke is supported by a postdoctoral fellowship of the Foundation for Scientific Research-Flanders (Belgium).

References Ashrafi K, Chang FY, Watts JL, Fraser AG, Kamath RS, Ahringer J and Ruvkun G. (2003). Nature, 421, 268–272. Aza-Blanc P, Cooper C, Wagner K, Batalov S, Deveraux QL and Cooke MP. (2003). Mol. Cell, 12, 627–637. Berns K, Hijmans EM, Mullenders J, Brummelkamp TR, Velds A, Heimerikx M, Kerkhoven RM, Madiredjo M, Nijkamp W, Weigelt B, Agami R, Ge W, Cavet G, Linsley PS, Beijersbergen RL and Bernards R. (2004). Nature, 428, 431–437. Bernstein E, Caudy AA, Hammond SM and Hannon GJ. (2001). Nature, 409, 363–366. Brummelkamp TR, Bernards R and Agami R. (2002). Science, 296, 550–553. Brummelkamp TR, Nijman SMB, Dirac AMG and Bernards R. (2003). Nature, 424, 797–801. Cerutti H. (2003). Trends Genet., 19, 39–46. Chang FH, Lee CH, Chen MT, Kuo CC, Chiang YL, Hang CY and Roffler S. (2004). Nucleic Acids Res., 32, e33. Colosimo A, Goncz KK, Holmes AR, Kunzelmann K, Novelli G, Malone RW, Bennett MJ and Gruenert DC. (2000). Biotechniques, 29, 314–318. Crnkovic-Mertens I, Hoppe-Seyler F and Butz K. (2003). Oncogene, 22, 8330–8336. Denli AM and Hannon GJ. (2003). Trends Biochem. Sci., 28, 196–201. Dieterich C, Wang H, Rateitschak K, Luz H and Vingron M. (2003). Nucleic Acids Res., 31, 55–57.

Dykxhoorn DM, Novina CD and Sharp PA. (2003). Nat. Rev. Mol. Cell. Biol., 4, 457–467. Elbashir SM, Harborth J, Lendeckel W, Yalcin A, Weber K and Tuschl T. (2001a). Nature, 411, 494–498. Elbashir SM, Martinez J, Patkaniowska A, Lendeckel W and Tuschl T. (2001b). EMBO J., 20, 6877–6888. Fire A, Xu S, Montgomery MK, Kostas SA, Driver SE and Mello CC. (1998). Nature, 391, 806–811. Gschwind A, Fischer OM and Ullrich A. (2004). Nat. Rev. Cancer, 5, 361–370. Hannon GJ. (2002). Nature, 418, 244–251. Kamath RS, Fraser AG, Dong Y, Poulin G, Durbin R, Gotta M, Kanapin A, Le Bot N, Moreno S, Sohrmann M, Welchman DP, Zipperlen P and Ahringer J. (2003). Nature, 421, 231–237. Kiger A, Baum B, Jones S, Jones M, Coulson A, Echeverri C and Perrimon N. (2003). J. Biol., 2, 27. Kumar R, Conklin DS and Mittal V. (2003). Genome Res., 13, 2333–2340. Lee NS, Dohjima T, Bauer G, Li H, Li MJ, Ehsani A, Salvaterra P and Rossi J. (2002). Nat. Biotechnol., 20, 500–505. Lee SS, Lee RY, Fraser AG, Kamath RS, Ahringer J and Ruvkun G. (2003). Nat. Genet., 33, 40–48. Liu Q, Rand TA, Kalidas S, Du F, Kim HE, Smith DP and Wang X. (2003). Science, 301, 1921–1925.

Oncogene

RNAi using cell arrays D Vanhecke and M Janitz

8358 Lum L, Yao S, Mozer B, Rovescalli A, Von Kessler D, Nirenberg M and Beachy PA. (2003). Science, 299, 2039–2045. Martinez J and Tuschl T. (2004). Genes Dev., 18, 975–980. Martinez LA, Naguibneva I, Lehrmann H, Vervisch A, Tchenio T, Lozano G and Harel-Bellan A. (2002). Proc. Natl. Acad. Sci. USA, 99, 14849–14854. McManus MT, Petersen CP, Haines BB, Chen J and Sharp PA. (2002). RNA, 8, 842–850. McManus MT and Sharp PA. (2002). Nat. Rev. Genet., 3, 737–747. Mishina YM, Wilson CJ, Bruett L, Smith JJ, Stoop-Myer C, Jong S, Amaral LP, Pedersen R, Lyman SK, Myer VE, Kreider BL and Thompson CM. (2004). J. Biomol. Screen, 9, 196–207. Miyagishi M and Taira K. (2002). Nucleic Acids Res. Suppl., 2, 113–114. Mousses S, Caplen NJ, Cornelison R, Weaver D, Basik M, Hautaniemi S, Elkahloun AG, Lotufo RA, Choudary A, Dougherty ER, Suh E and Kallioniemi O. (2003). Genome Res., 13, 2341–2347. Murphy D. (2002). Adv. Physiol. Educ., 26, 256–270. Ooi SL, Shoemaker DD and Boeke JD. (2003). Nat. Genet., 35, 277–286. Paddison PJ, Caudy AA, Bernstein E, Hannon GJ and Conklin DS. (2002). Genes Dev., 16, 948–958. Paddison PJ and Hannon GJ. (2003). Curr. Opin. Mol. Ther., 5, 217–224. Paddison PJ, Silva JM, Conklin DS, Schlabach M, Li M, Aruleba S, Balija V, O’Shaughnessy A, Gnoj L, Scobie K, Chang K, Westbrook T, Cleary M, Sachidanandam R, McCombie WR, Elledge SJ and Hannon GJ. (2004). Nature, 428, 427–431. Paul CP, Good PD, Winer I and Engelke DR. (2002). Nat. Biotechnol., 20, 505–508. Pothof J, van Haaften G, Thijssen K, Kamath RS, Fraser AG, Ahringer J, Plasterk RH and Tijsterman M. (2003). Genes Dev., 17, 443–448. Price JH, Goodacre A, Hahn K, Hodgson L, Hunter EA, Krajewski S, Murphy RF, Rabinovich A, Reed JC and Heynen S. (2002). J. Cell. Biochem. Suppl., 39, 194–210.

Oncogene

Redmond TM, Ren X, Kubish G, Atkins S, Low S and Uhler MD. (2004). Mol. Cell. Proteomics (Epub ahead of print). Schwarz DS, Hutvagner G, Du T, Xu Z, Aronin N and Zamore PD. (2003). Cell, 115, 199–208. Shirane D, Sugao K, Namiki S, Tanabe M, Iino M and Hirose K. (2004). Nat. Genet., 36, 190–196. Silva JM, Mizuno H, Brady A, Lucito R and Hannon GJ. (2004). Proc. Natl. Acad. Sci. USA, 101, 6548–6552. Somma MP, Fasulo B, Cenci G, Cundari E and Gatti M. (2002). Mol. Cell. Biol., 13, 2448–2460. Stark GR, Kerr IM, Williams BR, Silverman RH and Schreiber RD. (1998). Annu. Rev. Biochem., 67, 227–264. Tavernarakis N, Wang SL, Dorovkov M, Ryazanov A and Driscoll M. (2000). Nat. Genet., 24, 180–183. Trougakos IP, So A, Jansen B, Gleave ME and Gonos ES. (2004). Cancer Res., 64, 1834–1842. van de Wetering M, Oving I, Muncan V, Pon Fong MT, Brantjes H, van Leenen D, Holstege FC, Brummelkamp TR, Agami R and Clevers H. (2003). EMBO Rep., 4, 6 09–615. van Huijsduijnen RH, Bombrun A and Swinnen D. (2002). Drug Discov. Today, 19, 1013–1019. Vanhecke D and Janitz M. (2004). Encyclopedic Reference of Genomics and Proteomics in Molecular Medicine. Ganten D and Ruckpaul K (eds). Springer-Verlag: Heidelberg, (in press). Wang JC, Walker A, Blackwell TK and Yamamoto KR. (2004). J. Biol. Chem., 279, 29270–29277. Williams BR. (1999). Oncogene, 18, 6112–6120. Wolf DE. (2003). Methods Cell. Biol., 73, 319–336. Yang D, Buchholz F, Huang Z, Goga A, Chen CY, Brodsky FM and Bishop JM. (2002). Proc. Natl. Acad. Sci. USA, 99, 9942–9947. Yoshikawa T, Uchimura E, Kishi M, Funeriu DP, Miyake M and Miyake J. (2004). J. Control Rel., 96, 227–232. Zemanov L, Schenk A, Valler MJ, Nienhaus GU and Heilker R. (2003). Drug Discov. Today, 8, 1085–1093. Zheng L, Liu J, Batalov S, Zhou D, Orth A, Ding S and Schultz PG. (2004). Proc. Natl. Acad. Sci. USA, 101, 135–140. Ziauddin J and Sabatini DM. (2001). Nature, 411, 107–110.