Functional analysis of the yeast genome

Functional analysis of the yeast genome

771 Functional analysis of the yeast genome Elizabeth A Winzeler* and Ronald W Davist The release of the complete genome sequence of the yeast Saccha...

578KB Sizes 12 Downloads 116 Views

771

Functional analysis of the yeast genome Elizabeth A Winzeler* and Ronald W Davist The release of the complete genome sequence of the yeast Saccharomyces cerevisiae has ushered in a new phase of genome research in which sequence function will be assigned. The goal is to determine the biological function of each of the >6,000 open reading frames in the yeast genome. Innovative approaches have been developed that exploit the sequence data and yield information about gene expression levels, protein levels, subcellular localization and gene function for the entire genome.

Addresses Department of Biochemistry, Stanford University School of Medicine, Stanford, California 94305-5307, USA *e-mail: [email protected] te-mail: [email protected] Current Opinion in Genetics & Development 1997, 7:771-776 http://biomednet.com/elecref/095943?X00700771 © Current Biology Ltd ISSN 0959-437X Abbreviations ORF open reading frame SAGE serial analysis of gene expression

Introduction T h e yeast genome consists of 13,389kb and encodes -6,000 predicted open reading frames (ORFs) [1°,2°°]. T h o u g h analysts have assigned a function to many of the proteins encoded by the predicted ORFs [1",2°'], there is clearly a need for further experimental development. As of July 1997, biological data had been collected for only -46% of the predicted proteins. Although the function of another 22% can be inferred on the basis of homology to other experimentally characterized proteins, the role of 32% remains obscure [3"]. Determining the function of these uncharacterized ORFs would greatly benefit the larger biomedical community: yeast is the first eukaryotic organism to have its genome fully sequenced. As of N o v e m b e r 1996, 20% of all positionally-cloned genes mutated in human disease processes arc known to have homologues in yeast [4°]. As the number of sequenced human genes grows, so will the number of yeast genes with human homologues [4°,5]. Yeast is inexpensive to cultivate, can exist in either haploid or a diploid states, and is extremely facile to genetic manipulation. To understand as much as possible about each and every gene encoded by the yeast genome is an essential scientific endeavor. Ambitious projects have been initiated with this end in mind. This review describes recent work devoted to assigning function to the yeast genome. Expression analysis One way to acquire functional information about a genome is to analyze gene expression. Northern blot analysis has

been performed on many of the O R F s in chromosomes III, VI, and XI [6°,7-9]. Although limited data about a transcript's abundance can also be obtained, the chief advantage of Northern analysis is that a transcript's size can be measured. Northerns may also be used to identify new potential O R F s and to uncover genes that encode RNAs which are neither translated nor polyadenylated. Some RNAs cannot, however, be unambiguously assigned to genes and examining expression for all genes with gel-based Nottherns under a larger number of conditions will be difficult. An innovative alternative to Northern blots for measuring transcript abundance is SAGE (serial analysis of gene expression) [10]. In SAGE, polyadenylated R N A is isolated from yeast cells and then, using a series of enzymatic manipulations, each messenger RNA molecule is reduced to an 11 bp c D N A fragment which serves as an identifier. T h e s e small fragments are ligated to one another and the concatenated product is sequenced. By comparing these fragment sequences to the yeast genomic sequence, it is generally possible to assign each fragment to a specific location within the yeast genome. T h e frequency with which each tag appears in the concatenated sequence indicates transcript abundance. Using SAGE, it has been shown that 76% of the O R F s in the yeast genome produce R N A molecules that are detected at least once [11"°]. For this analysis, 60,000 tags were analyzed: 20,000 from cells in GI, 20,000 from cells in S phase, and 20,000 from cells in G 2 phase. T h e method was sensitive enough to detect RNAs that were transcribed at levels of <0.3 copies per cell in yeast [11°°]. Although SAGE depends on the presence of a polyA tail on an RNA, it makes no assumption about what an O R F is and can therefore function as a vehicle for discovering new ORFs. 160 small non-annotated O R F s were identified in the yeast study [11"]. T h e chief disadvantage of SAGE is that a large number of tags need to be sequenced to detect rare R N A molecules. In some cases it can be difficult to assign a tag to an ORF, especially if the tag maps to a region that lies far outside of a predicted O R E Although clearly superior to Northern analysis for assaying global gene expression, it is again difficult to examine a large number of different conditions using SAGE. Array-based hybridization technology offers an efficient way for monitoring gene expression under many conditions (Figure 1). In this massively parallel method, miniature ordered collections of nucleic acid probes are synthesized on [12-14] or attached to [15,16,17 °] a solid surface. T h e s e arrays can contain from 1,000 up to 100,000 single or double-stranded probes ranging in length from 20 to several thousand nucleotides. After labeling with a fluorescent marker, RNA targets are hybridized to the

772

Genomes and evolution

array. T h e hybridization signal for a given probe reflects the expression level of the corresponding ORF and the levels of thousands of different ORFs can be monitored simultaneously.

Figure 1

(a)

Multiple 25 nucleotide probes per gene

Gene 1

(b)

Gene 2

©

Label mRNA or cDNA • "v

m. • i

SPS4-~"

Alternatively, scientists at Affymetrix Inc. have synthesized probes to the whole yeast genome in high-density oligonucleotide arrays on chemically treated glass slides using a combination of oligonucleotide chemistry and photolithography (DJ Lockhart, L Wodicka, personal communication). These arrays contain 20 or more 25met oligonucleotide probes for every annotated ORF in the yeast genome. These 25mer probes were selected from the sequence of each ORF to be as unique as possible and to have good hybridization properties. When fluorescently-labeled RNA or cDNA is hybridized to the chip, the averaged intensity over all probes is highly reproducible and quantitative [20°°]. Using RNA from yeast grown in rich media, scientists at Affymetrix have detected transcripts for >85% of predicted ORFs, some at levels of one copy per 20 cells (DJ Lockhart, L Wodicka, personal communication).

Isolate mRNA from yeast ©

(c)

include every known ORF encoded by the yeast genome (J DeRisi, personal communication).

'

q ..if

~ b J, r a l p h . , , ~

Meiosis 0 hour

m.m ~ * e" n " "

r•

~

Meiosis 12 hours C u r r e n t O p i n i o n in G e n e t i c s & D e v e l o p m e n t

Expression analysis using Affymetrix high-density arrays. (a) 20 or more 25mer oligonucleotide probes for every annotated ORF are synthesized on a glass slide in a contiguous arrangement. (b) RNA is purified from yeast and is either labeled directly, or converted to cDNA and then labeled. The labeled target is fragmented before hybridization to the array and hybridization is detected using a modified confocal laser scanning device [20"']. (c) Close-up example of the fluorescent signal observed after hybridization of labeled cDNA derived from mRNA isolated from yeast during meiosis. The collection of probes specific to the SPS4 gene [43], encoding a spore-specific protein, is boxed, indicating induction of this gene during sporulation.

Two types of arrays have been used for gene expression analysis in yeast. First, researchers at Stanford University have used the yeast genome sequence and a highly efficient 96-well oligonucleotide synthesizer to generate oligonucleotide primers that permit the amplification of entire coding regions from yeast [18,19"]. T h e amplified ORFs are then placed at defined locations on a microscope slide using a robotic microarraying device [15]. In a preliminary study, 40% of the genome was arrayed and probed with RNA isolated from yeast grown under different conditions [19°°]. Many genes could be classified on the basis of their response to stimuli, such as heat or cold shock. This work has now been extended to

Although these arrays are useful for characterizing patterns of gene expression, they can also be used to probe genome content. Duplications, deletions, recombination hotspots and origins of DNA replication may be identified. Arrays will be the tool of choice for assaying genetic diversity. In the future, the entire yeast genome could be synthesized as a series of oligonucleotides in a high-density array format. Arrays may also become dense enough to allow resequencing of an entire genome in a single hybridization [21"]. The yeast

proteome

Scientists have coined the term 'proteomics' to describe analysis of the combined set of all protein products encoded by an organism's genome [22]. T h e behavior of many of the proteins in the proteome can be monitored using two-dimensional gel electrophoresis [23] in a technique that is, in some ways, analogous to the massively parallel array technology used for characterizing global gene expression. It is estimated that 25% to 35% of the soluble proteins encoded by the yeast genome can be separated and resolved from one another on a polyacrylamide gel on the basis of molecular weight and is•electric point [24,25]. Once the identity of a spot is known, changes in the level of that p r o t e i n - - a s well as post-translational modifications such as phosphorylation or glycosylation--can be monitored on the gel. Prior to the completion of the yeast genome project, however, making the connection between different spots on the gel and cloned yeast genes was difficult and tedious. Now, the availability of the completed genome sequence, plus better software and more sensitive instruments, has greatly accelerated the process of pairing two-dimensional gel spots and yeast genes. Recently, Shevchenko and co-workers [26 °°] assigned unambiguous identities to 150 spots in the yeast proteome using a highly automated procedure. 90% of the spots

Functional analysis of

were identified by their tryptic digest peptide-mass spectra alone, though the remainder required partial Edman degradation sequencing using nano-electrospray tandem mass spectrometry for positive identification. Though not all proteins are sufficiently abundant or soluble for resolution on a two-dimensional gel, instrumentation is becoming more sensitive: even femtomole quantities of protein can be identified now by mass-spectrometry [27].

the yeast genome Winzeler and Davis

773

Figure 2 (a) Before selection

(b) X

Cell 1 ~ i i i i Cell 2

~ i i

Ceil 3

......

Cell 4

~_~ii~

Cell 5

,,,,,~!ili:

i!i~

Y

!

~ i i ~ . ~ l m ~ _

~ i~

...... ~

S u b c e l l u l a r localization The location of a protein in a cell also provides significant information about its possible function. Burns et al. [28] constructed 2,800 yeast strains carrying translational fusions of lacZ to random genes using transposon mutagenesis and then localized the [3-galactosidase fusion proteins using immunofluorescence microscopy. Distinct staining patterns were observed for 68% of the fusion proteins. 10% of fusions were localized to discrete subcellular locations. On the basis of frequency of cells expressing lacZ and assuming random integration, they estimated that 74.2% of the ORFs in yeast are expressed under vegetative growth conditions.

Cre-lox recombination sites and an HA epitope have also been introduced into the transposon mutagenesis cassette [29], allowing the lacZ gene to be excised and, in some cases, to leave behind a detectable epitope-tagged protein. Researchers at Yale University have generated 6,500 additional vegetatively-expressed fusion proteins, covering an estimated 25% of ORFs in the yeast genome and are planning to generate 30,000 for 95% coverage (P Ross-McDonald, M Snyder, personal communication). Localization data can also be obtained by constructing fusions to the green fluorescent protein [30].

(d) (C) After selection and outgrowth Cell 3 Cell 4

i ~ i

Cell 5 Cell 6

m,,,, -

-

~

All cells , ~ ' ~

ii Gene X (essential)

Gene Y (non-essential)

PCR product PCR product (A-Ty primers) (a-Ty primers)

Current Opinion in Genetics & Development

Genomic footprinting for two genes in yeast [38°',39]. (a) Tyl elements are induced to a level where there are five insertions per genome on average. Insertions occur throughout the haploid genome in both essential and non-essential genes. The presence of Tyl in a particular ORF is revealed by performing PCR using one fluorescently labeled primer (A or B) complementary to the region upstream of the ORF (gene X or Y) and another primer complementary to the Tyl element itself. (b) The distribution of PCR products for each primer pair as revealed by gel electrophoresis. (c) After selection and outgrowth for several generations, cells carrying insertions in the essential gene X will disappear from the population. (d) When PCR is repeated after selection using primers A and B, a large number of different-sized products will be observed for the non-essential gene (Y) whereas a footprint will be observed for the essential gene (X).

P r o t e i n - p r o t e i n interactions Like genes, proteins can be classified on the basis of their interaction with other proteins. One effective system for identifying such interaction is the yeast two hybrid system [31]. Ideally, one would like to map interactions between every protein in a genome with every other protein. As a test case, an exhaustive screen for interactors was performed on the bacteriophage T7 genome [32"]. Many known interactions and some novel interactions were detected. More efficient, high-throughput, two-hybrid screens have been developed with the goal of generating a yeast full-genome protein linkage map [33°°]. The use of array-based hybridization technology may speed the identification of positive clones in this case (R Cho, personal communication).

Large-scale deletion and m u t a t i o n a l analysis One of the most effective ways to determine the function of an ORF is to mutate it and then examine the phenotype of the mutant organism. Several studies are underway to systematically provide this type of data for every ORF in

the yeast genome. Deletion strains are easily generated in yeast by targeted homologous recombination using as little as 30bp of sequence homology [34,35] on replacement cassettes generated by PCR [36,37]. Such deletions have been generated for all the genes in chromosome VIII, as well as a large proportion of genes in chromosomes I, V, and XIII (M Johnston, personal communication). Similar mutational data can be obtained without constructing deletion strains. Using an elegant method called genomic footprinting [38°',39], functional data was obtained for each of the 268 ORFs in chromosome V. In genomic footprinting, insertional mutagenesis is performed en masse on a large population of yeast cells by overexpressing the Tyl transposon (Figure 2). After induction of Tyl, the mutagenized yeast cells are outgrown under a variety of selective conditions. The presence and location of a Tyl within a gene is revealed by performing PCR on that ORE In the starting population, a distribution of product sizes is observed, as a result of the fact that Tyl has integrated at multiple points within the gene in the total population. If a gene is essential for growth under

774

Genomes and evolution

Figure 3

F ¸

(a)

(b)

Plasmid DNA

V

Oligonucleotide array

(d)

(e)

Gene A

Gene B

Current Opinion in Genetics & Development

Deletion analysis in yeast. (a) A deletion cassette is generated by PCR using primers that are homologous to a plasmid-based selectable marker [37] at their 3" ends and flanking yeast sequence from the gene to be targeted at their 5' ends. In addition, one primer contains a 20 bp tag sequence. Each unique tag sequence is flanked by common PCR priming sites. Yeast is transformed with the PCR product resulting in homologous recombination and gene replacement at the targeted ORE (b) Different deletion strains are pooled together. Before selection, genomic DNA is isolated from the pool and the tags are amplified using fluorescently-labeled primers. Then the labeled amplicons are hybridized to high-density oligonucleotide arrays that contain probes complementary to the tags. Hybridization is detected using a modified laser confocal scanning device. (c) The pool of yeast is transferred to selective media and outgrown for eight generations. (d) During selective growth, genomic DNA is isolated from the pool at time points, tags are PCR-amplified and hybridized to the array. (e) Strains unable to survive the selection are identified by a loss of signal at defined locations on the array. As each unique probe location is associated with a different deleted gene, the gene responsible for the phenotype can be identified.

a particular selective condition, or at least contributes to the yeast cell's fitness, the number of cells carrying a Tyl insertion in that gene is reduced. 10% decreases in growth rate can be detected. In the chromosome V study [38°°], 61.6 % of genes showed a detectable phenotype under the seven conditions examined. T h e conditions tested included rich-glucose medium, minimal medium, lactate medium, high-temperature, caffeine, high-salt, and mating. T h e most c o m m o n phenotype observed was decreased growth rate in rich-glucose media, though in many cases the growth patterns were more revealing and specific. This analysis is being extended to the entire yeast genome (B Dunn, personal communication).

One shortcoming of both genomic footprinting and conventional deletion analysis is that a large amount of work is expended for each additional condition examined. Even with four-fold multiplexing of the P C R reactions, genomic footprinting of each new condition entails performing and analyzing the products from 1,500 P C R reactions for each time point. For conventional deletion analysis, each of the 6,000 deletion strains must be re-assayed independently. A method of examining a large number of specific stress conditions is to perform the selections in a large pool of deletion strains. If a unique 20bp tag is introduced into the deletion mutant at the time the mutant is

Functional analysis of the yeast genome Winzeler and Davis

constructed [40"'] (Figure 3), the fraction of a given strain in a population can be measured by PCR-amplifying the tags from genomic DNA and then hybridizing the tags to oligonucleotide arrays bearing the tag-complements as probes. As a demonstration of the efficacy of this approach, Shoemaker eta/. [40"'] have constructed tagged deletion strains for eleven auxotrophic genes. T h e strains were pooled and grown in minimal media containing different supplements. When the medium lacked adenine, the adenine auxotrophs disappeared from the population over time; when tryptophan was omitted, the tryptophan auxotrophs disappeared. This method should be expandable to much larger pools and thus allow rapid phenotypic analysis of deletion strains under exotic stress conditions: a similar method, known as signature tagging, has been used successfully to identify virulence genes in bacteria [41]. A consortium of European and North American laboratories, led by Mark Johnston of Washington University, has been formed to generate tagged deletion strains and, to date, -10% of the ORFs have been deleted. Although significant effort will be devoted to making the deletion strains, it should be possible to analyze the collection under a large variety of different selective conditions, once the collection is made, with little additional effort. If a sufficiently large number of conditions are examined, then there is a greater chance that function will be revealed. In addition, the collection--as well as individual deletion strains--will be available to the larger yeast community for use in countless other experiments.

Conclusions and future directions In addition to the availability of genome sequence data being a benefit to researchers by sparing them the relatively dull tasks of sequencing, mapping and subcloning, the availability also marks a point where researchers will be able to adopt new paradigms in experimental analyses. Researchers may now switch from the analysis of single genetic units to the analysis of a complete system [42"']. For example, scientists will be able to perturb the genome by deleting a single gene and then monitor how expression changes for every other gene in the genome, both at the RNA and protein level. Finally, as much of the functional data will be made available on-line, much more research will be performed over the Internet instead of on the laboratory bench. T h e value of the functional analysis projects described above will ultimately depend on how well the resulting data can be organized and analyzed. If the data is presented well, new biological insights may be obtained from the synthetic analysis of thousands of results obtained in dissimilar experiments.

Note added in proof Two papers of exceptional interest (referred to as J DeRisi, personal communication and DJ Lockhart, L Wodicka,

775

personal communication in the text, respectively), describing whole genome transcriptional analysis by hybridisation, have now been published [44"',45"'].

Acknowledgements We thank members of the Davis group for helpful comments on the manuscript and Barbara Dunn, Joe DeRisi, David J Lockhart, Lisa Wodicka, Ray Cho, Petra Ross-MacDonald and Mark Johnston for permission to cite unpublished data. Elizabeth Winzeler was supported by the John Wasmuth Fellowship in genomic analysis (HG00185-01). Yeast genome research in the Davis laboratory is supported by National Institutes of Health grant HGO1633-01.

References and recommended reading Papers of particular interest, published within the annual period of review, have highighligted as: • of special interest oo of outstanding interest 1. The yeast genome directory. Nature 1997, • 387:5. The definitive guide to the yeast genome sequencing project. 2. •o

Mewes HW, Albermann K, Bahr M, Frishman D, Gleissner A, Hani J, Heumann K, Kleine K, Maierl A, Oliver SG et al.: Overview of the yeast genome. Nature 1997, 387:7-65. This description of the yeast genome project details the international sequencing effort and some of the features of the yeast genome. 3. •

Mewes HW, Albermann K, Heumann K, Liebl S, Pfeiffer F: MIPS: a database for protein sequences, homology data and yeast genome information. Nucleic Acids Res 1997, 25:28-30. A guide to the data contained in the world wide web accessible MIPS yeast database. 4. •

Bassett DE Jr, Boguski MS, Spencer F, Reeves R, Kim S, Weaver T, Hieter P: Genome cross-referencing and XREFdb: implications for the identification and analysis of genes mutated in human disease. Nat Genet 1997, 15:339-344. Many yeast genes have homologues in genes involved in human disease. This work describes the identification of these homologues. 5.

Tugendreich S, Bassett DE Jr, McKusick VA, Boguski MS, Hieter P: Genes conserved in yeast and humans. Hum Mo/ Genet 1994, 3:1509-151 7.

6. •

Richard GF, Fairhead C, Dujon B: Complete transcriptional map of yeast chromosome Xl in different life conditions. J Mol Biol 1997, 268:303-321. The authors describe the analysis of transcripts from yeast chromosome XI using Northern blot hybridization. 7.

Yoshikawa A, Isono K: Construction of an ordered clone bank and systematic analysis of the whole transcripts of chromosome VI of Saccharornyces cerevisiae. Nucleic Acids Res 1991, 19:1189-1195.

8.

Yoshikawa A, Isono K: Chromosome III of Saccharomyces cerevisiae: an ordered clone bank, a detailed restriction map and analysis of transcripts suggest the presence of 160 genes. Yeast 1990, 6:383-401.

9.

Tanaka S, Isono K: Correlation between observed transcripts and sequenced ORFs of chromosome III of Saccharomyces cerevisiae. Nucleic Acids Res 1993, 21:1149-1153.

10.

Velculescu VE, Zhang L, Vogelstein B, Kinzler KW: Serial analysis of gene expression. Science 1995, 270:484-487.

11. o•

Velculescu VE, Zhang L, Zhou W, Vogelstein J, Basral MA, Bassett DE Jr, Hieter P, Vogelstein B, Kinzler KW: Characterization of the yeast transcriptome. Cell 1997, 88:243-251. This is the first paper in which expression levels for all genes in the yeast genome are analyzed and is a seminal report on whole genome functional analysis. The authors demonstrate that transcripts can be detected for 75% of the genes in the yeast genome using the SAGE method. In addition, they detect expression for 160 small non-annotated ORFs. 12.

Fodor SP, Read JL, Pirrung Me, Stryer L, Lu AT, Solas D: Lightdirected, spatially addressable parallel chemical synthesis. Science 1991,251:767-773.

776

Genomes and evolution

13.

Fodor SP, Rava RP, Huang XC, Pease AC, Holmes CP, Adams CL: Multiplexed biochemical assays with biological chips. Nature 1993, 364:555-556.

14.

Pease AC, Solas D, Sullivan EJ, Cronin MT, Holmes CP, Fodor SP: Light-generated oligonucleotide arrays for rapid D N A sequence analysis. Proc Nat/Acad Sci USA 1994, 91:50225026.

15.

Shalon D, Smith SJ, Brown PO: A DNA microarray system for analyzing complex DNA samples using two-color fluorescent probe hybridization. Genome Res 1996, 6:639-645.

16.

DeRisi J, Penland L, Brown PO, Bittner ML, Meltzer PS, Ray M, Chen Y, Su YA, Trent JM: Use of a cDNA microarray to analyse gene expression patterns in human cancer. Nat Genet 1996, 14:457-460.

17.

Schena M, Shalon D, Heller R, Chai A, Brown PO, Davis RW: Parallel human genome analysis: microarray-based expression monitoring of 1000 genes. Proc Nat/Acad Sci USA 1996,



93:10614-10619.

This recent work describes the use of cDNA microarrays for gene expression analysis and for gene discovery. 18.

Lashkari DA, Hunicke-Smith SP, Norgren RM, Davis RW, Brennan T: An automated multiplex oligonucleotide synthesizer: development of high-throughput, low-cost D N A synthesis. Proc Nat/Acad Sci USA 1995, 92:7912-7915.

29.

Ross-Macdonald P, Sheehan A, Roeder GS, Snyder M: A multipurpose transposon system for analyzing protein production, localization, and function in Seccheromyces cerevisiae. Proc Nat/Acad Sci USA 1997, 94:190-195.

30.

Niedenthal RK, Riles L, Johnston M, Hegemann JH: Green fluorescent protein as a marker for gene expression and subcellular localization in budding yeasL Yeast 1996, 12:773786.

31.

Fields S, Song O: A novel genetic system to detect protein-protein interactions. Nature 1989, 340:245-246.

32. •

Bartel PL, Roecklein JA, SenGupta D, Fields S: A protein linkage map of Escherichia coil bacteriophage TZ Nat Genet 1996, 12:72-77. The authors are able to determine a complete protein linkage map for an organism with a small genome, the phage TT. The interaction of every protein in the genome with every other protein therein is documented. 33. •.

Fromont-Racine M, Rain JC, Legrain P: Toward a functional analysis of the yeast genome through exhaustive two- hybrid screens. Nat Genet 1997, 16:277-282. The authors of this paper describe a strategy by which two-hybrid analysis could be extended to generate a complete protein linkage map for the yeast genome. 34.

Rothstein R: Targeting, disruption, replacement, end allele rescue: integrative D N A transformation in yeast. Methods Enzymo/1991, 194:281-301.

35.

Manivasakam P, Weber SC, McEIver J, Schiestl RH: Microhomology mediated PCR targeting in Seccheromyces cerevisiae. Nucleic Acids Res 1995, 23:2799-2800.

36.

Lockhart D, Dong H, Byrne M, Follettie K, Gallo M, Chee M, Mittmann M, Wang C, Kobayashi M, Horton H, Brown E: Expression monitoring by hybridization to high-denisty oligonucleotide arrays. Nat Biotechnol 1996, 14:1675-1680. This paper describes the use of high-density oligonucleotide arrays to monitor gene expression. Similar arrays have been constructed for full-genome expression monitoring in yeast (see Note added in proof).

Baudin A, Ozier-Kalogeropoulos O, Denouel A, Lacroute F, Cullin C: A simple and efficient method for direct gene deletion in Saccharomyces cerevisiae. Nucleic Acids Res 1993, 21:33293330.

37.

Wach A, Brachat A, Pohlmann R, Philippsen P: New heterologous modules for classical or PCR-based gene disruptions in Saccharomyces cerevisiae. Yeast 1994, 10:17931808.

21. •-

38. •-

19. ••

Lashkari D, DeRisi J, McCusker J, Namath A, Gentile C, Hwang S, Brown P, Davis R: Yeast genome microarrays for parallel genetic and gene expression analysis of the yeast genome. Proc Nat/Acad Sci USA 1997, 94:13057-13062. The first paper describing the use of microarrays to study gene expression and genome content in yeast. 20. •-

Chee M, Yang R, Hubbell E, Berno A, Huang XC, Stern D, Winkler J, Lockhart DJ, Morris MS, Fodor SP: Accessing genetic information with high-density DNA arrays. Science 1996, 274:610-614. This paper demonstrates the potential array-based technology has to revolutionize genome analysis. In a demonstration, the 16.6 kb human mitochondrial genome was resequenced by hybridization. 22.

Wilkins M, Pasquali C, Appel R, Ou K, Golaz O, Sanchez J, Yan J, Gooley A, Hughes G, Humphery-Smith Iet a/.: From proteins to proteomes-large-scale protein identification by 2-dimensional electrophoresis and amino-acid analysis. Bio-Technology 1996, 14:61-65.

23.

O'Farrell PH: High resolution two-dimensional electrophoresis of proteins. J Bio/Chem 1975, 250:4007-4021.

24.

Garrels JI, Futcher B, Kobayashi R, Latter GI, Schwender B, Volpe T, Warner JR, McLaughlin CS: Protein identifications for a Saccharomyces cerevisiae protein database. Electrophoresis 1994, 15:1466-1486.

25.

Boucherie H, Dujardin G, Kermorgant M, Monribot C, Slonimski P, Perrot M: Two-dimensional protein map of Saccharomyces cerevisiee: construction of a gene-protein index. Yeast 1995, 11:601-613.

26. °°

Shevchenko A, Jensen ON, Podtelejnikov AV, Sagliocco F, Wilrn M, Vorm O, Mortensen P, Shevchenko A, Boucherie H, Mann M: Linking genome and proteome by mass spectrometry: largescale identification of yeast proteins from two dimensional gels. Proc Nat/Acad Sci USA 1996, 93:14440-14445. This paper describes a highly-automated procedure for generating a twodimensional gel map of the yeast proteome and recent progress toward that goal. 27.

28.

Wilm M, Shevchenko A, Houthaeve T, Breit S, Schweigerer L, Fotsis T, Mann M: Femtomole sequencing of proteins from polyacrylamide gels by nano-electrospray mass spectrometry. Nature 1996, 379:466-469. Burns N, Grimwade B, Ross-Macdonald PB, Choi EY, Finberg K, Roeder GS, Snyder M: Large-scale analysis of gene expression, protein localization, and gene disruption in Saccharomyces cerevisiae. Genes Dev 1994, 8:1087-1105.

Smith V, Chou KN, Lashkari D, Botstein D, Brown PO: Functional analysis of the genes of yeast chromosome V by genetic footprinting. Science 1996, 274:2069-2074. An elegant study assaying the phenotypic effects of Tyl transposon insertions in all genes on yeast chromosome V. 39.

Smith V, Botstein D, Brown PO: Genetic footprinting: a genomic strategy for determining a gene's function given its sequence. Proc Nat/Acad Sci USA 1995, 92:6479-6483.

40. •-

Shoemaker DD, Lashkad DA, Morris D, Mittmann M, Davis RW: Quantitative phenotypic analysis of yeast deletion mutants using a highly parallel molecular bar-coding strategy. Nat Genet 1996, 14:450-456. An explanation of how yeast deletion strains can be constructed, tagged and used in parallel quantitative phenotypic analysis. 41.

Hensel M, Shea JE, Gleeson C, Jones MD, Dalton E, Holden DW: Simultaneous identification of bacterial virulence genes by negative selection. Science 1995, 269:400-403.

42. °,

Wolfe KH, Shields DC: Molecular evidence for an ancient duplication of the entire yeast genome. Nature 1997, 387:708713. This work is an example of what can be achieved through the reanalysis of sequence data and provides an indication of the direction of genomic research. 43.

Garber AT, Segall J: The SPS4 gene of Saccharomyces cerevisiae encodes a major sporulation-specific mRNA. Mo/ Cell Biol 1986, 6:4478-4485.

44. •.

DeRisi JL, lyer VR, Brown PO: Exploring the metabolic and genetic control of gene expression on a genomic scale. Science 1997, 278:680-686. This paper describes how genome-wide transcript profiles change in yeast after a shift from fermentation to respiration, and how metabolic reprogramruing can be inferred from the analysis of the data. 45. °•

Wodicka L, Dong H, Mittman M, Ho M, Lockhart D: Genomewide expression monitoring in Saccharomyces cerevisiae. Nat Biotechnol 1997, 15:1359-1367. In this work, the authors describe determining the abundance of messenger RNA molecules for all annotated genes in the yeast genome under two different growth conditions using high-density oligonucleotide arrays.