WO1999045021A1

WO1999045021A1 - Cell cycle regulated genes

Info

Publication number: WO1999045021A1
Application number: PCT/US1999/004705
Authority: WO
Inventors: Raymond J. Cho; Michael J. Campbell; Lisa Wodicka; David J. Lockhart; Ronald W. Davis
Original assignee: Affymetrix, Inc.
Priority date: 1998-03-03
Filing date: 1999-03-03
Publication date: 1999-09-10
Also published as: AU2892199A

Abstract

Progression through the eukaryotic cell cycle is known to be both regulated and accompanied by periodic fluctation of the expression levels of numerous genes. We report here the genome-wide characterization of mRNA transcript levels during the cell cycle of the budding yeast Saccharomyces cerevisae. Cell cycle-dependent periodicity was found for 422 of the 6,220 monitored transcripts. Fewer than a quarter of the 422 genes identified in our screen have been previously implicated in cell cycle period-specific activities, and more than 40 % have no known biological role. Surprisingly, nearly 30 % of the 422 genes were found directly adjacent to other genes that displayed induction in the same cell cycle phase, suggesting a mechanism for local chromosomal organization in global mRNA regulation. Analysis of regions upstream of these genes indicates that most cell cycle-dependent transcription is driven by previously uncharacterized promoter and enhancer elements. Using statistical analysis of nucleotide frequencies in regions upstream of coinduced genes, candidate cell cycle-dependent regulatory elements were identified.

Description

CELLCYCLEREGULATED GENES

This application claims the benefit of co-pending provisional application Serial No. 60/076,658, filed March 3, 1998, which is incorporated by reference herein.

TECHNICAL FIELD OF THE INVENTION This invention is related to growth and replication regulatory genes and genes involved in growth and replication regulation. BACKGROUND OF THE INVENTION

The events of DNA replication, chromosome segregation, and mitosis define a fundamental periodicity in eukaryotic cell cycle. Precise coordination of the unidirectional transitions between these stages is critical to cell integrity and survival.

Loss of appropriate cell cycle regulation leads to genomic instability and is believed to play a role in the etiology of both hereditary and spontaneous cancers. Cell cycle periodicity has been observed for many cellular processes, including control of mRNA transcription, responsiveness to external stimuli , and subcellular localization of proteins. Genetic studies have revealed that the activity of cell cycle regulatory proteins are required for normal DNA repair, meiosis, and multicellular development. These observations suggest that all eukaryotic cells experience important physiological changes during the cell cycle, and that diverse biological events depend on maintenance of this periodicity. The numerous biological changes associated with the cell cycle make it an attractive model for the study of genome-wide regulation of gene activity. Parallel identification of all the genes in a genome that are coordinately regulated during such a process provides a consistent internal standard for comparison of gene activity over time and makes it possible to search statistically for candidate regulatory sequences.

Although the cell cycle-dependent regulation of activity of many individual genes has been studied, comprehensive results for a genome are likely to reveal novel functional and physical organization in coordinate gene regulation. Large scale analysis of gene activity also makes it possible to assess the extent to which conclusions drawn from the study of the regulation of individual genes may be applied to the entire genome.

One of the key mechanisms of gene regulation takes place on the level of mRNA transcription. There is a need in the art for a complete and exhaustive analysis of the transcription pattern of yeast genes as a function of time during the yeast cell cycle. This would provide a model for regulation in more complex organisms, such as humans. In addition, genes identified as important to the yeast cell cycle may be useful as therapeutics, diagnostics, or drug targets. Similarly the human homologues of such genes will likely be therapeutically or diagnostically important. SUMMARY OF THE INVENTION

It is an object of the present invention to provide a set of nucleotide probes. It is another object of the present invention to provide a method of identifying candidate therapeutic agents.

It is an object of the present invention to provide a method of determining phase of cell cycle of yeast cells.

These and other objects of the invention are achieved by providing a set of polynucleotide probes. Each of said probes comprises at least 12 nucleotides. Greater than 50 % of the probes in the set comprise portions of yeast genes which are cell cycle regulated. Cell cycle regulated yeast genes are defined as genes whose expression varies by more than 2-fold during the cell cycle of yeast.

According to another aspect of the invention a method is provided for identifying compounds which affect the cell cycle of yeast. Synchronized yeast cells are contacted with a test compound. RNA is isolated from the yeast cells. The amount of particular mRNA species in the RNA isolated from the yeast cells is determined using a set of polynucleotide probes. Each of said probes comprises at least 12 nucleotides. Greater than 50 % of the probes in the set comprise portions of yeast genes which are cell cycle regulated. Cell cycle regulated yeast genes are defined as genes whose expression varies by more than 2-fold during the cell cycle of yeast. A test compound is identified as a candidate drug if it is found to affect the amount of the particular mRNA species.

According to another aspect of the invention a method is provided for determining cell cycle of a culture of synchronized yeast cells. RNA from synchronized yeast cells is isolated. The amount of particular mRNA species in the RNA isolated from the yeast cells is determined using probes. Each of said probes comprises at least 12 nucleotides. Greater than 50 % of the probes in the set comprise portions of yeast genes which are cell cycle regulated. Cell cycle regulated yeast genes are defined as genes whose expression varies by more than 2-fold during the cell cycle of yeast. The relative amounts of the particular mRNA species determined is characteristic of phase of cell cycle of the yeast.

These and other aspects of the invention provide the art with a wealth of information on cell cycle regulation in eukaryotic cells. This information can be used for drug discovery as well as for analytic purposes. BRIEF DESCRIPTION OF THE DRAWINGS

Fig. la Phase contrast photographs of yeast cells in various phases of the cell cycle as indicated. Fig. lb . Graph of the percentage of cells that are unbudded (purple) small budded (yellow) , and large budded (green) versus time. Fig. lc Graph of the percentage of cells that are premitotic (red), mitotic (blue) and postmitotic (green) versus time. At 110 minutes, virtually all cells completed mitosis.

Fig.2 Close-up views of high-density oligonucleotide arrays containing 25-mer probes for nearly every gene in the Saccharomyces cerevisiae genome, following hybridization with labeled cDNA from synchronized cells. The first column shows array features containing probes to the CJNl gene (highlighted in red) from different time points. The second column shows array features containing probes to the YML027w open reading frame. Transcript levels for both genes reached their maxima during late G, phase.

Fig.3 Chromosomal positions of genes whose abundance is cell cycle period-specific.

Fig. 4 Graphs of normalized transcript level divided by mean value against time for Fig.

4a transcripts which peaked in both late G_! and G₂ ( yellow ), transcripts which peaked in both S and M (red),Fig. 4 b transcripts which peaked singly in S (pink), transcripts which peaked singly in M (green ), Fig.4c transcripts for the CLNl, CLN2, CLBl, and CLB2 cyclin genes, Fig. Ad transcripts for MCM genes (mustard), and transcripts for genes involved in DNA replication (blue).

Figs. 5_ι-5o. Phenotypic categorization of mRNAs which are regulated with the cell cycle.

Fig. 6. Functional classifications of cell cycle-dependent transcripts which display cell cycle-specific periodicity. Biological functions of characterized genes were determined from the published literature and the MIPS database. Transcripts which peaked twice in one cell cycle are also listed.

Fig. 7 Major functional groups of cell cycle-de pendent transcripts. ORF loci and names are listed for some genes detected as periodically transcribed by visual inspection of mRNA fluctuation patterns. Bioche mical functions were determined from the published literature and the MIPS database. Some of these transcriptional periodicities have been previously described.

Fig. 8 Upstream regulatory elements in cell cycle-dependent transcription. Known and putative regulatory sequences and their frequency of occurrence in genes which display cell cycle-dependent mRNA Auction. Frequencey of occurrence is also shown for a random set of 300 genes from the yeast genome which do not show periodic mRNA Auction. Percent of genes in each category containing the regulatory sequence is listed in parentheses.

DETAILED DESCRD?TION An exhaustive analysis of gene expression has been carried out in yeast cells which have been synchronized. The analysis shows that 5.7% of the yeast genes vary their expression according to the phase of the cell cycle. The criterion used for determining change is a greater than 2-fold change. Three hundred thirty five genes were identified which demonstrated consistent minima and maxima of transcript levels. This analysis permits interesting cell regulatory genes to be identified on the basis of their expression patterns.

Sets of polynucleotide probes are useful for testing the effects of compounds and extracts as pharmacological agents and for determining the phase of the yeast cell cycle. The probes for each gene or mRNA may comprise at least 10, 12, 14, 16, 18, 20, 22, or 25 nucleotides, and are typically single stranded. Double stranded probes can also be used. The probes may comprise sense or antisense sequences, so long as they are derived from transcribed portions of the genome. The probes may be DNA or RNA and can also contain nucleotide analogues or non-nucleotide moieties, such as radiolabels, fluorescent labels, enzymes, etc. At least 50% of the probes in the sets comprise portions of genes whose expression varies by more than 2-fold during the cell cycle of yeast. Preferably at least 60, 75, 90, 95, 97, or 98 percent of the probes in the sets are from such cell cycle varying genes. The probes which are not such genes may be controls for hybridization, which are irrelevant or mismatched at one or more nucleotides. The sets may be of various sizes, depending on the purpose for which they are to be used. Typically the sets comprise at least 10, 20, 30, 40, 50, 100, 200, 300, 313, or 355 probes. The sets of probes can be used for any type of nucleic acid hybridization reactions known in the art. They may be used for solution hybridization or on a solid phase. They may be used with samples which comprise RNA or cDNA They can be used with RNA which is isolated directly from cells or which is made in vitro using cDNA templates. Solid phases which may be used include microtiter dishes, arrays, filter blots, etc. One particularly useful type of solid support is an array in which the probes are synthesized onto discrete locations on the solid support. See, e.g., Chee et al, Science 274, 610-614 (1996). Such arrays can contain many discrete probes in a very small area, for example greater than 1,000 or 10,000 in an area of 1.6 cm². It is preferred that synchronized cells be used in the practice of the present invention. Cells can be synchronized using any techniques known in the art, including the use of conditional mutations and/or changes in the growth conditions of the cells.

Test compounds which can be tested using the sets of probes of the present invention may be, for example, simple organic or inorganic compounds, natural or synthetic polymers, proteins, oligonucleotides, polypeptides, semisynthetic derivatives of natural products, natural product extracts. Mixtures of compounds or test substances may be tested, as can libraries of test compounds. Combinatorial libraries of compounds may be used, as can libraries of cloned nucleic acids or proteins they produce. Any substances which can be contacted with yeast cells can be tested. Similarly, forms of radiation can be tested for their effects using the sets of probes of the present invention.

The test compounds may be those which are being screened for toxic, mutagenic, carcinogenic, or therapeutic value. Compounds or treatments which cause a change in the pattern of expression of the monitored genes are potentially useful to alter cell cycle of eukaryotes. Such treatments may be useful as antifungal agents. Such assays may also be useful as prescreens for identifying anti-cancer or anti-viral therapies.

The amount of hybridization of a test sample of RNA or DNA to a set of probes can be determined using any or the techniques or methods which are known in the art. For example, the test sample can be labeled with a radiolabel or a fluorescent label. After hybridization and rinsing, the amount of the label can be quantitated using appropriate instrumentation, such as a scintillation counter, an X-ray film, a fluorimeter. The amount of hybridization of the sets of probes to the samples collected from cells before and after treatment with a test compound, or treated and untreated can be readily compared. Similarly, the relative amounts of hybridization to a particular gene probe can be determined and compared to the amount of that gene's expression in the various phases of the cell cycle to determine what the phase of the cell cycle is. Any of the genes which are shown herein to be cell cycle regulated can be used for these purposes.

The above disclosure generally describes the present invention. A more complete understanding can be obtained by reference to the following specific examples which are provided herein for purposes of illustration only, and are not intended to limit the scope of the invention.

EXAMPLE 1 DETERMINATIONOF CELL CYCLEREGULATED GENES

Availability of complete sequence for the Saccharomyces cerevisiae genome has made it possible to quantitate mRNA transcript levels for virtually every yeast gene. In this study, commercially available high-density oligonucleotide arrays were used to quantitate mRNA transcript levels in synchronized yeast cells at regular intervals during the cell cycle. DNA oligonucleotide probes are directly synthesized on these arrays without individual manipulation or PCR amplification, minimizing the potential for cross-hybridization or clone error. To obtain synchronous yeast culture, cdc28-13 cells were arrested in late G, at 37

°C, and the cell cycle was reinitiated by shifting cells to 25 °C. Cells were collected at seventeen time points taken at 10 minute intervals, covering nearly two full cell cycles. Cells exhibited over 95% synchrony throughout the time course, as determined by bud size and nuclear position (Fig. 1). Poly(A)⁺ RNA was isolated from each sample, converted to cDNA, labeled, and hybridized to yeast whole genome oligonucleotide expression arrays (Fig. 2). To obtain cells synchronized in a different way, an isogenic strain bearing the cdcl5-2 allele, which enables arrest in late G₂, was used to generate a second time course. Transcript levels from the cdc!5-2 arrest time were also measured by hybridization to oligonucleotide arrays.

7 Some differences in transcript levels could result from the shift between the restrictive and permissive temperatures and from the state of cell cycle arrest. To avoid temperature-induced effects unrelated to cell cycle progression, analysis was focused on data from cdc28-13 time points taken more than forty minutes past the point of release from arrest.

The 1348 genes whose normalized mRNA level changed by more than 2-fold during the time course were visually examined for periodicity of expression in both the cdc28-13 and cdcl5-2 strains. Four hundred twenty-two genes were identified that demonstrated consistent periodic changes in transcript level (Fig. 3). This number represents 6.8% of all yeast genes and agrees with previous estimates of the number of genes in S. cerevisiae that display cell cycle-dependent transcription. Periodically expressed genes were distributed fairly randomly across the genome, and every chromosome contained at least one cell cycle regulated gene. A database containing a list of these genes, their relative mRNA fluctuations, and their functional classifications can be found at the WWW site http://sequence.www.stanford.edu/group/informatics/ yeastcycle.html.

The time course was divided into early G„ late G„ S, G₂, and M phase, based on the morphological stages of the cell cycle (Fig. 1A). One hundred thirty-five of these 422 transcripts peaked in late G„ while only 56 transcripts peaked during M phase (Table 1). Transcripts which peaked in late G, displayed particularly sharp rates of accumulation and decay, while transcripts which peaked in S generally displayed a less dramatic induction pattern. More than half of the transcripts which peaked in late G„ including the CLNl and CLN2 cyclins, displayed a minor peak in G₂. Thirteen out of 76 transcripts which peaked in S also displayed minor Mpeaks (Fig. 4). The presence of minor peaks may indicate that a transcript is affected by more than one cell cycle-dependent regulatory sequence. An additional 33 of the 422 identified genes were induced in two different cell cycle phases, but did not display a predominant peak. Genes displaying two transcriptional peaks were classified separately (Table 1).

EXAMPLE 2 CELL CYCLE REGULATORY ELEMENTS

8 Identification and characterization of upstream regulatory sequences are critical to elucidating global mechanisms of transcriptional regulation. Several upstream regulatory sequences involved in cell cycle-dependent transcription have already been identified, including the late G₍ elements MCB (Mlul cell cycle box) and SCB (Swi4/6 cell cycle box) and the early G, element ECB (early cell cycle box). To determine the proportion of cell cycle-dependent mRNA transcripts that involve these regulatory elements, the 500 bp upstream of each of the 422 genes were searched for MCB, SCB, and ECB consensus sequences. More than two-thirds of genes induced in late G, were downstream from an MCB or SCB sequence (Table 3), and almost a third of genes upregulated during early G, were downstream from an ECB consensus sequence.

However, more than 150 genes in the genome were downstream from an MCB sequence but did not display fluctuations in mRNA levels. Likewise, over 50 genes were downstream of an ECB sequence but were not induced in early G,. One explanation for this observation is that the sequences surrounding regulatory elements can negatively modulate their effects on transcription. Alternatively, these instances may represent transcripts that are too stable to display fluctuations in level during the cell cycle. Upstream regions of all genes were also searched for other known yeast regulatory sequences, including the ABF1 -binding site and the RAP 1 -binding site. None of these sites were found with disproportionate frequency upstream of one class of cell cycle regulated genes.

Other cell cycle period-specific transcription factors such as Swi5 do not have a highly conserved binding sequence, making it difficult to accurately search genomic sequence for possible sites of action. However, it is unlikely that these binding sequences alone could completely account for the large proportion of genes in which a known regulatory element could not be found. In addition, known cell cycle regulatory sequences were observed rarely in the promoter regions of genes induced in G₂, M, or S phase. Therefore, although known cell cycle regulatory sequences strongly influence G_! transcription in the yeast genome, it is likely that the majority of upstream elements conferring cell cycle-specific transcription have yet to be identified. The generation of a comprehensive list of coregulated genes makes it possible to statistically analyze a large set of promoter regions for previously undetected regulatory elements. A database of potential gene promoter sequences was created by extracting the 500 bp upstream of the translational start site of every gene in the genome. This data set was then searched for hexanucleotide and hepanucleotide sequences that occurred with disproportionate frequency in the upstream regions of one set of cell cycle-regulated genes. The nucleotides surrounding these short sequences were visually inspected to determine a longer consensus sequence. For example, the sequence 5' GTAAACA 3' was found upstream of nearly 40% of the genes induced in G₂and M, but upstream of less than 20% of genes induced at other times. After visual inspection of several G₂ and M gene promoter regions, this sequence was expanded to 5' AAAANGTAAACAA 3'. A search of non-coding sequence revealed that this sequence was found upstream of 14% of genes induced in G₂, but upstream of less than 1% of other genes in the genome (Table 3). Another method of identifying candidate regulatory elements is examination of

DNA sequences that have been previously implicated in cell cycle-dependent transcription, but for which a strong consensus sequence has not been established. For example, it has been shown that the MCM1 transcription factor plays a direct role in the induction of certain genes during G₂ and M. Examining the core MCM1 binding site 5' CCYWWWNNGG 3 ' upstream of genes induced in M, we were able to expand the core sequence to 5' CCYAAANNGGNNAAA 3'. This sequence was observed in the 1200 bp upstream of eight genes induced in G₂and M, but not closer than 2000 bp to other genes in the genome. Furthermore, this site closely resembles the constitutively bound MCM1 sites found upstream of the G₂ genes CLBl, CLB2, and SWI5. Because no single sequence was found near a majority of genes induced in G₂ and M, it is possible that a number of separate elements may be responsible for periodicity of mRNA abundance during these phases. Like the ECB sequence, G₂ and M regulatory elements maybe relatively rare, but highly specific in determining periodicity of transcript levels. Alternatively, regulatory sequences that affect transcription during these periods may be highly degenerate. Both methods described for identifying candidate regulatory elements

10 will become increasingly relevant as more genome-wide transcription data is generated. Further investigation will be needed to assess the role of candidate regulatory sequences suggested by these experiments.

EXAMPLE 3 GENOMIC ORGANIZATION

The chromosomal position of genes can strongly influence their transcription, as observed in the silencing of genes in telomeric regions. Little evidence was observed for a direct correlation between telomeric or centromeric gene location and mRNA fluctuation during the cell cycle. However, nearly 30% of all genes displaying periodic transcript levels were positioned directly adjacent to another gene induced in the same cell cycle phase (Fig. 3). The proportion of cell cycle regulated genes which would occupy adjacent positions by random chance is less than 3%. More than half of these gene pairs were transcribed divergently on opposite strands, many with fewer than 1500 bp bases separating their 5' ends. Because many eukaryotic transcription factor-binding sites are either nearly palindromic or are active on both strands, it is likely that these gene pairs are regulated by the same upstream sequence. Only 12 of the 110 adjacent gene pairs that were divergently displayed different patterns of mRNA fluctuation. Therefore, in genomes with limited intergenic sequence, sharing of upstream regulatory elements may be an important determinant of global mRNA regulation. It was also observed that almost 10% of all genes that displayed cell cycle-dependent transcript levels were directly adjacent to another gene transcribed in the same direction and induced at the same time. The high frequency of these adjacent gene pairs on the same strand suggests that a significant proportion of cell cycle-dependent mRNA fluctuation may result from uncharacterized enhancer elements. It is also possible that the clustering of cell cycle regulated genes results from local positional effects that are not sequence-dependent. For example, the restructuring of chromosomes during cell cycle events such as DNA replication may affect the transcription of genes at specific chromosomal positions.

11 EXAMPLE 4 BIOCHEMICAL FUNCTION OF GENES WHICH ARE CELL CYCLE REGULATED

The biochemical functions of genes displaying periodic mRNA fluctuation were examined. Consistent with previous studies, we observed cycle-dependent changes in transcript level for the CLN and CLB cyclin families, transcription factors, and gene products involved in DNA replication and packaging. (Fig. 4C). The resolution of these experiments was sufficient to distinguish induction of the MCM genes and CDC 6, which are involved in formation of the prereplication complex (pre-RC), from the DNA polymerase subunits and DNA replication factors induced in late Gj (Fig 4D). As expected, genes encoding constituents of a protein complex were generally coregulated. The DNA replication factors RFA1, RFA2, and RFA3 displayed nearly identical patterns of mRNA fluctuation. However, mRNA levels of components of the spindle pole body were induced in different cell cycle phases, perhaps reflecting distinct temporal roles for these genes.

Periodic mRNA fluctuation was also observed in functional classifications of genes not previously associated with the cell cycle. For example, transcripts for the FAA1, FAA3, and ELOl enzymes, which participate in fatty acid biosynthesis, peaked during G,. Many of the nuclear-encoded mitochondrial enzymes required for glycolysis and oxidative phosphorylation were induced in early G, with very similar patterns of mRNA fluctuation. None of the transcripts for these mitochondrial genes peaked outside of G,.

EXAMPLE 5

CELL CYCLE DEPENDENCY OF PROTEOLYTIC SUBSTRATES

Because proteolysis is known to be a critical factor in regulating progression of the cell cycle, transcripts of both proteolytic effectors and substrates were examined for periodic changes in transcript levels. No periodic Auctuation was observed for any transcripts encoding constituents of the anaphase promoting complex (APC) or ubiquitin-dependent degradation pathway, which are involved in the proteolytic

12 degradation of key cell cycle regulators. As previously reported, the transcripts of all known proteolytic substrates displayed cell cycle-dependent periodicity. However, using resolution of these experiments, it was observed that transcripts of the APC substrates ASE1 and CLB2 displayed identical kinetics of decay during G while the transcript of the APC substrate PDS 1 showed a distinct Auctuation, reaching its lowest level during S phase. Interestingly, the Auctuations of these transcripts exactly parallel the reported degradation of their gene products. Therefore, it is possible that the same mechanism underlies the attenuation of gene activity on both the mRNA and protein level.

EXAMPLE 6

PREDICTING FUNCTION OF PREVIOUSLY UNCHARACTERIZED GENES

It has been proposed that mRNA and protein expression patterns may provide clues to the function of previously uncharacterized genes. To assess the likelihood that uncharacterized genes found in our screen play a cell cycle period-specific role, we examined the correlation between mRNA Auctuation and gene function. We found that more than 60% of the characterized genes which displayed cell cycle-dependent mRNA fluctuation have been previously implicated in cell cycle period-specific biological activities. Therefore, mRNA regulation is a strong indicator of biological function in the cell cycle, and it is likely that many of the uncharacterized genes in this screen have functions related to the cell cycle. However, periodicity of mRNA abundance was observed in fewer than 25% of all known CDC genes and genes known to be involved in mitosis, DNA replication, or other cell cycle-specific biological roles. Many of the genes which do not display periodic transcript levels are known to be modulated at the post-translational level. These results emphasize the need for multiple approaches in elucidating the function of uncharacterized genes. It is likely that additional aspects of coordinate mRNA regulation await discovery from this data set. We encourage the reader to explore the database, which can be viewed at the internet address given above.

13 References

1. C. Wittenberg, K. Sugimoto, S.I. Reed, Cell 62, 225-37 (1990).

2. L.J. Oehlen, J.D. McKinney, F.R. Cross, Mol Cell Bio 16, 2830-7 (1996).

3. B. Zanolari, H. Riezman, Mol Cell Bio 11, 5251-8 (1991).

4. L.J. Oehlen, F.R. Cross, Genes Dev 8, 1058-70 (1994).

5. R. Scully, et al., Cell 90, 425-35 (1997).

6. R.B. Painter, B.R. Young, Proc Natl Acad Sci USA 11, 7315-7 (1980).

7. T. A. Weinert, L.H. Hartwell, Science 241, 317-22 (1988).

8. T. Weinert, Science 111, 1450-1 (1997).

9. J.K. Jang, L. Messina, M.B. Erdman, T. Arbel, R.S. Hawley, Science 268,

1917-9 (1995).

10. M.H. Nerlhac, et al., Development 111, 815-22 (1996).

11. P. Gonczy, B.J. Thomas, S. DiΝardo, Cell 11, 1015-25 (1994).

12. B.J. Thomas, D.A. Gunning, J. Cho, L. Zipursky, Cell 11, 1003-14 (1994).

13. X. Dong, et al., Genes Dev 11, 94-105 (1997).

14. B.J. Thomas, et al., Genes Dev 11, 1289-98 (1997).

15. L.H. Hartwell, M.B. Kastan, Science 266, 1821-8 (1994).

16. T. Hunter, J. Pines, Cell 79, 573-82 (1994).

17. T.C. Wang, et al., Nature 369, 669-71 (1994).

18. C.J. Sherr, J.M. Roberts, Genes Dev 9, 1149-63 (1995).

19. T. Wolfel, et al., Science 169, 1281-4 (1995).

20. M. Hall, G. Peters, Adv Cancer Res 68, 67-108 (1996).

14 21. C. Norbury, P. Nurse, Annu Rev Biochem 61, 441-70 (1992).

22. K. Nasmyth, Curr Opin Cell Biol 5, 166-79 (1993).

23. D.O. Morgan, Nature 374, 131-4 (1995).

24. E.A. Nigg, Bioessays 17, 471-80 (1995).

25. J.H. Toyn, AL. Johnson, J.D. Donovan, W.M. Toone, L.H. Johnston, Genetics

145, 85-96 (1997).

26. J.L. DeRisi, V.R. Iyer, P.O. Brown, Science 278, 680-686 (1997).

27. L. Wodicka, H. Dong, M. Mittmann, M.-H. Ho, D.J. Lockhard, Nature

Biotechnology 15, 1359-1367 (1997).

28. D.A Lashkari, J.H. McCusker, R.W. Davis, Proc Natl Acad Sci USA 94, 8945-7

(1997).

29. Frozen cell pellets were thawed by vortexing with 5 ml of phenol (USB

Laboratories, Cleveland, OH) and 5 ml of 75 mM NH₄OAc, 10 m MEDTA. Cells were

broken by vortexing vigorously for 5 min with 5 g of 0.5 mm glass beads in 30 ml

polypropylene tubes (Nalgene, Rochester, NY). Extraction with phenol was followed by an extraction with an equal volume of phenolxhloroform. Following ethanol

precipitation of nucleic acid, poly(A)+RNA was purified from total RNA with an

Oligotex dT-column selection step (Qiagen, Chatsworth, CA). Purified poly(A)+

appeared undegraded on an agarose gel. Ribosomal bands were substantially reduced,

and the final poly(A)+ RNA yield was approximately 1.2% of total RNA. Reverse

transcription reactions were performed using 20 μg of poly(l)+ RNA, 2 nmoles of oligo

dT21 primer, 10 mMDTT, 1st Strand Buffer (Gibco Life Technologies, Gaithersburg,

15 MD), 400 M of each dNTP (New England bioLabs), and 4000 units of Superscript II reverse Transcriptase (Gibco). RNA and primer were annealed for 10 min at 65°C, and the reaction was incubated at 42°C for 60 min. For second strand synthesis, the following components were added to reverse transcription reactions: 120 :1 5X 2nd strand buffer (Gibco), 110 nmoles of each dNTP, 12 units RNASE H (Gibco), 160 units

E. coli DNA polymerase I (Gibco), and 40 units E. coli DNA ligase (New England Biolabs, Beverly, MA). The 600 :1 reaction volume was incubated at 16°C for 180 min. 30 units of T4 DNA polymerase (Gibco) were added to each reaction for 5 min at 16°C. Reactions were extracted with an equal volume of phenolxhloroform. Phase-Lock Gel (5 Prime-3 Prime, Inc., Boulder, CO) was used for all organic extractions to increase

DNA recovery and decrease the potential for contamination with material from the organic interface. Double-stranded cDNA was ethanol precipitated and resuspended in 30 :1 of distilled water. cDNA was fragmented to an average length of 50 bp with addition of 3.5 :1 One-Phor-All buffer (Pharmacia Biotech, Piscataway, NJ), 2.2 :1 25 mM CoCl₂ (Boehringer Mannheim, Indianapolis, IN), and 0.15 units DNAse I (Gibco).

Reactions were incubated at 37°C for 5 min and terminated by incubation in a boiling water bath for 15 min. Following boiling, reactions were immediately chilled on ice. DNA fragments were 3' end-labeled by a 2 hr incubation at 37°C after addition of 125 nmoles biotin-N6-ddATP (DuPont NEN, Boston, MA) and 100 units of terminal transferase (Boehringer Mannheim).

30. All Saccharomyces cerevisiae strains used in this study were derived from the W101 genetic background. Strain K3445 contains the c_ϋc28-13 allele and strain K2944

16 contains the cdcl5-2 allele (kindly provided by K. Nasmyth). Strain K3445 was grown overnight in YPAD (yeast extract/adenine/peptone/glucose, Difco Laboratories, Detroit, MI) at 25°C to a density of 8.0 10₆ cells/ml and divided into 50 ml aliquots. All samples were transferred to a 37°C shaking waterbath for 165 min. The cell cycle was reinitiated by moving the flasks to a 25°C shaking waterbath. Every 10 min, one 50 ml sample was

centrifuged for 2 min at 25°C, frozen in liquid nitrogen, and stored at -80°C until use.

An identical time course was carried out for 50 ml cultures of strain K2944, and cells were frozen at -80°C until further use. 31. A set of four oligonucleotide arrays containing a total of more than 260,000

oligonucleotides complementary to 6,218 yeast genes were used for quantitation of

labeled cDNA (Affymetrix). End-labeled and fragmented cDNA was diluted to 200 :1

using a hybridization solution containing 1.0 MNaCl, 10 mM Tris-HCl (pH 7.6) and 0.005% Triton X-100 (referred to as ST-T). In addition, the solutions contained 0.1 mg ml unlabeled, sonicated herring sperm DNA (Promega). cDNA samples were heated

for 99°C for 2 min and cooled to room temperature before being placed in the

hybridization cartridge. Hybridizations were carried out at 42°C for 14 to 16 hr with

mixing on a rotisserie at 60 rpm. Following hybridization, the solutions were removed, the arrays were rinsed with 6XSSPE-T (0.9 M NaCl, 60 mM NaHjPO^ 6 mM EDTA, 0.005% Triton X-100 adjusted to pH 7.6), rinsed with 0.5XSSPE-T (75 mM NaCl, 5

mMNaH^O,,, 0.5 mMEDTA, 0.005% Triton X-100 adjusted to pH 7.6), and incubated

with 0.5XSSPE-T at 42°C for 15 minutes. Following washing, the hybridized

biotinylated DNA was fiuorescently labeled by incubating with 2 :g/ml

17 streptavidin-phycoerythrin (Molecular Probes, Eugene, OR) and 1 mg ml acetylated BSA (Sigma, St. Louis, MO) in 6XSSPE-T at 42°C for 10 min. Unbound streptavidin-phycoeythrin was removed by rinsing at room temperature prior to

scanning. The arrays were read at a resolution of 7.5 :m using especially designed

confocal scanner (Afϊymetrix, Santa Clara, CA) as described previously (27).

32. C. Koch, K. Nasmyth, Curr Opin CellBiol 6, 451-9 (1994).

33. O.M. Aparicio, D.M. Weinstein, S.P. Bell, Cell 91, 59-69 (1997).

34. R.W. King, R.J. Deshaies, J.M. Peters, M.W. Kirschner, Science 274, 1652-9

(1996).

35. R.J. Deshaies, Curr Opin Genet Dev 1, 7-16 (1997).

36. S. Irniger, K. Nasmyth, J Cell Sci, 1523-31 (1997).

37. F. Osaka, H. Seino, T. Seno, F. Yamao, Mol Cell Biol 17 3388-97 (1997).

38. AM. Page, P. Hieter, Cancer Surv 29, 133-50 (1997).

39. R. Verma, et al., Science 278, 455-60 (1997).

40. P.R. Yew, M.W. Kirschner, Science 111, 1672-6 (1997).

41. E. S. Lander, Science 274, 536-9 (1996).

42. K. Nasmyth, Cell 42, 225-35 (1985).

43. L. Breeden, K. Nasmyth, Cell 48, 389-97 (1987)

44. B.J. Andrews, I. Herskowitz, Nature 342, 830-3 (1989).

45. H. Althoefer, A. Schleiffer, K. Wassmann, A. Nordheim, G. Ammerer, Mol Cell

Biol 15, 5917-28 (1995).

18 46. We thank the labs of Frederick Cross and Kim Nasmyth for providing reagents. Helpful comments were provided by Tim Stearns and Jayant Pinto. This work was supported by an NIH Institutional Training Grant in Genome Science.

19

Claims

We Claim:

1. A set of polynucleotide probes, each of said probes comprising at least 12 nucleotides, wherein greater than 50 % of the probes comprise portions of yeast genes which are cell cycle regulated, wherein a cell cycle regulated yeast gene is a gene

whose expression varies by more than 2-fold during the cell cycle of yeast.

2. The set of probes of claim 1 which comprises at least 10 probes.

3. The set of probes of claim 1 which comprises at least 20 probes.

4. The set of probes of claim 1 which comprises at least 30 probes.

5. The set of probes of claim 1 which comprises at least 40 probes.

6. The set of probes of claim 1 which comprises at least 50 probes.

7. The set of probes of claim 1 which comprises at least 100 probes.

8. The set of probes of claim 1 which comprises at least 100 probes.

9. The set of probes of claim 1 which comprises at least 200 probes.

10. The set of probes of claim 1 which comprises at least 300 probes.

11. The set of probes of claim 1 which comprises at least 313 probes.

12. The set of probes of claim 1 which comprises at least 355 probes.

13. The set of probes of claim 1 wherein at least 60 % of said probes are cell

cycle regulated.

14. The set of probes of claim 1 wherein at least 75 % of said probes are cell

cycle regulated.

20

15. The set of probes of claim 1 wherein at least 90 % of said probes are cell cycle regulated.

16. The set of probes of claim 1 wherein at least 95 % of said probes are cell cycle regulated.

17. The set of probes of claim 1 wherein at least 97 % of said probes are cell

cycle regulated.

18. The set of probes of claim 1 wherein at least 99 % of said probes are cell cycle regulated.

19. The set of probes of claim 1 wherein the probes are attached to a solid

support.

20. The set of probes of claim 1 wherein the probes are arrayed on a solid

support in predetermined locations.

21. A set of at least 10 polynucleotide probes, each of said probes comprising at least 12 nucleotides, wherein the probes are arrayed on a solid support in

predetermined locations, wherein greater than 90 % of the probes comprise portions of

yeast genes which are cell cycle regulated, wherein a cell cycle regulated yeast gene is

a gene whose expression varies by more than 2-fold during the cell cycle of yeast.

22. A method of identifying compounds which affect the cell cycle of yeast,

comprising the steps of:

contacting synchronized yeast cells with a test compound;

isolating RNA from the yeast cells;

21 detent-ining the amount of particular mRNA species in the RNA isolated from the yeast cells using the set of polynucleotide probes of claim 1; wherein a test compound which affects the amount of the particular mRNA species is a candidate drug.

23. A method of determining cell cycle stage of a collection of yeast cells, comprising the steps of: isolating RNA from yeast cells; determining the amount of particular mRNA species in the RNA isolated from the yeast cells using the set of polynucleotide probes of claim 1, wherein the relative amounts of the particular mRNA species is characteristic of cell cycle stage of

the yeast.

24. The method of claim 23 wherein the yeast cells are synchronized.

22