Dissecting the regulatory circuitry of a eukaryotic genome

Abstract

A method of identifying cellular regulatory circuits which employ at least one component of a subcomplex of regulatory proteins within the RNA II polymerase holoenzyme which behaves as a signal processor for gene-specific regulators (at least one component of a eukaryotic transcription initiation apparatus) and of determining the set of components of the apparatus which are responsible for regulation of each gene and the set of genes which are coordinately controlled by each transcription factor.

Description

BACKGROUND OF THE INVENTION

[0001] Much of biological regulation occurs at the level of transcription initiation. Genes contain promoter sequences which are bound by transcriptional activators and repressors (Struhl, K. (1995) Annu Rev Genet 29, 651-74; Ptashne, M. and Gann, A. (1997) Nature 386, 569-77). Activators recruit the transcriptional initiation machinery, which for protein-coding genes consists of RNA polymerase II and at least 50 additional components (Orphamides et al. (1996) Genes Dev 10, 2657-83; Roeder, R. G., (1996) Trends Biochem Sci 21, 327-35; Greenblatt, J. (1997) Curr Opin Cell Biol 9, 310-9; Hampsey, M. (1998) Microbiology and Molecular Biology Reviews 62, 465-503; Myer, V. and Young, R. A. (1998) J. Biol. Chem. 273, 27757-27760). The transcriptional initiation machinery includes factors which bind to DNA, cyclin-dependent kinases which regulate polymerase activity, and acetylases and other enzymes which modify chromatin (Burley, S. K., and Roeder, R. G. (1996) Annu Rev Biochem 65, 769-99; Kingston, R. E. et al., Genes and Development 10, 905-20; Roth, S. Y. and Allis, C. D. (1996) Cell 87, 5-8; Sgeger, D. J. and Wovleman, J. L. (1996) Bioessays 18, 875-84, Tsukiyama, T. and Wu, C. (1997) Curr. Opin. Genet. Dev. 7, 182-91; Hengartner C. J. et al., (1998) Genes and Development 9, 897-910).

[0002] The understanding of eukaryotic gene expression remains limited in several ways. The complete set of transcriptional regulators has yet to be identified. How these regulators interact with and regulate components of the transcriptional machinery is not yet clear. The functions of just a fraction of the components of the transcriptional machinery are understood, and then only with respect to a small set of genes. Cells must adjust genome expression to accommodate changes in their environment and in their programs of growth control and development, but precisely how to coordinate remodeling of genome expression is accomplished for signal transduction pathways or for the cell cycle clock has yet to be learned.

SUMMARY OF THE INVENTION

[0003] Described herein are results of genome-wide expression analysis, which was carried out to identify the key components of the transcription initiation machinery in a eukaryote, in order to dissect the regulatory circuitry of the genome. Key components of the transcription initiation machinery (key components of the RNA polymerase II transcriptional machinery) were identified in yeast, as described herein. Assessment of the requirement for key components was carried out using high density oligonucleotide arrays (HDAs) (Wodicka, L. et al (1997) Nat. Biotech., 15, 1359-67) to determine the genome-wide effects of mutations in components of the transcriptional machinery. At any given promoter, the transcriptional machinery might include any or all of the following, among others: the RNA polymerase II core enzyme, the general transcription factors (GTFs), the core Srb/mediator complex, the Srb10 CDK complex, the Swi/Snf complex and the SAGA complex. The components of the transcription apparatus which were the focus of this study were selected because they are among the key subunits of the major multiprotein complexes which have roles in transcription of protein-coding genes. One or more subunits of each of these components has been investigated for its role in genome-wide gene expression through the use of mutations which affect either the function or the physical presence of the subunit.

[0004] Results showed that components of the RNA polymerase II holoenzyme, the general transcription factor TFIID and the SAGA chromatin modification complex have roles in expression of distinct sets of genes. They further showed that the Rpb1 subunit of core RNA polymerase II, the Srb4 subunit of the Srb/mediator complex and the Kin28 subunit of TFIIH are generally required for transcription of protein-coding genes. Two were found to be required for more than half, but not all, genes (Tfa1, Taf17). Most components investigated thus far were necessary for transcription of less than a fifth of the genome (Srb5, Med6, Srb10, Swi2, Taf145, Gcn5). In this latter group, the evidence indicates that Srb5, Med6, and Taf145 have predominantly positive roles, Srb10 has an almost exclusively negative role, and Swi2 and Gcn5 can have either a positive or a negative role in gene expression.

[0005] Work described herein shows that distinct sets of genes require the function of distinct components of the transcription machinery. Thus, coordinate regulation of large sets of genes can be accomplished by affecting the function of specific components of the transcription machinery. It follows that functional relationships exist among some genes within the sets of genes whose regulation is accomplished in this manner. Results described herein also revealed an unanticipated level of regulation that is available to the cell in addition to that provided by gene-specific regulators; the expression of specific sets of genes can be regulated by affecting the availability or function of a specific component of the general machinery. Results also showed a novel mechanism for co-ordinate regulation of specific sets of genes when cells encounter nutrient deprivation or limitation and evidence that the ultimate targets of signal transduction pathways can be identified within the initiation apparatus.

[0006] In one embodiment, the present invention is a method of determining regulatory interrelationships among genes in a cell. The method comprises the steps of:

[0007] (a) hybridizing a transcription indicator of a test cell to a set of nucleic acid probes;

[0008] (b) hybridizing a transcription indicator of a control cell to the set of nucleic acid probes,

[0009] wherein the transcription indicators are selected from the group consisting of mRNA, cDNA and cRNA, wherein the test cell contains a mutant component of the general transcription machinery and the control cell is the wild-type isogenic counterpart of the test cell;

[0010] (c) detecting amounts of the transcription indicators which hybridize to each of said set of nucleic acid probes; and

[0011] (d) identifying a gene as a member of the regulatory pathway of the general transcription factor if hybridization of the transcription indicator of the test cell to a probe comprising a portion of the gene is higher or lower than hybridization using a transcription indicator from the control cell.

[0012] In various embodiments of the method, the difference in hybridization between the control and the test cell varies. There can be, for example, at least a 2-fold difference in hybridization between the control and the test cell, at least a 3-fold difference, at least a 5-fold difference or at least a 10-fold difference in hybridization between the control and the test cell. In various embodiments of the method, the mutant component of the general transcription machinery is a mutual general transcription factor, such as a temperature sensitive mutant, a point mutant or a deletion mutant. The mutant component of the general transcription machinery can be, for example, a component of RNA polymerase II holoenzyme. The mutant component of the general transcription machinery can be a component necessary to reconstitute promoter-dependent transcription in vitro with core RNA polymerase II. Also the subject of this invention is a pair of isogenic eukaryotic cells which comprises a test cell which contains a mutant component of the general transcription machinery and a control cell which is the wild-type isogenic counterpart of the test cell. Such pairs can include a test cell in which the mutant component of the general transcription machinery is a mutant general transcription factor. They also can include a test cell in which the mutant component of the general transcription machinery is a temperature sensitive mutant, a point mutant or a deletion mutant. In such pairs, the mutant component of the general transcription machinery can be a component of RNA polymerase II holenzyme; the mutant component of the general transcription machinery can be one which is necessary to reconstitute promoter-dependent transcription in vitro with core RNA polymerase II.

[0013] The invention further relates to a method of studying the effects of drugs on cells. The method comprises:

[0014] (a) contacting a cell with a drug; and

[0015] (b) determining the effect of the drug on the cell by assessing expression of one or more of the genes which are determined to be members of the regulatory pathway of the general transcription factor according to methods described herein.

[0016] A further embodiment of the invention is a method of identifying a cellular regulatory circuit which employs a component of a subcomplex of regulatory proteins within the RNA polymerase II holoenzyme, referred to as the transcription initiation apparatus.

[0017] The method comprises:

[0018] (a) comparing genome expression signature during cellular responses to environmental or other stimuli with the genome expression signature produced by a defect in the transcription initiation apparatus; and

[0019] (b) determining differences between the two genome expression signatures and relating the differences to the defect in the transcription initiation apparatus, thereby identifying a component of the transcription initiation apparatus which is responsible for regulation of genes in the cells.

[0020] In various embodiments the cellular regulatory circuit is a yeast cell regulatory circuit, a primate (e.g., human) or other vertebrate cell regulatory circuit or a non-vertebrate cell regulatory circuit.

[0021] Thus, genome-wide expression analysis provides insights into the transcriptional regulatory circuitry of eukaryotic cells, as well as the foundation and context for interpreting mechanistic studies in control of gene expression.

BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The file of this patent contains at least one drawing executed in color. Copies of the patent with color drawing(s) will be provided by the Patent and Trademark Office upon request and payment of the necessary fee.

[0023] FIGS. 1A-1B show genes which go up in Srb5 mutants.

[0024] FIGS. 2A-2B show genes which go up in Srb5 mutants.

[0025] FIGS. 3A-3U show genes which go down in Srb5 mutants.

[0026] FIGS. 4A-4U show genes which go down in Srb5 mutants.

[0027] FIGS. 5A-5E show genes which go up in Sin4 mutants.

[0028] FIGS. 6A-6E show genes which go up in Sin4 mutants.

[0029] FIGS. 7A-7C show genes which go down in Sin4 mutants.

[0030] FIGS. 8A-8C show genes which go down in Sin4 mutants.

[0031] FIGS. 9A-9C show genes which go up in Gcn5 mutants.

[0032] FIGS. 10A-10C show genes which go up in Gcn5 mutants.

[0033] FIGS. 11A-11F show genes which go down in Gcn5 mutants.

[0034] FIGS. 12A-12F show genes which go down in Gcn5 mutants.

[0035] FIGS. 13A-13C show genes which go up in Srb2 mutants.

[0036] FIGS. 14A-14C show genes which go up in Srb2 mutants.

[0037] FIGS. 15A-15F show genes which go down in Srb2 mutants.

[0038] FIGS. 16A-16F show genes which go down in Srb2 mutants.

[0039] FIGS. 17A-17F show genes which go up in Swi2 mutants.

[0040] FIGS. 18A-18F show genes which go up in Swi2 mutants.

[0041] FIGS. 19A-19D show genes which go down in Swi2 mutants.

[0042] FIGS. 20A-20D show genes which go down in Swi2 mutants.

[0043] FIGS. 21A-21B show genes which go up in TAF145 (45 min 37 deg) mutants.

[0044] FIGS. 22A-22P show genes which go down in TAF 145 (45 min 37 deg) mutants.

[0045] FIGS. 23A-23E show genes which go up in Srb10 mutants.

[0046] FIGS. 24A-24E show genes which go up in Srb10 mutants.

[0047] FIGS. 25A-25B show genes which go down in Gal11 mutants.

[0048] FIGS. 26A-26B show genes which go down in Gal11 mutants.

[0049] FIGS. 27A-27C show genes which go up in Gall 1 mutants. FIGS. 28A-28C show genes which go up in Gal11 mutants.

[0050] FIG. 29 shows genes which go up in Med6 mutants.

[0051] FIGS. 30A-30E show genes which go down in Med6 mutants.

[0052] FIGS. 31A-31E show genes which go down in Med6 mutants.

[0053] FIGS. 32A-32E show genes which go down in Med6 mutants.

[0054] FIGS. 33A-33F show genes which are affected in Srb10 mutants; all of these genes go up and the list is in rank order of degree affected. See especially column headed Gene and column headed Fold up.

[0055] FIGS. 34A-34L show genes which are affected in SWI2 mutants; those in

[0056] FIGS. 34A-34F go up and those in FIGS. 34G-34L go down. See especially column headed Fold up and column headed Fold down.

[0057] FIGS. 35A-35E show genes which go down in TAF145 mutants. See especially columns headed Gene and % of WT expression.

[0058] FIG. 36 is a schematic representation of a model of RNA polymerase II transcription initiation machinery which, as depicted here, encompasses more than 85 polypeptides in 10 (sub) complexes: core RNA polymerase II (RNAPII) consists of 12 subunits; TFIIH, 9 subunits; TFIIE, 2 subunits; TFDIIF, 3 subunits; TFIID, 14 subunits; core SRB/mediator, more than 16 subunits; Swi/Snf complex, 11 subunits; Srb10 kinase complex, 4 subunits and SAGA, 13 subunits (see http://www.wi.mit.edu/young/expression.html site for more details). As detailed in Table 1, representative subunits of these complexes were chosen for analysis of genome-wide transcription dependence.

[0059] FIGS. 37A-37D show genome-wide expression data for selected components of the RNA polymerase II holoenzyme; data reflecting the change in mRNA levels when a mutant is compared to its isogenic wild type counterpart is presented in a grid format. In the grid, the upper left grid square represents the left-most gene on chromosome I, and the squares to its right represent adjacent genes, proceeding in a linear fashion through chromosome I, then II, the III, etc., until the last gene on the right arm of chromosome XVI is reached at the bottom of the grid.

[0060] FIG. 37A shows results for Rpb1; 37B shows results for Med6; 37C shows results for Srb10; and 37D shows results for Swi2.

[0061] FIGS. 38A and 38B are Venn diagrams illustrating the genome-wide dependence on key components of the transcription machinery. FIG. 38A illustrates that RNA polymerase II holoenzyme components show distinct patterns of genome control. It is a Venn diagram depicting Srb5-, Swi2-, Srb10- and Med6-dependent genes (small circles) in relation to the whole transcriptome (Rpb 1-, Srb4- and Kin28 dependent, large circle). The numbers under each subunit name are the sum of genes whose expression depends on that subunit. FIG. 38B illustrates genome control patterns of components of TFIID and SAGA.

[0062] FIGS. 39A and 39B present results showing that Srb5 is required for expression of pheromone response genes. FIG. 39A shows the pheromone response genes whose expression is reduced in the absence of Srb5. FIG. 39B is a graph showing that cells lacking Srb5 are defective in mating. The mating efficiencies for mutant strains are expressed as a percentage of the mating efficiency of an isogenic wildtype strain. For comparison, strains with mutations in two components of the mating signal transduction pathway (FUS3 and STE2) are included.

[0063] FIGS. 40A-40C present results showing that Srb10 CDK represses genes elevated during response to nutrient starvation. FIG. 40A is a subset of 173 genes whose expression is depressed in cells lacking Srb10 kinase activity. FIG. 40B is a Venn diagram showing the number of genes which are depressed during the nutrient deprivation which occurs during the diauxic shift and the fraction of these which are depressed in cells lacking Srb10 kinase activity. FIG. 40C is a graph which shows that Srb10 protein is depleted from cells as they enter the diauxic shift. The graph shows the growth curve of a yeast strain allowed to grow to stationary phase (33 hours).

DETAILED DESCRIPTION OF THE INVENTION

[0064] Detailed information and databases supporting all aspects of the work described herein, including experimental procedures, genetic reagents, HDA technology and data analysis can be found on the Internet at http://www.wi.mit.edu/young/expression.html. The entire contents of this Web site are incorporated herein by reference.

[0065] As described herein, HDAs were used to determine the effects of mutations on key components of the RNA polymerase II transcriptional machinery genome-wide in eukaryotic cells and, as a result, to assess the requirements for these components. As described in the examples which follow, the levels of all detectable mRNA species in yeast were determined using HDAs. Results showed that transcripts from 80% of expressed yeast genes exist at steady state levels of 0.1 to 2 molecules/cell.

[0066] Dependence of genome expression on key components of the transcriptional machinery was assessed, using mutations which affect either the function or the physical presence of one or more subunits of machinery components (RNA polymerase II core enzyme, the general transcriptional factors (GTFs), the core Srb/mediator complex, the Srb10 CDK complex, the Swi/Snf complex and the SAGA complex). Specifically described in the examples is work which resulted in determination of the levels of all detectable mRNA species in yeast, which is useful in evaluating the degree to which these levels depend on any one component of the transcription apparatus. Also described in the examples is assessment of the roles of components of the transcriptional machinery in genome-wide gene expression, using yeast as a eukaryotic model. As described, one or more subunits of the RNA polymerase II core enzyme, the general transcription factors (GTFs), the core Srb/mediator complex, the Srb10 CDK complex, the Swi/Snf complex and the SAGA complex have been investigated for their roles in genome-wide expression. This was carried out through the use of mutations which affect either the function or the physical presence of the subunit being assessed (see Table 1). The work described herein was carried out using yeast, but a similar approach (in which mutations which affect either the function or physical presence of one or more subunits of the transcription initiation machinery are used to assess dependence of genome expression on machinery components) can be used in other eukaryotic cells, including cells from vertebrates (e.g., cells of human and other primate origin, murine, canine, feline and bovine origin) and cells from non-vertebrates (e.g., cells from worms and flies).

[0067] Results showed that the Rpb1 subunit of core RNA polymerase II, the Srb4 subunit of the Srb/mediator complex and the Kin28 subunit of the general transcription factor TFIIH are generally required for transcription of protein-coding genes. Results also showed that only a subset of genes is dependent on Med6, Srb5, Srb10, Swi2, TAFII45, TAFII17 and Gcn5. The sets of genes whose expression requires various RNA polymerase II holoenzyme components are compared in the Venn diagram of FIG. 38A. The sets of genes whose expression requires various TFIID and SAGA components are shown in the Venn diagram of FIG. 38B. Together, these diagrams show how distinct sets of genes require the function of distinct components of the transcription machinery.

[0068] Thus, coordinate regulation of large sets of genes can be accomplished by affecting the function of specific components of the transcriptional machinery. For example, FIG. 38A shows, in addition to the three key RNAPII holoenzyme components which are generally required for transcription of protein-coding genes, at least four other components which regulate expression of subsets of genes. Specifically, Srb5 regulates expression of 698 genes; Med6 regulates expression of 506 genes; Swi2 regulates expression of 329 genes and Srb10 regulates expression of 173 genes. Coordinate regulation of genes in each of these sets of genes can be effected by altering (enhancing or reducing/repressing) function or activity of the respective regulating components (e.g., Srb5, Med6, Swi2, Srb10). Even broader coordination can be accomplished by altering (enhancing or reducing/repressing) function of one or more of the three RNAPII holoenzyme components (Rpb1, Srb4, Kin28) shown to be generally required for transcription of protein-coding genes. It is interesting to note the “overlap” of genes regulated; that is, the fact that expression of some genes is regulated by two components (e.g., 62 genes are regulated by Srb5 and Med6; 86 by Srb5 and Swi2; 12 by Srb5 and Srb10; 30 by Srb10 and Swi2 and 28 by Swi2 and Med6). FIG. 38B presents comparable information for TFIID and SAGA components. In addition to Rpb1, Srb4 and Kin28, other components regulate a large number of genes: TAFII17, 3180 genes; TAFII145, 766 genes; and Gcn5, 268 genes. Here, too, overlap is evident: TAFII17 and TAFII145 regulate 562 genes in common; TAFII17 and Gcn5 regulate 50 genes in common; TAFII145 and Gcn5 regulate 7 genes in common and the three regulate 15 genes in common.

[0069] The following is a summary of results described in greater detail herein.

[0070] General Factors

[0071] Rpb1 and Srb4 proteins are generally required for expression of protein-coding genes, and they are both associated tightly and exclusively with RNA polymerase II and the mediator complex, respectively (Koleske, A. J. and Young, R. A. (1994) Nature 368, 29970-7; Kim, Y. J. et al., (1994) Cell 77, 599-608; Myers, L. C. et al., Genes Dev 12, 45-54). Therefore, it is reasonable to infer that RNA polymerase II and the core mediator complex are generally required for transcription. Assuming that the function of Kin28 is restricted to TFIIH, the data obtained with the Kin28 mutant demonstrates that TFIIH is a general factor. The expression of 54% of yeast genes is as dependent on Tfa1 as it is on Rpb1, supporting the idea that TFIE is directly involved in expression of at least 54% of protein-coding genes. Without knowing the contribution of Tfa2, the other subunit of TFIE, one cannot eliminate the possibility that TFIIE has roles at additional genes.

[0072] SRB/Mediator Complex

[0073] The SRB/mediator core complex is essential for general transcription, as evidenced by the requirement for Srb4, but components such as Srb5 and Med6 have roles at specific subsets of genes. These results are consistent with the proposal that the Srb/mediator complex is recruited to promoters of most genes together with RNA polymerase II, where it acts in a manner analogous to a signal processor with the capacity to integrate the combinatorial effects of multiple inputs from gene-specific transcriptional activators and repressors. (Koleske, A. J. and Young, R. A. (1994) Nature 368, 466-9; Kim, Y. J. et al., (1994) Cell 77, 599-608; Koh, S. S. et al., (1998) Cell 1, 895-904; Myers, L. C. et al., Genes Dev 12, 45-54; Sun, X. et al., (1998) Molecular Cell 2,1-11)

[0074] Srb10 CDK Complex

[0075] The function of the Srb10 CDK complex can be defined by the kinase itself, since loss-of-function mutations in any of the four components of this complex produce identical phenotypes (Hengartner, C. J. et al (1995) Genes and Development 9, 897-910). The Srb10 kinase is a negative regulator of a substantial fraction of genes which are repressed when cells grow vegetatively in rich media and are induced as cells experience nutrient deprivation. The genes regulated by Srb10 include those which are critical for the morphological change which permits foraging for nutrients and stress responses. Srb10 isphysically depleted from cells as they enter the diauxic shift, providing a mechanism for derepression of this set of genes. Srb10 in wild type cells is, thus, responsible for repressing this set of genes when cells are in exponential growth on glucose, but no longer performs this function as cells enter the diauxic shift.

[0076] Swi/Snf Complex

[0077] If the function of the Swi/Snf complex is ATP-dependent remodeling of chromatin, (Laurent, B. C. et al., (1993) Genes Development 7, 583-91; Cote, J. et al., (1994) Science 265, 53-60), then the effects observed herein due to the Swi2 ATPase mutation should represent the dependence of genome-wide expression on the entire Swi/Snf complex. The results indicate that a greater number of genes is negatively regulated by Swi/Snf than is positively regulated. This is surprising in view of the model that Swi/Snf-catalyzed remodeling of chromatin facilitates activator binding. It is possible that chromatin remodeling may facilitate binding of negative factors as well as positive factors. An alternative possibility is suggested by recent data indicating that the Swi/Snf complex can remodel chromatin in both directions: it can convert a repressive nucleosome structure towards a more accessible state and vice versa (Schnitzler, G. et al (1998) Cell 94, 17-27). It is thus possible that Swi/Snf helps produce a nucleosome structure conducive to transcription at some promoters, and a structure which is repressive at others.

[0078] TFIID and SAGA

[0079] The general transcription factor TFIID and the SAGA complex share two features: they both contain a subunit capable of histone acetylation (TAFII145 in the case of TFIID and Gcn5 in the case of SAGA) and they share multiple subunits, among which is the histone H3-like TAF, TAFII17 (Grant, P. A. et al (1998) Cell 94, 45-53). As summarized in FIGS. 39A-39B, the results indicate that Gcn5, TAFII145 and TAFII17 are necessary for expression of 5%, 16% and 67% of yeast genes, respectively. Two models can account for this data: one posits that TAFII17 functions exclusively within the TFIID and SAGA complexes, and the other that TAFII17 is a component of one or more additional complexes. If TAFII17 functions exclusively within the TFIID and SAGA, then TAFII145 and Gcn5 do not fully represent the functions of the two complexes, since the sum of genes which require TAFII145 and Gcn5 function is much smaller than the number of genes which require TAFII17. In this model, one or both complexes contain subunits which make different contributions to gene expression, as might be expected if different subunits are targets of different transcriptional activators and repressors. The results can also be accommodated in a second model, in which TAFII17 is a component of one or more complexes in addition to TFIID and SAGA. The results described here lay a useful foundation for the additional experiments necessary to gain a fuller understanding of the roles of TFIID and SAGA subunits in gene expression.

[0080] The data presented herein, in conjunction with that of previous studies, reveal several striking similarities between TAFII145 and prokaryotic sigma factors. First, both sigma factors and TAFII145 are components of the general transcription machinery. Second, many sigma factors are required for the expression of a related subset of genes; similarly, it has been shown that TAFII145 appears to be required for expression of a set of genes involved in chromosomal synthesis and G1/S progression. Finally, both sigma factors and TAFII145 act through core promoter elements by direct DNA contacts.

[0081] An unexpected finding of the work described herein is the role Srb5 has in pheromone response. It was striking that many of the genes whose mRNA levels are most dramatically affected by the loss of Srb5 fall into the pheromone response pathway. The 15 genes involved in the pheromone response which are expressed at substantially lower levels in the absence of Srb5 are shown in FIG. 39A. Dramatic effects are seen in genes involved in mating factor production and export; the expression of MFA1 and MFA2, the two genes encoding mating pheromone a-factor, are down 28-fold and 11-fold, respectively. Additional genes involved in maturation (STE13) and export (STE6) of mating factor are expressed at substantially lower levels than in the cognate wild type. Furthermore, several components of the signal transduction pathway that responds to mating pheromone are expressed at reduced levels in the Srb5 mutant. These genes include the receptor for pheromone (STE2), subunits of the signaling G-protein (GPA1), and the transcription factor which is itself the target of the signaling response and directly regulates subsequent gene expression (STE12).

[0082] The genome-wide expression profile for the Srb5 mutant suggests that these cells should exhibit a defect in mating efficiency, a phenotype which was not previously suspected or investigated. Indeed, quantitative mating assays show that Srb5 mutant does have a significant defect in mating (FIG. 39B). The mating defect was more pronounced than that due to mutations in Fus3, a MAP kinase required for cell cycle arrest and cell fusion during mating, but less pronounced than that due to mutations in STE12. The defect in mating deficiency exhibited by the Srb5 mutant may reflect coordinate regulation of the set of pheromone response genes identified through genome-wide expression analysis.

[0083] The present invention is illustrated by the following examples, which are not intended to be limiting in any way. Detailed information and databases supporting all aspects of this study can be found on the Internet at http://www.wi.mit.edu/young/expression.html. The entire content of this web site, including the content of all linked sites (e.g., hypertext) is expressly incorporated herein by reference in its entirety.

EXAMPLE 1

Technology, Genetic Reagents, Experimental Protocols, and Analysis Technology

[0084] Affymetrix GeneChip high-density oligonucleotide arrays (HDAs) were used in this study. The yeast genome HDAs are described in detail in Wodicka, L. et al., Nature Biotechnology 15:1359-1367 (1997)). The arrays can detect as few as 0.1 mRNA molecules/cell; the dynamic range over which detection is accurate is approximately 0.1-100 mRNA molecules/cell.

[0085] With the Genechip arrays, the entire yeast genome is covered by four HDAs. In total, 6181 ORFs are present within this set. Each gene is represented on the HDA by 20 25-mer oligos that match the sequence of the message (perfect match oligos) and 20 oligos that are identical but differ by one base (mismatch oligos). Expression levels are calculated by subtracting the signal of a mismatch from its perfect match partner and averaging the difference for each oligo pair for a given gene. The average difference value is a measure of the expression level of that gene. Based on various criteria (e.g. consistent behavior of each oligo pair) a Present or Absent call is also returned (see information provided by Affymetrix and also by Wodicka, L. et al., Nature Biotechnology 15:1359-1367 (1997)).

[0086] Genetic Reagents

[0087] Two types of mutations have proven to be useful for determining which genes depend on the function of a component of the transcription apparatus. For essential components of the apparatus (e.g. Rpb1), temperature-sensitive mutations are valuable because they permit conditional growth of cells and they allow the investigator to examine effects on gene expression at any point after factor inactivation. For non-essential components, we have used either point mutations which knock out the catalytic function of known enzymatic activities (e.g. Srb10), or complete deletion mutations (e.g. SRB5).

[0088] Strain List 1 Subunit Type Strain Genotype GCN5 Wildtype FY86 MAT&agr;, his3&Dgr;200, ura3-52, leu2&Dgr;1 Mutant FY1370 MAT&agr;, his3&Dgr;200, ura3-52, leu2&Dgr;1, gen5&Dgr;::HIS3 KIN28 Wildtype GF1047 MAT&agr;, ura3, trp1, leu2, lys2 Mutant GF2092 MAT&agr;, ura3, trp1, leu2, his3, kin28-ts3 MED6 Wildtype YCL4 MAT&agr;, ura3-52, leu2&Dgr;1, his3&Dgr;200, trp1&Dgr;63, ade2-101, lys2-801, med6&Dgr;::LEU2 (AMP, URA3, CEN, MED6) Mutant YCL8 MAT&agr;, ura3-52, leu2&Dgr;1, his3&Dgr;200, trp1&Dgr;63, ade2-101, lys2-801, med6&Dgr;::LEU2 (AMP, HIS3, CEN, med6-ts) RPB1 Wildtype Z579 MATa, ura3-52, leu2-3, 112, his3&Dgr;200, srb4&Dgr;2::HIS3 [RY2884(AMP, CEN, LEU2, SRB4+)] Mutant Z460 MAT&agr;, ura3-52, leu2-3, 112, his3&Dgr;200, his4-912, lys2-128, rpb1&Dgr;187::HIS3 [pRP1-1U(AMP, CEN, LEU2, rpb1-1)] SRB4 Wildtype Z579 MATa, ura3-52, leu2-3, 112, his3&Dgr;200, srb4&Dgr;2::HIS3 [RY2844 (AMP, CEN, LEU2, SRB4+)] Mutant Z628 MATa, ura3-52, leu2-3, 112, his3&Dgr;200, srb4&Dgr;2::HIS3 [RY2882 (AMP, CEN, LEU2 srb4-138)] SRB5 Wildtype Z579 MATa, ura3-52, leu2-3, 112, his3&Dgr;200, srb4&Dgr;2::HIS3 [RY2844 (AMP, CEN, LEU2, SRB4+)] Mutant Z651 MATa, ura3-52, leu2-3, 112, his3&Dgr;200, srb5&Dgr;1::hisG::URA3::his G SRB10 Wildtype Z579 MATa, ura3-52, leu2-3, 112, his3&Dgr;200, srb4&Dgr;2::HIS3 [RY2844 (AMP, CEN, LEU2, SRB4+)] Mutant Z690 MATa, ura3-52, leu2-3, 112, his3&Dgr;200, srb10-3::hisG SWI2 Wildtype Cy397 MAT&agr;, leu2&Dgr;1, his3&Dgr;200, lys2-801, ade2-101, HO-lacZ::HO, swi2&Dgr;::HIS3, SWI2::URA3 Mutant Cy396 MAT&agr;, leu2&Dgr;1, his3&Dgr;200, lys2-801, ade2-101, HO-lacZ::HO, swi2&Dgr;::HIS3, swi2(K798A)::URA3 TAF145 Wildtype YSW87 MATa, ade2-101, his3&Dgr;200, leu2&Dgr;1, lys2-801, ura3&Dgr;99, GAL2, GAL3, taf145::LEU2 [pSW104(AMP, CEN, TRP1, TAF145)] Mutant YSW93 MATa, ade2-101, his3&Dgr;200, leu2&Dgr;1, lys2-801, ura3&Dgr;99, GAL2, GAL3, taf145::LEU2 [AMP, CEN, HIS3, taf145-ts2)] TAF17 Wildtype LY740 MATa, ade2-1, his3-11, 15, leu2-3, 112, trp1-1, ura3-1, can1-100, taf17::LEU2 [Lp20 (CEN, HIS3, TAF17)] Mutant LY761 MATa, ade2-1, his3-11, 15, leu2-3, 112, trp1-1, ura3-1, taf17::LEU2 [CEN, HIS3, taf17ts-2] TFA1 Wildtype Z888 MAT&agr;, ade2-1, leu2-3, 112, trp1-1, ura3-1, can1-100, tfa1::ADE2 [pSK492(AMP, CEN, TRP1, HA-TFA1)] Mutant Z889 MAT&agr;, ade2-1, leu2-3, 112, trp1-1, ura3-1, can1-100, tfa1::ADE2 [ptfa1-21(AMP, CEN, TRP1 HA-tfa1-21)]

[0089] Experimental Protocols

[0090] Two independent experiments were performed for each wild type versus mutant comparison. Individual mRNA levels were scored if the computer algorithm analyzing the scanned results (Wodicka, L. et al., Nature Biotechnology 15:1359-1367 (1997)) returned a “Present” call in both the two wild type and the two mutant expression profiles for that gene or if the expression levels of that gene changed in the same direction and were greater than background levels in both wild type and mutant comparisons. A decrease was called if an mRNA dropped more than two-fold in both comparisons.

[0091] Expression profiles were determined by growing yeast cultures to mid-log phase, isolating total RNA, spiking in control RNA for normalization and isolating poly-A RNA. This was used to generate double stranded complementary DNA (dscDNA) that in turn was used to generate biotin labeled copy RNA (cRNA; the oligo used for dscDNA synthesis contains a T7 RNA polymerase promoter). 1-3 mg of polyA RNA thus resulted in synthesis of approximately 60 mg cRNA. This was fragmented and hybridized to the oligonucleotide arrays, arrays were washed, stained, washed and scanned.

[0092] The protocols included here are set up to monitor transcription factor mutants that result in significant loss of mRNA (e.g. rpb1-1 that results in a 5-fold reduction of mRNA 45 minutes after temperature shift). For mutants that result in less severe loss of mRNA the culture size and RNA preparations can be significantly scaled down and normalization can be performed through assuming that the bulk RNA undergoes no change. The protocols describe how cells were grown and harvested, how total RNA and polyA RNA were prepared. All subsequent steps were carried out exactly as described in the Affymetrix product information supplied with the Genechips (the leaflets are part numbers 700187 Rev 1 and 700163 Rev 1 5/98).

[0093] Protocol 1: Cell Growth and Harvesting

[0094] 1. Streak out strains from permanent stock. Wait 2-3 days until colonies form. (In case of slow-growing mutants, streak out mutant strains first, then streak out wild-type strains a day or so later.) Always start cultures from freshly streaked plates to avoid revertants.

[0095] 2. Pick 3 or 4 colonies from each strain to start 10 ml cultures. Grow up overnight at 30° C. If mutant strains grows considerably slower, start mutant strain culture first (morning) and wild-type strain in the evening.

[0096] 3. Determine the OD600 of all cultures. The set of 3-4 cultures from each individual strain should result in similar OD600's. If not, there is a problem with this strain: test phenotypes. If the small cultures result in consistent OD600's then proceed. In morning dilute 2 of the wild-type and 2 of the mutant 10 ml cultures into >200 ml of pre-warmed YPD (and pre-warmed flask) to an OD600 of 0.1 or 0.05. If the mutant strain grows considerably slower it is advisable to start the wild-type strain at a 2×lower density. Be sure all of the media used was made the same day. Incubate at 30° C. for 4-8 hr until the OD600˜0.5. (N.B. pairwise comparisons of expression profiles of identical strains grown to even only marginally different OD's show significant differences.)

[0097] For ts Strains Only

[0098] 4. Dilute 1:2 in 200 ml of pre-warmed YPD so that the temperature-shift temperature is correct (e.g., shifting to 37° C., pre-warn YPD to 44° C. and prewarm final culture flasks and measuring flasks to 37° C.). Use 500 ml flask to measure out correct volumes, then pour into 2 L flask and incubate at the appropriate temperature. Heat shock for exactly 45 minutes. Stagger heat shocking the individual cultures by at least 20 min.

[0099] For All Strains

[0100] 5. Harvesting must take place as quickly as possible and identically for all strains whose expression profiles are to be compared. Differences in harvesting (pH water, length of time in centrifuge etc) will be reflected in differences in the expression profiles. For each culture: pour into 250 ml disposable centrifuge tubes (pre-cooled on ice). Remember to take an aliquot for determination of the OD600 before centrifugation. Centrifuge in GSA rotor 3 min at 3500 rpm (make sure adaptors are in place and rotor is precooled). Write down final volume and OD600 of cells.

[0101] 6. Immediately decant supernatant into sink. If not performed at once the pellet will not hold together (do not need to remove all liquid).

[0102] 7. Resuspend in 6 ml of ice cold ddH2O by pipetting up and down, transfer to next tube and continue. Pipette into 15 ml Falcon tube (pre-cooled on ice). Centrifuge in table top centrifuge 2.5 minutes at 2600 rpm.

[0103] 8. Aspirate off supernatant. Place in liquid nitrogen container (for at least 2 minutes). Store at −80° C. Steps 5-8 will take approximately 15 min depending on the acceleration and braking rates of the centrifuges. It is not advisable to harvest more than two cultures at a time.

[0104] Protocol 2: Total RNA Preparation

[0105] 1. Take cells out of-80° C. freezer and add 3 ml of acid phenol-chloroform-isoamylalcohol (125:24:1, pH 4.7; Sigma P-1 944), pre-warmed for 10 minutes at 65° C. and 3 ml of TES (10 mM Tris pH 7.5, 10 mM EDTA, 0.5% SDS; autoclaved) per 200 OD600 units to the RNA. For 50 OD600 units add 0.7 ml of Phenol-Chloroform+0.7 ml of TES.

[0106] 2. Vortex tube at highest setting at ˜20° angle, to resuspend pellet.

[0107] 3. Incubate at 65° C. in waterbath for 1 hr. Incubate in beaker filled with hot water (move beaker to vortexer, etc.). Vortex 20 seconds (10 seconds upright, 10 seconds at 20° angle) every 10 minutes.

[0108] 4. Vortex 20 seconds, then aliquot into 4×1.5 ml Eppendorf tubes. Spin 20 minutes at 14,000 rpm at 4° C.

[0109] 5. Extract with 750 &mgr;l of acid phenol-chloroform-isoamylalcohol per tube. Vortex 20 seconds (10 seconds up, 10 seconds angle). Spin 10 minutes at 4° C.

[0110] 6. Extract with 24:1 chloroform:Isoamyl alcohol (Sigma C-0549). Vortex 20 seconds, spin ˜10 minutes room temperature.

[0111] 7. Transfer aqueous phase to tube with 50 &mgr;l of 3M Sodium Acetate (NaOAc) pH 5.2. Add 1 ml of 100% EtOH (pre-cooled to −20° C.), fill to top of tube. Store at −20° C. for longer than ½ hr.

[0112] 8. Spin down RNA for 5 minutes in a microcentrifuge at room temperature. Aspirate (leave a little left).

[0113] 9. Wash with 500 &mgr;l of 80% ETOH (temperature ˜−20° C.), no shaking, just add. Spin down for 1 minute.

[0114] 10. Aspirate, remove last bit with a pipette. Let air dry 1 minute.

[0115] 11. Add DEPC treated water to samples to resuspend. To calculate amount to add: expect 20 &mgr;g RNA for a heat-shocked culture or 30 &mgr;g RNA for a 30° C. grown culture per unit OD600.

[0116] 12. Measure OD260/OD280 accurately (e.g. by measuring 5 &mgr;l of a 1:20 dilution).

[0117] 13. Make aliquots, snap-freeze and store at −80° C.

[0118] Resuspension of Samples:

[0119] Add enough DEPC treated ddH2O to get final concentration ˜15 mg/ml.

[0120] Add majority to tubes, rinse sides of tube by pipetting, resuspend by pipetting (try to transfer clumps of RNA to single tube then resuspend).

[0121] Vortex briefly

[0122] Wash other tubes with remaining DEPC-treated ddH2O and combine.

[0123] Pipette up and down until no particles are left (20-40 times).

[0124] Vortex (speed 3-4) 10 seconds (centrifuge to collect at bottom of tube), and repeat pipetting up and down 20-40 times.

[0125] Protocol 3: Total mRNA Preparation

[0126] 1. For strains in which you do not expect a general loss of mRNA use 1 mg of RNA (total). If you expect a general loss of mRNA (srb4 ts, rpb 1 ts) use 2 or 3 mg total RNA.

[0127] 2. Measure OD260/OD280 of all total RNA preps accurately and in parallel (e.g. by measuring 5 &mgr;l of a 5:100 dilution). Accurate determination of the amount of total RNA is essential for correct normalization of results when normalizing with the spiked controls (see polyA controls, below).

[0128] 3. Place total RNA with 5 &mgr;l of poly-A controls per mg total RNA into each tube.

[0129] 4. For 1 mg total RNA: add ODB buffer (Qiagen Oligotex kit) to make 920 &mgr;l total volume. Note: Be sure ODB+OLI Buffers are completely dissolved (may need to heat tube a little).

[0130] 5. Add 230 &mgr;l of OL1 buffer to each tube.

[0131] 6. Heat Oligotex bead suspension to 37° C+mix (vortex) immediately before adding 70 &mgr;l to each tube. Mix gently by pipetting up and down.

[0132] 7. Place tube at 65° C. for 3 min. Mix, let sit at room temperature for 10 min with occasional mixing.

[0133] 8. Spin 2 min at 14,000 g at room temperature. Aspirate supernatant (leave ˜50 &mgr;l in tube).

[0134] 9. Add 600 &mgr;l OW1 wash buffer. Resuspend beads by pipetting gently. Spin 2 min at RT and aspirate. Repeat for a total of three washes (1×600 &mgr;l OW1, 2×600 &mgr;l OW2). Remove remainder with pipette after last wash.

[0135] 10. Resuspend beads in 175 &mgr;l DEPC—dH2O. Resuspend by pipetting up and down.

[0136] 11. Place tube in 65° C. waterbath for 1 min. Spin down 1 min at 14,000 g RT. Transfer cluate to fresh tube.

[0137] 12. Repeat elution with 175 &mgr;l DEPC—dH2O. Pool eluates.

[0138] 13. To get rid of any beads: spin pooled eluates 5 min at 14,000 rpm at RT. Remove 330 &mgr;l (know exact amount), transfer to new tube.

[0139] 14. Measure OD260/OD280 of 30 &mgr;l of eluate diluted with 270 &mgr;l dH2O.

[0140] 15. To remainder add {fraction (1/10)} volume (30 &mgr;l) 3M NaOAc pH 5.2, 1 &mgr;l 20 mg/ml glycogen, 675 &mgr;l Ethanol (−20° C.).

[0141] 16. Place 30′ at −80° C. or overnight at −20° C.

[0142] 17. Spin 30′ at 14,000 rpm, 4° C. Remove supernatant with pipette.

[0143] 18. Wash with 140 &mgr;l of 80% Ethanol (−20° C.). Spin, remove with pipette. Remove wash completely.

[0144] 19. Resuspend pellet with appropriate volume of DEPC—dH2O to give a final concentration of 0.5 mg/ml.

[0145] 20. Snap freeze, store at −80° C.

[0146] PolyA Controls:

[0147] PolyA tagged lys, phe, thr, trp and dap T7/T3 IVT expression constructs were obtained from the ATCC (#'s 87482, 87483, 87484, 87485, 87486, respectively). The polyA tagged RNA was generated with NotI digested template DNA and Ambion's Megascript T3 IVT kit according to their instructions. Transcripts were purified using the Qiagen RNeasy kit. An undiluted stock of these controls was prepared by mixing the various transcripts: 133 &mgr;g lys, 58.5 &mgr;g phe, 24.5 &mgr;g dap, 8.8 &mgr;g thr, 3.3 &mgr;g trp in 225 &mgr;l DEPC water. This was aliquoted in 22.5 &mgr;l amounts. A diluted stock was made by adding 477.5 &mgr;l DEPC water to the undiluted stock and this was again aliquoted. The diluted stock was spiked into the total RNA, 5 &mgr;l per mg total RNA. The final amounts of the controls in the total RNA were then: 4 trp, 13.3 thr, 40 dap, 133.3 phe, 400 lys (pmol/mg). By using theses same controls in every experiment all experiments can be normalized to equivalent amounts of total RNA.

[0148] Analysis

[0149] Controls

[0150] Various controls were used to monitor the quality of the experiment and to permit accurate comparisons between experiments. We verify that the yield of total RNA obtained from wild type and mutant cells is very similar. Equivalent amounts of total RNA from each strain are used to make target preparations (stable RNAs such as rRNA and tRNA account for approximately 96% of total RNA). Known amounts of five different in vitro transcribed polyA-tagged B. subtilis RNAs are added to the RNA preparation prior to polyA selection (control features on the HDA are designed to detect the levels of each of these RNA species). The amounts of different B. subtilis RNA molecules added to the RNA preparation represent the maximum range of values expected for the levels of various yeast mRNA species. The values of these controls obtained by scanning and returned by the computer algorithm were used to normalize the values for each yeast mRNA in a given experiment. In this manner, the levels of all mRNA species are normalized to a constant level of total RNA from each strain. Most experiments involve a comparison between the values obtained for a mutant cell and its isogenic wild type counterpart. Additional controls applied to these experiments come from the analysis of duplicate data sets. For example, a comparison is made between the values obtained for the wild type cells in the two independent experiments. Similarly, a comparison is made between the values obtained for the mutant cells in the two independent experiments. Data is used only when the comparison reveals a high correlation between the two data sets from the duplicate experiment.

[0151] For any experiment involving a comparison between a mutant and its isogenic wild type counterpart, the HDAs used were from the same lot number.

[0152] Reproducibility

[0153] To assess the reproducibility of the HDA technology, RNA was harvested from two independent wild type colonies and the two RNA preparations were hybridized to two different HDAs on two separate days. The results were plotted on a histogam in which the ratio of levels of each mRNA measured from the two independent experiments (fold change) was plotted against the number of genes whose mRNAs have been counted in that category. For 99% of all genes scored, the expression values produced for each gene in the two experiments were within 1.7-fold. The genes whose expression values scored outside this narrow range in the two experiments tended to be genes whose expression levels are near the limits of detection (0.1 mRNA molecules/cell). Thus, the technology generates highly reproducible results.

[0154] Genes Scored

[0155] For each component analyzed, four expression profiles were collected: two independent wild type and two independent mutant profiles. A given gene was scored in our analysis if a “Present” call had been returned in each of the four expression profiles or if the mRNA level was altered consistently in both wild type versus mutant comparisons. See Table 1, below.

[0156] Fold Change

[0157] For all genes scored, the fold change was calculated by dividing the mutant value by the wild type value. If this number was less than one the (negative) reciprocal is listed (e.g. 0.75, or a drop of 25% from wild type is reported as either 1.3 fold down or −1.3 fold change). The reported fold changes are the average of the two independent experiments.

[0158] A change was deemed significant and reported in the lists containing genes >2-fold down (or up) based on the following criteria: the gene was scored, the fold change was more than 2-fold in the two independent experiments, and the change in the values was above background values in both comparisons.

[0159] Genome Dependence

[0160] Genome Dependence: Constitutive Mutants

[0161] An estimate of the genome-wide requirement for a transcription factor was made for the constitutive mutant experiments (deletion or enzymatically dead mutants) by adding the number of genes in the 2-fold up and 2-fold down list and dividing this by the total of scored genes. This number is expressed as a percentage and represents the portion of genes in the genome whose expression is significantly affected by the mutation.

[0162] Genome Dependence: Temperature Sensitive Mutants

[0163] An estimate of the genome-wide requirement for transcription factors analyzed through temperature sensitive mutations was made by estimating the number of genes whose transcription showed the same dependence on a given transcription factor as on core RNA polymerase itself. The portion of genes in the genome whose expression is equivalently dependent on RPB1 and the factor is expressed as a percentage. This was calculated as described below.

[0164] Genes Compared to Rpb1

[0165] Those mRNAs whose levels are called in both the rpb1-1 experiment and in the experiment involving ts factor X were collected. Transcripts whose levels drop less than 2-fold in the rpb1-1 experiment are not included in the analysis because the decay rates are not sufficient to provide a meaningful comparison.

[0166] Genes Equivalent Dependence

[0167] To determine the number of genes whose expression is as dependent on a transcription factor as it is on core RNA polymerase II itself, we used the following criteria. For the subset of rpb1-1 transcripts where the average decrease was greater than two-fold (Genes Compared to Rpb1), a comparison was made between the apparent half-life in rpb1-1 and that observed in the other ts factor experiment. A gene is determined to be equivalently dependent on Rpb1 and the factor of interest, and is incorporated into the “Genes Equivalent Dependence” group, if its transcript decays to a level which is within one apparent half-life of the decrease observed with rpb1-1 cells.

[0168] Assessment of Primary/Secondary Effects

[0169] Constitutive Mutants

[0170] Components of the transcription apparatus which are not essential for cell viability on YPD media can be inactivated or depleted in various ways. In this study, two types of inactivating mutations have been used, complete deletions, and point mutations which knock out a specific catalytic activity.

[0171] With constitutive mutants, the results obtained are the sum of primary and secondary effects of factor inactivation. However, there are several approaches to elucidate what effects are primary and what are secondary.

[0172] Identifying Mini-Circuits Due to Altered Levels of Transcription Factors

[0173] Data obtained from a single experiment can identify potential mini-circuits. The effects on transcript levels can be examined by Functional Category, and the effects on gene-specific transcription activators and repressors can be found in the “Transcription” list. For example, the levels of transcripts for STE12 decrease 4.3 in the SRB5 deletion mutant. Ste12 is a transcription factor that binds to the pheromone response element (PRE) to regulate genes required for mating, and the transcript levels of these genes are all decreased in the SRB5 deletion mutant. As another example, transcript levels for the MET28 transcription activator increase in the SWI2 mutant, and the level of transcripts from all genes under Met28 control also increase.

[0174] Cluster Analysis

[0175] Data obtained from multiple experiments can help dissect primary from secondary effects, and we are developing informatics tools that should facilitate this process. For example, if an effect on a set of pheromone-response genes is found upon inactivation of two different factors, we generate models that assume that one of the two results is a consequence of a secondary effect. Then we ask whether any gene product which is known to affect the expression of these genes has an altered mRNA expression level in one of the two datasets.

[0176] Data from Activator and Repressor Mutants

[0177] There are 2-3 hundred DNA-binding transcriptional regulators in yeast. Genes for many of these regulators can be efficiently deleted and the effects examined on HDAs. Comparison of these results with those from the general transcription apparatus will help identify primary effects. This information will also suggest which components of the transcription apparatus are direct targets of the transcriptional regulator under study.

[0178] Temperature-Sensitive Mutants

[0179] The reduction in mRNA levels observed in temperature sensitive (ts) mutants soon after a temperature shift (i.e., 45 minutes) is likely due to primary effects because of the time required to produce most secondary effects involves a substantial reduction in both a transcript and its translation product. Nonetheless, the results obtained in this type of experiment must be regarded as the sum of primary and secondary effects of factor inactivation. We have devised a method to identify the set of genes whose change in expression is almost certainly a direct consequence of the loss of function of the temperature sensitive factor. This involves comparing data from ts inactivation of RNA polymerase II with that obtained by ts inactivation of any other factor. The fold-reduction in all transcripts has been determined for the RNA polymerase II temperature sensitive mutant rpb1-1 relative to its isogenic wild-type counterpart 45 minutes after a shift to 37° C. The same data is collected for another ts mutant exposed to identical experimental conditions. Comparison of the two data sets reveals the set of transcripts with equivalent decay kinetics in rpb1-1 and another ts mutant. This method identifies the set of genes whose expression is equivalently dependent on the factor of interest and RNA polymerase II itself.

EXAMPLE 2

Determination of the Levels of All Detectable mRNA Specks in Yeast

[0180] Knowledge of the levels of all detectable mRNA species in yeast is useful for evaluating the degree to which these levels depend on any one component of the transcription apparatus. To obtain this information and to assess the reproducibility of the HAD technology, RNA was harvested from two independent wild type cultures and compared using two sets of HDAs on two separate days (Example 1 above). The HDAs used here can score mRNA levels for up to 6181 genes. This is a more accurate representation of the transcriptome than that previously determined because it is better able to score mRNA species which are expressed at very low levels (5460 genes were scored using HDAs, whereas 4465 genes were scored with SAGA). It is particularly valuable to have information on transcripts from genes expressed at low levels because many of the regulatory components of the cell are expressed at low levels.

[0181] Of the 5460 genes whose mRNA levels were accurately determined and compared in both experiments, 99% of the mRNAs differed no more than 1.7 fold, and only 35 transcripts (0.65) showed more than a two-fold change. In order to prevent these minimal variations from influencing the results, all experiments were performed in duplicate. The levels determined for the 5460 transcripts in wild type yeast cells and additional information derived from this experiment are described above. The SAGA method has previously been used to determine values for 4465 transcripts, the results of which has been termed the yeast transcriptome. (Velculescu, V. E. et al (1997) Cell, 88, 243-51). The sensitivity of the HDA technology permitted a determination of the levels of many additional gene products, and revealed that transcripts from 80% of expressed yeast genes exist at steady state levels of 0.1 to 2 molecules/cell.

EXAMPLE 3

Assessment of the Role of Components of Transcriptional Machinery in Genome-Wide Gene Expression

[0182] At any one promoter, the transcriptional machinery might include the RNA polymerase II core enzyme, the general transcription factors (GTFs), the core Srb/mediator complex, the Srb10 CDK complex, the Swi/Snf complex, and the SAGA complex, among others (FIG. 36). One or more subunits of each of these components has been investigated for its role in genome-wide gene expression through the use of mutations which affect either the function or the physical presence of the subunit (Table 1). Loss-of-function mutations in various components of the transcription apparatus were constructed or obtained from various investigators (See Study Design on the Web site for details). Two types of mutations have proven to be useful in this study. For essential components of the apparatus, temperature-sensitive (ts) mutations are valuable because they allow the investigator to examine effects on gene expression at any point after inactivating the factor. Point mutations which knock out the catalytic function of known enzymatic activities or complete deletion mutations were used to study non-essential components. In each experiment, a mutant cell and its isogenic wild-type counterpart are grown to mid-log phase, the two populations are harvested, RNA is prepared, and hybridization to HDAs is carried out, all in duplicate. Dependence on Core RNA Polymerase II

[0183] To determine the genome-wide dependence of gene expression on core RNA polymerase II, RNA was isolated from an rpb1-1 temperature sensitive ts cell and its wild type counterpart 45 minutes after a shift to the nonpermissive temperature and was hybridized to HDAs. Because rpb1-1 cells shut down transcription of protein-coding genes immediately after a temperature shift, these cells have been used as described here and by other investigators to determine the half-life of various yeast mRNAs (Nonet M. et al. (1987) Mol Cell Biol 7, 1602-1 1; Herrick D. et al (1990) Mol Cell Biol 10, 2269-84). The 45 minute time point was used for the analysis of all ts mutants in this study because it is sufficiently long to detect a significant (i.e. a two-fold or more) loss of mRNA levels for 94% of detectable gene products without any loss of rRNA (Nonet M. et al., (1987) Mol Cell Biol 7, 1602-11). In addition, the 45 minutes time point is short enough to minimize the potentially complicating effects of cell cycle arrest and cell death.

[0184] The results of genome wide expression analysis of the rpb1-1 mutant as compared to an isogenic wild type strain are shown in a grid format in FIG. 37A. The grid shows the change in mRNA level for each gene, beginning with the left most gene on chromosome I and proceeding in a linear fashion, left to right, through chromosome I and II, then m, etc., until the last gene on the right arm of chromosome IVI is reached at the lower right hand corner. 5735 genes were scored in this analysis. The vast majority of mRNAs are reduced more than two-fold in the mutant cells relative to wild type cells, and this reduction provides an apparent half-life for each of the mRNA species (see Yeast mRNA Population on the Web site). The value determined with this approach is an approximation, but is useful for comparative purposes. Comparison of this data with that obtained for another ts factor identifies the set of genes whose expression is equivalently dependent on RNA polymerase II and the factor of interest.

[0185] There is a set of genes whose mRNAs are not significantly reduced in the mutant cells. These consist of genes that have stable messages as well as genes whose mRNA levels are slightly elevated in the mutant cells relative to wild-type. In this latter group are many known heat shock or stress response genes (e.g. SSA4, SSA3, HSP26, HSP30, HSP42 and SSL2), plus additional ORFs of unknown but perhaps related function. Similar results were obtained using ts mutants in other general transcription factors.

[0186] Dependence on Srb/Mediator Core Subunits

[0187] The Srb/mediator complex is tightly associated with RNA polymerase II in a complex which has been termed the holoenzyme (Koleske A. J., and Young, R. A. (1994) Nature 368, 466-9; Kim, Y. J. et al., (1994) Cell 77, 599-608). Srb4 is an essential component of the Srb/mediator complex (Thompson, C. M. et al., (1993) Cell 73, 1361-75; Kim, Y. J. et al., (1994) Cell 77, 599-608); Hengartner, et al., (1995) Genes and Development 9, 897-910). A ts mutant in Srb4 (srb4-138) was previously used to obtain evidence that several protein-coding genes require the function of Srb4, and are thus likely to have the holoenzyme form of RNA polymerase II recruited to their promoters (Thompson, C. M., and Young, R. A. (1995) Proc Natl Acad Sci USA 92, 4587-90). Genome-wide expression analysis provides a more rigorous test of the model that expression of all protein-coding genes is dependent on Srb4. The experiment was carried out with the same protocol used with the Rpb1 ts mutant. Of the 5361 genes whose mRNA expression levels could be compared (i.e., those that had a greater than two-fold decrease in the experiment with Rpb1 ts and were scored in the Srb4 ts experiment), 93% showed a decrease that closely fit the decreased observed in the Rpb1 ts experiment. Of the mRNAs that did not closely fit the Rpb1 ts decay, only 2 could be found that reproducibly showed large differences in their decay in the two experiments performed. Furthermore, the set of genes whose mRNAs are not significantly reduced in the Rpb1 ts mutant exhibit the same behavior in the Srb4 ts experiment. The results indicate that genome-wide expression is a dependent on Srb4 as it is on core RNA polymerase II (see Genome-Wide Expression Data on the Web site for details). Srb4 is associated tightly and exclusively with the RNA polymerase II holoenzyme (Koleske, A. J., and Young, R. A. (1994) Nature 368, 466-9; Kim, Y. J. et al., (1994) Cell 77, 599-608; and Myers, L. C. et al., (1998) Genes Development 12, 45-54). Thus, it is reasonable to infer Srb4-containing RNA polymerase II holoenzyme is generally required for transcription.

[0188] Med6 is another essential component of the Srb/Mediator complex and appears to be physically associated with Srb4 (Li, Y. et al., (1995) Proc Natl Acad Sci USA 92, 1064-68; Myers, L. C. et al., (1998) Genes Development 12, 45-54; Lee, T. I. et al., (1998). A Med6 ts mutant has been generated and used to demonstrate that Med6 is necessary for full induction of GAL, SUC2, MFA1 and PKY1 genes, but is not required for expression of several others. (Lee, Y. C. et al., (1997) Mol Cell Biol 17,4622-32). The genome-wide dependence of gene expression on Med6 was determined with this Med6 ts strain as described above for Rpb1. The results indicate that the expression of 10% of yeast genes is as dependent on Med6 as it is on Rpb1 (FIG. 37B; see the Web site for detailed information).

[0189] The reduction in mRNA levels observed in ts mutants soon after a temperature shift (i.e., 45 minutes) is likely a consequence of primary effects due to factor inactivation because the time required to produce most secondary effects involves a substantial reduction in both a transcript and its translation product. Nonetheless, the results obtained in this type of experiment must be regarded as the sum of primary and secondary effects. To identify the set of genes whose change in expression is most likely a direct consequence of the loss of function of the ts factor, data from ts inactivation of RNA polymerase II was compared with that obtained by ts inactivation of any other factor. Comparison of the two data sets reveals the transcripts with equivalent decay kinetics in rpb1-1 and the other ts mutant (see Technology, Protocols and Data Analysis on the Web site for details). For those genes affected by ts disruption of Med6 where such a comparison could be made, the mRNAs of 506 genes decreased with similar kinetics in the Med6 and Rpb1 experiments. Thus, the expression of 10% of yeast genes is as dependent on Med6 as it is on Rpb1. These 506 genes are most likely to have a direct requirement for Med6 function. The genes whose transcript levels do not fit the Rpb1 kinetics could have a direct, but partial, requirement for Med6 function, or the effects observed at these genes are a secondary consequence of some other gene's altered mRNA levels. The 506 genes identified which require Med6 function to the same extent as Rpb1 function are those at which promoter-associated transcriptional regulators are most likely to function through interactions with Med6.

[0190] Srb5 is a component of the Srb/Mediator complex whose function is also not known (Thompson, C. M. et al., (1993) Cell 73,1361-75; Kim, Y. J. et al., (1994) Cell 77,599-608); Koleske, A. J., and Young, R. A., (1994) Nature 368, 466-9; Hengartner, C. J. et al., (1995) Genes and Development 9, 897-910; Myers, L. C. et al., (1998) Genes and Development 12, 45-54). To determine the genome-wide dependence of gene expression on Srb5, a strain lacking an SRB5 gene and its wild type counterpart were compared (see the Web site for detailed information). The results indicate that 16% of all genes require Srb5 function for their expression. With the SRB5 deletion strain and other constitutive mutants analyzed here, it is not possible to distinguish between results which are a direct consequence of the loss of Srb5 function and those which are due to a secondary effect such as the loss of another transcriptional regulator. Nonetheless, these results provide important information in that they reveal the complete set of genes which are directly or indirectly affected by loss of Srb5 function. It was striking that expression of many genes central to the pheromone response pathway are dramatically affected by the loss of Srb5, as discussed herein.

[0191] Dependence on SRb10 CDK Complex

[0192] Srb10 is cyclin dependent kinase which is part of a holoenzyme subcomplex containing Srb8, 9, 10 and 11 proteins (Liao, S. M. et al., (1995) Nature 374, 193-6; Hengartner, C. J. et al., (1995) Genes and Development 9, 897-910). Srb10 and its associated proteins have been proposed to form a negative regulatory complex which functions through phosphorylation of the RNA polymerase II complex CTD (Hengartner, C. J. et al., (1998) Molecular Cell 2, 45-53). To determine how gene expression depends on Srb10, RNA was isolated from an Srb10 point mutant which lacks catalytic activity and the expression profile was compared to that of its wild type counterpart. The results are shown in a grid format in FIG. 37C. Of the 5626 genes which were scored, 173 gene products showed 2-fold or greater increases in mRNA levels in the mutant relative to the wild type. This indicates that Srb10 is normally a negative regulator of these 173 genes (approximately 3% of the genome).

[0193] It is notable that nearly half of these genes are derepressed during the nutrient deprivation which occurs during the diauxic shift. (DeRisi, J. et al., (1997) Science 278, 680-86). (FIGS. 40A-40C) Yeast cells undergo a diauxic shift as nutrients are depleted in culture, and a variety of genes which enable the cell to survive nutrient-limiting conditions are derepressed (Johnston, M., and Carlson, M. (1992) Gene Expression p. 193; Yin, Z. et al., (1996) Molecular Microbiology 20, 751-64). These include genes involved in dimorphic morphology (nutrient starved cells alter their morphology to permit foraging for nutrients) and stress responses (starved cells are apparently better able to survive nutrient deprivation when stress proteins are elevated). Srb10 in wild type cells is most likely responsible for repressing this set of genes when cells are in exponential growth on glucose, but no longer performs this function as cells enter the diauxic shift. Coordinate regulation of this set of genes could be accomplished by eliminating the function of Srb10 as cells enter the diauxic shift.

[0194] To determine whether Srb10 is physically lost from cells as they enter the diauxic shift, cells containing an epitope-tagged Srb10 protein were grown in YPD media and sampled at various times during the growth curve. (FIG. 40C). Cell lysates were prepared from each sample and the levels of Srb10 were assayed by Western blot. Results showed that Srb10 is physically depleted as cells enter the diauxic phase of growth. This result is consistent with evidence that the levels of Srb11, the cyclin partner of Srb10, are reduced when cells are exposed to the limiting nutrient environment in sporulation media. (Cooper, K. F. et al., (1997) EMBO J. 16, 4665-75.) It may also explain why a form of yeast holoenzyme purified from commercially available yeast cells lacks the Srb10/Srb11 kinase/cyclin pair (Li, Y. et al., (1995) Proc Natl Acad Sci USA 92, 10864-8; Myers, L. C. et al., (1998) Genes Development 12, 45-54), as these cells are typically grown past mid-log phase. The results thus indicate that the nutrient starvation response is mediated, in part, through the physical loss of the Srb10 CDK from the holoenzyme. This novel mechanism provides one example of how coordinate regulation of gene expression can be accomplished through regulation of components of the general initiation machinery.

[0195] FLO11, which encodes a cell wall protein which is highly expressed in pseudohyphal cells, is expressed at 15-fold higher levels when Srb10 function is lost (FIG. 40A). The dramatic increase in the expression of FLO11 and other genes whose products are involved in the dimorphic shift led Applicants to determine whether the absence of Srb10 function produces a pseudohyphal phenotype. Both copies of the SRB10 gene were deleted from a diploid strain which is generally used to assay this phenotype, and colony morphology was examined under the microscope. Results demonstrated that the loss of Srb10 causes cells to grow preferentially in a pseudohyphal form. This again shows that expression analysis is useful for predicting unexpected phenotypes. More importantly, specific signal transduction pathways control the dimorphic shift (Madhani, H. D., and Fink, G. R., (1998) Trends Cell Biology 8, 348-53), and these results suggest that one of the ultimate targets of these pathways is the Srb10 kinase.

[0196] Dependence on Swi/Snf

[0197] Swi2 ATPase activity plays an essential role in the ability of the Swi/Snf complex to remodel chromatin (Laurent, B. C. et al., (1993) Genes Development 7, 583-91; Cote, J. et al., (1994) Science 265, 53-60; Khavari, P. A. et al., (1993) Nature 366, 170-4). This activity is thought to facilitate activator and transcription apparatus binding to promoter regions for a small number of genes, thereby overcoming repression by nucleosomes at those promoters (Cote, J. et al., (1994) Science 265, 53-60; Imbalzano, A. N. et al., (1994) Nature 370, 481-5; Kwon, H. et al., (1994) Nature 370, 477-81; Burns, L. G. and Peterson, C. L., (1997) Mol Cell Biol 17, 4811-9). Consequently, it was expected that a small number of genes would be reduced in expression levels in the Swi2/Snf2 mutant. To determine the genome-wide dependence of gene expression on the Swi/Snf complex, RNA was isolated from a Swi2/Snf2 point mutant which lacks ATPase activity and its wild type counterpart and the two RNA preparations were hybridized to HDAs. The surprising result was that a greater number of genes appear to be negatively regulated by Swi/Snf than are positively regulated (FIG. 37D; see the Web site for detailed information). The data show that 203 gene products were elevated 2-fold or more in the mutant relative to the wild type, while just 126 transcripts decreased 2-fold or more (See Genome-Wide Expression Data on the Web site). As described herein, this result may be explained by recent data indicating that the Swi/Snf complex can catalyze chromatin remodeling in either direction (Schnitzler, G. et al., (1998) Cell 94, 17-27).

[0198] Dependence on General Transcription Factors

[0199] The general transcription factors are necessary to reconstitute promoter-dependent transcription in vitro with core RNA polymerase II. These factors include TFIID, TFIIB, TFIIF, TFIIE and TFIIH. Among these factors, TFIIE and TFIIH are of particular interest because numerous reports have suggested that they are in fact not generally required for gene expression (Parvin, J. D. et al., (1992) Cell 68, 1135-44; Serizawa, H. et al., (1993) Nature 363, 371-4; Timmers, H. (1994) EMBO J. 13, 391-9; Holstege, F. C. et al., (1995) EMBO J. 14, 810-9; Sakurai, H. et al., (1997) J Biol Chem 272, 15936-42; Kuldell, N. H., and Buratowski, S., (1997) Mol Cell Biol 17, 5288-98; Tijerina, P., and Sayre, M. H., (1998) J Biol Chem 273, 1107-13). Genome-wide expression analysis was carried out on a Kin28 ts cell and its isogenic wild type counterpart using the same experimental protocol used for the Rpb1 ts mutant. Kin28, a CDK subunit of TFIIH, is an RNA polymerase II CTD kinase which is involved in the transition from initiation to elongation (Dahmus, M. (1996) J Biol Chem. 271, 19009-19012). The results reveal that Kin28 is generally required for expression of protein-coding genes (see Genome-Wide Expression Data on the Web site). TFIIE is thought to facilitate certain functions of TFIIH. In contrast to the results obtained with Kin28, analysis of genome-wide expression with a Tfa1 ts mutant shows that only 54% of yeast genes require the largest subunit of TFIIE to the same extent as core RNA polymerase II (see Genome-Wide Expression Data on the Web site).

[0200] The TBP-associated factors (TAFIIs) of TFID are especially interesting because they have been postulated to play important roles in promoter selectivity and gene activation (Burley, S. K., and Roeder, R. G., (1996) Annu Rev Biochem 65, 769-99; Verrijzer, C. P., and Tijan, R., (1996) Trends Biochem Sci 21, 338-42; Lee, T. I., and Young, R. A., (1998) Genes and Development 12, 1398-1408). A ts mutation in the TFIID subunit TAFII145 (Walker, S. S. et al., (1997) Cell 90, 607-14) was used to determine the genome-wide dependence of gene expression on this TAF. Of the 5441 genes which were scored, 1618 genes products were reduced by 2-fold or greater on average in the two comparisons made, 45 minutes after temperature shift. For those genes where a comparison with the Rpb1 experiment could be made, 16% showed a dependence on TAFII145 that was similar to their dependence on Rpb1 (see Genome-Wide Expression on the Web site for details). Interestingly, a large number of genes involved in functions associated with progression through the cell cycle are among the genes most likely to have a direct requirement for TAFII145 function. The TAFII145 ts mutant has a cell cycle phenotype: it arrests growth in G1-S after cells are shifted to the nonpermissive temperature. Previous studies showed that several G1-S cyclin genes are expressed at reduced levels in these cells, perhaps accounting for the cell cycle arrest phenotype TAFII145. (Walker, S. S. et al., (1997) Cell 90, 607-14) A subset of the genes that have a direct requirement for TAFII145 function and which are involved in functions associated with progression through the cell cycle are listed in Table 2. For example a significant decrease in mRNA levels was observed for Ctr9, which is required for expression of G1 cyclins Cln1 and Cln2. In addition, genes which are involved in DNA repair and DNA synthesis are dependent on TAFII145 function. Thus, the G1/S arrest phenotype of TAFII145 mutants may be due to multiple defects in cyclin and chromosome synthesis which occur during this period of the cell cycle.

[0201] Analysis of which genes depend on TAFII17, a histone H3-like TAF which is shared by TFIID and SAGA complexes, for their expression was also carried out. RNA was isolated from a TAFII17 temperature sensitive cell (TAF17-ts) and its wild type counterpart 45 minutes after a shift to the nonpermissive temperature and was hybridized to HDAs. Of the yeast genes identified in the TAFII17 experiment and appropriate for comparison, 67% are as dependent on TAFII17 function as they are on Rpb1, and are thus most likely to have a direct requirement for TAFII17 function (see Genome-Wide Expression on the Web site for details). This indicates that TAFII17 is critical for the expression of a much larger portion of the transcriptome than TAFII145. The presence of TAFII17 in two different complexes may account for this observation.

[0202] Dependence on Gcn5 Subunit of SAGA

[0203] The recent discovery that certain TAFs are components of both the TFIID general transcription factor and the SAGA complex (Grant, P. A. et al., (1998) Cell 94, 45-53) makes it particularly interesting to compare the effects of a mutation in a component specific to each complex (TAFII145 in the case of TFIID and Gcn5 in the case of SAGA) with those of a mutation in a component shared by the two complexes (TAFII17). The expression profile of a GCN5 deletion mutant was compared with its isogenic counterpart (see Genome-Wide Expression on the Web site for details). Of the 4912 genes which were scored, 185 transcripts were reduced by 2-fold or more and 83 increased by 2-fold or more.

[0204] The Gcn5 results indicate that this component of SAFA is necessary for normal expression of no more than 5% of yeast genes. Expression of 16% of protein-coding genes depends on the TAFII145 subunit of TFIID to the same extent they depend on Rpb1. In contrast, the expression of 67% of yeast genes depends on the function of the TAF1117 subunit shared by SAGA and TFIID. 2 TABLE 1 Transcriptional Machinery Fraction of genes dependent on sub- Complex and Subunit Features unit function RNA Polymerase II Rpb1 Largest subunit, mRNA catalysis, 100% contains CTD Srb/mediator (core) Srb4 Target of Ga14 activator  93%* Srb5 Unknown function  16% Med6 Role in activation of some genes  10% Srb CDK complex Srb10 CTD kinase, negative regulator  3% Swi/Snf Swi2 ATP-dependent chromatin  6% remodeling General Transcription Factors TFIID (TAFII145) Large TBP-associate factor,  16% histone acetylase (TAFII17) Component of both TFIID and  67% SAGA TFIIE (Tfa1) Promoter opening  54% TFIIH (Kin28) CTD kinase  87%* SAGA Gen5 Histone acetylase  5% TAFII17 Component of both TFIID and  67% SAGA *Srb4 and Kin28 results were essentially identical to Rpb1, but because of the stringency applied by the fit algorithm, a minimal estimate is produced.

[0205] 3 TABLE 2 Genes That Require Taf145 Function FOLD RE- GENE DESCRIPTION DUCTION Cell Cycle *DDCI DNA damage checkpoint protein 10  YER066W Similar to CDC4, which degrades G1 cyclins 9 SPO1 Possible role in spindle pole body 8 duplication *LTE1 GDP/GTP exchange factor 8 *MKK2 Kinase involved in cell wall integrity 8 *BIM1 Possible role in early spindle pole body 8 assembly *MDM1 Involved in mitochondrial segregation 7 *CTR9 Required for normal expression of G1 7 cyclins *PAC1 Possible role in spindle pole body orientation 6 *SCP160 Involved in control of chromosome 6 transmission CDC13 Telomere binding protein 6 *TOP3 DNA topoisomerase III 5 *TRX1 Thioredoxin I 5 ARD1 N-acetyltransferase 5 *SCC2 Required for sister chromatid cohesion 5 *CLB2 G2/M cyclin 5 *KIP2 Kinesin related protein 5 *MEC1 Cell cycle checkpoint protein 4 RAD9 DNA repair checkpoint protein 4 *SPC98 Spindle pole body component 4 *BCK1 Kinase involved in cell wall integrity 4 DNA Repair *RAD3 Involved in nucleotide excision repair 8 *YHR031C Possible role in chromosome repair 7 *RAD5 Involved in DNA repair 6 *HSM3 Involved in mismatch repair 6 *RAD50 Involved in recombinational repair 5 *EXO1 Involved in mismatch repair 5 *MSH3 Involved in mismatch repair 5 YER041W Similar to DNA repair protein, Rad2 5 REV1 Involved in translesion DNA synthesis 4 HDF2 Involved in DNA end-joining repair pathway 4 MSH6 Involved in mismatch repair 4 DNA Synthesis *MCM3 Involved in replication initiation, MCM/P1 13  family RLF2 Chromatin assembly complex, subunit 2 9 *MCM6 Involved in replication initiation, MCM/P1 9 family REV7 DNA polymerase subunit zeta 7 *MIP1 Mitochrondial DNA-directed DNA 6 polymerase *CDC47 Involved in replication initiation, MCM/P1 6 family *CDC5 Kinase 5 *CDC46 Involved in replication initiation, MCM/P1 5 family *RFC1 DNA replication protein RFC large subunit 5 *CAC2 Chromatin assembly complex, subunit 1 5 *Gene exhibits equivalent dependence on Taf145 and Rbp1 for normal expression

[0206] While this invention has been particularly shown and described with reference to preferred embodiments thereof, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as described herein and/or as defined by the appended claims.

Claims

1. A pair of isogenic eukaryotic cells comprising a test cell which contains a mutant component of the general transcription machinery and a control cell which is the wild-type isogenic counterpart of the test cell.

2. The pair of isogenic eukaryotic cells of claim 1 wherein the mutant component of the general transcription machinery is a mutant general transcription factor.

3. The pair of isogenic eukaryotic cells of claim 1 wherein the mutant component of the general transcription machinery is a temperature sensitive mutant.

4. The pair of isogenic eukaryotic cells of claim 1 wherein the mutant component of the general transcription machinery is a point mutant.

5. The pair of isogenic eukaryotic cells of claim 1 wherein the mutant component of the general transcription machinery is a deletion mutant.

6. The pair of isogenic eukaryotic cells of claim 1 wherein the mutant component of the general transcription machinery is a component of RNA polymerase II holoenzyme.

7. The pair of isogenic eukaryotic cells of claim 1 wherein the mutant component of the general transcription machinery is a necessary to reconstitute promoter-dependent transcription in vitro with core RNA polymerase II.

8. A method of studying the effects of drugs on cells comprising the steps of:

(a) contacting a cell with a drug; and

(b) determining the effect of the drug on the cell by assessing expression of one or more of the genes which are determined to be members of the regulatory pathway of the general transcription factor according to claim 1.

9. A method of identifying a cellular regulatory circuit which employs a component of a subcomplex of regulatory proteins within the RNA polymerase II holoenzyme, referred to as the transcription initiation apparatus, comprising:

(a) comparing genome expression signature during cellular responses to environmental or other stimuli with the genome expression signature produced by a defect in the transcription initiation apparatus; and

(b) determining differences between the two genome expression signatures and relating the differences to the defect in the transcription initiation apparatus, thereby identifying a component of the transcription initiation apparatus which is responsible for regulation of genes in the cells.

10. The method of claim 9 wherein the cellular regulatory circuit is a yeast cell regulatory circuit.