Rap1p Requires Gcr1p and Gcr2p Homodimers to Activate Ribosomal Protein and Glycolytic Genes, Respectively

Corresponding author: George M. Santangelo, Department of Biological Sciences, University of Southern Mississippi, Hattiesburg, MS 39406-5018., george.santangelo{at}usm.edu (E-mail)

Communicating editor: M. HAMPSEY

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Efficient transcription of ribosomal protein (RP) and glycolytic genes requires the Rap1p/Gcr1p regulatory complex. A third factor, Gcr2p, is required for only the glycolytic (specialized) mode of transcriptional activation. It is recruited to the complex by Gcr1p and likely mediates a change in the phosphorylation state and/or conformation of the latter. We show here that leucine zipper motifs in Gcr1p and Gcr2p (1LZ and 2LZ) are each specific to one of the two activation mechanisms—mutations in 1LZ and 2LZ impair transcription of RP and glycolytic genes, respectively. Although neither class of mutations causes more than a mild growth defect, simultaneous impairment of 1LZ and 2LZ results in a severe synthetic defect and a reduction in the expression of both sets of genes. Intracistronic complementation by point mutations in the charged e and g positions confirmed that Gcr1p/Gcr1p and Gcr2p/Gcr2p homodimers are the forms required for the different roles of the activator complex. Direct heterodimerization between 1LZ and 2LZ apparently does not occur. Dichotomous Rap1p activation and its striking requirement for distinct homodimeric subunits give cells the capacity to switch between coordinated and uncoupled RP and glycolytic gene regulation.

GROWTH control in Saccharomyces cerevisiae appears to be mediated by the transcriptional activation function of Rap1p/Gcr1p, which regulates expression of glycolytic and translational component genes through the upstream Rap1p binding site (UAS_RPG) in each (TORNOW et al. 1993 ; SHORE 1994 ). The auxiliary factor Gcr2p (UEMURA and FRAENKEL 1990 ), which apparently does not contact DNA directly, is essential to the specialized mode of activation that also requires the Gcr1p binding site in DNA (the CT box; ZENG et al. 1997 ). This Gcr2p/CT box-dependent transcription is observed for most glycolytic promoters. In contrast, ribosomal protein genes and other translational component genes do not have CT boxes (HUIE et al. 1992 ) and do not require Gcr2p. In these genes Gcr1p recruitment to the Rap1p DNA binding site is sufficient for activation (SANTANGELO and TORNOW 1990 ; TORNOW et al. 1993 ; ZENG et al. 1997 ). It therefore seems likely that the mild growth defect of gcr2 cells is caused by impairment of the specialized component of Gcr1p function (ZENG et al. 1997 ). In gcr1 cells, most if not all Rap1p activation is lost, and a severe growth defect results. Importantly, Gcr2p is inert in the absence of Gcr1p; its removal from gcr1 cells causes little or no worsening of the defects in cellular growth or activation of target genes.

We previously eliminated several models for GCR2 function in glycolytic gene expression. Gcr2p does not improve Rap1p/Gcr1p complex formation, strengthen Gcr1p binding to the CT box, or help to maintain wild-type steady-state levels of Gcr1p or Rap1p (ZENG et al. 1997 ). Like Gcr1p (TORNOW et al. 1993 ), Gcr2p contains an activation domain (UEMURA and JIGAMI 1992 ; our unpublished data). However, despite ample evidence that it physically associates with Gcr1p in vivo (see below), Gcr2p fails to augment either Gcr1p activation at UAS_RPG elements in the absence of a CT box or Gcr1lexAp activation at a lexA operator (ZENG et al. 1997 ). Since the current paradigm predicts that activation should be synergistic or at least additive under these circumstances (PTASHNE and GANN 1997 ), this argues against the proposal (UEMURA and JIGAMI 1992 ) that Gcr2p generically provides Gcr1p with an activation domain. At either Gcr1lexAp-bound lexA operators or at Rap1p/Gcr1p-bound UAS_RPG sites in the absence of a CT box, perhaps Gcr2p is somehow made activation incompetent (for example, see HALBACH et al. 2000 ), or its recruitment by Gcr1p is blocked. We therefore cannot rule out the possibility that the Gcr2p activation domain might be dispensable even in specialized Gcr1p function at glycolytic genes.

Several lines of evidence suggest a new though still poorly defined mechanism for GCR2 function. After Gcr2p is recruited by Rap1p/Gcr1p to the UAS_RPG/CT box sites in glycolytic genes, it may make an essential structural change in the complex (ZENG et al. 1997 ). First, two-hybrid analysis (UEMURA and JIGAMI 1992 ) and coimmunoprecipitation that we report here indicate that Gcr1p and Gcr2p bind to each other. Second, the requirements for the CT box (BAKER 1991 ), for the CT-binding domain of Gcr1p (HUIE et al. 1992 ), and for Gcr2p are equivalent. Removal of any or all of these components eliminates CT-dependent transcription (the CT box effect) without affecting Gcr1p function at GCR2-, CT-independent promoters (TORNOW et al. 1993 ; ZENG et al. 1997 ). Third, the loss of the CT box effect, and the accompanying mild growth defect of gcr2 cells, can be suppressed by numerous lesions in GCR1 (UEMURA and JIGAMI 1995 ; ZENG et al. 1997 ). Many of these mutations map to the central serine/proline (SP)-rich region of Gcr1p, which appears to be heavily phosphorylated (ZENG et al. 1997 ; our unpublished data). Fourth, phosphorylation of Gcr1p, in particular the occurrence of a hyperphosphorylated form of the protein, is stimulated by Gcr2p (ZENG et al. 1997 ). The precise nature of the critical Rap1p/Gcr1p alteration(s) at CT boxes remains to be discovered, as does an explanation for the high frequency with which mutations in GCR1 cause gcr2 bypass. A thorough understanding of this unusual transcriptional activation complex, including its dichotomous behavior at different groups of target genes, may therefore hinge upon an elucidation of the molecular interactions between subunits.

We have found that Gcr2p, like Gcr1p (DEMINOFF et al. 1995 ), contains an excellent match to leucine zipper (LZ) motifs. We had expected 1LZ to function primarily in Gcr2p recruitment, due to the discovery of 2LZ and the fact that in-frame deletion of the Gcr1p leucine zipper (the GCR1^D allele; Fig 1) causes a gcr2-like (mild) growth defect (DEMINOFF et al. 1995 ). A direct prediction of this idea was that combination of the GCR1^D and gcr2 mutations should result in a defect no worse than that caused by either mutation alone. Interestingly, simultaneous deletion of 1LZ and Gcr2p instead leads to a severe growth defect similar to that of gcr1 cells. In the synthetically defective GCR1^D gcr2 strain, Rap1p/Gcr1p activation of both translational component genes (due to removal of 1LZ) and glycolytic genes (due to GCR2 deletion) is eliminated. This synthetic phenotype can be recapitulated by threonine replacements confined to the respective hydrophobic surfaces of 1LZ and 2LZ, confirming that simultaneous interference with Gcr1p/Gcr1p and Gcr2p/Gcr2p homodimer formation suffices to eliminate Rap1p/Gcr1p activation of most if not all target genes.

View larger version (32K): In this window In a new window Download PPT slide   Figure 1. Diagram of Gcr1p and Gcr2p. The N terminus (open box), central Gcr1p-homologous region (2H; shaded box), and C-terminal leucine zipper (2LZ, checkered box) are indicated within the predicted primary sequence of Gcr2p (a total of 534 residues). In Gcr1p (785 residues), hypomutable regions A, B1, B3, and C (solid boxes) are indicated; region D (checkered box) coincides with 1LZ (DEMINOFF et al. 1995 ). The central serine/proline (SP)-rich region (shaded box), which shares homology with the 2H region of Gcr2p, and the CT box-binding domain (open box) are also shown.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Yeast methods:
The following strains were used in this study: SD5, MAT leu2-3,112 his6 ura3-52 trp1::LEU2 gcr2::ura3 gcr14::URA3; SD6, same as SD5 but with GCR1^D1 at the GCR1 locus and Ura3^-; XZ12, MATa leu2-3,112 ura3-52 leu2::pHL199(LEU2) trp1::HIS3 gcr2::URA3. In SD5, GCR1 was replaced with URA3 in a spontaneous Ura3^- version of DFY643 (UEMURA and FRAENKEL 1990 ). The GCR1^D1 allele integrated into SD6 lacks codons 223–308 (DEMINOFF et al. 1995 ). Yeast cells were grown in liquid or on solid (2% agar) synthetic complete drop-out (SC) media or in rich (YEP) media, supplemented with 2% glucose (D) unless otherwise noted. The Yeastmaker kit from CLONTECH (Palo Alto, CA) was used for yeast transformation. When transformants were grown in SCD, selection for the plasmid was accomplished by dropping out the appropriate nutrient.

Plasmid construction:
One multicopy [YEpTRP, TRP1 2µ ori Amp^r pMB1 ori (TORNOW et al. 1993 )] and two low-copy vectors [YCpTXZ, TRP1 CEN ARS ori Amp^r pMB1 ori (ZENG et al. 1997 ); and Ycp50, URA3 CEN ARS ori Amp^r Tet^r pMB1 ori (ROSE et al. 1987 )] were used in this study. YCpTXZ-GCR2 was constructed by ligating a fragment of GCR2, extending from 300 bp upstream of the translational start codon to the genomic EcoRI site downstream of the stop codon, into the plasmid. Mutated variants of GCR2 were also carried on YCpTXZ. GCR2^N encodes the translational start codon in its proper location relative to upstream signals, followed immediately by codon 223 and the remainder of GCR2. The gcr2^2H allele replaces codons 223–481 with a linker sequence (GCGGCCGCC); gcr2^LZ was a spontaneous isolate with a frameshift deletion of a single (CG) base pair in codon 509. Beginning at residue 509, the predicted (out-of-frame) amino acid sequence of gcr2^LZ is lys-arg-arg-ser-gly-stop. In the four point-mutated versions of GCR2 (L2T, E2R, R2E, and E2R/R2E), codon 507 is changed from CAG to AGC, creating a Q to S amino acid change that does not affect Gcr2p function (not shown). YCpTXZ-(gcr2^E2R + gcr2^R2E) carries both alleles in tandem. Care was taken to ensure that the results obtained with this plasmid, as well as those obtained with the combination of plasmids bearing GCR1^E2K and GCR1^K2E (see below), did not result from recombination between the two alleles (see RESULTS). The 2x hemagglutinin (HA) tag, inserted between GCR2 codons 336 and 337, and all point mutations described herein, were introduced using mutagenic oligonucleotides and either the polymerase chain reaction or the pAlter kit (Promega, Madison, WI). In each case, the resulting mutation was confirmed by DNA sequence analysis. Standard methods were used to introduce HA-tagged GCR2 downstream of the GCR1 gene in YEpTRP-GCR1, -mycGCR1 (described previously in TORNOW et al. 1993 ), and -mycGCR1^L2T (see below); a version of GCR2 encoding an N-terminal fusion to the Escherichia coli lexAp DNA-binding domain was similarly introduced.

GCR1^D has been described previously (DEMINOFF et al. 1995 ). GCR1^D, GCR1^E2K, and GCR1^L2T were carried on YCpTXZ, while GCR1^K2E was carried on YCp50, each as a fragment comprising sequences between the genomic SalI and XhoI sites, as is wild-type GCR1 on all plasmids used in this study. GCR1^E2K and GCR1^K2E contain a 2x myc tag and behave indistinguishably from untagged versions (our unpublished data). A set of four plasmids was constructed by ligating the SalI-XhoI fragment of GCR1 or (GCR1^L2T) into the XhoI site of YCpTXZ-GCR2 (or -GCR2^L2T); YCpTXZ-(GCR1^D + GCR2) was constructed in an analogous way. A myc-tagged version of GCR1^L2T was ligated into YEpTRP in the same orientation as wild-type GCR1. The point mutations in GCR1^L2T were also recombined with the fusion gene GCR1lexA [also carried on YEpTRP; see TORNOW et al. 1993 for a detailed description of myc- and lexA-tagged GCR1 alleles].

Primer extension, immunoprecipitation, Western blotting, and band retardation:
Oligonucleotides specific for glycolytic genes, ribosomal protein genes, or ACT1 were labeled, primer extension reactions were done, and products were analyzed by using standard methods (TRIEZENBERG 1992 ). Immunoprecipitation with -myc antibody was done as described previously (TORNOW et al. 1993 ). Western blotting (BURNETTE 1981 ) with -myc, -HA, or -lexA antibody, the latter of which was the generous gift of Clyde Denis, was done with the Pierce kit (Pierce Chemical, Rockford, IL). Band retardation was done as described in DEMINOFF et al. 1995 . Extracts of XZ12 transformants, which are nearly isogenic with respect to the two other strains used in this study (SD5 and SD6), were used in experiments that involved Western blotting with -HA antibody or in band retardation analysis.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

2LZ is required for Gcr2p function:
We previously reported that Gcr1p contains an excellent match to LZ motifs and demonstrated that it is both necessary and sufficient for Gcr1p homodimerization (DEMINOFF et al. 1995 ). Since leucine zippers are also known to mediate heteromultimer formation (LANDSCHULZ et al. 1988 ; GENTZ et al. 1989 ; O'SHEA et al. 1989 ), and Gcr2p is recruited by Gcr1p to target promoters, we looked for an LZ motif in Gcr2p and found one in its C terminus. We tested the requirement for this motif by constructing GCR2 genes that lack either the nonessential N-terminal codons, the central region that shares homology with GCR1 [UEMURA and JIGAMI (1992), which we designate 2H], or the C-terminal (28-codon) segment that encodes 2LZ (Fig 1). All of the resulting gene products were stably expressed in gcr2 cells (not shown). 2LZ was required for GCR2 function, as was the 2H region; as expected (UEMURA and JIGAMI 1992 ), the N-terminal region was dispensable (Fig 2). The role of the 2H region is unclear, though it does appear to be required for transcriptional activation by Gcr2p (not shown). The remainder of our experiments with GCR2 explored the 2LZ requirement.

View larger version (63K): In this window In a new window Download PPT slide   Figure 2. Gcr2p function requires its leucine zipper (2LZ) and Gcr1p-homologous region (2H). XZ12 (GCR1 gcr2) cells, transformed with a low-copy TRP1 plasmid (YCpTXZ or YCpTXZ-GCR2 derivative), were streaked onto SCD plates and grown for 2 days at 30°. Each plasmid carries the indicated GCR2 allele: GCR2, wild type; GCR2N, codons 2–222 deleted; gcr22H, the region homologous to GCR1 deleted (codons 223–481); gcr2LZ, the C-terminal 26 codons that specify the Gcr2p leucine zipper deleted; Vector, no GCR2 gene.

To confirm that the essential property of the 2LZ motif is its capacity to form an -helical leucine zipper structure in vivo, we disrupted its predicted hydrophobic interface by changing two leucines and one isoleucine to threonine (I497T + L500T + L504T; L2T, boxed residues in Fig 3A). In two other GCR2 products, we disrupted potential salt bridges (dashed lines in Fig 3A; HU et al. 1993 ) by changing either glutamic acid residues to arginine (E503R + E510R; E2R) or arginines to glutamic acid (R508E + R515E; R2E). None of these three mutated products is predicted to be capable of forming either a parallel Gcr2p/Gcr2p homodimer (Fig 3A) or a parallel Gcr1p/Gcr2p heterodimer (Fig 3B).

View larger version (87K): In this window In a new window Download PPT slide   Figure 3. Helical wheel diagrams of the Gcr1p and Gcr2p leucine zippers (1LZ and 2LZ). (A) A Gcr2p/Gcr2p homodimer is shown; note that GCR2 mutations cause changes in both subunits. Clustered point mutations in 2LZ were made to alter hydrophobic residues (in the a and d positions of the -helix; boxes) or charged residues (in the e and g positions; outline font). Each mutated codon is numbered adjacent to the corresponding residue. In the allele referred to in the text as gcr2L2T, codons for three hydrophobic amino acids (I497, L500, and L504) were changed to threonine (T) codons. In two other alleles, codons for residues that may mediate salt bridge interactions (dashed lines) were changed to those that specify the opposite charge: gcr2R2E (R508E and R515E) and gcr2E2R (E503R and E510R). (B) The Gcr1p/Gcr2p leucine zippers are depicted to interact as a heterodimeric coiled coil; the Gcr2p subunit is described in A, above. Note that mutations in GCR1 or GCR2 affect only one of the two subunits; each mutated codon is numbered adjacent to the corresponding residue. In the allele referred to in the text as GCR1L2T, three leucine codons (L267, L274, and L281; boxes) were changed to threonine codons. Outline font indicates substitutions for charged (GCR1E2K; E270K and E277K) or polar and charged (GCR1K2E; S275E and K282E) residues in GCR1 products. (C) A Gcr1p homodimer is shown, with substitutions as described in B. Note that GCR1 mutations cause changes in both subunits. A double point mutation in GCR2 affecting residues only in the g position (gcr2E2R) or only in the e position (gcr2R2E) is predicted to destabilize both the heterodimer and the homodimer. Combination of all four lesions in the same molecule (GCR2E2R+R2E; see text) predicts disruption of the heterodimeric but not the homodimeric interaction.

Reciprocal charge switches in 1LZ and 2LZ complement intracistronically and by second site reversion:
The L2T, E2R, or R2E mutations in GCR2 each lead to a growth defect equivalent to that of gcr2 strains (Fig 4); the steady-state levels of these mutated GCR2 products are indistinguishable from that of wild-type Gcr2p (not shown). This confirms that the predicted 2LZ structure is required for Gcr2p function. Surprisingly, when charges are reversed in both the e and g positions in the same molecule (Gcr2p^E2R+R2E), which should destabilize a Gcr1p/Gcr2p heterodimer but not a Gcr2p/Gcr2p homodimer (compare Fig 3B with 3A), function is restored (Fig 4).

View larger version (67K): In this window In a new window Download PPT slide   Figure 4. Effect of substitutions for hydrophobic or charged residues in 2LZ. XZ12 (GCR1 gcr2) cells, transformed with a low-copy TRP1 plasmid (YCpTXZ or YCpTXZ-GCR2 derivative), were streaked onto SCD plates and grown for 2 days at 30°. Each plasmid contained the indicated allele(s), as follows: GCR2, wild type; gcr2L2T, point mutations I497T, L500T, and L504T; gcr2E2R, point mutations E503R and E510R; gcr2R2E, point mutations R508E and R515E; GCR2E2R+R2E, E503R, E510R, R508E, and R515E; gcr2E2R + gcr2R2E, two alleles in tandem, one with E503R and E510R, the other with R508E and R515E (see text); Vector, no GCR2 gene.

The gcr2^E2R, gcr2^R2E, and GCR2^E2R+R2E mutations should all destabilize a heterodimeric interaction with wild-type Gcr1p by setting up charge repulsions (Fig 3B). Since GCR2^E2R+R2E is a functional allele, the charged residues in positions e and g appear to participate in Gcr2p homodimerization (Fig 3A) rather than Gcr1p/Gcr2p heterodimerization (Fig 3B). We tested this idea by taking advantage of the inability of either the gcr2^E2R or the gcr2^R2E product to complement the gcr2 defect. A low-copy plasmid was constructed that contained both gcr2^E2R and gcr2^R2E (in tandem; see MATERIALS AND METHODS). Transformation with this plasmid restored wild-type growth to gcr2 cells by intracistronic complementation (Fig 4). Reisolation of the complementing plasmid from yeast cells, followed by DNA sequence analysis of the rescued alleles, demonstrated that the Gcr2⁺ phenotype did not result from recombination between gcr2^E2R and gcr2^R2E (not shown). Intracistronic complementation between these two GCR2 alleles rules out the idea that coiled-coil formation between 1LZ and 2LZ makes a significant contribution to Gcr1p/Gcr2p function and suggests instead that position e and g charged residues in 2LZ support Gcr2p homodimerization.

We next used an analogous genetic approach to test whether 1LZ participates in Gcr1p/Gcr2p heterodimerization or Gcr1p/Gcr1p homodimerization. A definitive answer was obtained in the SD6 background (see below). As expected, 1LZ could be inactivated by point mutations that disrupt potential salt bridges between residues in the e and g positions [E270K + E277K (E2K) or S275E + K282E (K2E); Fig 3C and Fig 5]. In complete agreement with the results of 2LZ analysis, intracistronic complementation was observed between GCR1^E2K and GCR1^K2E (Fig 5). This result independently rules out the possibility that 1:1 heteromeric leucine zipper formation between 1LZ and 2LZ plays a role in Gcr1p/Gcr2p function.

View larger version (71K): In this window In a new window Download PPT slide   Figure 5. Intracistronic complementation between GCR1 alleles encoding reciprocal mutations in charged amino acids in 1LZ. The numbered segments of this SCD plate contain colonies of SD6 cells transformed with low-copy plasmids, as follows: 1, YCp50-GCR1 + YCpTXZ-GCR2; 2, YCp50-GCR1 + YCpTXZ; 3, YCp50 + YCpTXZ-GCR1E2K; 4, YCp50-GCR1K2E + YCpTXZ; 5, Ycp50 + YCpTXZ; or 6, YCp50-GCR1K2E + YCpTXZ-GCR1E2K. See the text and Fig 3B for the amino acid substitutions encoded by these alleles. The transformants were grown for 3 days at 30°.

Gcr1p/Gcr2p complex formation does not require 1LZ:
Since 2LZ function does not require coiled-coil formation with 1LZ (and vice versa), we tested whether 1LZ is needed at all for association between Gcr1p and Gcr2p. If not, it should be possible to detect Gcr1p/Gcr2p complexes in the absence of 1LZ. We had previously shown that deletion of 1LZ leads to loss of the slow- but not the fast-migrating complex between the fusion protein Gcr1lexAp and the lexA operator in band retardation analysis (DEMINOFF et al. 1995 ). These and other data demonstrated that removal of 1LZ eliminates Gcr1p homodimerization. We found that the GCR1^L2T product (L267T + L274T + L281T; boxed residues in Fig 3C) also fails to homodimerize according to the band retardation assay (Fig 6, lanes 6–9) but forms complexes with Gcr2p just as effectively as wild-type Gcr1p according to the coimmunoprecipitation assay (Fig 7, compare lanes 1 and 2). Unless the use of overexpressed polypeptides results in artifactual complex formation, association between Gcr1p and Gcr2p does not require 1LZ. In contrast, we have not yet detected Gcr1p/Gcr2p complexes in the absence of 2LZ (see below).

View larger version (114K): In this window In a new window Download PPT slide   Figure 6. Point mutations in 1LZ impair Gcr1p homodimerization. The fusion protein Gcr1lexAp (TORNOW et al. 1993 ), with an intact (wt, lanes 2–5), point-mutated (L2T, lanes 6–9), or deleted (D, lanes 10–13) leucine zipper, was expressed from the multicopy plasmid YEpTRP. GCR1L2T encodes the substitutions L267T, L274T, and L281T in the hydrophobic surface of 1LZ (see Fig 3B). Band retardation analysis was done with 0.3 µg (lanes 2, 6, and 10), 1 µg (lanes 3, 7, and 11), 3 µg (lanes 4, 8, and 12), or 10 µg (lanes 5, 9, and 13) of total protein from each strain. Probe alone was loaded in lane 1 to mark the position of free (F) lexAOP DNA.

View larger version (30K): In this window In a new window Download PPT slide   Figure 7. 1LZ is not required for Gcr1p/Gcr2p coimmunoprecipitation. Epitope-tagged (myc or HA) or untagged (-) versions of Gcr1p and Gcr2p were carried in tandem on the multicopy plasmid YEpTRP. The myc-tagged variant of Gcr1p was wild type (WT, lanes 1 and 3) or contained substitutions (L267T + L274T + L281T) for hydrophobic amino acids in 1LZ (L2T, lane 2; see Fig 3B). Protein from cell extracts was immunoprecipitated with -myc antibody. Eluted proteins were separated by SDS-PAGE, blotted to nitrocellulose, and subjected to Western analysis with -myc (top) or -HA (middle) antibody. Western analysis of HA-tagged Gcr2p from the crude extracts is shown at the bottom (input).

Simultaneous impairment of 1LZ and 2LZ creates a synthetic growth defect:
As part of the analysis of GCR2 function, we constructed a mutant strain in which gcr1 gcr2 (SD5) cells were transformed with a low-copy plasmid carrying GCR1^D (the 1LZ-deleted allele; DEMINOFF et al. 1995 ), expecting that combination of the gcr2 and GCR1^D mutations would result in a defect no worse than that caused by either mutation alone (see Introduction). We were intrigued to find that this GCR1^D gcr2 genotype (Fig 8A, segment 1), like the equivalent SD6 genotype (see MATERIALS AND METHODS), results in a synthetic growth defect nearly as severe as that of Gcr1^- strains (Fig 8A, segments 2 and 3; note that adding GCR2 to the gcr1 background does not improve growth). As expected, addition of both GCR1 and GCR2 to SD5 cells on the low-copy plasmid YCpTXZ restores wild-type growth (Fig 8A, segment 4), and addition of YCpTXZ-GCR1 or YCpTXZ-(GCR1^D+GCR2) partially improves the growth rate (Fig 8A, segments 5 and 6, respectively). Western analysis demonstrated that steady-state levels of the GCR1^D product are unaffected by removal of Gcr2p (not shown). Reduced Gcr1p levels therefore cannot explain the synthetic phenotype.

View larger version (166K): In this window In a new window Download PPT slide   Figure 8. Impairment of 1LZ and 2LZ causes a synthetic defect. (A) The double mutant SD5 (gcr1 gcr2) was transformed with YCpTXZ derivatives that conferred each of the indicated genotypes or with vector alone (see text for details). The equivalent mild growth defect of GCR1D GCR2 and GCR1 gcr2 cells, relative to wild type (GCR1 GCR2), is apparent by streaking onto SCD plates and incubating at 30° for 2 days. In contrast, the GCR1D gcr2 strain, which is genotypically analogous to the nearly isogenic strain SD6 (see MATERIALS AND METHODS), grows almost as slowly as gcr1 GCR2 cells. Note that the latter are phenotypically indistinguishable from the double null. (B) The growth defects described above are recapitulated at the level of point mutations in 1LZ and 2LZ. Strain SD5 (gcr1gcr2) was transformed with vector (YCpTXZ) alone or by carrying a combination of GCR1 and GCR2 alleles (either WT or with the GCR1L2T mutations, gcr2L2T mutations, or both; see Fig 3A and Fig C). Cells were streaked onto SCD plates and incubated at 30° for 2 days.

The effect of point mutations in 1LZ and 2LZ suggested that the synthetic defect might be explained by the combined loss of Gcr1p/Gcr1p and Gcr2p/Gcr2p homodimers. If so, it should be possible to reconstitute the defect with point mutations confined to the predicted hydrophobic surfaces of 1LZ and 2LZ. We therefore combined the GCR2 allele containing the L2T mutation in 2LZ with the analogous GCR1^L2T allele. As expected, the latter inactivates 1LZ (Fig 6) and causes a mild growth defect (Fig 8B); i.e., it phenocopies GCR1^D (DEMINOFF et al. 1995 ). Combination of GCR1^L2T and gcr2^L2T did indeed recapitulate the synthetic defect (Fig 8B).

We next sought an explanation for the synthetic phenotype by quantitating transcription of Rap1p/Gcr1p target genes. In isolation, deletion of 1LZ (GCR1^D) resulted in decreased mRNA levels of the ribosomal protein genes RPS2, RPS14A, and RPL11B (Fig 9, compare lanes 1 and 3), and had little or no effect on glycolytic transcripts (compare lanes 6 and 8). As shown previously (TORNOW et al. 1993 ; ZENG et al. 1997 ), the isolated gcr2 mutation had the opposite effect: there was little or no decrease in ribosomal protein gene transcription (Fig 9, compare lanes 1 and 2), while levels of the glycolytic transcripts ENO1, ENO2, and PYK1 were reduced (compare lanes 6 and 7). Combination of the GCR1^D and gcr2 mutations appears to result in a simple composite phenotype resembling that of gcr1 strains (Fig 9, compare lanes 4 and 5 and lanes 9 and 10). The decrease in steady-state levels of enolase, pyruvate kinase, and other glycolytic enzymes is equally severe in gcr2 and gcr1 backgrounds (UEMURA and FRAENKEL 1990 ; our unpublished data), confirming that this composite phenotype does not result from a synthetic loss of glycolytic enzyme activity.

View larger version (49K): In this window In a new window Download PPT slide   Figure 9. Removal of 1LZ or Gcr2p leads to distinct gene expression defects. Low-copy plasmids were introduced into strain SD6 to generate the indicated genotype (WT, wild type; D, deletion derivative lacking 1LZ; -, null) and GCR2 (WT, wild type; -, null). Ten micrograms of total RNA from each transformant was analyzed by primer extension. Lanes 1–5, primers capable of quantitating transcripts initiating in RPS2, RPS14A, RPL11B, or ACT1 (as a control) were used. Lanes 6–10, primers capable of quantitating transcripts initiating in ENO1, ENO2, PYK1, or ACT1 (as a control) were used.

Mutations in 2LZ eliminate Gcr1p/Gcr2p complex formation:
We next took advantage of a set of constructs, originally designed to analyze Gcr2p activation, which encode the lexAp DNA-binding domain fused to the N terminus of Gcr2p. We used this series of fusion constructs to test whether mutations in Gcr2p, specifically in 2LZ, eliminate coimmunoprecipitation with Gcr1p. Extracts from strains expressing myc-tagged or untagged Gcr1p and, on the same high-copy vector, variants of lexA-Gcr2p, all of which are expressed and stable (not shown), were immunoprecipitated with -myc antibody. The products encoded by lexA-GCR2 wild type, -GCR2^N, and -gcr2^2H were all coimmunoprecipitated by the antibody in the presence of the myc tag on Gcr1p (Fig 10, lanes 1–3), but not in its absence (lane 8). However, one deletion and several inactivating point mutations in 2LZ (lanes 4–7) eliminated association with Gcr1p. This is consistent with the idea that Gcr1p/Gcr2p complex formation is required for Gcr2p function.

View larger version (45K): In this window In a new window Download PPT slide   Figure 10. Mutations in 2LZ eliminate Gcr1p/Gcr2p association. Extracts from SD5 cells, transformed with a multicopy vector that carries both GCR1 [myc-tagged (myc, lanes 1–7) or with no tag (NT, lane 8)] and variants of lexA-GCR2 (see text and Fig 3A for a description of each GCR2 deletion and point mutation shown), were immunoprecipitated with -myc antibody. The immunoprecipitates were then analyzed by Western blotting using -lexA antibody, the results of which are shown. Each lexA-GCR2 variant encodes a stable product, as determined by Western blotting of the crude extracts used for immunoprecipitation (not shown). The full-length version of the R2E variant was unstable; a stable N derivative was used for this analysis. Its predicted position in lane 7 is indicated by an open arrow. The shaded arrow indicates the position of wild-type lexA-Gcr2p.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

We report here a molecular analysis of the leucine zipper motifs in Gcr1p (1LZ) and Gcr2p (2LZ). The L2T point mutations essentially substitute a hydroxyl for two methyl groups (Thr replacing Leu or Ile) at each of three positions in the hydrophobic interface of 1LZ or 2LZ (Fig 3). These L2T alleles cause a mild growth defect, comparable to that resulting from deletion of each corresponding leucine zipper domain (Fig 2, Fig 4, and Fig 8). The phenotypic consequences of the L2T lesions are therefore consistent with inactivation of the predicted -helical structures in Gcr1p and Gcr2p. Homodimerization of Gcr1p, and its loss due to L2T mutations in 1LZ, can be observed directly with band retardation analysis (DEMINOFF et al. 1995 ; Fig 6). Unfortunately, this approach is unsuccessful with Gcr2p. We have also tried several other biochemical methods to detect Gcr2p homodimerization, but they are hampered by the fact that, in gcr1 backgrounds, Gcr2p is rapidly degraded after cell lysis (unpublished data). The gcr1 background must be used in these experiments to avoid the possibility that Gcr1p can tether Gcr2p subunits to each other, which would mimic true multimerization. Nevertheless, the results of charge switch mutagenesis at the e and g positions in 1LZ and 2LZ (Fig 4 and Fig 5), including reciprocal second site reversion (GCR2^E2R+R2E) and intracistronic complementation (gcr2^E2R + gcr2^R2E and GCR1^E2K + GCR1^K2E), confirm the existence of the predicted 1LZ and 2LZ -helices (Fig 3).

Charge switching also determined the coiled-coil forms that contribute to GCR1 and GCR2 function in vivo. Both reciprocal second site reversion and intracistronic complementation are inconsistent with 1:1 heterodimerization between 1LZ and 2LZ. Conversely, all attempts at intergenic suppression were unsuccessful (gcr2^E2R or gcr2^R2E do not suppress the synthetic defect in SD6 cells containing GCR1^K2E or GCR1^E2K, respectively; not shown). These data suggest that 1LZ does not form a coiled coil with 2LZ (Fig 3B), or at least that such heteromultimerization does not contribute to function. Instead, both leucine zippers appear to homodimerize, generating Gcr1p and Gcr2p homodimers (Fig 3A and Fig C, respectively) that represent the entire contribution of each structure to the function of the corresponding polypeptide. Together with the band retardation assay shown in Fig 6, the simplest interpretation of which is that the LZ-mutated Gcr1p variants fail to dimerize, the data also suggest that Gcr1p can probably bind DNA as a monomer. Interaction with the CT box in DNA is essential to the specialized activation mechanism (ZENG et al. 1997 ), which is affected neither by deletion of 1LZ nor by the LZ-disruptive GCR1^L2T lesions (Fig 9 and our unpublished data).

On the basis of theoretical considerations, the data also appear to rule out the possibility that functional complexes involve 1:1 heterodimeric coiled-coil formation between 1LZ or 2LZ and leucine zippers in other proteins. Imagine replacing either 1LZ or 2LZ in Fig 3B with any heterologous leucine zipper: the conclusions drawn from reciprocal second site reversion and/or intracistronic complementation are identical. Charge switching, preferably confirmed by biochemical analysis, should therefore be a generally applicable method to distinguish homodimerization from heterodimerization. A possible limitation is that three-stranded or four-stranded coiled coils might be difficult to resolve with this genetic test; tetrameric coiled coils have been crystallized and may have functional significance in other systems (HARBURY et al. 1993 ). Nevertheless, it is unlikely that 1LZ and 2LZ are unusual structures in their sensitivity to charge switch analysis. Since e and g positions in coiled coils are known to harbor both attractive and repulsive effects on dimerization (HU et al. 1993 ), most leucine zipper complexes should be decipherable with this method. Charge switch analysis of leucine zippers should be possible even if disruption of the -helix in question causes no obvious phenotype; fusion of an LZ to the basic segment of a bZIP motif and measurement of the DNA-binding capacity of the resulting chimera were used successfully in an analogous study of the LR domain in Vpr, a 15 kD late viral gene product of HIV-1 (WANG et al. 1996 ).

Despite the progress in understanding 1LZ and 2LZ function, the surfaces in each polypeptide that mediate Gcr1p/Gcr2p complex formation remain unidentified. In Gcr1p, the residues that specify interaction with Gcr2p appear to lie entirely outside of 1LZ (which coincides with hypomutable region D; Fig 1) or are located redundantly within 1LZ and elsewhere: coimmunoprecipitation of Gcr2p by wild-type Gcr1p and the GCR1^L2T product are indistinguishable (Fig 7). Furthermore, the Gcr2p-dependent genes ENO1, ENO2, and PYK1 are affected little, if at all, in the GCR1^D GCR2 background (Fig 9), which also suggests that Gcr2p recruitment occurs in the absence of 1LZ. It is tempting to speculate that one or more of the remaining GCR1 hypomutable regions (A, B1, B3, or C; Fig 1) harbor the Gcr2p contact domain. For example, point mutations in hypomutable region A eliminate Gcr2p-dependent protection of GCR1 products made unstable by the Q15R lesion (our unpublished data). However, the possibility of an essential interaction between Gcr2p and residues C-terminal to 1LZ cannot yet be eliminated.

We have not yet successfully coimmunoprecipitated Gcr1p/Gcr2p complexes in the absence of 2LZ function (Fig 10). It is therefore possible that 2LZ is required for contact with Gcr1p; i.e., it contacts a domain in Gcr1p outside of 1LZ. The most likely way for this to occur would be through positions b, c, or f in 2LZ, or some combination thereof; these residues should remain exposed after 2LZ homodimerization. Of course, the idea that interaction with Gcr1p explains the requirement for 2LZ homodimers in GCR2 function remains only one of several possibilities.

A recent microarray filter hybridization experiment (LOPEZ and BAKER 2000 ) in which cells were grown on a nonfermentable carbon source (glycerol plus lactate) failed to identify several Gcr1p-dependent genes that we have consistently found are 2- to >20-fold less well transcribed in several different gcr1 backgrounds (RPS2, RPS14A, RPS3, RPL11B, RPL30, TEF1, TEF2, and ADH1; SANTANGELO and TORNOW 1990 ; TORNOW and SANTANGELO 1990 ; TORNOW et al. 1993 ; ZENG et al. 1997 ; Fig 9 and our unpublished data). ADH1 is a particularly striking example—it is even more strongly Gcr1p dependent in cells grown on pyruvate (>20-fold) than in those grown on glucose (7- to 8-fold), yet it fell below the 2-fold benchmark chosen by Lopez and Baker for GCR1 dependence. It would appear that their filter hybridization assay is not well suited for quantitative analysis; otherwise it is difficult to account for the dramatic discrepancy between their results and previously published data obtained with more sensitive methodologies. Perhaps a microchip analysis in which glucose, pyruvate, and/or other nonfermentable compounds are each tested as the carbon source would yield more accurate quantitation (BARTOSIEWICZ et al. 2000 ) and a clearer picture of the genome-wide effect of GCR1 deletion.

Our data rule out the idea that the gcr1 effect on translational component gene expression is an indirect consequence of a slow growth rate (HUIE et al. 1992 ), since GCR1^D GCR2 strains are only mildly growth defective (Fig 8A; DEMINOFF et al. 1995 ) but exhibit a gcr1-like loss of RPS2, RPS3, RPS14A, RPL11B, and RPL30 and TEF1 and TEF2 transcription (Fig 9, compare lanes 3 and 5; our unpublished data). Moreover, reduced glycolytic enzyme levels are insufficient to explain the gcr1 growth phenotype, since similar reductions in gcr2 cells have a much milder effect on growth (UEMURA and FRAENKEL 1990 ). Therefore, although the central role of UAS_RPG-bound Rap1p in glycolytic gene expression may indeed be to recruit Gcr1p/Gcr2p complexes to the adjacent CT box and permit a specialized mode of activation, it is equally true that Rap1p independently stimulates transcription of translational component genes by recruiting Gcr1p in the absence of a CT box. The disparity in the requirements for Gcr1p and Gcr2p domains in these two mechanisms of activation could not be more striking (Table 1). We therefore remain convinced (SANTANGELO and TORNOW 1990 ; TORNOW et al. 1993 ; ZENG et al. 1997 ) that Rap1p potentiates transcription at UAS_RPG sites by doing more than just clearing chromatin so that other activator proteins can bind DNA (YU and MORSE 1999 ).

The mechanistic dichotomy of Rap1p/Gcr1p function at translational component and glycolytic genes suggests that independent control of these two classes of regulatory targets is a useful feature of growth control. Recent analysis of genome-wide expression patterns suggests that, under the conditions tested so far, most translational component and glycolytic genes are coregulated. For example, coordinated downshifts appear to occur during the diauxic transition, heat-shock response, and sporulation time courses (DERISI et al. 1997 ; EISEN et al. 1998 ). It remains to be seen if a regulatory scenario can be found in which the CT-dependent and CT-independent modes of Rap1p/Gcr1p activation respond differentially. Regardless, it is probably significant that the Gcr1^- condition appears to be a unigenic synthetic phenotype: the combined Gcr1p-stimulated functions of translational component and glycolytic genes have a synergistic effect on both the growth rate (Fig 8) and the protein synthesis rate (our unpublished data). This arrangement should allow fine-tuned growth regulation in response to nutrient availability and other stimuli over a wide range. We are therefore particularly interested in the possibility that phosphorylation of Gcr1p or Gcr2p (or both) through the Ras signal transduction pathway is an important feature of Rap1p transcriptional activation (see KLEIN and STRUHL 1994 ; ZENG et al. 1997 ). Given its demonstrated effect on Cln expression (TOKIWA et al. 1994 ), the Ras pathway may therefore represent the predominant regulatory link between the growth and cell cycles.

These results extend our working model for Gcr2p function, that a Gcr2p-dependent alteration to Rap1p/Gcr1p complexes mediates Gcr1p activation in the presence of a combinatorial UAS_RPG/CT box element (ZENG et al. 1997 ). The data that we report here, especially the identification of a synthetic growth defect resulting from a combination of mutations that inactivate 1LZ (loss of translational component gene expression) and 2LZ (loss of glycolytic expression), confirm the idea that Rap1p activates transcription at UAS_RPG through two distinct mechanisms. An updated version of our working model is depicted in Fig 11. This model explains why the gcr1 and gcr2 mutations, which result in an equivalent decrease in glycolytic enzyme levels, cause growth defects that differ greatly in magnitude. In the absence of a CT box adjacent to UAS_RPG, 1LZ (unlike the CT-binding domain or Gcr2p) is essential; in the presence of a CT box, those requirements are neatly reversed. A full understanding of these observations should elucidate other important features of this system, perhaps including the mystery surrounding the other duality in Rap1p function: its capacity to act as both an activator and a repressor of transcription.

View larger version (53K): In this window In a new window Download PPT slide   Figure 11. A model for general and specialized Rap1p/Gcr1p transcriptional activation. Rap1p activation of translational component and glycolytic genes requires recruitment of Gcr1p and Gcr2p homodimers, respectively. An upstream Rap1p DNA-binding site (UASRPG) is required in both the general (translational component) and specialized (glycolytic) mechanisms; the Gcr1p DNA-binding site (CT box) is required only in specialized activation. The leucine zipper motifs in Gcr1p (1LZ) and Gcr2p (2LZ) mediate homodimerization in each protein (see text). The depicted absence of a Gcr1p dimer from the specialized mechanism is for the sake of simplicity and does not necessarily imply that Gcr2p is normally recruited by monomers and not homodimers of Gcr1p. Arrows indicate the separate requirements for 1LZ in the general mechanism and for 2LZ in the specialized mechanism; they are not meant to imply that these are the activation domains in Gcr1p and Gcr2p. N, the N-terminal hypomutable regions of Gcr1p, which are required in both mechanisms and appear to collaboratively form the Gcr1p activation domain (TORNOW et al. 1993 ; DEMINOFF et al. 1995 ).

	ACKNOWLEDGMENTS

We thank Joanne Tornow for her invaluable help and advice throughout the beginning stages of this work. We also thank Martha Sparrow and Linda Cimbora for outstanding technical support, and the other members of the Santangelo laboratory, past and present, for helpful discussions. This work was supported by a National Science Foundation grant (MCB-9974636) to G.M.S.

Manuscript received November 6, 2000; Accepted for publication February 9, 2001.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

BAKER, H. V., 1991  GCR1 of Saccharomyces cerevisiae encodes a DNA binding protein whose binding is abolished by mutations in the CTTCC sequence motif. Proc. Natl. Acad. Sci. USA 88:9443-9447[Abstract/Free Full Text].

BARTOSIEWICZ, M., M. TROUNSTINE, D. BARKER, R. JOHNSTON, and A. BUCKPITT, 2000  Development of a toxicological gene array and quantitative assessment of this technology. Arch. Biochem. Biophys. 376:66-73[Medline].

BURNETTE, W. N., 1981  "Western blotting": electrophoretic transfer of proteins from sodium dodecyl sulfate-polyacrylamide gels to unmodified nitrocellulose and radiographic detection with antibody and radioiodinated protein A. Anal. Biochem. 112:195-203[Medline].

DEMINOFF, S. J., J. TORNOW, and G. M. SANTANGELO, 1995  Unigenic evolution: a novel genetic method localizes a putative leucine zipper that mediates dimerization of the Saccharomyces cerevisiae regulator Gcr1p. Genetics 141:1263-1274[Abstract].

DERISI, J. L., V. R. IYER, and P. O. BROWN, 1997  Exploring the metabolic and genetic control of gene expression on a genomic scale. Science 278:680-686[Abstract/Free Full Text].

EISEN, M. B., P. T. SPELLMAN, P. O. BROWN, and D. BOTSTEIN, 1998  Cluster analysis and display of genome-wide expression patterns. Proc. Natl. Acad. Sci. USA 95:14863-14868[Abstract/Free Full Text].

GENTZ, R., F. J. RAUSCHER, III, C. ABATE, and T. CURRAN, 1989  Parallel association of Fos and Jun leucine zippers juxtaposes DNA binding domains. Science 243:1695-1699[Abstract/Free Full Text].

HALBACH, T., N. SCHEER, and W. WERR, 2000  Transcriptional activation by the PHD finger is inhibited through an adjacent leucine zipper that binds 14-3-3 proteins. Nucleic Acids Res. 28:3542-3550[Abstract/Free Full Text].

HARBURY, P. B., T. ZHANG, P. S. KIM, and T. ALBER, 1993  A switch between two-, three-, and four-stranded coiled coils in GCN4 leucine zipper mutants. Science 262:1401-1407[Abstract/Free Full Text].

HU, J. C., N. E. NEWELL, B. TIDOR, and R. T. SAUER, 1993  Probing the roles of residues at the e and g positions of the GCN4 leucine zipper by combinatorial mutagenesis. Protein Sci. 2:1072-1084[Medline].

HUIE, M. A., E. W. SCOTT, C. M. DRAZINIC, M. C. LOPEZ, and I. K. HORNSTRA et al., 1992  Characterization of the DNA-binding activity of GCR1: in vivo evidence for two GCR1-binding sites in the upstream activating sequence of TPI of Saccharomyces cerevisiae. Mol. Cell. Biol. 12:2690-2700[Abstract/Free Full Text].

KLEIN, C. and K. STRUHL, 1994  Protein kinase A mediates growth-regulated expression of yeast ribosomal protein genes by modulating RAP1 transcriptional activity. Mol. Cell. Biol. 14:1920-1928[Abstract/Free Full Text].

LANDSCHULZ, W. H., P. F. JOHNSON, and S. L. MCKNIGHT, 1988  The leucine zipper: a hypothetical structure common to a new class of DNA binding proteins. Science 240:1759-1764[Abstract/Free Full Text].

LOPEZ, M. C. and H. V. BAKER, 2000  Understanding the growth phenotype of the yeast gcr1 mutant in terms of global genomic expression patterns. J. Bacteriol. 182:4970-4978[Abstract/Free Full Text].

O'SHEA, E. K., R. RUTKOWSKI, W. F. STAFFORD, III, and P. S. KIM, 1989  Preferential heterodimer formation by isolated leucine zippers from fos and jun. Science 245:646-648[Abstract/Free Full Text].

PTASHNE, M. and A. GANN, 1997  Transcriptional activation by recruitment. Nature 386:569-577[Medline].

ROSE, M. D., P. NOVICK, J. H. THOMAS, D. BOTSTEIN, and G. R. FINK, 1987  A Saccharomyces cerevisiae genomic plasmid bank based on a centromere-containing shuttle vector. Gene 60:237-243[Medline].

SANTANGELO, G. M. and J. TORNOW, 1990  Efficient transcription of the glycolytic gene ADH1 and three translational component genes require the GCR1 product, which can act through TUF/GRF/RAP binding sites. Mol. Cell. Biol. 10:859-862[Abstract/Free Full Text].

SHORE, D., 1994  RAP1: A protean regulator in yeast. Trends Genet. 10:408-412[Medline].

TOKIWA, G., M. TYERS, T. VOLPE, and B. FUTCHER, 1994  Inhibition of G1 cyclin activity by the Ras/cAMP pathway in yeast. Nature 371:342-345[Medline].

TORNOW, J. and G. M. SANTANGELO, 1990  Efficient expression of the Saccharomyces cerevisiae glycolytic gene ADH1 is dependent upon a cis-acting regulatory element (UAS_RPG) found initially in genes encoding ribosomal proteins. Gene 90:79-85[Medline].

TORNOW, J., X. ZENG, W. GAO, and G. M. SANTANGELO, 1993  GCR1, a transcriptional activator in Saccharomyces cerevisiae, complexes with RAP1 and can function without its DNA binding domain. EMBO J. 12:2431-2437[Medline].

TRIEZENBERG, S. J., 1992 Primer extension, pp. 4.8.1–4.8.5 in Current Protocols in Molecular Biology, edited by V. B. CHANDA. John Wiley & Sons, New York.

UEMURA, H. and D. G. FRAENKEL, 1990  gcr2, a new mutation affecting glycolytic gene expression in Saccharomyces cerevisiae. Mol. Cell. Biol. 10:6389-6396[Abstract/Free Full Text].

UEMURA, H. and Y. JIGAMI, 1992  Role of GCR2 in transcriptional activation of yeast glycolytic genes. Mol. Cell. Biol. 12:3834-3842[Abstract/Free Full Text].

UEMURA, H. and Y. JIGAMI, 1995  Mutations in GCR1, a transcriptional activator of Saccharomyces cerevisiae glycolytic genes, function as suppressors of gcr2 mutations. Genetics 139:511-521[Abstract].

WANG, L., S. MUKHERJEE, O. NARAYAN, and L.-J. ZHAO, 1996  Characterization of a leucine-zipper-like domain in Vpr protein of human immunodeficiency virus type 1. Gene 178:7-13[Medline].

YU, L. and R. H. MORSE, 1999  Chromatin opening and transactivator potentiation by RAP1 in Saccharomyces cerevisiae. Mol. Cell. Biol. 19:5279-5288[Abstract/Free Full Text].

ZENG, X., S. J. DEMINOFF, and G. M. SANTANGELO, 1997  Specialized Rap1p/Gcr1p transcriptional activation through Gcr1p DNA contacts requires Gcr2p, as does hyperphosphorylation of Gcr1p. Genetics 147:493-505[Abstract].

This article has been cited by other articles:

				M. Bendjennat and P. A. Weil The Transcriptional Repressor Activator Protein Rap1p Is a Direct Regulator of TATA-binding Protein J. Biol. Chem., March 28, 2008; 283(13): 8699 - 8710. [Abstract] [Full Text] [PDF]

				K. A. Garbett, M. K. Tripathi, B. Cencki, J. H. Layer, and P. A. Weil Yeast TFIID Serves as a Coactivator for Rap1p by Direct Protein-Protein Interaction Mol. Cell. Biol., January 1, 2007; 27(1): 297 - 311. [Abstract] [Full Text] [PDF]

				G. M. Santangelo Glucose Signaling in Saccharomyces cerevisiae Microbiol. Mol. Biol. Rev., March 1, 2006; 70(1): 253 - 282. [Abstract] [Full Text] [PDF]

				B. B. Menon, N. J. Sarma, S. Pasula, S. J. Deminoff, K. A. Willis, K. E. Barbara, B. Andrews, and G. M. Santangelo Reverse recruitment: The Nup84 nuclear pore subcomplex mediates Rap1/Gcr1/Gcr2 transcriptional activation PNAS, April 19, 2005; 102(16): 5749 - 5754. [Abstract] [Full Text] [PDF]

				I. Gaziova, P. C. Bonnette, V. C. Henrich, and M. Jindra Cell-autonomous roles of the ecdysoneless gene in Drosophila development and oogenesis Development, June 1, 2004; 131(11): 2715 - 2725. [Abstract] [Full Text] [PDF]

				K. A. Willis, K. E. Barbara, B. B. Menon, J. Moffat, B. Andrews, and G. M. Santangelo The Global Transcriptional Activator of Saccharomyces cerevisiae, Gcr1p, Mediates the Response to Glucose by Stimulating Protein Synthesis and CLN-Dependent Cell Cycle Progression Genetics, November 1, 2003; 165(3): 1017 - 1029. [Abstract] [Full Text] [PDF]

				Z. Wang, D.-Z. Wang, G. C. T. Pipes, and E. N. Olson Myocardin is a master regulator of smooth muscle gene expression PNAS, June 10, 2003; 100(12): 7129 - 7134. [Abstract] [Full Text] [PDF]

				G. Fourel, T. Miyake, P.-A. Defossez, R. Li, and E. Gilson General Regulatory Factors (GRFs) as Genome Partitioners J. Biol. Chem., October 25, 2002; 277(44): 41736 - 41743. [Abstract] [Full Text] [PDF]