- Email: [email protected]
Conservation of a putative inhibitory domain in the GAL4 family members
Gene 184 (1997) 229–235
Conservation of a putative inhibitory domain in the GAL4 family members Olivier Poch * U.P.R. Me´canismes Mole´culaire de la Division Cellulaire et du De´veloppement, Institut de Biologie Mole´culaire et Cellulaire, 15 rue Rene´ Descartes, 67084 Strasbourg, France Received 24 May 1996; revised 11 July 1996; accepted 12 July 1996
Abstract The GAL4 family members are fungal transcriptional activators composed of several functional domains: a characteristic cysteine-rich DNA-binding domain common to all members, a dimerization domain, various transactivation domains generally exhibiting a high acidic content and a highly variable central region supposed to be involved in regulation and in effector recognition. We report here that the central region of the GAL4 family members share eight conserved motifs embedded in a large functional domain of 225 up to 405 residues. This domain may also be present in four proteins belonging to another family of transcriptional activators sharing a C2H2-type zinc finger. Analysis of the biochemical data available on the well-studied GAL4 protein suggests that this domain may be involved in the regulation of the activity of the protein, particularly in an inhibitory function. This hypothesis is further supported by deletion and site-directed mutagenesis experiments on other GAL4 family members. The mean secondary structure prediction performed on the eight motifs strongly suggests that the inhibitory activity may be mediated by hydrophobic interactions linked to the presence of amphipathic a-helices. Keywords: Fungal transcriptional activator; GAL4 family member; GAL4; PUT3; LEU3; PPR1; PDR1; PDR3; ADR1; Inhibitory domain; Sequence conservation; Saccharomyces cerevisiae
1. Introduction The GAL4 family members (GFM ) are fungal transcriptional regulatory proteins (Dahwale and Lane, 1993; Svetlov and Cooper, 1995) which share a characteristic cysteine-rich DNA-binding domain. They are involved in the regulation of numerous anabolic and catabolic pathways (i.e., galactose, arginine, proline, maltose and leucine metabolisms) as well as in the regulation of various pathways such as pleiotropic drug resistance, chromosome segregation or sporulation processes ( Table 1). Frequently, the activity of these transcriptional activators is known to be stimulated by the presence of a small molecule such as a metabolic
* Corresponding author. Abbreviations: S. cerevisae, Saccharomyces cerivisiae; S. carlbergensis, Saccharomyces carlbergensis; K. lactis, Kluyveromyces lactis; S. pombe, Schizosaccharomyces pombe; A. nidulans, Aspergillus nidulans; A. oryzae, Aspergillus oryzae; N. crassa, Neurospora crassa; C. albicans, Candida albicans; L. edodes, Lentinus edodes.
intermediate or a drug. In addition, with the totality of S. cerevisiae sequences available, a large set of potential GFM has been determined whose function is generally unknown (see Table 1). At the sequence level, GFM are large proteins (ranging from 465 up to 1502 residues) with a characteristic DNA-binding domain composed of a highly conserved 28-residue cysteine-rich motif (C6-type zinc finger) (Svetlov and Cooper, 1995) generally located in the most amino-terminal region. The three-dimensional structure of the cysteine-rich motif of two GFM (GAL4 and PPR1) revealed an essentially identical structure in which two Zn2+ions are coordinated to six invariant cysteine residues, with the rest of the amino acids folded around the Zn2Cys6 (Baleja et al., 1992; Kraulis et al., 1992; Marmorstein et al., 1992; Marmorstein and Harrison, 1994). In addition to a common DNA-binding domain, GFM share general functional organization with (in the immediate carboxy-terminal side of the zinc cluster) a coiled-coil dimerization domain with heptad repeats of hydrophobic residues (Reece and Ptashne, 1993; Zhang et al., 1993; Sua`rez et al., 1995). The
0378-1119/97/$17.00 Copyright © 1997 Elsevier Science B.V. All rights reserved PII S 03 7 8 -1 1 1 9 ( 9 6 ) 0 0 6 02 - 6
230
O. Poch/Gene 184 (1997) 229–235
transcriptional activation function is assured by various regions without extensive sequence conservation but generally with a high proportion of negatively charged residues. Frequently, the most potent activation region is located in the most carboxy-terminal portion of each activator while some regions, scattered in the central regions, can exhibit weaker activation functions (Zhou and Kohlhaw, 1990). Finally, transcriptional activation is regulated by the central regions which represent the major part of the proteins (more than 60% of the total mass of the factors) and are involved in more specialized functions such as recognition of peptide cofactors or metabolic intermediates. These central regions have been observed as highly variable when closely related GFM are compared and, up to now, only limited regions have been reported to be conserved between some GFM (Salmeron and Johnston, 1986; Wray et al., 1987; Chasman and Kornberg, 1990; Marczak and Brandriss, 1991).
Here we report that at least eight motifs composed of 161–167 residues are conserved in the variable central region of almost all reported GFM. In addition, these eight motifs could be detected in the sequences of four fungal transcriptional factors belonging to the C2H2-type zinc finger family further highlighting their general conservation in fungal transcriptional factors. The conservation of eight motifs spanning from 225 up to 405 residues strongly suggests their cooperation in a well-defined structural and functional domain. At the functional level, analysis of the mutational data available on various GFM and of biochemical results concerning the well-studied GAL4 factor strongly suggests that the conserved domain is involved in inhibitory function probably mediated by hydrophobic interactions. The secondary structures present in the domain, as deduced from the mean secondary structure predictions performed, are discussed in the context of their possible role in regulation of protein activity.
O. Poch/Gene 184 (1997) 229–235
231
Fig. 1. Multiple alignment of the putative inhibitory domains of the GAL4 family members (GFM ). This alignment was obtained by initial multiple alignment, performed by the PILEUP program (Devereux et al., 1984), of a set of transcriptional activator sequences (names underlined) and analysis of the pairwise sequence comparisons of closely related factors (i.e., Ppr1-Yeast and Andnauay-1; Gal4-Yeast and Lac9-Klula; Nira-Emeni and Nit4-Neucr; Put3-Yeast and Thi1-Schpo, Quta-Emeni and Qalf-Neucr). After manual adjustment, eight highly conserved regions were defined and then used to construct profiles (Gribskov et al., 1987, 1988) (‘fuzzy’ probes constituted from a group of aligned sequences) based on a single motif or by a concatenation of some or all of the eight motifs (named I–VIII ). Within a motif, the gap and gap-length penalties were defined as 5 and 0.2, respectively. One undetermined residue was introduced between each motif to allow the non-conserved interregions separating each motif to vary without constraint during the profile alignment process. Their values for gap and gap-length penalties were both defined as 0.03. These profiles were first validated and then used for further database scans of the Protein Information Resource (Pir), SwissProt (Sw) and translated EMBL ( TrEmbl ) protein databanks. The 47 GFM presented (upper part of the figure) and the four amino acid sequences belonging to the C2H2-type members ( lower part of the figure) were detected at a statistically significant level. The distances between the conserved motifs and distances from the termini are indicated. The consensus, at the top, shows residues conserved in more than 70% of the GFM sequences (bold in the alignment) and indicated by: @ for aromatic residues (F, Y, W ), h for bulky aliphatic or aromatic residues (I, L, M, V, F, Y, W ), G for small residues (P, A, G, S, T ), D for negatively charged residues (D, E, Q, N ). The secondary structure prediction (sec. struc.) shows the consensus of the predictions for individual GFM sequences obtained using the PHD program (Rost and Sander, 1993); a designates an a-helix; b designates a b-strand.
2. Experimental and discussion 2.1. Transcriptional activators presenting the eight conserved motifs Starting from the observation of limited sequence conservation in the central regions of a subset (Put3p,
Gal4p, Lac9p, Ppr1p, Pdr1p, QutAp and Leu3p) of GFM (Salmeron and Johnston, 1986; Chasman and Kornberg, 1990; Marczak and Brandriss, 1991), we initiated a computer analysis combining various sequence analysis tools such as pairwise sequence comparisons, multiple sequence comparisons and profile scans (see legend of Fig. 1). This allowed us to delineate
232
O. Poch/Gene 184 (1997) 229–235
Table 1 Gene names, organisms and functions of the GAL4 family members (GFM ) present in the protein databases and the four C2H2-type members detected by the profile scans Gene name or ORF
Organism
Database source and mnemonic
Accession No.
Regulation of
GAL4 LAC9 PPR1 CYP1 LEU3 PDR1 PDR3 CHA4 CAT8 PUT3 DAL81 CBF3B MAL3R/MAL33 MAL63 TEA1 ARGR2 LYS14 FUN43/YAL051W SIP4 YBL066C YBR150C YBR033W YCR106W 9740.13 9461.10 YER184C YFL052W YHR178W YIL130W YJL206c YKR064W YKL038W YKL222C L8003.10 L8479.13 L0584 YM9711.08 00938 9677.10 NTF1 SPAC1F7.11c QUTA UaY amdR NIRA NIT4 qalF priBc SUC1 ADR1 ZMS1 amdA Sc9367.02
S. cerevisiae K. lactis S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. carlsbergensis S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. pombe S. pombe A. nidulans A. nidulans A. oryzae A. nidulans N. crassa N. crassa L. edodes C. albicans S. cerevisiae S. cerevisiae A. nidulans S. cerevisiae
Sw:Gal4-Yeast Sw:Lac9-Klula Sw:Ppr1-Yeast Sw:Cyp1-Yeast Sw:Leur-Yeast Sw:Pdr1-Yeast Sw:Pdr3-Yeast Sw:Cha4-Yeast Sw:Cat8-Yeast Sw:Put3-Yeast Sw:Da81-Yeast Sw:Cb32-Yeast Sw:Ma3r-Yeast Sw:Ma6r-Yeast Sw:Tea1-Yeast Pir : S49627 Sw:Ly14-Yeast Sw:Yaf1-Yeast Sw:Sip4-Yeast Sw:Ybg6-Yeast Sw:Yb00-Yeast Sw:Ybo3-Yeast Sw:Ycz6-Yeast TrEmbl:Scl9740-9 TrEmbl:Sc94611-10 Sw:Ye14-Yeast Sw:Yff2-Yeast Sw:Yhx8-Yeast Sw:Yin0-Yeast Sw:Yju6-Yeast Sw:Yk44-Yeast Sw:Ykd8-Yeast Sw:Ykw2-Yeast Pir:S50366 Pir:S51403 Pir:S50966 Pir:S54020 Pir:S50406 TrEmb1:Sc19677-11 Sw:Thi1-Schpo Sw:Yakb-Schpo Sw:Quta-Emeni Sw:Uay-Emeni Sw:Amdr-Aspor Sw:Nira-Emeni Sw:Nit4-Neucr Sw:Qa1f-Neucr Sw:Prib-Lened Sw:Suc1-Canal Sw:Adr1-Yeast Sw:Zms1-Yeast TrEmbl:Andbp-1 TrEmbl:Sc9367-2
P04386 P08657 P07272 P12351 P08638 P12383 P33200 P43634 P39113 P25502 P21657 P40969 P39157 P10508 P47988 S49627 P40971 P39720 P46954 P34228 P38114 P38073 P25611 U28374 U33007 P39961 P43551 P38699 P40467 P39529 P36023 P32862 P35995 S50366 S51403 S50966 S54020 S50413 U25841 P36598 Q09922 P10563 P49413 Q06157 P28348 P28349 P11638 P49412 P33181 P07248 P46974 L28810 Z49274
Galactose metabolism Lactose metabolism Pyrimidine metabolism iso-1 and iso-2 cytochrome c expression Leucine biosynthesis Pleiotropic drug resistance Pleiotropic drug resistance Serine and threonine catabolism Gluconeogenic pathway Proline metabolism GABA, urea, arginine and allantoin catabolism Chromosome segregation Maltose metabolism Maltose metabolism Ty1 enhancer activator Arginine metabolism Lysine biosynthesis Unknown, chr. I seq. Unknown, chr. X seq. Sporulation processes, chr. II seq. Unknown, chr. II seq. Unknown, chr. II seq. Unknown, chr. III seq. Unknown, chr. IV seq. Unknown, chr. IV seq. Unknown, chr. V seq. Unknown, chr. VI seq. Unknown, chr. VIII seq. Unknown, chr. IX seq. Unknown, chr. X seq. Unknown, chr. XI seq. Unknown, chr. XI seq. Unknown, chr. XI seq. Unknown, chr. XII seq. Unknown, chr. XII seq. Unknown, chr. XII seq. Unknown, chr. XIII seq. Unknown, chr. XV seq. Unknown, chr. XVI seq. Thiamine biosynthesis Unknown, chr. I Quinate metabolism Purine catabolism Amide catabolism Nitrate assimilation Nitrate assimilation Quinate metabolism Unknown Sucrose utilization Alcohol dehydrogenase repressor Unknown Acetamidase gene Unknown
Underlined and bold gene names (ARGR2 and LYS14) indicate the two GFM sequences that are not detected at a statistically significant level by profile scans. Mnemonics are from the SwissProt database (Sw) Release 26.3, the translated EMBL ( TrEmbl ) Release 45, and from the Protein Information Resource (Pir) release 46. Abbreviations: chr., chromosome; seq., sequencing project.
O. Poch/Gene 184 (1997) 229–235
eight conserved blocks (motifs I–VIII ) involving 161–167 residues and separated by more divergent spacer regions (Fig. 1). Considering the inter-motif sequences, the eight concatenated motifs may constitute an entire domain which can span from 225 residues for the shortest sequence (Lelipri2-1 in Fig. 1) up to 405 residues for the longest one (Cyp1-Yeast). The profile scans performed with a concatenation of the eight motifs could detect at higher scores 47 out of 49 GFM available in the protein databases including numerous sequences of unknown function arising from the yeast sequencing project ( Table 1). Two transcriptional activators, ArgrIIp and Lys14p, could not be detected at significant levels. It should be pointed out that, among GFM, the Lys14p and ArgrIIp proteins appear exceptional as seen by the functional analysis performed on these proteins. In contrast to other yeast regulatory proteins, 92% of ArgrIIp is necessary for its anabolic repression function and 80% is necessary for its catabolic activator function (Qui et al., 1991) suggesting that almost the entire protein is required for efficient regulation. Similarly, internal deletion experiments on the Lys14p failed to identify limited segments which were sufficient for the activation function ( Feller et al., 1994). In this context, the absence of the eight motifs in these two unusual proteins is in good agreement with the biochemical data and further emphasizes the significance of the proposed alignment. The profile scans performed on all protein databases detect at an ‘interesting’ score level (Gribskov et al., 1987) of approx. 2 S.D.s above the mean, four fungal transcriptional activators exhibiting a DNA-binding motif distinct from the typical cysteine-rich motif present in the GFM sequences, namely a C2H2-type zinc finger (four last sequences in Fig. 1). A refined analysis of the sequences of these C2H2-type members strongly supports the idea of a structural link between the four proteins detected and the GFM. Indeed, among the C2H2-type members, amdAp, Zms1p, Adr1p and Sc9367p appear quite remarkable by two specific structural features: first, with respect to their length (880, 1380, 1323 and 1133 amino acid long, respectively), these proteins are longer than most of the C2H2-type members (from 211 up to 914 amino acids) and second, their DNA-binding domain is located in the aminoterminal region (amino acid positions 25, 153, 106 and 33 for the amdAp Zms1p Adr1p and Sc9367p, respectively). This location is unusual since the DNA-binding domains of most of the C2H2-type members are positioned in the central or in the most carboxy-terminal region of the proteins. Thus, these results strongly suggest that the four atypical fungal C2H2-type members are structurally linked to the GFM. These sequence relationships probably arise from a modular evolution which supposes that either an ancestral C2H2-type member had acquired the region encompassing the eight
233
motifs or that the four proteins have evolved from ancestral GFM by the acquisition of a C2H2-type zinc finger.
2.2. Sequence analysis of the eight conserved motifs The eight matched regions correspond to those of highest homology when closely related GFM are compared separately enhancing the significance of the sequence similarities found. Analyzing the GFM sequences detected through the profile scans, we introduced one limited gap at position 17 of motif I to increase the similarity. The resulting alignment includes 53 conservatively maintained residues in more than 70% of the GFM sequences. Among these conserved residues, numerous hydrophobic and aromatic residues (45 out of 53) are observed reflecting the overall composition of the central regions of the GFM which are mainly hydrophobic. The longest stretch of conserved amino acids is constituted by five consecutive residues present in motif VI (positions 7–11) defining a pattern of two basic residues followed by an aliphatic-aromatic-aliphatic sequence. In addition, at position 17 of the motif, a negatively charged residue is frequently observed implying that among the 53 conservatively maintained residues, all the charged residues are clustered in motif VI. Global inspection of Fig. 1 revealed a great variability in the lengths separating the different motifs. Such variability was also observed in closely related pairwise sequences (e.g. Ppr1-Yeast/Andnauy-1; Cyp1-Yeast/Yk44-Yeast or Pdr3-Yeast-Ye14-Yeast in Fig. 1) which can vary widely in their inter-motif distances. In addition, it should be pointed out that the amino acid compositions of the largest inter-motif sequences are frequently biased with a high content in acidic residues or in small neutral residues such as proline, glycine or serine residues. Biased compositions are noteworthy in the sequences separating motifs IV and V and motifs V and VI in the Cyp1-Yeast and Scd94611-10 amino acid sequences which exhibit an extremely high content in acidic residues. The variability and importance of the inter-motif distances clearly reflect an extended plasticity of the central region of the GFM with preservation of the eight motifs scattered between variable sequences in which more specialized functions, such as transactivation for example, may take place. Finally, the most conserved motifs are IV, V and VI which correspond roughly to the three regions previously appointed by Marczak and Brandriss (1991). Among the five remaining motifs, motif III is the least conserved while hydrophobic residues preserved in motifs IV, VII and VIII frequently display a 3–4 periodicity. In addition to GFM already recognised as closely related, analysis of the percentage of sequence conservation existing between the distinct members allows to cluster one group composed of 27
234
O. Poch/Gene 184 (1997) 229–235
sequences (27 upper sequences from Ybg6-Yeast to Qalf-Neucr in Fig. 1). 2.3. Structural and functional analysis of the alignment The preservation in such widely divergent proteins of eight concatenated motifs, spanning from 225 up to 405 residues, suggests their cooperative implication in a welldefined structural and functional domain. At the structural level, the mean secondary structure predictions performed (see legend of Fig. 1) suggest that the motifs are mainly composed of a-helices. The only non-a structure predicted is located in the most conserved motif VI, in which a b-turn-b structure is strongly predicted. This b-turn-b structure may maintain the highly conserved acidic residue (position 17) near the turn structure. It should be noted that in most of the predicted a-helices, a 3–4 periodicity of the conservatively maintained hydrophobic residues can be observed (motifs I, II, IV, VII and VIII ). In motifs IV, VII and VIII, conserved heptad repeats are observed suggesting the existence of stable amphipathic a-helices preserved throughout the whole domain. The presence of these amphipathic a-helices strongly suggests that hydrophobic interactions may take place in the conserved domain. Such hydrophobic interactions may play an important role in maintaining the general topology of the domain and/or in interactions with other proteins involved in the regulation of the activity of these fungal transcriptional activators (see below). At the functional level, the numerous functional analyses performed on the well-studied GAL4 protein strongly suggest the involvement of the central region in regulation of the transcriptional activity (Gill and Ptashne, 1987; Johnston et al., 1987; Ma and Ptashne, 1987a,b; Johnston and Dover, 1988; Chasman and Kornberg, 1990; Griggs and Johnston, 1991). Recently, it has been shown that the conserved domain defined by the eight motifs coincides almost exactly to a region of Gal4p corresponding to the so-called ‘inhibitory domain’ (Stone and Sadowski, 1993). These authors proposed that an inhibitory domain may bind an inhibitory protein through a major conformational change induced by the fixation of glucose on an adjacent glucose response domain. The inhibitory domain (ID) delineated by Stone and Sadowski (1993) corresponds to positions 236–614 of the GAL4 protein which fully encompasses the eight motifs shown in Fig. 1. This large ID could be subdivided in three consecutive smaller regions (ID1–ID3) corresponding to residues 320–412, 412–478 and 554–585, respectively. With respect to the eight motifs, ID1 corresponds almost exactly to the three most conserved motifs IV, V and VI since it begins one residue upstream from motif IV and ends five residues downstream. ID2 is located between motifs VI and VII while ID3 begins 21 residues upstream from motif VIII
up to the proline at position 17 of this motif. The correlation observed between these biochemical data concerning the central region of Gal4p and the eight motifs delineated by sequence analysis further highlights the significance of the proposed alignment. As a corollary, this strongly suggests that the domain delineated by the eight motifs may be involved in an inhibitory function common to all the GFM and C2H2-type members sharing these regions. The fact that this domain may be involved in negative regulation is also consistent with studies on the LEU3 transactivator. A large deletion of the central region of Leu3p renders the protein constitutively active ( Friden et al., 1989; Sze et al., 1993) suggesting that an inhibitory effect has been deleted from the central region. The functional importance of the domain encompassing the eight motifs is in agreement with internal deletion experiments on various GFM. In the HAP1/CYP1 transactivator (CYP1-yeast) (Pfeifer et al., 1989) or in the amdR (Aoamdr-1) (Parsons et al., 1992), internal deletions of all or some of the motifs led to the loss of inducibility, highlighting the requirement of this region for a full and wellregulated function of the transactivators. With respect to site-directed mutagenesis, it is noteworthy that a single replacement in Put3p of the highly conserved glycine by an aspartate residue in position 15 of motif II leads to a noninducible allele. Finally, functional dissection of the ADR1 protein (Cook et al., 1994) has revealed that the region encompassing the eight motifs (from position 642 up to 1323) may contain latent activation regions as well as inhibitory function. More refined deletion experiments suggest that the inhibitory function may reside between residues 1068 and 1323 which corresponds to motifs VII and VIII. Taken together, these results clearly advance the importance of the conserved domain presented here in the regulation of the transactivators and most probably in the inhibition of their activity. Frequently, this inhibitory function is proposed to be mediated by the existence of hydrophobic interactions. These interactions may take place either within a single protein, between regions of the inhibitory domain and other functional regions (e.g. DNA-binding, dimerization, transactivation or effector domains) or by interaction with a putative inhibitory protein as suggested by Stone and Sadowski (1993). Whatever the exact nature of the mechanisms, it is striking to note that the presence of numerous predicted amphipathic a-helices may allow hydrophobic interactions via the hydrophobic faces of the helices. The finding and delineation of a conserved putative inhibitory domain will greatly aid the analysis of mutagenesis data and functional dissection of these proteins. In addition, this alignment suggests that an inhibitory mechanism common to all GFM and possibly to some other yeast transcriptional factors (i.e., C2H2-type mem-
O. Poch/Gene 184 (1997) 229–235
bers) may be implicated in the regulation of numerous fungal metabolic processes.
Acknowledgement I wish to thank Prof. A. Goffeau and Prof. C. Jacq for constant encouragement and helpful discussions concerning certain proteins analyzed in this article. I am pleased to thank Dr. S.A. Johnston for his stimulating interest. Special thanks are due to Dr. A. Delahodde for constant support and fruitful discussions. I am deeply endebted to Barbara Winsor for pertinent comments on the manuscript, critical reading and constant encouragement. I thank the staff of the IBMC and especially Dr. M. Zerbib and A. Mouchaboeuf for constant computer assistance. This work was supported by grants from the Centre National de Recherche Scientifique.
References Baleja, J.D., Marmorstein, R., Harrison, S.C. and Wagner, G. (1992) Solution structure of the DNA-binding domain of Cd2-GAL4. Nature 356, 450–453. Chasman, D.I. and Kornberg, R.D. (1990) Gal4 protein: purification, association with Gal80 protein, and conserved domain structure. Mol. Cell. Biol. 10, 2916–2923. Cook, W.J., Chase, D., Audino, D.C. and Denis, C.L. (1994) Dissection of the ADR1 protein reveals multiple, functionally redundant activation domains interspersed with inhibitory regions: evidence for a repressor binding to the ADR1c region. Mol. Cell. Biol. 14, 629–640. Dahwale, S.S. and Lane, A.C. (1993) Compilation of sequence-specific DNA-binding proteins implicated in transcriptional control in fungi. Nucleic Acids Res. 21, 5537–5546. Devereux, J., Haeberli, P. and Smithies, O. (1984) A comprehensive set of sequence analysis programs for the VAX. Nucleic Acids Res. 12, 387–395. Feller, A., Dubois, E., Ramos, F. and Pierard, A. (1994) Repression of the genes for lysine biosynthesis in Saccharomyces cerevisiae is caused by limitation of Lys14-dependent transcriptional activation. Mol. Cell. Biol. 14, 6411–6418. Friden, P., Reynolds, C. and Schimmel, P. (1989) A large internal deletion converts yeast LEU3 to a constitutive transcriptional activator. Mol. Cell. Biol. 9, 4056–4060. Gill, G. and Ptashne, M. (1987) Mutants of GAL4 protein altered in an activation function. Cell 51, 121–126. Gribskov, M., McLachlan, A.D. and Eisenberg, D. (1987) Profile analysis: detection of distantly related proteins. Proc. Natl. Acad. Sci. USA 84, 4355–4358. Gribskov, M., Homyak, M., Edenfield, J. and Eisenberg, D. (1988) Profile scanning for three-dimensional structural patterns in protein sequences. Comput. Appl. Biosci. 4, 61–66. Griggs, D.W. and Johnston, M. (1991) Regulated expression of the GAL4 activator gene in yeast provides a sensitive genetic switch for glucose repression. Proc. Natl. Acad. Sci. USA 88, 8597–8601.
235
Johnston, M. and Dover, J. (1988) Mutational analysis of the GAL4-encoded transcriptional activator protein of Saccharomyces cerevisiae. Genetics 120, 63–74. Johnston, S.A., Salmeron, J.M. and Dincher, S.S. (1987) Interaction of positive and negative regulatory proteins in the galactose regulon of yeast. Cell 50, 143–146. Kraulis, P.J., Raine, A.R.C., Gadhavi, P.L. and Laue, E.D. (1992) Structure of the zinc-containing DNA-binding domain of GAL4. Nature 356, 448–450. Ma, J. and Ptashne, M. (1987a) The carboxy-terminal 30 amino acids of GAL4 are recognized by GAL80. Cell 50, 137–142. Ma, J. and Ptashne, M. (1987b) Deletion analysis of GAL4 defines two transcriptional activating segments. Cell 48, 847–853. Marczak, J.E. and Brandriss, M.C. (1991) Analysis of constitutive and noninducible mutations of the PUT3 transcriptional activator. Mol. Cell. Biol. 11, 2606–2619. Marmorstein, R. and Harrison, S.C. (1994) Crystal structure of a PPR1-DNA complex: DNA recognition by proteins containing a Zn Cys binuclear cluster. Genes Dev. 8, 2504–2515. 2 6 Marmorstein, R., Carey, M., Ptashne, M. and Harrison, S.C. (1992) DNA recognition by GAL4: structure of a protein-DNA complex. Nature 356, 408–414. Parsons, L.M., Davis, M.A. and Hynes, M.J. (1992) Identification of functional regions of the positively acting regulatory gene amdR from Aspergillus nidulans. Mol. Microbiol. 6, 2999–3007. Pfeifer, K., Kim, K.-S., Kogan, S. and Guarente, L. (1989) Functional dissection and sequence of yeast HAP1 activator. Cell 56, 291–301. Qui, H.F., Dubois, E. and Messenguy, F. (1991) Dissection of the bifunctional ARGRII protein involved in the regulation of arginine anabolic and catabolic pathways. Mol. Cell. Biol. 11, 2169–2179. Reece, R.J. and Ptashne, M. (1993) Determinants of binding-site specificity among yeast C6 zinc cluster proteins. Science 261, 909–911. Rost, B. and Sander, C. (1993) Prediction of protein secondary structure at better than 70% accuracy. J. Mol. Biol. 232, 584–599. Salmeron, J.M. and Johnston, S.A. (1986) Analysis of the Kluyveromyces lactis positive regulatory LAC9 reveals functional homology to, but sequence divergence from, the Saccharomyces cerevisiae GAL4 gene. Nucleic Acids Res. 14, 7767–7781. Stone, G. and Sadowski, I. (1993) GAL4 is regulated by a glucoseresponsive functional domain. EMBO J. 12, 1375–1385. Sua`rez, T., Queiroz, M.V.D., Oestreicher, N. and Scazzocchio, C. (1995) The sequence and binding specificity of UaY, the specific regulator of the purine utilization pathway in Aspergillus nidulans, suggest an evolutionary relationship with PPR1 protein of Saccharomyces cerevisiae. EMBO J. 14, 1453–1467. Svetlov, V.V. and Cooper, T.G. (1995) Review: compilation and characteristics of dedicated transcription factors in Saccharomyces cerevisiae. Yeast 11, 1439–1484. Sze, J., Remboutsika, E. and Kohlhaw, G.B. (1993) Transcriptional regulator Leu3 of Saccharomyces cerevisiae: separation of activator and repressor functions. Mol. Cell. Biol. 13, 5702–5709. Wray, L.V., Witte, M.M., Dickson, R.C. and Riley, M.I. (1987) Characterization of a positive regulatory gene, LAC9, that controls induction of the lactose-galactose regulon of Kluyveromyces lactis: structural and functional relationships to GAL4 of Saccharomyces cerevisiae. Mol. Cell. Biol. 7, 1111–1121. Zhang, L., Bermingham-McDonogh, O., Turcotte, B. and Guarente, L. (1993) Antibody-promoted dimerization bypasses the regulation of DNA binding by the heme domain of the yeast transcriptional activator HAP1. Proc. Natl. Acad. Sci. USA 90, 2851–2855. Zhou, K.M. and Kohlhaw, G.B. (1990) Transcriptional activator LEU3 of yeast. Mapping of the transcriptional activation function and significance of activation domain tryptophans. J. Biol. Chem. 265, 17409–17412.
Conservation of a putative inhibitory domain in the GAL4 family members Olivier Poch * U.P.R. Me´canismes Mole´culaire de la Division Cellulaire et du De´veloppement, Institut de Biologie Mole´culaire et Cellulaire, 15 rue Rene´ Descartes, 67084 Strasbourg, France Received 24 May 1996; revised 11 July 1996; accepted 12 July 1996
Abstract The GAL4 family members are fungal transcriptional activators composed of several functional domains: a characteristic cysteine-rich DNA-binding domain common to all members, a dimerization domain, various transactivation domains generally exhibiting a high acidic content and a highly variable central region supposed to be involved in regulation and in effector recognition. We report here that the central region of the GAL4 family members share eight conserved motifs embedded in a large functional domain of 225 up to 405 residues. This domain may also be present in four proteins belonging to another family of transcriptional activators sharing a C2H2-type zinc finger. Analysis of the biochemical data available on the well-studied GAL4 protein suggests that this domain may be involved in the regulation of the activity of the protein, particularly in an inhibitory function. This hypothesis is further supported by deletion and site-directed mutagenesis experiments on other GAL4 family members. The mean secondary structure prediction performed on the eight motifs strongly suggests that the inhibitory activity may be mediated by hydrophobic interactions linked to the presence of amphipathic a-helices. Keywords: Fungal transcriptional activator; GAL4 family member; GAL4; PUT3; LEU3; PPR1; PDR1; PDR3; ADR1; Inhibitory domain; Sequence conservation; Saccharomyces cerevisiae
1. Introduction The GAL4 family members (GFM ) are fungal transcriptional regulatory proteins (Dahwale and Lane, 1993; Svetlov and Cooper, 1995) which share a characteristic cysteine-rich DNA-binding domain. They are involved in the regulation of numerous anabolic and catabolic pathways (i.e., galactose, arginine, proline, maltose and leucine metabolisms) as well as in the regulation of various pathways such as pleiotropic drug resistance, chromosome segregation or sporulation processes ( Table 1). Frequently, the activity of these transcriptional activators is known to be stimulated by the presence of a small molecule such as a metabolic
* Corresponding author. Abbreviations: S. cerevisae, Saccharomyces cerivisiae; S. carlbergensis, Saccharomyces carlbergensis; K. lactis, Kluyveromyces lactis; S. pombe, Schizosaccharomyces pombe; A. nidulans, Aspergillus nidulans; A. oryzae, Aspergillus oryzae; N. crassa, Neurospora crassa; C. albicans, Candida albicans; L. edodes, Lentinus edodes.
intermediate or a drug. In addition, with the totality of S. cerevisiae sequences available, a large set of potential GFM has been determined whose function is generally unknown (see Table 1). At the sequence level, GFM are large proteins (ranging from 465 up to 1502 residues) with a characteristic DNA-binding domain composed of a highly conserved 28-residue cysteine-rich motif (C6-type zinc finger) (Svetlov and Cooper, 1995) generally located in the most amino-terminal region. The three-dimensional structure of the cysteine-rich motif of two GFM (GAL4 and PPR1) revealed an essentially identical structure in which two Zn2+ions are coordinated to six invariant cysteine residues, with the rest of the amino acids folded around the Zn2Cys6 (Baleja et al., 1992; Kraulis et al., 1992; Marmorstein et al., 1992; Marmorstein and Harrison, 1994). In addition to a common DNA-binding domain, GFM share general functional organization with (in the immediate carboxy-terminal side of the zinc cluster) a coiled-coil dimerization domain with heptad repeats of hydrophobic residues (Reece and Ptashne, 1993; Zhang et al., 1993; Sua`rez et al., 1995). The
0378-1119/97/$17.00 Copyright © 1997 Elsevier Science B.V. All rights reserved PII S 03 7 8 -1 1 1 9 ( 9 6 ) 0 0 6 02 - 6
230
O. Poch/Gene 184 (1997) 229–235
transcriptional activation function is assured by various regions without extensive sequence conservation but generally with a high proportion of negatively charged residues. Frequently, the most potent activation region is located in the most carboxy-terminal portion of each activator while some regions, scattered in the central regions, can exhibit weaker activation functions (Zhou and Kohlhaw, 1990). Finally, transcriptional activation is regulated by the central regions which represent the major part of the proteins (more than 60% of the total mass of the factors) and are involved in more specialized functions such as recognition of peptide cofactors or metabolic intermediates. These central regions have been observed as highly variable when closely related GFM are compared and, up to now, only limited regions have been reported to be conserved between some GFM (Salmeron and Johnston, 1986; Wray et al., 1987; Chasman and Kornberg, 1990; Marczak and Brandriss, 1991).
Here we report that at least eight motifs composed of 161–167 residues are conserved in the variable central region of almost all reported GFM. In addition, these eight motifs could be detected in the sequences of four fungal transcriptional factors belonging to the C2H2-type zinc finger family further highlighting their general conservation in fungal transcriptional factors. The conservation of eight motifs spanning from 225 up to 405 residues strongly suggests their cooperation in a well-defined structural and functional domain. At the functional level, analysis of the mutational data available on various GFM and of biochemical results concerning the well-studied GAL4 factor strongly suggests that the conserved domain is involved in inhibitory function probably mediated by hydrophobic interactions. The secondary structures present in the domain, as deduced from the mean secondary structure predictions performed, are discussed in the context of their possible role in regulation of protein activity.
O. Poch/Gene 184 (1997) 229–235
231
Fig. 1. Multiple alignment of the putative inhibitory domains of the GAL4 family members (GFM ). This alignment was obtained by initial multiple alignment, performed by the PILEUP program (Devereux et al., 1984), of a set of transcriptional activator sequences (names underlined) and analysis of the pairwise sequence comparisons of closely related factors (i.e., Ppr1-Yeast and Andnauay-1; Gal4-Yeast and Lac9-Klula; Nira-Emeni and Nit4-Neucr; Put3-Yeast and Thi1-Schpo, Quta-Emeni and Qalf-Neucr). After manual adjustment, eight highly conserved regions were defined and then used to construct profiles (Gribskov et al., 1987, 1988) (‘fuzzy’ probes constituted from a group of aligned sequences) based on a single motif or by a concatenation of some or all of the eight motifs (named I–VIII ). Within a motif, the gap and gap-length penalties were defined as 5 and 0.2, respectively. One undetermined residue was introduced between each motif to allow the non-conserved interregions separating each motif to vary without constraint during the profile alignment process. Their values for gap and gap-length penalties were both defined as 0.03. These profiles were first validated and then used for further database scans of the Protein Information Resource (Pir), SwissProt (Sw) and translated EMBL ( TrEmbl ) protein databanks. The 47 GFM presented (upper part of the figure) and the four amino acid sequences belonging to the C2H2-type members ( lower part of the figure) were detected at a statistically significant level. The distances between the conserved motifs and distances from the termini are indicated. The consensus, at the top, shows residues conserved in more than 70% of the GFM sequences (bold in the alignment) and indicated by: @ for aromatic residues (F, Y, W ), h for bulky aliphatic or aromatic residues (I, L, M, V, F, Y, W ), G for small residues (P, A, G, S, T ), D for negatively charged residues (D, E, Q, N ). The secondary structure prediction (sec. struc.) shows the consensus of the predictions for individual GFM sequences obtained using the PHD program (Rost and Sander, 1993); a designates an a-helix; b designates a b-strand.
2. Experimental and discussion 2.1. Transcriptional activators presenting the eight conserved motifs Starting from the observation of limited sequence conservation in the central regions of a subset (Put3p,
Gal4p, Lac9p, Ppr1p, Pdr1p, QutAp and Leu3p) of GFM (Salmeron and Johnston, 1986; Chasman and Kornberg, 1990; Marczak and Brandriss, 1991), we initiated a computer analysis combining various sequence analysis tools such as pairwise sequence comparisons, multiple sequence comparisons and profile scans (see legend of Fig. 1). This allowed us to delineate
232
O. Poch/Gene 184 (1997) 229–235
Table 1 Gene names, organisms and functions of the GAL4 family members (GFM ) present in the protein databases and the four C2H2-type members detected by the profile scans Gene name or ORF
Organism
Database source and mnemonic
Accession No.
Regulation of
GAL4 LAC9 PPR1 CYP1 LEU3 PDR1 PDR3 CHA4 CAT8 PUT3 DAL81 CBF3B MAL3R/MAL33 MAL63 TEA1 ARGR2 LYS14 FUN43/YAL051W SIP4 YBL066C YBR150C YBR033W YCR106W 9740.13 9461.10 YER184C YFL052W YHR178W YIL130W YJL206c YKR064W YKL038W YKL222C L8003.10 L8479.13 L0584 YM9711.08 00938 9677.10 NTF1 SPAC1F7.11c QUTA UaY amdR NIRA NIT4 qalF priBc SUC1 ADR1 ZMS1 amdA Sc9367.02
S. cerevisiae K. lactis S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. carlsbergensis S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. cerevisiae S. pombe S. pombe A. nidulans A. nidulans A. oryzae A. nidulans N. crassa N. crassa L. edodes C. albicans S. cerevisiae S. cerevisiae A. nidulans S. cerevisiae
Sw:Gal4-Yeast Sw:Lac9-Klula Sw:Ppr1-Yeast Sw:Cyp1-Yeast Sw:Leur-Yeast Sw:Pdr1-Yeast Sw:Pdr3-Yeast Sw:Cha4-Yeast Sw:Cat8-Yeast Sw:Put3-Yeast Sw:Da81-Yeast Sw:Cb32-Yeast Sw:Ma3r-Yeast Sw:Ma6r-Yeast Sw:Tea1-Yeast Pir : S49627 Sw:Ly14-Yeast Sw:Yaf1-Yeast Sw:Sip4-Yeast Sw:Ybg6-Yeast Sw:Yb00-Yeast Sw:Ybo3-Yeast Sw:Ycz6-Yeast TrEmbl:Scl9740-9 TrEmbl:Sc94611-10 Sw:Ye14-Yeast Sw:Yff2-Yeast Sw:Yhx8-Yeast Sw:Yin0-Yeast Sw:Yju6-Yeast Sw:Yk44-Yeast Sw:Ykd8-Yeast Sw:Ykw2-Yeast Pir:S50366 Pir:S51403 Pir:S50966 Pir:S54020 Pir:S50406 TrEmb1:Sc19677-11 Sw:Thi1-Schpo Sw:Yakb-Schpo Sw:Quta-Emeni Sw:Uay-Emeni Sw:Amdr-Aspor Sw:Nira-Emeni Sw:Nit4-Neucr Sw:Qa1f-Neucr Sw:Prib-Lened Sw:Suc1-Canal Sw:Adr1-Yeast Sw:Zms1-Yeast TrEmbl:Andbp-1 TrEmbl:Sc9367-2
P04386 P08657 P07272 P12351 P08638 P12383 P33200 P43634 P39113 P25502 P21657 P40969 P39157 P10508 P47988 S49627 P40971 P39720 P46954 P34228 P38114 P38073 P25611 U28374 U33007 P39961 P43551 P38699 P40467 P39529 P36023 P32862 P35995 S50366 S51403 S50966 S54020 S50413 U25841 P36598 Q09922 P10563 P49413 Q06157 P28348 P28349 P11638 P49412 P33181 P07248 P46974 L28810 Z49274
Galactose metabolism Lactose metabolism Pyrimidine metabolism iso-1 and iso-2 cytochrome c expression Leucine biosynthesis Pleiotropic drug resistance Pleiotropic drug resistance Serine and threonine catabolism Gluconeogenic pathway Proline metabolism GABA, urea, arginine and allantoin catabolism Chromosome segregation Maltose metabolism Maltose metabolism Ty1 enhancer activator Arginine metabolism Lysine biosynthesis Unknown, chr. I seq. Unknown, chr. X seq. Sporulation processes, chr. II seq. Unknown, chr. II seq. Unknown, chr. II seq. Unknown, chr. III seq. Unknown, chr. IV seq. Unknown, chr. IV seq. Unknown, chr. V seq. Unknown, chr. VI seq. Unknown, chr. VIII seq. Unknown, chr. IX seq. Unknown, chr. X seq. Unknown, chr. XI seq. Unknown, chr. XI seq. Unknown, chr. XI seq. Unknown, chr. XII seq. Unknown, chr. XII seq. Unknown, chr. XII seq. Unknown, chr. XIII seq. Unknown, chr. XV seq. Unknown, chr. XVI seq. Thiamine biosynthesis Unknown, chr. I Quinate metabolism Purine catabolism Amide catabolism Nitrate assimilation Nitrate assimilation Quinate metabolism Unknown Sucrose utilization Alcohol dehydrogenase repressor Unknown Acetamidase gene Unknown
Underlined and bold gene names (ARGR2 and LYS14) indicate the two GFM sequences that are not detected at a statistically significant level by profile scans. Mnemonics are from the SwissProt database (Sw) Release 26.3, the translated EMBL ( TrEmbl ) Release 45, and from the Protein Information Resource (Pir) release 46. Abbreviations: chr., chromosome; seq., sequencing project.
O. Poch/Gene 184 (1997) 229–235
eight conserved blocks (motifs I–VIII ) involving 161–167 residues and separated by more divergent spacer regions (Fig. 1). Considering the inter-motif sequences, the eight concatenated motifs may constitute an entire domain which can span from 225 residues for the shortest sequence (Lelipri2-1 in Fig. 1) up to 405 residues for the longest one (Cyp1-Yeast). The profile scans performed with a concatenation of the eight motifs could detect at higher scores 47 out of 49 GFM available in the protein databases including numerous sequences of unknown function arising from the yeast sequencing project ( Table 1). Two transcriptional activators, ArgrIIp and Lys14p, could not be detected at significant levels. It should be pointed out that, among GFM, the Lys14p and ArgrIIp proteins appear exceptional as seen by the functional analysis performed on these proteins. In contrast to other yeast regulatory proteins, 92% of ArgrIIp is necessary for its anabolic repression function and 80% is necessary for its catabolic activator function (Qui et al., 1991) suggesting that almost the entire protein is required for efficient regulation. Similarly, internal deletion experiments on the Lys14p failed to identify limited segments which were sufficient for the activation function ( Feller et al., 1994). In this context, the absence of the eight motifs in these two unusual proteins is in good agreement with the biochemical data and further emphasizes the significance of the proposed alignment. The profile scans performed on all protein databases detect at an ‘interesting’ score level (Gribskov et al., 1987) of approx. 2 S.D.s above the mean, four fungal transcriptional activators exhibiting a DNA-binding motif distinct from the typical cysteine-rich motif present in the GFM sequences, namely a C2H2-type zinc finger (four last sequences in Fig. 1). A refined analysis of the sequences of these C2H2-type members strongly supports the idea of a structural link between the four proteins detected and the GFM. Indeed, among the C2H2-type members, amdAp, Zms1p, Adr1p and Sc9367p appear quite remarkable by two specific structural features: first, with respect to their length (880, 1380, 1323 and 1133 amino acid long, respectively), these proteins are longer than most of the C2H2-type members (from 211 up to 914 amino acids) and second, their DNA-binding domain is located in the aminoterminal region (amino acid positions 25, 153, 106 and 33 for the amdAp Zms1p Adr1p and Sc9367p, respectively). This location is unusual since the DNA-binding domains of most of the C2H2-type members are positioned in the central or in the most carboxy-terminal region of the proteins. Thus, these results strongly suggest that the four atypical fungal C2H2-type members are structurally linked to the GFM. These sequence relationships probably arise from a modular evolution which supposes that either an ancestral C2H2-type member had acquired the region encompassing the eight
233
motifs or that the four proteins have evolved from ancestral GFM by the acquisition of a C2H2-type zinc finger.
2.2. Sequence analysis of the eight conserved motifs The eight matched regions correspond to those of highest homology when closely related GFM are compared separately enhancing the significance of the sequence similarities found. Analyzing the GFM sequences detected through the profile scans, we introduced one limited gap at position 17 of motif I to increase the similarity. The resulting alignment includes 53 conservatively maintained residues in more than 70% of the GFM sequences. Among these conserved residues, numerous hydrophobic and aromatic residues (45 out of 53) are observed reflecting the overall composition of the central regions of the GFM which are mainly hydrophobic. The longest stretch of conserved amino acids is constituted by five consecutive residues present in motif VI (positions 7–11) defining a pattern of two basic residues followed by an aliphatic-aromatic-aliphatic sequence. In addition, at position 17 of the motif, a negatively charged residue is frequently observed implying that among the 53 conservatively maintained residues, all the charged residues are clustered in motif VI. Global inspection of Fig. 1 revealed a great variability in the lengths separating the different motifs. Such variability was also observed in closely related pairwise sequences (e.g. Ppr1-Yeast/Andnauy-1; Cyp1-Yeast/Yk44-Yeast or Pdr3-Yeast-Ye14-Yeast in Fig. 1) which can vary widely in their inter-motif distances. In addition, it should be pointed out that the amino acid compositions of the largest inter-motif sequences are frequently biased with a high content in acidic residues or in small neutral residues such as proline, glycine or serine residues. Biased compositions are noteworthy in the sequences separating motifs IV and V and motifs V and VI in the Cyp1-Yeast and Scd94611-10 amino acid sequences which exhibit an extremely high content in acidic residues. The variability and importance of the inter-motif distances clearly reflect an extended plasticity of the central region of the GFM with preservation of the eight motifs scattered between variable sequences in which more specialized functions, such as transactivation for example, may take place. Finally, the most conserved motifs are IV, V and VI which correspond roughly to the three regions previously appointed by Marczak and Brandriss (1991). Among the five remaining motifs, motif III is the least conserved while hydrophobic residues preserved in motifs IV, VII and VIII frequently display a 3–4 periodicity. In addition to GFM already recognised as closely related, analysis of the percentage of sequence conservation existing between the distinct members allows to cluster one group composed of 27
234
O. Poch/Gene 184 (1997) 229–235
sequences (27 upper sequences from Ybg6-Yeast to Qalf-Neucr in Fig. 1). 2.3. Structural and functional analysis of the alignment The preservation in such widely divergent proteins of eight concatenated motifs, spanning from 225 up to 405 residues, suggests their cooperative implication in a welldefined structural and functional domain. At the structural level, the mean secondary structure predictions performed (see legend of Fig. 1) suggest that the motifs are mainly composed of a-helices. The only non-a structure predicted is located in the most conserved motif VI, in which a b-turn-b structure is strongly predicted. This b-turn-b structure may maintain the highly conserved acidic residue (position 17) near the turn structure. It should be noted that in most of the predicted a-helices, a 3–4 periodicity of the conservatively maintained hydrophobic residues can be observed (motifs I, II, IV, VII and VIII ). In motifs IV, VII and VIII, conserved heptad repeats are observed suggesting the existence of stable amphipathic a-helices preserved throughout the whole domain. The presence of these amphipathic a-helices strongly suggests that hydrophobic interactions may take place in the conserved domain. Such hydrophobic interactions may play an important role in maintaining the general topology of the domain and/or in interactions with other proteins involved in the regulation of the activity of these fungal transcriptional activators (see below). At the functional level, the numerous functional analyses performed on the well-studied GAL4 protein strongly suggest the involvement of the central region in regulation of the transcriptional activity (Gill and Ptashne, 1987; Johnston et al., 1987; Ma and Ptashne, 1987a,b; Johnston and Dover, 1988; Chasman and Kornberg, 1990; Griggs and Johnston, 1991). Recently, it has been shown that the conserved domain defined by the eight motifs coincides almost exactly to a region of Gal4p corresponding to the so-called ‘inhibitory domain’ (Stone and Sadowski, 1993). These authors proposed that an inhibitory domain may bind an inhibitory protein through a major conformational change induced by the fixation of glucose on an adjacent glucose response domain. The inhibitory domain (ID) delineated by Stone and Sadowski (1993) corresponds to positions 236–614 of the GAL4 protein which fully encompasses the eight motifs shown in Fig. 1. This large ID could be subdivided in three consecutive smaller regions (ID1–ID3) corresponding to residues 320–412, 412–478 and 554–585, respectively. With respect to the eight motifs, ID1 corresponds almost exactly to the three most conserved motifs IV, V and VI since it begins one residue upstream from motif IV and ends five residues downstream. ID2 is located between motifs VI and VII while ID3 begins 21 residues upstream from motif VIII
up to the proline at position 17 of this motif. The correlation observed between these biochemical data concerning the central region of Gal4p and the eight motifs delineated by sequence analysis further highlights the significance of the proposed alignment. As a corollary, this strongly suggests that the domain delineated by the eight motifs may be involved in an inhibitory function common to all the GFM and C2H2-type members sharing these regions. The fact that this domain may be involved in negative regulation is also consistent with studies on the LEU3 transactivator. A large deletion of the central region of Leu3p renders the protein constitutively active ( Friden et al., 1989; Sze et al., 1993) suggesting that an inhibitory effect has been deleted from the central region. The functional importance of the domain encompassing the eight motifs is in agreement with internal deletion experiments on various GFM. In the HAP1/CYP1 transactivator (CYP1-yeast) (Pfeifer et al., 1989) or in the amdR (Aoamdr-1) (Parsons et al., 1992), internal deletions of all or some of the motifs led to the loss of inducibility, highlighting the requirement of this region for a full and wellregulated function of the transactivators. With respect to site-directed mutagenesis, it is noteworthy that a single replacement in Put3p of the highly conserved glycine by an aspartate residue in position 15 of motif II leads to a noninducible allele. Finally, functional dissection of the ADR1 protein (Cook et al., 1994) has revealed that the region encompassing the eight motifs (from position 642 up to 1323) may contain latent activation regions as well as inhibitory function. More refined deletion experiments suggest that the inhibitory function may reside between residues 1068 and 1323 which corresponds to motifs VII and VIII. Taken together, these results clearly advance the importance of the conserved domain presented here in the regulation of the transactivators and most probably in the inhibition of their activity. Frequently, this inhibitory function is proposed to be mediated by the existence of hydrophobic interactions. These interactions may take place either within a single protein, between regions of the inhibitory domain and other functional regions (e.g. DNA-binding, dimerization, transactivation or effector domains) or by interaction with a putative inhibitory protein as suggested by Stone and Sadowski (1993). Whatever the exact nature of the mechanisms, it is striking to note that the presence of numerous predicted amphipathic a-helices may allow hydrophobic interactions via the hydrophobic faces of the helices. The finding and delineation of a conserved putative inhibitory domain will greatly aid the analysis of mutagenesis data and functional dissection of these proteins. In addition, this alignment suggests that an inhibitory mechanism common to all GFM and possibly to some other yeast transcriptional factors (i.e., C2H2-type mem-
O. Poch/Gene 184 (1997) 229–235
bers) may be implicated in the regulation of numerous fungal metabolic processes.
Acknowledgement I wish to thank Prof. A. Goffeau and Prof. C. Jacq for constant encouragement and helpful discussions concerning certain proteins analyzed in this article. I am pleased to thank Dr. S.A. Johnston for his stimulating interest. Special thanks are due to Dr. A. Delahodde for constant support and fruitful discussions. I am deeply endebted to Barbara Winsor for pertinent comments on the manuscript, critical reading and constant encouragement. I thank the staff of the IBMC and especially Dr. M. Zerbib and A. Mouchaboeuf for constant computer assistance. This work was supported by grants from the Centre National de Recherche Scientifique.
References Baleja, J.D., Marmorstein, R., Harrison, S.C. and Wagner, G. (1992) Solution structure of the DNA-binding domain of Cd2-GAL4. Nature 356, 450–453. Chasman, D.I. and Kornberg, R.D. (1990) Gal4 protein: purification, association with Gal80 protein, and conserved domain structure. Mol. Cell. Biol. 10, 2916–2923. Cook, W.J., Chase, D., Audino, D.C. and Denis, C.L. (1994) Dissection of the ADR1 protein reveals multiple, functionally redundant activation domains interspersed with inhibitory regions: evidence for a repressor binding to the ADR1c region. Mol. Cell. Biol. 14, 629–640. Dahwale, S.S. and Lane, A.C. (1993) Compilation of sequence-specific DNA-binding proteins implicated in transcriptional control in fungi. Nucleic Acids Res. 21, 5537–5546. Devereux, J., Haeberli, P. and Smithies, O. (1984) A comprehensive set of sequence analysis programs for the VAX. Nucleic Acids Res. 12, 387–395. Feller, A., Dubois, E., Ramos, F. and Pierard, A. (1994) Repression of the genes for lysine biosynthesis in Saccharomyces cerevisiae is caused by limitation of Lys14-dependent transcriptional activation. Mol. Cell. Biol. 14, 6411–6418. Friden, P., Reynolds, C. and Schimmel, P. (1989) A large internal deletion converts yeast LEU3 to a constitutive transcriptional activator. Mol. Cell. Biol. 9, 4056–4060. Gill, G. and Ptashne, M. (1987) Mutants of GAL4 protein altered in an activation function. Cell 51, 121–126. Gribskov, M., McLachlan, A.D. and Eisenberg, D. (1987) Profile analysis: detection of distantly related proteins. Proc. Natl. Acad. Sci. USA 84, 4355–4358. Gribskov, M., Homyak, M., Edenfield, J. and Eisenberg, D. (1988) Profile scanning for three-dimensional structural patterns in protein sequences. Comput. Appl. Biosci. 4, 61–66. Griggs, D.W. and Johnston, M. (1991) Regulated expression of the GAL4 activator gene in yeast provides a sensitive genetic switch for glucose repression. Proc. Natl. Acad. Sci. USA 88, 8597–8601.
235
Johnston, M. and Dover, J. (1988) Mutational analysis of the GAL4-encoded transcriptional activator protein of Saccharomyces cerevisiae. Genetics 120, 63–74. Johnston, S.A., Salmeron, J.M. and Dincher, S.S. (1987) Interaction of positive and negative regulatory proteins in the galactose regulon of yeast. Cell 50, 143–146. Kraulis, P.J., Raine, A.R.C., Gadhavi, P.L. and Laue, E.D. (1992) Structure of the zinc-containing DNA-binding domain of GAL4. Nature 356, 448–450. Ma, J. and Ptashne, M. (1987a) The carboxy-terminal 30 amino acids of GAL4 are recognized by GAL80. Cell 50, 137–142. Ma, J. and Ptashne, M. (1987b) Deletion analysis of GAL4 defines two transcriptional activating segments. Cell 48, 847–853. Marczak, J.E. and Brandriss, M.C. (1991) Analysis of constitutive and noninducible mutations of the PUT3 transcriptional activator. Mol. Cell. Biol. 11, 2606–2619. Marmorstein, R. and Harrison, S.C. (1994) Crystal structure of a PPR1-DNA complex: DNA recognition by proteins containing a Zn Cys binuclear cluster. Genes Dev. 8, 2504–2515. 2 6 Marmorstein, R., Carey, M., Ptashne, M. and Harrison, S.C. (1992) DNA recognition by GAL4: structure of a protein-DNA complex. Nature 356, 408–414. Parsons, L.M., Davis, M.A. and Hynes, M.J. (1992) Identification of functional regions of the positively acting regulatory gene amdR from Aspergillus nidulans. Mol. Microbiol. 6, 2999–3007. Pfeifer, K., Kim, K.-S., Kogan, S. and Guarente, L. (1989) Functional dissection and sequence of yeast HAP1 activator. Cell 56, 291–301. Qui, H.F., Dubois, E. and Messenguy, F. (1991) Dissection of the bifunctional ARGRII protein involved in the regulation of arginine anabolic and catabolic pathways. Mol. Cell. Biol. 11, 2169–2179. Reece, R.J. and Ptashne, M. (1993) Determinants of binding-site specificity among yeast C6 zinc cluster proteins. Science 261, 909–911. Rost, B. and Sander, C. (1993) Prediction of protein secondary structure at better than 70% accuracy. J. Mol. Biol. 232, 584–599. Salmeron, J.M. and Johnston, S.A. (1986) Analysis of the Kluyveromyces lactis positive regulatory LAC9 reveals functional homology to, but sequence divergence from, the Saccharomyces cerevisiae GAL4 gene. Nucleic Acids Res. 14, 7767–7781. Stone, G. and Sadowski, I. (1993) GAL4 is regulated by a glucoseresponsive functional domain. EMBO J. 12, 1375–1385. Sua`rez, T., Queiroz, M.V.D., Oestreicher, N. and Scazzocchio, C. (1995) The sequence and binding specificity of UaY, the specific regulator of the purine utilization pathway in Aspergillus nidulans, suggest an evolutionary relationship with PPR1 protein of Saccharomyces cerevisiae. EMBO J. 14, 1453–1467. Svetlov, V.V. and Cooper, T.G. (1995) Review: compilation and characteristics of dedicated transcription factors in Saccharomyces cerevisiae. Yeast 11, 1439–1484. Sze, J., Remboutsika, E. and Kohlhaw, G.B. (1993) Transcriptional regulator Leu3 of Saccharomyces cerevisiae: separation of activator and repressor functions. Mol. Cell. Biol. 13, 5702–5709. Wray, L.V., Witte, M.M., Dickson, R.C. and Riley, M.I. (1987) Characterization of a positive regulatory gene, LAC9, that controls induction of the lactose-galactose regulon of Kluyveromyces lactis: structural and functional relationships to GAL4 of Saccharomyces cerevisiae. Mol. Cell. Biol. 7, 1111–1121. Zhang, L., Bermingham-McDonogh, O., Turcotte, B. and Guarente, L. (1993) Antibody-promoted dimerization bypasses the regulation of DNA binding by the heme domain of the yeast transcriptional activator HAP1. Proc. Natl. Acad. Sci. USA 90, 2851–2855. Zhou, K.M. and Kohlhaw, G.B. (1990) Transcriptional activator LEU3 of yeast. Mapping of the transcriptional activation function and significance of activation domain tryptophans. J. Biol. Chem. 265, 17409–17412.