Isolation of the ace1 Gene Encoding a Cys 2 -His 2 Transcription Factor Involved in Regulation of Activity of the Cellulase Promoter cbh1 of Trichoderma reesei *

A genetic selection method was developed for the cloning of positive-acting transcriptional regulatory genes in Saccharomyces cerevisiae . The method was applied for the isolation of activators of Trichoderma reesei ( Hypocrea jecorina ) cellulase genes. Activator genes were isolated from a T. reesei expression cDNA library on the basis of the ability of their translation products to activate transcription from the full-length T. reesei cbh1 promoter coupled to the S. cerevisiae HIS3 gene and to support the growth of the yeast colonies in the absence of histidine. Among the clones obtained was the ace1 gene encoding a novel polypeptide, ACEI, that contains three zinc finger motifs of Cys 2 -His 2 type. Possible ACEI homologues were found among expressed sequence tags of Aspergillus and Neurospora . The ability of ACEI to bind to the cbh1 promoter was further confirmed in the yeast one-hybrid system. In vitro binding and gel mobility shift assays revealed several binding sites for the ACEI protein in the cbh1 promoter. Disruption of the ace1 gene in T. reesei resulted in retarded growth of the fungus on a cellulose-containing medium, on which cellulases are normally highly expressed.

The filamentous fungus Trichoderma reesei is well known for efficient production of cellulolytic enzymes and its powerful capacity to hydrolyze cellulose into glucose. It is an excellent cellulolytic model organism, and the cellulolytic system of T. reesei has become the best characterized among filamentous fungi in many respects. The enzymatic properties and threedimensional structures (1)(2)(3)(4) of T. reesei cellulases, as well as the carbon source dependent regulation of cellulase gene expression (for a recent review, see Ref. 5), have been studied in detail by several groups. The activity of cellulase genes is controlled at the level of transcription: the genes are repressed in the presence of glucose and highly induced when cellulose, its derivatives, or certain oligosaccharides, such as sophorose, are present. Glucose repression is mediated by the CREI protein (6 -8), which has been shown to act directly on the promoter of the gene encoding the major cellulase cellobiohydrolase I (CBHI) (9). CREI also mediates repression of a number of other genes coding for enzymes involved in degradation of hemicellulose and cellulose (10). In addition to cre1, no other genes for transcription factors have been described in T. reesei. Based on the gene expression data, it can be concluded that a distinct induction pathway for cellulases must exist in T. reesei that is needed for high level transcription (11,12), but the genes responsible for transcription activation are completely unknown. Furthermore, there is lack of knowledge about possible target sequences for cellulase-specific transcription activators from any filamentous fungi. The aim of this study was to identify genes encoding transcription activators involved in regulation of the activity of cellulase genes. For this purpose, we developed a genetic selection system that allowed us to isolate novel transcription activators from T. reesei cDNA expression library based on their ability to bind and activate transcription from the T. reesei cbh1 promoter in Saccharomyces cerevisiae.

EXPERIMENTAL PROCEDURES
Strains-Escherichia coli strains JS4 and DH5␣ were used for library and plasmid constructions, respectively. Strain TOP10FЈ was used as a host for the pZErO-1 based vectors and strain BL21(DE3)LysS was used as a host for production of glutathione Stransferase (GST) 1
Media and Culture Conditions-Host yeast strains were grown in yeast/peptone dextrose medium. Synthetic selection media (SC) lacking the appropriate nutrients were used for the plasmid-carrying strains (18). SC-Leu-His plates supplemented with 45 mM 3-aminotriazole were used in the one-hybrid activation test. Plates used in yeast electroporation contained 1 M sorbitol.
The culture conditions used for T. reesei Rut-C30 to induce the production of hydrolytic enzymes and the preparation of the cDNA * Financial support was provided by the Foundation for Biotechnical and Industrial Fermentation Research, Helsinki Graduate School in Biotechnology and Molecular Biology, Roal Oy, and the Technology Development Center. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AF190793.
Isolation Method for Activator Genes-The reporter plasmid pAS3 was constructed as follows. pRS315 (21), the yeast single-copy vector containing the LEU2 marker, was digested with the restriction enzymes BamHI and SalI. The HIS3 reporter gene of S. cerevisiae was cloned by PCR using the cosmid p3030 (22) 3 as a template and sequence-specific primers complementary with the HIS3 promoter sequences starting from 78 bp upstream of the initiator ATG (AAAGGAT-CCTTATACATTATATAAAGTAATG; BamHI underlined) and the HIS3 terminator sequences (ATATAGTCGACCTCGGGGACACCAAA-TATGG; SalI underlined), creating terminal BamHI and SalI sites to facilitate cloning of the fragment into pRS315. The HIS3 gene of the resulting pAS1 plasmid begins 78 bp upstream from the initiator ATG and contains a minimal promoter including the TATA box, and it should not support growth in the absence of histidine. The resulting plasmid pAS1 was digested with SacI and XbaI, located immediately upstream of the HIS3 gene, and a PCR product containing the 1.15-kb cbh1 promoter fragment upstream from and excluding the cbh1 TATA box (from Ϫ1281 to Ϫ133 upstream of ATG) was ligated in front of the HIS3 gene producing pAS3. The primers used for cbh1 promoter amplification were (Ϫ1281) (ATACCCGGGAGCTCATTCCCGAAAAAACTCGG; SacI underlined) and (Ϫ133) (ATTCCCGGGTCTAGACACATTCGCTG-ACTTTGCC; XbaI underlined), and the template was pMLO16 (9).
The polylinker region present in pAS1 caused some leaky expression of the HIS3 gene, and a more proper negative control plasmid pMS95 was constructed as follows. Plasmid pAS1 was digested with SacI and a 1.4-kb SacI fragment from a nonrelevant cDNA (5Јend of a glutamate receptor cDNA from rat) was ligated in front of the HIS3 gene.
The T. reesei expression cDNA library of 10 5 independent clones was constructed in E. coli as described (19). RNA from the T. reesei Rut-C30 strain was used as a template. The yeast multicopy vector pAJ401 used contains the yeast URA3 marker and the constitutive PGK promoter of yeast, which provides expression signals for the cDNA insert. The DBY746 yeast strain harboring the pAS3 reporter plasmid was transformed by electroporation according to the manufacturer's instructions (Bio-Rad). Electroporation with the total of 40 g of the plasmid pool gave a library of 10 6 yeast cells. In order to screen Hisϩ colonies, the library was plated on SC-Leu-Ura-His plates to a density of 10 6 cells per plate and grown at 30°C for 5 days. 0.004% of the transformed yeast cells could grow in the selection conditions. Plasmid DNA was isolated from the growing colonies and transformed to the DBY746 yeast strain and the strains with the reporter construct and the negative control plasmid. Plasmids that supported growth on the SC-Leu-Ura-His plates only together with the reporter plasmid pAS3 were analyzed further.
Transformation Procedures-E. coli was transformed by electroporation according to the manufacturer's instructions (Bio-Rad). Yeast transformation was done either with electroporation (Bio-Rad) or by using the improved lithium acetate method of Gietz et al. (23). Protoplast transformation of T. reesei (20) was performed by selecting transformants for growth on acetamide as the sole nitrogen source.
Nucleic Acid Hybridizations-T. reesei chromosomal DNA and total RNA were isolated as described earlier (24,25). Southern and Northern hybridizations were carried out using standard methods.
The full-length ace1 cDNA was isolated from a cellulase-induced T. reesei Rut-C30 ZAP library (26) using a 300-bp PCR fragment from the 5Ј end of the original non-full-length cDNA as a probe by plaque hybridization according to the ZAP manual (Stratagene). The genomic ace1 copy was isolated from the cosmid library of T. reesei VTT-D-80133 (27) using the ace1 coding sequence as a probe in colony hybridization (28). A 7-kb HindIII fragment was subcloned into the HindIII site of the pZErO-1 vector (Invitrogen) resulting in the plasmid pAS34.
Production of the ACEI Zinc Finger Region in E. coli and in Vitro Binding Studies-The DNA binding domain of ACEI was produced as a GST fusion in E. coli for binding studies. To construct the GST-ACEI fusion expression vector, a DNA fragment encompassing the ACEI DNA binding domain (amino acids 382-528) was amplified by PCR with the primers 5Ј-AGCGCGGATCCATGCGGTCCATGGCCCGCCG-3Ј and 5ЈAGCCGGAATTCGTAGCTGGGCGTGGAGGAAG-3Ј (BamHI and EcoRI restriction sites underlined) and cloned in-frame into the BamHI-EcoRI-cut pGEX-2T plasmid (Amersham Pharmacia Biotech), resulting in plasmid pARO18. E. coli BL21(DE3) cells transformed with pARO18 were grown to A 600 ϭ 0.7, and production of the GST-ACEI 382-582 fusion protein was induced by addition of isopropyl-␤-D-thiogalactopyranoside to 1 mM. Cells were harvested and broken by sonication. Cell debris was removed by centrifugation. The fusion protein contained in the supernatant was affinity-purified on a glutathione-Sepharose 4B column (Amersham Pharmacia Biotech) and eluted in 5 mM glutathione, 50 mM Tris-HCl (pH 8.0). For some experiments, the ACEI 382-582 domain was cleaved from the GST moiety by thrombin (20 units of thrombin/mg of fusion protein), and the cleavage was confirmed by SDS-PAGE.
DNA Labeling for Binding Reactions-Complementary oligonucleotides were annealed producing a hybrid with a recessed 3Ј end in the coding strand that was filled in with [␣-32 P]dCTP using the Klenow fragment of DNA polymerase. Fragments longer than 100 bp were amplified by PCR using sequence-specific primers, digested with appropriate restriction enzymes, gel-purified, and labeled as described above.
One-Hybrid Experiment-Binding of ACEI 382-582 to a 170-bp fragment of the cbh1 promoter was assessed in by using the Matchmaker one-hybrid system (CLONTECH). For the construction of a targetreporter yeast strain, a cbh1 promoter fragment corresponding to nucleotides from Ϫ843 to Ϫ676 was amplified by PCR with primers GAG-AGAGAGCTCCTGGAAAATACAAACCAATGGC (forward) and GAGA-GAGAGCTCAGAAACAAACGTGGGGAAGTG (reverse) flanked by SacI sites (underlined). The PCR product was cloned into the SacI site of pHisi-1 (CLONTECH). The resulting plasmid, pPL2, contained two tandem copies of the insert in reverse orientation. pPL2 and pHisi-1 were linearized and transformed into the S. cerevisiae strain YM4271 by the lithium acetate method, and the transformants were selected on SC-His plates for integration of the target-reporter construct into the HIS3 locus. For the in vivo binding experiments, the DNA binding domain of ACEI from pARO18 (see above) was cloned into BamHI-EcoRI cut pGAD10 expression vector (CLONTECH) of the Matchmaker two-hybrid system, which generates a hybrid protein that contains the GAL4 activation domain in the N terminus. The GAL4ad-ACEI 382-582 expression construct was named pARO20. Subsequently, the Leu-selectable plasmids pARO20 and pGAD10 were transformed into the pPL2/YM4271 and pHisi-1/YM4271 reporter yeasts, and transformants were selected on SC-His-Leu plates. The test plates for GAL4ad-ACEI 382-582 binding and activation were SC-His-Leu supplemented with 45 mM 3-aminotriazol to prevent leaky expression from HIS3.
Disruption of the ace1 Gene-For disruption of the ace1 gene from the fungal genome, plasmid pAS38 was constructed as follows. The protein coding region of the ace1 chromosomal gene, except for the last 150 bp, was removed from the pAS34 plasmid by digestion with the restriction enzymes BglII and NarI. This region was replaced with the amdS gene of Aspergillus nidulans excised from the p3SR2 plasmid with the restriction enzymes SphI and XbaI. Prior to ligation, the fragments were filled in with the T4 DNA polymerase (Roche Molecular Biochemicals). Plasmid pAS38 was digested with the restriction enzyme HindIII, and the disruption cassette containing the amdS transformation marker flanked with fragments of about 2.5 kb from the 5Јand 3Јends of the ace1 gene was isolated from the gel and transformed to Trichoderma. The transformants were screened by Southern hybridization for the replacement of the ace1 gene.
Other Methods-All procedures not described above were carried out using standard methods (e.g. Ref. 28).

Isolation of the ace1 Gene by Genetic Selection in Yeast-A
genetic selection method was developed for cloning of positiveacting transcriptional regulatory genes in S. cerevisiae, and it was applied for the isolation of activators of cellulase genes of T. reesei. First, a yeast reporter plasmid was constructed in which the promoter of the T. reesei major cellulase gene cbh1 was fused to a promoterless S. cerevisiae HIS3 gene, thus bringing HIS3 expression under the control of the cbh1 promoter. A 1.15-kb promoter fragment located immediately up-stream of the cbh1 TATA box was inserted in front of the yeast HIS3 reporter gene, which lacked upstream regulatory sequences but contained the TATA box and the sequences downstream. The resulting LEU2-selectable single copy yeast plasmid pAS3 was transformed into the yeast strain DBY746. The reporter strain DBY746-pAS3 could not grow on media lacking histidine, showing that the Trichoderma promoter could not drive expression of the reporter gene by itself. In order to exclude the possibility that HIS3 was not expressed because the S. cerevisiae glucose repressor protein MIG1, which recognizes similar sequence elements as present in the cbh1 promoter, repressed the reporter construct, pAS3 was transformed into a mig1 deletion strain. The pAS3 construct remained silent, demonstrating that MIG1 did not repress the promoter and that yeast-encoded factors alone could not activate the reporter construct.
In order to isolate T. reesei cDNAs that would activate transcription from the cbh1-HIS3 construct, a cDNA of T. reesei was introduced into the reporter yeast. The cDNA library of T. reesei grown in cellulase-inducing conditions was prepared into a URA3-selectable multicopy yeast expression vector pAJ401 under the strong constitutive PGK promoter (19). A subset of the expression library containing 10 5 independent clones was transformed into the reporter yeast strain DBY746-pAS3, resulting in 10 6 yeast colonies growing on media lacking leucine and uracil. The colonies were scraped from the plates and screened on plates lacking histidine, leucine, and uracil. Growing colonies were detected with the frequency of 4:100 000. Plasmids were recovered from the growing colonies and retransformed into the reporter yeast strain, into the yeast strain containing the negative control plasmid pMS95, and into the host strain DBY746. Most of the plasmids supported growth of all the strains on media lacking histidine. These clones contained the T. reesei his3 gene as verified by partial sequencing followed by homology comparison of the open reading frame against the yeast and Neurospora his3 genes. 15% of the plasmids could not support growth of any of the strains and thus represented false positives. One of the remaining plasmids (pAS27) contained a 2-kb cDNA insert and supported slow growth of the reporter strain but not of the negative control or the host strain on media lacking histidine (Fig. 1). The cDNA was studied further, and the gene was named ace1 (activator of cellulase expression). ace1 cDNA Codes for a DNA-binding Protein-Sequencing of the ace1 cDNA from the pAS27 plasmid revealed a 1943-bp cDNA with an open reading frame of 491 amino acids starting from the first ATG codon in the insert. Northern hybridization using the cDNA as a probe gave two signals of about 3.2 and 3.0 kb in length (data not shown) and thus showed that the cDNA was not full-length. Therefore, the full-length cDNA of the ace1 gene was isolated from a library prepared in ZAP from the same induced Trichoderma cDNA as used in the initial screening. A 300-bp PCR fragment from the 5Ј end of the original cDNA was used as a probe in plaque hybridization. The resulting plasmid pAS28 contained a cDNA insert of 3223 bp, which is in good accordance with the estimated size of the longest mRNA. The open reading frame of 733 amino acids starting from the first ATG codon of the cDNA maintains the frame of the original open reading frame and contains 242 additional amino acids. There is a rather long, 611-bp, nontranslated 5Ј leader sequence in the cDNA. The existence of three in-frame stop codons before the first ATG in the cDNA confirms that the plasmid contains the whole protein coding sequence of the ace1 gene.
The ace1 gene sequence and the deduced protein sequence are shown in Fig. 2. Amino acids 387-403 form a putative bipartite nuclear targeting signal RRKKNATPEDVAPKKCR (basic residues are shown in boldface), fitting well to the consensus (29). Partially overlapping with the nuclear targeting signal follows an area containing three zinc fingers of the Cys 2 -His 2 type. The original PROSITE pattern C 2 H 2 recognizes the first Cys 2 -His 2 finger of ACEI, and an extended pattern (C 2 H 2 can; Ref. 30) recognizes the third finger that contains a serine instead of the conserved hydrophobic amino acid of the original pattern. The middle finger of ACEI has a 15amino acid-long loop instead of the usual 12 amino acids between the conserved zinc-coordinating second Cys and first His. The distribution of acidic and bulky nonpolar residues between amino acids 373-388 and 666 -682 suggests that amphipatic ␣-helices characteristic of several transcription activation domains may be formed (31,32).
Sequence Comparisons of the ACEI Protein-The deduced ACEI protein sequence was compared with the sequence data banks by using the BLAST program. The first finger showed highest similarity to zinc fingers of fungal origin, e.g. to those of the S. cerevisiae sulfite resistance gene FZF1 (33), the Aspergillus developmental regulator BRLA (34), homologues of the CREA glucose repressor from several fungi, the meiotic regulator RIM1/RIM101 of S. cerevisiae (35), and RIM101 of Yarrowia lipolytica (36) that shows homology to the A. nidulans pH-regulator PACC (37). The second finger showed similarity to the meiotic inhibitor RME1 from yeast (38). RME1 also contains three zinc fingers, the second of which has a 15-amino acid central loop as ACEI, but the amino acid sequences of the fingers are quite different. The third finger showed similarity to the MSS1 protein of S. cerevisiae (39) involved in glucoamylase gene expression. Outside the zinc finger region, no meaningful similarities with known proteins could be detected. DNA and protein sequence comparisons against the A. nidulans and Neurospora crassa expressed sequence tag data bases detected four Aspergillus clones (n8h08a1.r1, c3b04a1.r1, w4e06a1.r1, g6h12a1.r1), three of which are partially overlapping, and one Neurospora clone (b306ne.f1). All five of these clones have clear similarity with T. reesei ACEI, suggesting that homologues of ace1 are expressed in other filamentous fungi. Amino acid sequence alignment of these clones with T. reesei ACEI is shown in Fig. 3.
Characterization of the Genomic ace1 Locus-Chromosomal DNA isolated from different T. reesei strains, including hypercellulolytic and cellulase-negative strains (15,14), was subject to Southern hybridization in stringent conditions using the full-length cDNA of ace1 as a probe (data not shown). ace1 appeared to be a single-copy gene in the genome. Identical   FIG. 1. Activation of HIS3 transcription by ACEI through the  T. reesei cbh1 promoter in S. cerevisiae DBY746. Yeast colonies each carrying two of the following plasmids in different combinations, the reporter plasmid pAS3 (cbh1-HIS3), the ace1 cDNA clone pAS27 (PGK-ace1), empty cDNA vector pAJ401 (PGK promoter), and the negative reporter construct pMS95 (gluR-HIS3), were streaked onto SC-Ura-Leu plates and replicated onto fresh SC-Ura-Leu-His and SC-Ura-Leu plates. Only yeast colonies containing plasmids pAS3 and pAS27 grew in the absence of histidine.
bands were detected from all the strains studied (see under "Experimental Procedures"), indicating that deletion, duplication, or any major rearrangement of the ace1 activator locus is not responsible for the different amounts of cellulase enzymes produced by the cellulase-negative or -overproducing strains.
In order to clone the genomic copy of the ace1 gene, a chromosomal cosmid library of T. reesei was screened using a PCR fragment of the coding sequence as a probe in colony hybridization. The genomic gene (Fig. 2) was subcloned, and sequencing revealed three introns of 229, 63, and 65 bp. The first intron is located 5Ј to the predicted protein coding region. The 5Ј noncoding region of ace1 does not appear to contain a TATA box, but a CT-rich sequence is present immediately upstream of the cDNA start site. Nine putative binding sites for the glucose repressor protein CREI, 5ЈSYGGRG, are present within the sequenced 1-kb promoter region.

ACEI Protein Binds to the cbh1 Promoter in Vitro and in
Vivo-Because the cbh1 promoter region used in the initial screening was about 1.15 kb, further experiments were required to map more precisely the region to which ACEI binds. The putative DNA binding domain of ACEI (ACEI 382-582 ) was produced in E. coli as a GST fusion for in vitro protein-DNA binding studies. At the beginning, four cbh1 promoter fragments about 300 bp each were amplified by PCR and assayed for binding to the GST-ACEI 382-582 fusion protein. GST-ACEI 382-582 bound to fragments from Ϫ133 to Ϫ392 (fragment 1), Ϫ621 to Ϫ941 (fragment 2), and Ϫ1116 to Ϫ1420 (fragment 4), but not to Ϫ1177 to Ϫ886 (fragment 3). The best results were obtained with the region Ϫ621 to Ϫ941 (fragment 2) (data not shown). Based on DNA sequence features, it was assumed that this region in the cbh1 promoter could be especially important for the regulation of the activity of the cbh1 promoter because it contains repeated nucleotide motifs that are possible targets for transcription factors. There are, for example, three GTGGGG repeats that are binding sites for CREI and mediate glucose repression (9), three CCAAT repeats, two GGCAAA repeats, and three GGCTAA repeats. The following gel mobility shift assay results indicated that the GST-ACEI 382-582 protein bound in vitro to the 170-bp cbh1 promoter fragment 2P, corresponding to the region between Ϫ843 and Ϫ676, and that an overlapping fragment, 2K, corresponding to sequences Ϫ763 to Ϫ667, did not compete for binding (Fig. 4). It was also shown that GST-ACEI 382-582 bound to a 120-bp fragment 2A from Ϫ843 to Ϫ726 (data not shown). Based on these data, it was concluded that a binding site for GST-ACEI 382-582 would be localized between nucleotides Ϫ843 and Ϫ763.
After having roughly identified regions recognized by GST-ACEI 382-582 , oligonucleotide probes were prepared for binding assays (Table I). A series of 30 -40-bp double-stranded oligonucleotides covering the whole 170-bp 2P region was made. Among these, ACEI 382-582 bound to the 36-mer oligonucleotide Ϫ818 2C (Table I and Fig. 5). A series of mutations was generated into the sequence (Table I) in groups of six and subsequently in groups of three bases in order to clarify which nucleotide changes altered the binding properties. Results of the gel mobility shift assays with mutated oligonucleotides used as probes (Fig. 5, A, lanes 7-11, and C, lanes 3-8) or as competitors (Fig. 5B, lanes 5-8) indicated that the mutations in the nucleotides corresponding to Ϫ804 ATGCCTAAA Ϫ796 (complement, Ϫ796 TTTAGGCAT Ϫ804 ) in the native sequence reduced binding of ACEI 382-582 , suggesting that ACEI 382-582 contacts some of these bases.
Interestingly, 13 copies of the GGC(T/A)AA hexanucleotide can be found in the 1.15-kb cbh1 promoter fragment present in the reporter construct pAS3. Therefore, other oligonucleotides representing these GGC(T/A)AA sequences, five of which occur inside and eight outside the 2P fragment, were also used as probes in gel shift assays. ACEI 382-582 also bound to the four oligonucleotides 1D, 1C, 1A, and 3B, which contain aGGCAAA sequences located at Ϫ151, Ϫ253, Ϫ400, and Ϫ1018 in the cbh1 promoter, respectively (Fig. 6, lane 2; data shown for Ϫ151 GGCAAA, fragment 1D), but not to any of the four (g/c/ t)GGCAAA sequences (Table I, oligonucleotides 3A, 1B, 2J, and 2H; data not shown). The hexanucleotide Ϫ151 GGCAAA was mutated into Ϫ151 AATCCC, which abolished binding of GST-ACEI 382-582 (Fig. 6, lane 10) indicating that the Ϫ151 GGCAAA repeat is critical for binding of ACEI 382-582 . A mutation generated elsewhere in the oligonucleotide did not affect binding of ACEI 382-582 (Fig. 6, lane 12). Competition assays with the native and mutated oligonucleotides confirmed the result (Fig.  6, lanes 3-8). Furthermore, ACEI 382-582 did not bind to any of the five GGCTAA sequences present in the 1.15-kb cbh1 promoter fragment (Table I,  Based on the binding data obtained thus far, a common feature in the oligonucleotides giving a positive binding reaction was a 5ЈAGGCA sequence. Therefore oligonucleotides 4A, 1E, 1F, and 3D, representing the remaining AGGCA sequences in the cbh1 promoter, were prepared and tested for ACEI binding. Oligonucleotides 4A, 1E, and 1F gave a positive binding result, but 3D did not (Table I, data not shown). A summary of ACEI-binding and nonbinding sequences is shown in Fig. 7.
Most binding reactions described here were performed with both the GST-ACEI 382-582 fusion and with the thrombincleaved protein preparation releasing the ACEI 382-582 from the GST fusion partner. The thrombin-cleaved ACEI 382-582 -DNA complexes migrated faster in the gel than the larger GST fusions; otherwise, identical binding results were obtained with both protein preparations.
In parallel with the in vitro binding experiments, we applied the yeast one-hybrid system to assess whether ACEI binds in vivo to the defined 170-bp cbh1 promoter region between Ϫ843 and Ϫ676. This fragment was amplified by PCR, and it was cloned in front of the S. cerevisiae HIS3 gene into the vector pHisi-1. The resulting target-reporter construct pPL2 was integrated into the YM4271 yeast genome at the HIS3 locus. As a control, pHisi-1 was integrated into the genome of the YM4271 yeast. The plasmid pARO20 expressing the S. cerevisiae GAL4 activation domain-ACEI DNA binding domain fusion protein (GAL4ad-ACEI 382-582 ) was then transformed into the reporter yeast. The pPL2 yeast transformed with pARO20 grew on SC-Leu-His plates, indicating that the GAL4ad-ACEI 382-582 fusion protein bound to the cbh1 promoter target (Ϫ843-686) and activated transcription of the HIS3 gene ( Fig.   FIG. 2-continued  8). This result is in accordance with the in vitro binding data. The pPL2 reporter yeast transformed with the control vector pGAD10 expressing the GAL4ad without any DNA binding domain did not grow on the test plates, nor did the negative control strains containing the integrated construct pHisi-1 and pARO20 or pGAD10 (Fig. 8).
Effect of the Disruption of the ace1 Gene-In order to study the role of ace1 in Trichoderma, a strain deleted for the ace1 gene was constructed. The ace1 disruption cassette contained in pAS38 included 2.5 kb of the 5Ј region and 2.5 kb of the 3Ј region of the chromosomal ace1 gene, but the protein coding region (except for the last 150 bp) was removed and replaced by the A. nidulans amdS gene. The disruption cassette was transformed into T. reesei, and the transformants were screened for replacement of ace1 by the amdS gene by Southern analysis (data not shown). The desired gene replacement occurred at a frequency of 50%, which indicated that ace1 is not an essential gene.
Whether disruption of ace1 affects the ability of the fungus to grow on cellulose was studied by plating the ace1 disruptants and the host ALKO2221 as single spore colonies on minimal medium containing either glucose or cellulose as the carbon source. The disruptants grew normally on glucose medium as compared with the host, but on cellulose medium, the diameter of ace1 disruptant colonies was smaller than that of the host strain (Fig. 9). Even though the effect of ace1 disruption was observed on cellulose, it is clear that ace1 disruption did not render Trichoderma completely incapable of using cellulose as a carbon source under the conditions tested. DISCUSSION We have shown earlier that the production of cellulolytic enzymes by the fungus T. reesei is subject to transcriptional regulation by the available carbon source (12). In glucosecontaining media, cellulase genes are repressed by CREI (7, 10). In media containing cellulose or its derivatives, cellulase transcription is very strongly induced. The observations that FIG. 3. Amino acid sequence alignment of T. reesei ACEI and the homologous sequences found in A. nidulans and N. crassa expressed sequence tag sequence data banks. The A. nidulans sequence showing similarity to the amino acids 68 -374 of ACEI was assembled from three overlapping clones (n8h08a1.r1, w4e06a1.r1, and g6h12a1.r1), and the A. nidulans sequence having similarity with the amino acids 462-600 of ACEI represents the clone c3b04a1.r1. The N. crassa sequence corresponds to clone b306ne.f1. Identical (*) amino acids and amino acids with high (:) or low (.) similarity are indicated. The alignment was made using the Clustal W (1.74) software. on sorbitol or glycerol alone, no cellulase mRNAs were detected and that the addition of small amounts of the known inducer of cellulase genes, sophorose, into sorbitol or glycerol cultures caused strong induction of the genes suggested that a distinct induction mechanism must exist. This prompted us to proceed toward isolation of genes mediating this regulation.
From our previous deletion analysis of the cbh1 promoter, one could not clearly define regions required for activation because no dramatic differences between different deletion derivatives with respect to their inducibility by sophorose were observed (9). Because there was no knowledge of components involved in activation of cbh1 transcription, a method was required that would allow isolation of transcription activator genes without any previous knowledge of the important DNA sequence elements or of the nature of the activator genes and proteins. Our approach selects for both binding and activation properties of the trans-acting factor and uses a large promoter fragment as the target. This enables simultaneous isolation of activators with different binding specificities. Furthermore, the method is likely to yield activators but not repressors or other DNA-binding proteins without activation properties, which is an advantage as compared with alternative screening methods that are based on binding only, such as one-hybrid or Southwestern, and that also require detailed knowledge on the DNA target. A further complication of one-hybrid screening may be caused by the fact that the GAL4 activation domain is at the N terminus of the fusion, and if the activator cDNA contains an in-frame translation stop codon upstream of the first ATG of the cDNA, those clones will be missed. However, the probability of getting such transcription factors, the function of which requires more than one polypeptide encoded by different genes, is very low using our method or the one-hybrid system. We have demonstrated that a full-length promoter can be used in cloning of regulatory genes. The method should be generally applicable for different promoters and organisms. Limitations may be encountered in such cases, where the promoter in question contains functional target sequences for some yeastencoded factors that dominate the regulation in yeast in such a  4 and 7), or 170-fold (lanes 5 and 8) molar excess over the labeled DNA.

Regions of the cbh1 promoter used in gel mobility shift assays
The 5Ј and 3Ј end points in the cbh1 promoter relative to the initiator ATG and the entire coding strand sequences in case of short oligonucleotides are shown. Changes introduced to the oligonucleotides as compared to the native sequence are underlined. Binding to the GST-ACEI 382-582 protein is indicated (ϩ).
Ϫ172 Ϫ133 ATTCCCTGGAATGAGCTAGTAGGCAAAGTCAGCGAATGTG ϩ way that leakage from the promoter occurs or that the activator function of the searched component is prevented. ace1 was isolated on the basis of the ability of its translation product to activate transcription from the cellulase cbh1 promoter in yeast. In addition to ace1, a second activator gene, ace2, was isolated using the same approach. 4 ace1 is a novel gene encoding a DNA-binding protein that belongs to the Cys 2 -His 2 class of transcription factors. It is the first reported Trichoderma gene with activation properties. Furthermore, except for the glucose repressor cre1 and Aspergillus xylanase activator xlnR, which also affects expression of other hemicellulases and endoglucanases (40,41), no regulatory genes controlling expression of fungal cellulases have been described thus far.
ACEI has one typical Cys 2 -His 2 zinc finger and two more unusual ones. The yeast genome contains approximately 50 proteins that have Cys 2 -His 2 zinc fingers (30), and relatively few of them contain more than two fingers, in contrast to Cys 2 -His 2 proteins of higher eukaryotes, which often contain several fingers. The yeast genome did not appear to contain a homologue of ace1. However, ace1 homologues were identified among A. nidulans and N. crassa expressed sequence tag sequences, suggesting that ACEI might be a regulatory protein specific for filamentous fungi.
According to the in vitro binding data, there are at least eight binding sites for ACEI in the cbh1 promoter. ACEI recognized all AGGCAAA sites and some AGGCA sites preceded by a relatively A-T rich region. It is possible that other variants of the binding sequence exist and will be found, e.g. in other promoters regulated by ACEI. Identical or very closely related sequences occur, e.g. in the promoters of T. reesei cellulase genes egl1, egl5, and cbh2 and the xylanase gene xyn1.
Our results also showed that the disruption of the ace1 gene caused retardation in the radial growth of the colonies on cellulose-containing plates representing conditions under which cellulolytic enzymes need to be formed in order to enable growth. However, growth was not totally prevented, which indicates that significant cellulase expression occurs even in the absence of ace1. It is not possible to conclude yet to what extent ace1 regulates cellulase expression and what is the contribution of each of the several putative binding sites in the cbh1 promoter identified in the in vitro binding assays. It is clear that the presence of one ACEI binding site at Ϫ152 is not enough for cbh1 induction on cellulose because truncated cbh1 promoter derivatives less than 210 bp, containing one ACEI binding site, are not induced by cellulose (11). 4 On the other  Table I for coding strand sequences of the double stranded oligonucleotides used as probes or competitor DNA in the binding reactions. Oligonuclotide 1D represents the native cbh1 promoter sequence between nucleotides Ϫ172 and Ϫ133, 1D2 is mutated in the Ϫ151 GGCAAA sequence, and 1D3 is mutated elsewhere in the oligonucleotide. Unlabeled competitor oligonucleotides were added into binding reactions in 50-fold (lanes 3, 5, and 7) Comparison of suggested ACEI binding and ACEI nonbinding sequences in the cbh1 promoter. The data are combined from gel shift assays. The GGC triplet found in all oligonucleotides in either strand is aligned centrally, and the 5Ј and 3Ј end points relative to the initiator ATG are shown. G and C nucleotides are shaded in dark gray, A nucleotides in medium gray, and T nucleotides in light gray. The complete sequences of the oligonucleotides are shown in Table I. FIG. 5. Determination of the nucleotides involved in binding of ACEI in the cbh1 promoter sequence between the nucleotides ؊818 and ؊780 upstream of ATG. See Table I for coding strand sequences of the double stranded oligonucleotides used as probes or as competitor DNA in the binding reactions. The wild type sequence (2C) was mutated in blocks of 6 (2C2-2C7) or 3 bp (2C8 -2C11). A, gel mobility shift assay of ACEI 382-582 binding to the 32 P-labeled probes 2C (wild type) and 2C2-2C7 (6-bp mutations). 1 ng of each probe and 250 ng of thrombin-cut GST-ACEI 382-582 protein preparation were used. B, gel mobility shift assay with the wild type sequence 2C as a probe and unlabeled competitor oligonucleotides 2C, 2C4, and 2C5, added in a 50or 200-fold excess. The amounts of probe and protein were as in A. C, gel mobility shift assay of ACEI 382-582 binding to the 32 P-labeled probes 2C (wild type) and 2C8 -2C11 (3-bp mutations). The amounts of probe and protein were as in A.
hand, the region from Ϫ373 to Ϫ210 in the cbh1 promoter has been reported to be required for cellulose induction (11). This region carries two ACEI binding sites. Thus, it would be tempting to speculate that ACEI might influence the activity of the cbh1 promoter through these sites on cellulose. In contrast to induction by cellulose, sophorose-induced transcription occurs from the truncated cbh1 promoter derivatives of 210, 184, and 161 bp (78, 52, and 29 bp upstream of the TATA box, respectively) in cultures grown on sorbitol (9). Whether the ACEI binding site Ϫ152 AGGCAAA has a regulatory role or whether induction by sophorose is mediated by another factor remain to be studied. Disruption of the ace1 gene in the promoter deletion strains would clarify the situation. It was recently reported (17) that a ATTGGGTAATA sequence, dissimilar to the in vitro ACEI binding sites, present in the T. reesei cellulase promoter cbh2 is needed for sophorose induction of cbh2 transcription, and it was suggested to consist of adjacent binding sites for the CCAAT-binding factor and another transcriptional activator. This type of sequence, however, is not present in the cbh1 promoter, unless several mismatches are allowed. Consequently, it remains to be studied how many regulatory factors are involved in cellulase regulation and to what extent each of these contribute to sophorose and cellulose mediated induction. Further investigations are in progress to clarify the role of ace1 in induction of cbh1 and other hydrolase genes. FIG. 8. In vivo one-hybrid assay for binding of GAL4 activation domain-ACEI DNA binding domain fusion protein to the cbh1 promoter region from ؊843 to ؊676. The target-reporter strains contain the pPL2 (cbh1 (-843-676) -HIS3) or the pHisi-1 (HIS3) construct integrated in the yeast genome. The gene encoding GAL4ad-ACEI 382-582 fusion protein was carried on the plasmid pARO20. pGAD10 (GAL4ad without a DNA binding domain) was used as a vector control. The colonies were streaked onto SC-Leu plates and replicated onto fresh selective SC-Leu-His medium containing 45 mM 3-aminotriazol and onto fresh SC-Leu medium. Plasmid pARO20 supported growth of the pPL2-containing yeast under the selective conditions. FIG. 9. Growth of the ace1-disrupted T. reesei strain (⌬ace1) and the host strain on plates containing either glucose or cellulose as the carbon source. The plates were photographed after 4 (glucose) or 7 days (cellulose) of cultivation.