Different inducibility of expression of the two xylanase genes xyn1 and xyn2 in Trichoderma reesei.

Regulation of formation of the extracellular xylanase system of Trichoderma reesei QM 9414 during growth on xylan, cellulose, and replacement onto a number of soluble inducers was investigated by Northern analysis of xyn1 and xyn2 transcripts and by the use of the Escherichia coli hph (hygromycin B-phosphotransferase-encoding) gene as a reporter. Whereas the xyn1 promoter is active in the presence of xylan and xylose, and virtually silenced in the presence of glucose, the xyn2 promoter enables basal transcription at a low level, but is enhanced in the presence of xylan and xylobiose and also of sophorose or cellobiose. The respective regulatory nucleotide regions were localized on a 221-base pair fragment and a 55-base pair fragment of the xyn1 and xyn2 5′-upstream noncoding sequences, respectively. Electrophoretic mobility shift assays, using cell-free extracts, identified induction-specific protein-DNA complexes: one complex of high mobility was observed under basal, noninduced conditions (glucose) with xyn2, which was in part replaced by a slow-migrating complex upon induction by xylan or sophorose. Both complexes bound to a CCAAT box. With xyn1, the induced complex also binds to a CCAAT box, but this binding is not observed in the presence of the carbon catabolite repressor Cre1, which binds to a nearby located consensus motif.

In contrast, the regulation of their formation in Trichoderma has not yet received sufficient attention: xylanases are generally produced together with cellulases during growth of the fungus on macromolecular substrates derived from plant polysaccharides, which inevitably always contain cellulose and xylan. The resulting xylanase to cellulase ratio has been shown to be directly proportional to the xylan to cellulose ratio in the growth substrate (Senior et al., 1989). These data would be consistent with results from induction studies, which showed that xylanase and cellulase biosynthesis in T. reesei is differentially regulated (Hrmova et al., 1986). In contrast, Royer and Nakas (1990) reported that efficient xylanase induction in Trichoderma longibrachiatum required the simultaneous presence of xylo-as well as cellooligosaccharides. In all these studies, the formation of xylanases was determined by enzyme assays only, and since at least endoglucanase I from T. reesei has xylanase activity also (Biely et al., 1991), these results are difficult to interpret.
The regulation of xylanase biosynthesis by T. reesei has not yet been studied on the molecular level. The formation of T. reesei cellulases, for which some data are as yet available, has been shown to be regulated on the level of transcription (El-Gogary et al., 1989;Abrahao-Neto et al., 1995), and it may be reasonable to assume that this is also the case with its xylanases. In this study, we will show that this is the case, and that the expression of xyn1 and xyn2 is regulated by different inducers derived from xylan and cellulose, but both processes involve CCAAT-binding protein-DNA complexes.

EXPERIMENTAL PROCEDURES
Microbial Strains-T. reesei QM 9414 was used throughout this study. Its maintenance and conditions for cultivation have been reported earlier (Kubicek et al., 1988). T. reesei TU-6 (Gruber et al., 1990a) was used for transformation. For experiments with soluble inducers, the replacement technique described by Sternberg and Mandels (1979) was used.
Plasmids and Manipulation of DNA-Plasmids pFG1 (Gruber et al., 1990b) and pLMRS3 (Mach et al. (1994), see below) were obtained from our department stock.
The pRAMB series of reporter plasmids was developed from plasmid pUC19and generally contained the E. coli hph (hygromycin B-phosphotransferase-encoding) gene fused to the T. reesei cbh2 3Ј-noncoding regions as reporter and the T. reesei pyr4 gene (Gruber et al., 1990b) as a marker for transformation. To construct pRAMB1, oligonucleotides CKT 085 (5Ј-ATCTCGAGGTCGACGCAAATGGCCTCAAGCAACTAC-G-3Ј) and CKT 086 (5Ј-ATTCTAGAGATGTTATTTGTGCGTGTTTTCC--3Ј), which represent sequences Ϫ534 to Ϫ513 and Ϫ1 to Ϫ23 in xyn1, respectively, were used to amplify a 534-bp 1 fragment from the xyn1 5Ј-noncoding sequences, thereby also generating additional XhoI/SalI and XbaI terminal sites. This fragment was used to replace a SalI/XbaI fragment of the pki1 promoter in pRLM EX 30 (Mach et al., 1994), thereby fusing it to the E. coli hph gene, followed by the T. reesei cbh2 terminator. Finally, a 2.7-kb SalI fragment containing the T. reesei pyr4 gene (Gruber et al., 1990b) was inserted into a single XhoI site, to yield pRAMB1.
Deletion Analysis-pRAMB1321 and pRAMB1100, pRAMB1 analogues carrying 5Ј deletion variants of the xyn1 5Ј-upstream sequences, were constructed as follows: pRAMB1 was digested with HpaI/NsiI, and the resulting 1.3-kb fragment (containing the hph structural gene and part of the xyn1 5Ј-upstream sequences) fused to a 6.3-kb SalI/NsiI fragment of pRAMB1 after filling in the SalI protruding ends with Sequenase version 2.0 (U. S. Biochemical Corp., Cleveland, OH). This plasmid contains the xyn1 5Ј-upstream sequences from Ϫ1 to Ϫ321 and was named pRAMB1321. To construct pRAMB1100, pRAMB1 was cleaved with EcoRI, the resulting 350-bp fragment was treated with Sequenase version 2.0 (U. S. Biochemical Corp., Cleveland, OH) and subsequently cleaved with XbaI. The resulting 100-bp fragment was then used to replace the xyn1 fragment in pRAMB1, by removal of the original xyn1 fragment by a SalI digestion, followed by the generation of blunt ends with Sequenase and subsequent cleavage with XbaI.
pRAMB2316, pRAMB2235, and pRAMB 2180, pRAMB2 analogues carrying 5Ј deletion variants of the xyn2 5Ј-upstream sequences were constructed as follows: pRAMB2316 was obtained via ligation of a 1.3-kb XhoI/NsiI fragment of pRAMB2 to a 6.3-kb SalI/NsiI fragment of pRAMB2. To construct pRAMB2235, pRAMB2 was digested with SalI/ XbaI, the resulting 7.6-kb fragment was isolated, the ends were made blunt with Sequenase version 2.0 (U. S. Biochemical Corp.), and the fragment religated to give pRAMB2235. For construction of pRAMB2180, a 1.2-kb HindIII/NsiI fragment of pRAMB2 was isolated, and the HindIII site was blunt-ended with Sequenase version 2.0 (U. S. Biochemical Corp.). This fragment was then ligated to a 6.3-kb SalI/ NsiI fragment of pRAMB2, whose SalI site had previously been bluntended with Sequenase version 2.0. All vector constructs were verified via sequencing by the dideoxynucleotide chain termination method (Sanger et al., 1977).
Transformation-E. coli transformations were carried out according to standard techniques (Sambrook et al., 1989). Transformation of T. reesei TU-6 was carried out as described by Gruber et al. (1990b).
Enzyme Assays-Xylanase activity was assayed as described previously (Törrönen et al., 1992), using Lenzing xylan (Lenzing AG, Austria) as substrate. Activities are given as units, 1 unit being defined as the release of 1 mol of reducing sugar per min under these conditions. Protein concentrations in the culture filtrate were determined by the dye-binding procedure (Bradford, 1976).
Hygromycin B Resistance Assay-T. reesei transformants were analyzed for hygromycin B resistance by means of a plate assay, which measures the increase in colony diameter at different hygromycin B concentrations. To obtain linear growth rates, transformants were pregrown on the same carbon source and agar plugs (4-mm diameter) overgrown with mycelium from the growing front were excised, placed in the middle of fresh plates, and incubated at 28°C in the dark. The rates of increase in colony diameter were calculated from at least three measurements during this incubation.
RNA Isolation and Analysis-T. reesei mycelium was harvested on a sinter funnel and ground to a fine powder under liquid nitrogen. Total cellular RNA was isolated as described by Chomczynski and Sacchi (1987). Following electrophoretic separation, RNA was blotted onto nylon membranes (Hybond -N, Amersham) and hybridized according to standard protocols (Sambrook et al., 1989) at 42°C for 20 h. Washing was performed with 2 ϫ SSC ϩ 0.1% (w/v) SDS at room temperature (2 ϫ 5 min) and at 60°C and 64°C, respectively (30 min). Slot-blot hybridization was performed in a PR 600 Slot Blot apparatus (Bio-Rad).
DNA Fragments, Preparation of Cell-free Extracts, and Electrophoretic Mobility Shift Assay (EMSA)-A xyn1 fragment, containing the regulatory nucleotide regions from Ϫ534 to Ϫ343, was obtained by cleavage of pRAMB1 with SalI/XbaI, isolation of the resulting 538-bp fragment, subsequent cleavage with NsiI, and isolation of a 200-bp fragment which was used in EMSA. A fragment of the regulatory region of xyn2 was obtained by polymerase chain reaction amplification, using the primers CKT 088 and CKT 144 (5Ј-ATCATTGATGAAAGGGA-3Ј), and pRAMB2 as a template. The resulting amplicon was cleaved with HindIII/XbaI, and the resulting 55-bp fragment was used for EMSA. Both fragments were end-labeled with the appropriate [␣-32 P]dNTPs, using Sequenase version 2.0 (U. S. Biochemical Corp.) and purified by nondenaturating polyacrylamide gel electrophoresis.
Preparation of cell-free T. reesei extracts following growth on various carbon sources and their use in EMSA was carried out described previously (Stangl et al., 1993). Binding was achieved by incubating 100 g of cell-free protein with 5 ng of labeled fragment.

RESULTS
Induction of Xylanase Activity-Preliminary studies in our laboratory showed that cultivation of T. reesei on xylan resulted in highest xylanase activities, whereas those on cellulose were considerably lower. Even lower activities were observed on lactose, and no xylanase activity at all could be detected upon cultivation on glucose (data not shown). Since xylan and cellulose are polysaccharides which cannot be taken up by the fungus, the induction of xylanase activity is likely mediated by   FIG. 1. Northern slot-blot analysis reporting the expression of  xyn1 and xyn1::hph (A) and xyn2 and xyn2::hph fusions (B), respectively. act1 indicates the actin controls. Slots corresponding to the different inducing or noninducing sugars (G, glucose; Xo, xylose; So, sophorose; Xb, ␤1,4-xylobiose) were loaded with total RNA in the following concentrations: 5, 2, 0.5, and 0.005 g. products of xylan catabolism. To investigate this, xylobiose, xylose, and metabolites of their further catabolism (xylitol, xylulose; cf. Witteveen et al. (1989)), as well as the cellulose degradation and transglycosylation products cellobiose and sophorose, the most effective cellulase inducer (Sternberg and Mandels, 1979), were added to washed, glycerol-pregrown mycelia. In these experiments, the cellulase-inducer sophorose induced the highest xylanase activities, and an equimolar mixture of cellobiose and xylobiose also was induced efficiently (Table I). Conversely, xylose as well as xylulose induced moderate xylanase activities, and xylobiose was only a poor inducer. Xylitol, arabitol, and arabinose, although active inducers in Aspergillus tubigensis (De Graaff et al., 1993), did not induce any xylanase activity in T. reesei.

The Expression of the Two Major Xylanase-encoding Genes xyn1 and xyn2 Is Mediated by Different Inducers-Northern
analysis was carried out to study expression of the two xylanase genes (Fig. 1). A different pattern of induction was apparent: whereas xyn1 mRNA was detected only in mycelia transferred to xylose, xyn2 mRNA was formed under all conditions tested, and xyn2 expression was also detected on glucose (Fig. 1). This indicates that the two xylanase genes are differently regulated and hence their translation products contribute to different extents to the activities induced by the various carbon sources.
Identification of Nucleotide Regions in the 5Ј-Upstream Sequences of xyn1 and xyn2 Responsible for Transcriptional Regulation-To identify nucleotide motifs responsible for the induction of xyn1 and xyn2 gene expression, we first fused the 538 and 844 bp of the respective promoters to the E. coli hph (hygromycin B phosphotransferase-encoding) gene as a reporter and studied whether the formation of hph-mRNA from this fusion parallels that of xyn1 and xyn2. The results, documented in Fig. 1, A and B, prove that this is the case and that the respective promoter fragments therefore carry the informa-tion required for the induction of transcription of these two genes.
To localize the promoters of the two xylanase genes in more detail, we performed 5Ј-deletions of the xyn1 and xyn2 5Јupstream fragments and investigated the effect of these deletions on the formation of hygromycin B resistance. Since there is not yet a targetted integration system available in T. reesei, a population of 8 -15 transformants was investigated for every deletion, in order to account for position and multicopy effects. In xyn1 (Fig. 2A), a removal of a region between Ϫ538 and Ϫ321 resulted in a complete loss of expression on every carbon source tested, except for very low growth on glucose in the presence of very low concentrations of hygromycin B, indicating a very low constitutive level of expression. This level of expression was still detectable when the 5Ј-sequences were further shortened to Ϫ100.
A similar analysis with xyn2 showed that a fusion of hph with a 235-bp promoter fragment resulted in a full pattern of regulation of hph expression, whereas virtually all regulation was lost in fusions with a 180-bp fragment (Fig. 2B). This suggests that the regulatory elements are located between Ϫ235 and Ϫ180. No constitutive level of expression of the 180-bp truncated promoter could be detected even at very low hygromycin B concentrations.
The sequence of the two nucleotide fragments shown to be relevant to xyn1 and xyn2 gene expression is given in Fig. 3, A  and B; both xyn1 and xyn2 contained consensus sequences for binding of the Cre1 catabolite repressor protein (Cubero and Scazzocchio, 1994;Strauss et al., 1995;Mach et al., 1996). Also CCAAT boxes, which have been shown to be functionally involved in the induction of cbh2 gene expression, 2 were present in both promoters. No other nucleotide sequences with sufficient similarity to known DNA-binding targets were detected. The low constitutive level of xyn2 may be due to a TATA box located at Ϫ90.
Binding of Proteins to the xyn1 and xyn2 Promoters-In order to provide evidence that the regions of xyn1 and xyn2 described above are responsible for the binding of respective transcriptional activators, cell-free extracts were prepared from T. reesei mycelia, grown under various inducing and noninducing conditions, and used in EMSA with the DNA-fragments determined as relevant in vivo. Different results were obtained for xyn1 and xyn2, respectively. The xyn1 promoter formed two protein-DNA complexes of relatively low mobility (i.e. high M r ) with cell-free extracts from xylan (or xylose, data not shown)-induced cultures. Addition of an excess of unlabeled oligonucleotide CKT AAT , which contains a CCAAT motif, resulted in the removal of these two complexes and in the formation of a single complex of almost as fast mobility as the free 200-bp fragment (Fig. 4A, track 3). Since the addition of a similar oligonucleotide (CKT TTT ), in which only the CCAAT motif was changed to CCTTT, had no effect, we conclude that the two inducing complexes are bound to the CCAAT box at Ϫ430. An even slower migrating DNA-protein complex was obtained with cell-free extracts from mycelia grown on glucose, whose appearance was not eliminated by the addition of an excess of oligonucleotide CKT AAT . Evidence is described elsewhere that this complex involves binding of the carbon catabolite repressor protein Cre1 to its target sequence (Mach et al., 1996). It is therefore concluded that transcriptional regulation of the xyn1 promoter involves at least the mutually exclusive binding of the CCAAT-binding and the Cre1-binding protein complex.
Protein binding to the xyn2 promoter showed a simpler picture (Fig. 4, B and C): extracts from noninduced mycelia produced one major, specific protein-DNA complex, which was accompanied by a number of faint, unspecific complexes. The formation of this major DNA-protein complex was eliminated by the addition of an excess of oligonucleotide CKT AAT but not CKT TTT and is therefore due to binding to the CCAAT box at Ϫ216. Upon incubation of the 55-bp DNA fragment with cellfree extracts from induced cultures (xylan, Fig. 4C, tracks 1-5; sophorose, Fig. 4B), one additional complex of slower mobility was observed. This complex was not seen in the presence of an excess of a cold CCAAT fragment and, therefore, also binds to the CCAAT box. However, whereas no competition was observed by the addition of nonspecific DNA, the addition of an excess of CKT TTT bearing the CCTTT motif equally prevented complex formation. Since the CCAAT/CCTTT boxes are the only regions of similarity between the 55-bp xyn2 promoter fragment and oligonucleotides CKT AAT and CKT TTT , we conclude that the induced protein complex binds DNA at the CCAAT motif in a different way than the constitutive protein complex.

DISCUSSION
Expression of xyn1 and xyn2 in T. reesei is regulated in a different way: whereas xyn1 expression is triggered by the presence of xylan and its final degradation product xylose only, that of xyn2 is also initiated upon growth on cellulose and supply of the cellulase-inducer sophorose. Differences in the FIG. 3. Nucleotide sequences of the xyn1 (A) and the xyn2 (B) 5-noncoding sequences. For areas identified as relevant for regulation of transcription, the sequence of the complementary strand is included. Within these, Cre1 target consensus sequences (5Ј-SYGGRG-3Ј; Cubero and Scazzocchio (1994)) are indicated by bold letters whereas CCAAT motifs are boxed. regulation of xyn1 and xyn2 transcription were also noted in the inducibility of xyn1 and xyn2 by xylan degradation products: whereas expression of xyn1 was observed upon incubation of mycelia with xylose but not with xylobiose, xyn2 expression was triggered by xylobiose only. Both compounds accumulate extracellularly during the breakdown of xylan, but xylose uptake occurs at a faster rate than xylobiose hydrolysis, 3 and the addition of only xylobiose thus will result in a slower supply of xylose to the fungus, than during growth on xylan or when xylose is added directly. The unique inducibility of xyn1 and xyn2 transcription by xylose and xylobiose, respectively, may reflect a triggering of xylan-dependent signal transduction by different intracellular steady-state concentrations of xylose (or one of its catabolites such as xylitol, xylulose etc.). The role of ␤-xylosidase in this process is apparent and warrants further investigation.
Using 5Ј-deletion analysis and EMSA, the nucleotide areas responsible for regulation of xyn1 and xyn2 transcription were localized. Evidence was obtained that the induction of xyn1 and xyn2 transcription involves the binding of nuclear proteins to a CCAAT box. CCAAT boxes have been observed in the regulatory regions of several fungal genes, and their function has been proven in Aspergillus nidulans amdR (van Heeswijck and Hynes, 1991) and yA (Aramayo and Timberlake, 1993) and in T. reesei cellobiohydrolase II-encoding gene cbh2. 2 In the latter case, site-directed mutagenesis of the CCAAT box revealed that this motif mediates induction by cellulose and sophorose, which fits perfectly to the present results. However, transcription of cbh2 is not triggered by xylan or its degradation products, and the protein components involved must therefore, at least in part, be different. While a number of diverse CCAAT-binding proteins has been described in mammalian cells (Chodosh et al., 1988;Benoist and Mathis, 1990), the HAP2/HAP3/HAP5 proteins of Saccharomyces cerevisiae (Forsburg and Guarente, 1989;McNabb et al., 1995) and the homologues in Schizosaccharomyces pombe (Olesen et al., 1991) and Kluyveromyces lactis (Mulder et al., 1994) are the only components known from fungi. The A. nidulans hapC (HAP3 homologue) has recently been cloned (Papagiannopoulos et al., 1995), and we thus assume its presence also in T. reesei by analogy. In A. nidulans amdR, the CCAAT box is responsible for the basal transcriptional level (van Heeswijck and Hynes, 1991), and the present data suggest that this also seems to be the case in T. reesei xyn2.
The simplest model to explain the regulation of xyn2 would be to postulate the binding of a Hap2/Hap3 (and eventually Hap5) complex to its promoter under basal conditions, which associates with additional components upon induction by xylan or cellulose. This model basically also can be applicable to the regulation of xyn1, yet is complicated by the apparent involvement of additional factors. The lack of binding of protein extracts from glucose-grown cultures to the CCAAT motif, and the concomitant observation of a Cre1-DNA complex would be a typical example of carbon catabolite repression. This coincides with our findings (Mach et al., 1996) that functional impairment of the Cre1 target sequence in xyn1 in vivo allows the gene to be constitutively expressed at a level comparable to that of basal xyn2 transcription. As the CCAAT box and the Cre1-binding consensus are separated by only 31 bp and the Cre1-binding site lies downstream of the CCAAT box, repression may act either by competition for binding sites or by inhibition of contact with the RNA-polymerase II initiation complex. Upon induction, binding of the CCAAT-protein complex instead of the Cre1-DNA complex is observed. We do not know whether this requires a functional inactivation of the Cre1 complex or a gain of binding strength of the CCAATbinding complex. From these data, we assume as a model for further work that the expression of xyn1, in contrast to that of xyn2, is regulated by competition of DNA-binding complexes for binding at the Cre1 consensus sites and the CCAAT box.