Triggering of cellulase biosynthesis by cellulose in Trichoderma reesei. Involvement of a constitutive, sophorose-inducible, glucose-inhibited beta-diglucoside permease.

We prepared [U-14C]cellobiose by cultivating Acetobacter pasteurianus in the presence of [U-14C]glucose and hydrolyzing the [U-14C]cellulose formed with beta-glucosidase-free cellulase from Trichoderma reesei. This 14C-labeled cellobiose was used to investigate the presence of an uptake system for cellobiose in T. reesei. Evidence was obtained for the presence of a high affinity (Km for cellobiose 0.3 microM) but low activity (2.5 milliunits/mg fungal dry weight) cellobiose permease. The permease is formed constitutively, but higher levels are formed after addition of sophorose (glucosyl-beta-1,2-diglucoside), a reputed cellulase inducer. The permease appears to be specific for beta-diglucosides, as the uptake of [U-14C]cellobiose is inhibited by sophorose, gentiobiose (glucosyl-beta-1,3-glucoside), and cellobiose. Under these conditions, cellooligodextrines (n, 4-7; final concentration, 1 mM) are not inhibitors. Glucose, but no other monosaccharides, inhibits the permease. The hypersecretory mutant T. reesei RUT C-30 exhibits elevated permease activities, whereas in T. reesei QM 9979, a mutant strain defective in the induction of cellulases by cellulose or sophorose, strongly reduced permease activities were demonstrated. The results stress a hitherto not recognized point of control in the induction of cellulases by T. reesei at the level of uptake of cellulose oligosaccharides.

ity (K,,, for cellobiose 0.3 pM) but low activity (2.5 milliunitdmg fungal dry weight) cellobiose permease. The permease is formed constitutively, but higher levels are formed after addition of sophorose (glucosyl-j3-1,2-diglucoside), a reputed cellulase inducer. The permease appears to be specific for 8-diglucosides, as the uptake of [U-'4C]cellobiose is inhibited by sophorose, gentiobiose (glucosyl-8-1,3-glucoside), and cellobiose. Under these conditions, cellooligodextrines (n, 4-7; final concentration, 1 DIM) are not inhibitors. Glucose, but no other monosaccharides, inhibits the permease. The hypersecretory mutant T. reesei RUT C-30 exhibits elevated permease activities, whereas in T. reesei QM 9979, a mutant strain defective in the induction of cellulases by cellulose or sophorose, strongly reduced permease activities were demonstrated. The results stress a hitherto not recognized point of control in the induction of cellulases by T. reesei at the level of uptake of cellulose oligosaccharides.
Despite the progress that has been made in the molecular biology of the Trichoderma reesei cellulase system (Penttila et al., 1991), it is still only poorly understood how the biosynthesis of cellulases is triggered by the extracellular, insoluble polysaccharide cellulose (Kubicek et al., 1993). The available evidence suggests that low levels of constitutively formed cellulases (particularly cellobiohydrolase 11) are responsible for the initial attack on the cellulose and thereby release cellobiose (Kubicek et al., 1988;El-Gogary et al., 1989;Messner et al., 1992). Further events in the inductory cascade are still speculative; however, cellobiose, in theory the most logical inducer of cellulases and in fact an inducer of cellulases in other fungi (Bisaria and Mishra, 1989), inhibits cellulase formation in T. reesei (Fritscher et al., 1990). In contrast, the * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U. disaccharide sophorose (glucosyl-@-1,2-glucoside) is a strong cellulase inducer (Sternberg and Mandels, 1979). Its formation in vivo has been shown (Vaheri et al., 1979) and may be due to the transglycosylating ability of either @-glucosidase (Schmid and Wandrey, 1987) or endo-@-1,4-glucanase (Claeyysens et al., 1990).
One of the uncertainties within this model is the mode of uptake of cellobiose and sophorose. Uptake of cellobiose has been demonstrated in some yeasts and Escherichia coli mutants (Freer and Greene, 1990;Schnetz et al., 1987) but not yet in filamentous fungi and may not be required for cellobiose metabolism in T . reesei because of the presence of an extracellular constitutively formed @-glucosidase (Umile and Kubicek, 1986;Chirico and Brown, 1987;Hofer et al., 1989).
We have previously studied the metabolism of cellobiose by T . reesei and concluded that both a permease as well as @glucosidase are present and involved (Fritscher et al., 1990). A more detailed study on the putative permease was, however, hampered by the lack of availability of radioactively labeled cellobiose and by the interference of @-glucosidase in the uptake assay.
Here we will describe the preparation of [U-'4C]cellobiose from Acetobacter pasteurianus [U-14C]cellulose, its use in the assay of the presence of a @-linked diglucoside permease of T . reesei, and the characterization of its major properties.

Trichoderma Strains and Culture Conditions-The
strain used throughout these studies was T. reesei QM 9414, which was obtained from ATCC (No. 26921). In selected experiments (when stated explicitly), the following strains were also used ( a ) a recombinant strain T . reesei RLMEx40#1, which carries an antisense copy of the bgll (pglucosidase-encoding) gene under the control of the pkil (pyruvate kinase-encoding) promoter and which does not form @-glucosidase during growth on glucose'; ( b ) the hypersecretory strain T. reesei ; and (c) the cellulase-negative mutant strain T. reesei QM 9979 (Mandels and Reese, 1960). The strains were maintained on malt agar. Cultures were grown by inoculating lo6 spores/ml in Mandels and Andreotti (1978) medium, with glucose (176, w/v) as a carbon source, in 1-1 flasks containing 250 ml of medium using a rotary shaker at 30 "C. A. pasteurianus was obtained from Dr. W. Steiner, Institute of Biotechnology, Graz, Austria.
Preparation of [U-'4C]Cellobiose-To prepare 14C-labeled cellobiose, we cultivated Acetobacter pasteurianus, a cellulase excreting bacterium, in 0.1-1 Erlenmeyer flasks containing 20 ml of glucoseyeast extract-peptone medium (Forng et al., 1989) for 3 days at 26 "C at 250 rpm. 1 ml of this culture was then used to inoculate 10 ml of the same medium in 100-ml flasks, yet contained only 2 g/liter of glucose and 2 pCi/ml of [U-'4C]glucose. Incubation was carried out ~ ~ ~~~ R. L. Mach, J. Straws, R. Gonzalez, and C. P. Kubicek, unpublished data. as described above for further 85 h. Thereafter, the bacterial suspension was then decanted, and the cellulose matrix formed was washed twice with 50 ml of dH,O. The cellulose was then boiled (2 X 30 min) in 2% (w/v) NaOH and thereafter washed successively with each 50 ml of dHnO, 1% (w/v) acetic acid, d H Z 0 , and finally 50 ml of sodium citrate buffer, pH 5. Centrifugation in an Eppendorf centrifuge (10 min) yielded approximately 1 ml of swollen cellulose, to which 40 pl of cellulase (prepared as described below) were added. The mixture was incubated at 37 "C over night and then centrifuged again. The supernatant was recovered, 2 volumes of chilled ethanol (70%) were added, and the mixture was incubated for 15 min at -70 "C. The precipitate formed was removed by centrifugation (as described above), and the supernatant was concentrated to 0.5 ml under reduced pressure, fractionated on Bio-Gel P-2 (17 X 1.1 cm), and equilibrated in 25 mM sodium citrate buffer, pH 5, at a flow rate of 0.3 ml/min. Fractions of 1 ml were collected and analyzed for oligosaccharides by TLC (Kennedy et al., 1979). Fractions containing purified [U-"C] cellobiose were lyophilized, taken up in 1 ml of 10% (w/v) ethanol, and stored at -20 "C. The specific radioactivity of [U-'4C]cellobiose was 37 mCi/mmol.
Assay of Cellobiose Uptake-16-h old mycelia of T. reesei, grown on 1% (w/v) glucose as the exclusive carbon source, were harvested by filtration at 0 "C, washed with 2 volumes of 25 mM sodium succinate buffer, pH 6.0, containing 6 mM MgC12, and finally suspended in the same buffer at a final density of 0.7 g of dry weight/ liter and incubated in 20-ml aliquots in 0.1 liter Erlenmeyer flasks in a water bath incubator (28 "C, 150 rpm) for 30 min. The uptake was then initiated by the addition of [U-'4CC]cellobiose at the final concentration indicated at the respective experiments. After appropriate times (i.e. 3, 6, and 9 min in the standard assay), 1 ml of the broth was pipetted to 0.2 ml of ice-cold 0.1 M cellobiose. The suspensions were filtered through Millipore HAWP membranes (0.45 pm), washed with 2 X 2 ml of 0.1 M cellobiose, and counted in 10 ml of scintillation fluid (Beckman Ready Value). Preliminary experiments were performed to ensure that uptake was linear with time and mycelial concentration.
When other compounds were added to the uptake assay, this was done 30-60 s before the addition of [U-'4C]cellobiose, except when indicated otherwise.
Transport of [U-'4C]glucose was carried out in the same way, using a final concentration of 50 p~ glucose.
Induction of @-Diglucoside Permease in T. reesei-To investigate the inducibility of @-diglucoside permease, T. reesei was pregrown in Mandels and Andreotti medium with glucose as the carbon source as described above. The mycelia was harvested by filtration (without suction), washed, transferred after 20 h to fresh medium lacking carbon source, peptone, and nitrogen source, distributed in 20-ml aliquots into 100-ml Erlenmeyer flasks, and incubated further at 200 rpm (30 "C). Mycelial densities were determined separately and varied between 0.7 and 1.4 g/liter. This had no effect on the results. After incubation for 1 h, the putative inducer was added in a total volume of 0.5 ml to give the final concentration as indicated. Thereafter 2ml samples were withdrawn in intervals as indicated, washed with cold medium lacking inducer, and used to assay @-diglucoside permease activity as described above, using incubation times of 4 and 8 min.
Determination of Cellulase and @-Glucosidase Activity-This was carried out as described previously (Kubicek, 1981).
Quantification of Cellobiohydrolase Z Formation-This was done by enzyme-linked immunosorbent assay, using a monoclonal antibody (CH 16) directed against the core of cellobiohydrolase I as described previously (Kolbe and Kubicek, 1990).
Preparation of @-Glucosidase-free Cellulase-@-Glucosidase was removed from Celluclast by adsorption to PBE94 polybuffer ion exchanger in 0.025 M Tris-HC1 buffer, pH 7.5, as described previously (Kubicek, 1987). The material, which does bind to the column under these conditions, contains more than 90% of the cellulase but less than 2% of the @-glucosidase activity present in the original preparation, respectively.
Purification of @-Glucosidase-@-Glucosidase I (IP 8.7, M, = 85,000) from T. reesei was purified as described by Hofer et al. (1989). The final preparation had a specific activity of 46 units/mg.
Isohtion of Cell Walls from T. reesei-Cell walls from T. reesei, grown on glucose as a carbon source for 18 h, were prepared as described by Messner and Kubicek (1990).

RESULTS
Demonstration of a Cellobiose Permease in T. reesei-When T. reesei was pregrown on glucose as a carbon source, the mycelia harvested, washed, replaced to succinate buffer, and [U-'4C]cellobiose added, the 14C-label was taken up by the mycelium (Fig. la). Using a mycelial density between 0.7 and 1 mg of dry weight/ml, the rate of uptake was linear for at least 15 min. Control experiments with isolated cell walls of T. reesei showed no binding of radioactive label, which proved that the uptake of [U-'4C]cellobiose was not the result of physical adsorptipn (data not shown). The uptake rate displayed a biphasic dependence on the cellobiose concentration (Fig. 16). We assumed that this may be due to an interference by the fungus' plasma membrane bound P-glucosidase. Since this enzyme has a K, of 1.5 mM for cellobiose and a V,,, of 0.080 units/mg fungal dry weight (Umile and Kubicek, 1986), it competes with the uptake system at higher cellobiose concentrations. To prove this assumption, several experiments were carried out: (a) classical glucose trapping by adding hexokinase/Mg-ATP to the uptake assay (De La Fuente and Sols, 1962)  which is not taken up by T. reesei); (b) addition of purified pglucosidase to the uptake assay to increase the total p-glucosidase activity present; (c) carrying out the uptake experiments with a recombinant of T. reesei, which contains an antisense-bgll -DNA under the T. reesei pki (pyruvate kinase) promoter and therefore does not form any 6-glucosidase in the presence of glucose'; (d), finally, using an inhibitor of pglucosidase, delta-gluconolactone, in these experiments (Table I) provided evidence for an interference of p-glucosidase with the cellobiose transport assay at higher cellobiose concentrations (0.1 mM). The only 2-fold increase in uptake by 2 units/ml of extra p-glucosidase at low cellobiose concentrations makes it unlikely that the low activities of constitutively present ,&glucosidase activity of T. reesei interfere with the assay under these conditions. This concentration was thus used routinely in all further experiments. Specificity of T. reesei Cellobiose Uptake-We made use of competition experiments to investigate the substrate specificity of T. reesei cellobiose permease (Table 11). They suggest that the permease is specific for cellobiose, laminaribiose, and sophorose and has no affinity for oligodextrins of higher chain length. Other disaccharides (sucrose, lactose, and xylobiose) had no effect on the transport of [U-'4C]cellobiose even in 103-fold higher concentrations, suggesting that they are not substrates of the permease. We thus propose the name "@linked diglucoside permease" for it and will use this term further throughout the paper.
Glucose, but not other monosaccharides such as galactose or mannose, inhibited [U-'4C]cellobiose uptake with a K, of 4.0 mM. To rule out the possibility that the @-linked diglucoside permease can transport glucose, we assayed [U-14C]glucose uptake in the presence of 1 mM cellobiose.
Properties of the P-Linked Diglucoside Permease-The uptake of cellobiose exhibited a pronounced pH dependence displaying the highest uptake rates at pH 5. To investigate whether cellobiose uptake was due to facilitated diffusion or to carrier-mediated transport, the effect of inhibitors of ATP formation (NaN3, 2,4-dinitrophenol) and of a ionophore (carbonyl cyanide p-trifluoromethoxyphenyldrazone) were studied (Table 111). The data obtained suggest that cellobiose uptake requires an active ATP gradient over the plasma membrane, which is directly coupled to H' . Stimulation of Cellobiose Uptake by Sophorose-In the com- Final &glucosidase concentration was 2 unitslml assay. e The p-diglucoside permease assay was carried out with the recombinant strain T. reesei R L M~~~O # l .

~-~~~l u~o~i d e Permease from T r i~~d e r m~ reesei
The &glucosidase activity of this strain at the time of use for the uptake experiments was less than 0.1 milliunits/mg fungal dry weight.
The A-gluconolactone concentration used was 1 mM in the assay.   petition experiments shown in Table 11, we observed that prolonged incubation in the presence of sophorose stimulated cellobiose uptake. We have therefore investigated the inducibility of @-diglucoside permease by sophorose in a resting cell system, which has previously been successfully used for studies on the induction of cellulases (Sternberg and Mandels, 1979). In this system, the addition of 0.1 mM sophorose promoted a 2-3-fold rise in p-diglucoside permease activity within 70-100 min (Fig. 2a). When an inhibitor of translation (cycloheximide) was added simultaneously with sophorose, this stimulation did not occur. The kinetics of stimulation show that the @-linked diglucoside permease is formed before cellulase (assayed as cellobiohydro~ase I, which accounts for the major portion of the cellulol~ic enzyme system) activity (Fig. 2b). From these results we conclude that sophorose stimulates the formation of cellobiose permease in T. reesei either by induction or by inducing an activator. Maximal stimulation was observed at 0.2 mM sophorose (Fig. 2c).

@-Linked Diglucoside Permease Activity in T. reesei Mutant
Strains-In view of the obvious importance of the p-linked diglucoside permease in the uptake of the inducer of cellulase formation, we have also investigated its activity in two mutant strains, Le. l". reesei RUT C-30 and QM 9979. The former is a hypersecretory strain, which secretes about 3-fold higher amounts of cellulase protein, whereas the latter mutant strain is unable to form cellulases on cellulose or by addition of sophorose and forms a complete cellulase system during growth on lactose. The permease activity of these two mutant strains deviated significantly from that of T. reesei QM 9414, whereas that of the hypersecretory strain RUT C-30 was about twice as high and that of QM 9979 was below 10% of it (Table IV). The synthesis of @glucoside permease in mutant QM 9979 could not be induced by sophorose (data not shown).

DISCUSSION
In the present study, we provide for the first time direct evidence of the presence of a @-&glucoside permease in T. reesei. It has a very high affinity for cellobiose (K,,,, 0.3 PM) and therefore can successfully compete with the constitutive &glucosidase for the common substrate cellobiose.
Because of the comparatively low V,,, (2.5 uersus 80 milliunits/mg fungal dry weight for permease and @-glucosidase, respectively), however, this situation is reversed at higher concentrations. Such properties of the putative permease had  previously been anticipated from inhibitor experiments (Fritscher et ul., 1990). Hence uptake of cellobiose is preferred to hydrolysis when its concentration is very low. It is tempting to speculate that this situation offers an advantage to the fungus. This appears to be the case during the early phase of contact of ~~i c~~e r~~ with cellulase. The activities of the constitutive cellulases are very low (0.025 units/108 conidia; Kubicek et ut., 1988) and hence only low concentrations of cellobiose will become available to the fungus at this stage.
To obtain some insights into the putative essential nature of the @-diglucoside permease in cellulase synthesis, we have also compared its activity in two mutant strains of T. reesei. The findings of elevated permease activities in strain T. reesei RUT C-30 are not unexpected. As this strain has been shown to be hypersecretory (Ghosh et al., 1984) and secretes all components of the cellulase enzyme mixture in elevated amounts, it is of little surprise that @-glucoside permease, which also has to be transported via a secretory pathway, occurs in higher amounts. The strongly decreased activities in mutant T. reesei QM 9979 are more interesting; however, this strain cannot be induced to form cellulases by cellulose or sophorose, but it forms a complete cellulase system during growth on lactose. Theoretically, these properties would exactly resemble the phenotype one would expect for a @glucoside permease negative mutant. However, although reduction of the permease was significant (over 90%) and its induction by sophorose did not occur, the low residual activity was clearly demonstrable. The inability to grow on cellulose therefore cannot be explained by a loss of the permease for the inducer. It may be speculated that the remaining activities are already low enough to enable the accumulation of concen-P-Diglucoside Permease from Trichoderma reesei trations of inducer sufficient for triggering P-glucoside permease and cellulase induction, but this clearly needs other experiments to be proven. The fact that the @-glucoside permease transcript is apparently one of the earliest mRNAs that accumulates upon induction by sophorose offers the possibility to clone its gene by differential hybridization techniques.
The finding that the activity of the @-glucoside permease is inhibited but not repressed by glucose is of particular importance for the regulation of cellulase synthesis. The information of whether cellulase formation is repressed by glucose or not is controversial. Recent findings on the presence of functional creA binding sites in the cbhl promoter' of T. reesei now provide conclusive for glucose repression of cellulase gene transcription. Our data, however, indicate that glucose also interferes with the uptake of the inducer of cellulase biosynthesis, which can be considered as a mechanism that aids in amplifying the negative glucose effect. Inhibition and/or repression of the respective permease have also been observed for other disaccharides whose hydrolases are glucose repressible (Dickson and Barr, 1983;Johnston, 1987).