INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Filamentous fungi are renowned for the efficient secretion of various enzymes, such as cellulases or amylases, in large amounts and are thus also considered as potentially attractive host systems for the production of biotechnologically relevant heterologous proteins. However, so far the overall levels of heterologous (nonfungal) proteins are still considerably lower than those obtained for homologous proteins (19, 22, 46).

There may be a large number of factors that influence the final level of a secreted protein, i.e., the regulation of transcription, mRNA stability, translational initiation and elongation, translocation, protein folding, intracellular transport, and processing (16, 29). Even after secretion of the proteins into the extracellular fluid, they may be degraded by extracellular proteases (6).

Several strategies have been developed to identify and eliminate these potential bottlenecks and thus to improve protein yields (1, 29). These include the use of highly inducible promoters, the introduction of a high gene copy number, the use of protease-deficient host strains, and the fusion of heterologous genes with an endogenous gene encoding a protein secreted at high levels such as glucoamylase (6).

In many cases, the production of heterologous proteins appears to be limited at the level of secretion (28, 36). In all organisms, secretory proteins are synthesized as preprotein precursors which are N-terminally extended by a signal peptide that targets them into the secretory pathway (40, 43). However, protein maturation in eukaryotes often requires additional proteolytic processing at later stages of the secretory pathway. In Saccharomyces cerevisiae, the maturation of -factor has been shown to involve three different specific peptidases: Kex1, Kex2, and a dipeptidylaminopeptidase (10, 21, 23). Calmels et al. (7) have reported that dibasic amino acid doublets, which resemble Kex2 target sites, predominate as target sequences in secretory proteins from filamentous fungi. However, apart from the demonstration of the cleavage of dibasic targets in homologous and heterologous proteins in vivo (7, 9, 32), there have been no studies on the detection and properties of a Kex2 homolog in filamentous fungi.

Here we demonstrate a dibasic endopeptidase activity in Trichoderma reesei and, with the aid of an irreversible inhibitor of it (p-amidinophenylmethylsulfonyl fluoride [pAPMSF]), we show that it constitutes an essential step in the secretion of proteins containing the respective target sequence.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Strains and transformation. The T. reesei strains used in this study are listed in Table 1. They were maintained on malt agar (containing 5 mM uridine in the case of T. reesei TU-6) and subcultured monthly. Transformants were obtained by cotransformation with plasmid pFG1, which carries the homologous pyr4 (previously termed pyrG) gene as a selectable marker (17, 18).

Medium and growth conditions. T. reesei was grown in conical flasks (1 liter) containing 250 ml of the medium described by Mandels and Andreotti (31), supplemented with the appropriate carbon source (1% [wt/vol]) and buffered with 50 mM sodium citrate at pH 5.0, at 30°C. Carbon sources were used as indicated for the respective experiments. To prepare cell extracts for enzyme assays, T. reesei QM 9414 was grown on 1% (wt/vol) xylose or 1% (wt/vol) glycerol. To induce xylanase or cellulase formation in resting mycelia, T. reesei QM 9414 was pregrown for 20 h with glycerol as the carbon source and then transferred to a resting-cell medium. To do this, the mycelium was washed with tap water and resuspended in minimal medium lacking a carbon source, and 3-ml aliquots were transferred to 10-ml bottles. Inducers (2 mM sophorose [30] and 4 mM xylose [27]) and inhibitors (PMSF and pAPMSF [26]; Sigma, Deisenhofen, Germany) were added to give the final concentration as indicated. Induction was started 1 h after the addition of the inhibitor, and incubation was continued for up to 20 h on a rotary shaker (250 rpm, 30°C).

Plasmids and manipulation of DNA. Plasmid pFG1 (18) was obtained from the stock of the Institute of Biochemical Technology and Microbiology, Technische Universität Wien, Vienna, Austria. Construction of the recombinant plasmid pUT740, which is a pUC19 (48) derivative containing a gene fusion consisting of the Aspergillus nidulans gpdA promoter, a synthetic oligonucleotide for a preprosequence (see Fig. 4), the phleomycin resistance gene (ble) from Streptoalloteichus hindustanus (11), and the A. nidulans trpC terminator, has already been described (8). Plasmids pUT964 through pUT967 were derivatives of pUT953, which is similar to pUT740 except that it contains 1,320 bp of the constitutive T. reesei Tr1 promoter (35). These plasmids were constructed by insertion of custom-designed oligonucleotides into the XhoI-NcoI site between the T. reesei promoter and the S. hindustanus ble structural gene.

Cell extracts for proteolytic enzyme assays and immunological analysis. Mycelia were harvested on a Buchner funnel and washed with ice-cold tap water; excess liquid was removed by blotting the material between filter paper sheets. Then 1 g of the blotted mycelium was suspended in 4 ml of 20 mM Tris-HCl buffer (pH 7.0) containing 0.1 mM PMSF, 1 mM EGTA, 1 mM 2-mercaptoethanol, and 1 µM pepstatin A or in 20 mM sodium acetate buffer (pH 5.5) containing the same additives. The suspension was further sonicated in an ice-water bath with a Sonoplus HD60 sonifier (Bandelin; four times for 90 s each with alternating 90-s cooling periods). The homogenate was centrifuged at 10,000 × g (15 min, 4°C), and the supernatant was kept for enzyme activity determinations.

Endopeptidase enzyme assays. Endopeptidase activity was measured as follows. First, 20 nmol of the substrate (N-tert-butoxycarbonyl [Boc]-GKR-4-methylcoumarin [MCA], Boc-LKR-MCA, Boc-VPR-MCA, and Boc-LRR-MCA, where GKR, LKR, VPR, and LRR indicate the peptide substrates in the amino acid one-letter code; Sigma, Deisenhofen, Germany) was incubated with 50 µl of the cell extract in 0.2 M Tris-HCl buffer (pH 7.0) or 0.2 M sodium acetate buffer (pH 5.5) and 1 mM CaCl₂ in a total volume of 250 µl at 37°C for 30 min. Thereafter, 3.0 ml of distilled water was added, and the amount of MCA released from the substrate was measured in a fluorescence spectrophotometer with excitation at 380 nm and emission at 460 nm. One unit of enzyme activity is defined as the amount of enzyme that can release 1 nmol of MCA from the substrate under the above assay conditions in 1 min.

Subcellular fractionation. T. reesei QM 9414 was grown for 24 h on 1% (wt/vol) glycerol as the carbon source. The washed mycelial mat was thoroughly mixed with an equal volume of glass beads (diameter, 0.5 mm) and 20 mM Tris-HCl buffer (pH 7.0) containing 1 mM 2-mercaptoethanol, 1 mM EGTA, 0.1 mM PMSF, 1 µM pepstatin A, and 1 M sorbitol and was then homogenized four times for 20 s each, with alternating 2-min cooling periods, in the 75-ml chamber of a Bead-Beater apparatus (Biospec Products, Bartlesville, Okla.). The homogenate was centrifuged at 5,000 × g for 12 min to remove the beads, nuclei, and cell wall material. Then, 8 ml of the supernatant was layered onto a 30-ml discontinuous sucrose gradient (17.1 to 34.2% [wt/vol]). The gradient was centrifuged for 2 h at 18,000 × g at 4°C and fractionated from the top into 30 fractions. One portion of each fraction was fixed for electron microscopy as described by Glaurt (15). Published methods were used to assay citrate synthase (41), acid phosphatase (24), and glucose-6-phosphate dehydrogenase (20).

Immunological analysis. Samples from the culture supernatants and from the cell extracts were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) followed by Western blotting on nitrocellulose membranes. SDS-PAGE was carried out in 10 and 15% polyacrylamide gel slabs, as described previously (25). The protein concentration was determined as described by Bradford (3) with bovine serum albumin as a standard.

Analytical chromatofocusing. Portions (0.5 ml) of the cell extract were dialyzed against 25 mM imidazole-HCl buffer, pH 7.0, and applied to a Mono-P column (Pharmacia Biotechnology, Uppsala, Sweden) connected to a fast protein liquid chromatography apparatus (Pharmacia) operating at 0.5 ml/min. After elution of all nonbound protein, an automatic pH gradient was started by rinsing the column with 40 ml of 10% (vol/vol) Polybuffer 74 (Pharmacia), pH 3.5. Fractions of 1 ml were collected and assayed for dibasic endopeptidase activity with Boc-GKR-MCA as a substrate.

Purification of secreted ShBLE and N-terminal sequencing. ShBLE was purified from the culture broth as described earlier (8). N-terminal sequencing was performed on an Applied Biosystems model 477A gas-phase sequencer equipped with an automatic on-line phenylthiohydantoin derivative analyzer.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Evidence for a dibasic endopeptidase activity in T. reesei. To demonstrate the presence of an endoproteinase with specificity for basic amino acid doublets in T. reesei, cell extracts were tested for activity against four fluorogenic substrates containing appropriate target sites (KR, RR, and PR [13, 39, 42]) (Table 2). The relative activities against the various substrates coincide well with the activities of yeast Kex2 (5, 33). In support of this similarity, the dibasic endopeptidase activity was optimal in the presence of 1 mM CaCl₂ (Table 2) and was inhibited by EDTA at pH 7.0 (data not shown). In contrast, however, it had a higher activity at pH 5.5 than at pH 7.0 and, unlike in yeast cells (5), this was not due to limited stability. Subcellular fractionation localized the highest activity at a bouyant density of 1.1 g/cm³, which was juxtaposed by the activity of acid phosphatase, an enzyme secreted by T. reesei (Fig. 1). The relative appearance of marker enzymes for cytosol (glucose-6-phosphate dehydrogenase) and mitochondria (citrate synthase) on the gradient suggested that the fraction with the highest dibasic target cleaving activity consisted of smaller vesicles and/or membrane fragments, and this was supported by electron microscopy (data not shown). The presence of acid phosphatase further indicates involvement of these vesicles in protein secretion. Hence, we conclude that this activity is localized on membranes and/or smaller subcellular particles in T. reesei.

View larger version (16K): [in this window] [in a new window]   FIG. 1.   Subcellular localization of the Kex2-like activity in T. reesei. T. reesei mycelia were homogenized as described in Materials and Methods, cell debris and nuclei were removed by centrifugation, and the supernatant was loaded onto a 30-ml discontinuous sucrose gradient (17.1 to 34.2% [wt/vol]). The gradient was centrifuged for 2 h at 18,000 × g and fractionated from the top into 30 fractions. Symbols: , the actual bouyant density of individual fractions; , the dibasic endopeptidase-like activity of individual fractions; , glucose-6-phosphate dehydrogenase activity; , citrate synthase activity; , acid phosphatase. All enzyme activities are given in arbitrary units (a.u.) only, so they could be plotted on the same graph.

View larger version (11K): [in this window] [in a new window]   FIG. 2.   Analytical chromatofocusing of the dibasic endopeptidase activity of T. reesei. Chromatofocusing was performed as described in Materials and Methods. Symbols: , pH of every second fraction; , dibasic endopeptidase activity (with Boc-LKR-MCA as a substrate).

Irreversible inhibition of enzyme activity with pAPMSF in vitro and in vivo. To confirm that the processing enzyme is a serine protease, we used pAPMSF, which specifically and irreversibly inactivates proteases with substrate specificity for positively charged amino acid side chains (26). A time course of inhibition showed that half-maximal inactivation appeared to have occurred after 30 to 40 min and that complete inactivation occurred within 2 h of incubation (data not shown); all data are thus reported for incubation times with the inhibitor of 1 h. Similar results were obtained when the activity was assayed with Boc-LKR-MCA, Boc-LRR-MCA, or Boc-GKR-MCA (Table 3). In contrast, 1 mM PMSF had no effect on T. reesei dibasic endopeptidase activity (Table 3).

Effect of pAPMSF and PMSF on maturation and secretion of cellulases and xylanases in vivo. The results described above showed that the dibasic endopeptidase can be inactivated in vivo by addition of pAPMSF, thus providing us with a tool to study whether this proprotein-processing activity is required for secretion of T. reesei proteins in vivo. To accomplish this, we chose the cellulases cellobiohydrolase I (CBH I) and CBH II and the xylanases XYN I and XYN II as model secretory proteins. A comparison of their nucleotide sequence-deduced putative amino acid sequences with that derived from N-terminal sequencing of the purified protein revealed different potential target sites for posttranslational proteolytic modification, e.g., -R-A--Q-, CBH I; -E-R--Q-, CBH II; -R-R--R--A-, XYN I; and -K-R--Q-, XYN II (Fig. 3A). Therefore, processing of XYN I and XYN II, but not of CBH II or CBH I, should occur by the dibasic endopeptidase and thus be inhibited by pAPMSF.

View larger version (34K): [in this window] [in a new window]   FIG. 3.   Effect of pAPMSF on formation and secretion of some cellulases and xylanases by T. reesei. (A) Preprosequences of the enzymes investigated. The arrows indicate the N-terminal amino acid of the secreted protein according to published evidence (2, 39, 42, 44). (B) Demonstration of the intra- and extracellular levels of these enzymes by SDS-PAGE and immunostaining with specific monoclonal antibodies. Experiments were carried out as described in Materials and Methods. The marks in the left margin indicate the positions of the unprocessed (higher mark) and processed (lower mark) enzyme form. Blots for extracellular and intracellular samples from the same protein are placed so that the relative positions of these two enzyme forms are comparable. For intracellular samples, equal amounts (40 µg) of protein were loaded onto the gels; equal volumes (20 µl) of the culture broth were loaded for the extracellular samples. Only a single mark is shown where only the processed form can be observed (CBH I and CBH II). Results shown are representative of at least three separate experiments.

Effect of pAPMSF and PMSF on the secretion of a heterologous protein fused to various propeptide sequences. The findings presented above are in accordance with an essential role of the pAPMSF-inhibitable dibasic endopeptidase in the secretory pathway of T. reesei proteins containing dibasic processing targets. However, a direct comparison of the results obtained with these four different proteins is hampered by the possibility that they have different rates and kinetics of secretion. In order to overcome this difficulty, we have reassessed the effect of pAPMSF on secretion of a single heterologous protein fused to different propeptide sequences. We have chosen the phleomycin-binding protein from S. hindustanus (ShBLE [11]) for this purpose and fused it to various prepro- sequences as shown in Fig. 4. The respective constructs were introduced into T. reesei TU-6 by cotransformation with the plasmid pFG1 containing the pyr4 gene as a selection marker (see Materials and Methods). Mitotically stable transformants were selected, and the copy number of plasmids integrated into the genome was determined to allow a comparison of the obtained results. Whenever possible, transformants with two integrated copies were chosen for further experiments.

View larger version (29K): [in this window] [in a new window]   FIG. 4.   Effect of pAPMSF on formation and secretion by T. reesei of the phleomycin-binding protein fused to different proprotein sequences. The N-terminal amino acid of the secreted protein was determined by Edman degradation and is marked by an arrow. The UT numbers of the respective recombinant strains are given in the left margin. The construct is shown on top (gpdA/Tr1, promoter; SS, signal sequence; PS, prosequence; Sh, the phleomycin-binding protein-encoding sequence; t, TrpC terminator). Sequences differing between the different constructs are underlined, and the putative cleavage targets are double underlined. Detection of intracellular and extracellular levels of ShBLE (13.7 kDa) by SDS-PAGE and immunostaining with a polyclonal antibody are shown to the right of the amino acid sequences. Columns: C, control (no inhibitor added); P, PMSF; A, pAPMSF. The markers on the left of the blots indicate the positions of the mature proteins. For intracellular samples, equal amounts (40 µg) of protein were loaded onto the gels; equal volumes (20 µl) of the culture broth were loaded for extracellular samples. The blots shown are typical for the results obtained in three independent experiments.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

The presence of a dibasic, proprotein-processing endopeptidase in filamentous fungi has been assumed previously on the basis of comparison of proprotein sequences of secretory proteins (7) and because of detection of free interleukin-6 of the correct size in extracellular culture supernatants of recombinant strains producing a fusion of interleukin-6 to CBH I, separated by a Kex2 cleavage site (8, 31). The data presented here provide experimental support for this assumption. The pAPMSF-inhibitable substrate specificity, the Ca²⁺ dependency of the reaction, and its localization in a vesicular fraction are all very similar to the properties of yeast Kex2 protease (5, 14, 34, 37). However, the pH optimum was strikingly different. Although attempts to clone a kex2 homolog from T. reesei have been carried out in several laboratories, they were unsuccessful (15a, 34a, 37a). It is thus uncertain whether the enzyme catalyzing the endopeptidase studied in this work actually justifies the name Kex2, and we are at this point referring to it consequently as a dibasic endopeptidase only.

The finding that pAPMSF inhibited the T. reesei dibasic endopeptidase irreversibly when added in vivo provided us with an alternative strategy to manipulate the activity of this enzyme and thus to study its effect on the secretion of proteins by T. reesei. We consider this approach even more straightforward, since the disruption of KEX2 in S. cerevisiae or Yarrowia lipolytica results in severe growth defects (13, 21), thus making the investigation of effects on protein secretion rather difficult. Furthermore, pAPMSF may offer an attractive means for isolating mutants overexpressing the dibasic endopeptidase.

Results from both homologous and heterologous systems show that pAPMSF efficiently blocks the secretion of proteins containing dibasic target sites. The simultaneous accumulation of intracellular protein forms with higher M_r values suggests that this is not due to a nonspecific general effect but rather that the inhibition of proprotein processing by the dibasic endopeptidase blocks furtheror at least decreases the rate ofprotein secretion. This is in contrast to results with Y. lipolytica, where alkaline protease was not processed, yet it was secreted (13). At first sight, these data also contrast with findings in Aspergillus niger and A. awamori that artificial gene constructs of restrictocin and cutinase lacking a prosequence ending with a dibasic amino acid motif were efficiently secreted (4, 47). These data suggest that processing by the dibasic endopeptidase in itself is not a prerequisite for secretion in filamentous fungi, yet if the respective target sequence is present in the protein its cleavage appears to be essential.

Secretion of CBH II, which contains a monobasic target site, was inhibited only by much higher concentrations of pAPMSF, but this effect was also seen with PMSF and thus was not specific. Thus, we conclude that this site is processed not by the dibasic endopeptidase but probably by other serine proteases. In yeast, Yap3p, which cleaves after a single R, which would perfectly match the sequence present in CBH II, is responsible for the processing of monobasic proprotein target sites (12). No intracellular accumulation of unprocessed CBH II was detected in the presence of pAPMSF. However, the prosequence of CBH II is very small (42), and the difference between the proprotein and the mature protein may not be detectable by SDS-PAGE. Because of the results with XYN I and II (see above) and with CBH I (see below), we consider it rather unlikely that pAPMSF exerts a general inhibitory effect on CBH II formation.

The processing of CBH I, whose apparent proprotein cleavage site does neither correspond with a dibasic endopeptidase or a Yap3 target site, was not at all affected by the addition of pAPMSF, thus providing support for the hypothesis that pAPMSF at the concentration used has no general effects on secretion. Intriguingly, the secretion of CBH I was enhanced by pAPMSF addition, the secretion of ShBLE UT965 was only detected in the presence of 5 mM pAPMSF, and inverse observations were made for the intracellular precursor pool. One explanation for these findings would be that dibasic endopeptidase-dependent and dibasic endopeptidase-independent secretory proteins compete for further transport subsequent to the dibasic endopeptidase processing step. If this assumption is correct, it would indicate that the prosequence can influence the rate of protein secretion, which would be an interesting focus for further studies, offering as it does the possibility to alter the efficacy of protein secretion in T. reesei and perhaps other fungi by an alteration of its processing peptidases.