Introduction

Aspergillus species are known for their capacity to secrete high levels of enzymes in their growth environment. Several of these secreted enzymes, produced in a large-scale submerged fermentation, have been widely used in the food and beverage industry for decades. The enzymes secreted during solid state fermentation by the so-called koji-molds like Aspergillus oryzae and A. sojae have been used for an even longer period for modification of rice, wheat kernels (WK) and soybean substrates for the production of fermented foods and alcoholic drinks. Only recently, solid state fermentation with filamentous fungi has gained increasing attention for production of specific enzymes (Pandey et al. 1999; te Biesebeke et al. 2002; Holker et al. 2004).

Although there has been some controversy, most studies suggest that protein secretion in filamentous fungi occurs around the apical and subapical region of the advancing hyphal tip (Gordon et al. 2000; Wösten et al. 1991; Muller et al. 2002; Archer and Pederby 1997). In A. oryzae, a correlation between the number of growing hyphal tips and protein secretion levels has been suggested (Amanullah et al. 2002; Muller et al. 2002). However, conflicting evidence about this correlation in other filamentous fungal cultivations has been presented (Bocking et al. 1999; McIntyre et al. 2001; Wessels 1993). To date, only morphological mutants with unknown genotypes and different, often pleiotropic, mutations altering hyphal morphology have been investigated (for references, see Conesa et al. 2001), whereas only specific gene disruptions that affect the formation of apical branches would ultimately establish whether protein secretion capacity is correlated to the number of growing tips.

In different studies, it has been shown that filamentous fungi grown on a solid substrate can secrete high titres of enzymes and, in some cases, even higher enzyme titres, compared to growth in a liquid medium (Holker et al. 2004; Machida 2002; Pandey et al. 1999 and references therein). Growth of A. oryzae on a solid substrate involves modification of the substrate by secreted enzymes and subsequently allows the growing hyphae to penetrate the substrate. Therefore, to study the correlation between the number of hyphal tips and protein secretion, solid state fermentation with A. oryzae appears to be an excellent cultivation method. In this paper, we describe two mutants, constructed by site-directed gene disruption, that are altered in the number of hyphal tips. Both mutant A. oryzae strains were analysed for branch frequency and amylase, glucoamylase and protease expression during growth on WK.

Materials and methods

Strains and media

A. oryzae 16868 was used for total RNA preparation and isolation of samples for amylase and protease activities measurements. Wheat kernels were pretreated and inoculated as described by Hoogschagen et al. (2001). Ground WK were prepared and pretreated as described by te Biesebeke et al. (2004). The wheat-based solid medium (2% WSM) supplemented with 10 mM uridine was prepared as previously described (te Biesebeke et al. 2004) and used to determine the amount of biomass formed during growth at 30°C at 98% humidity in a climate-controlled incubator (VEA Instruments, Houten, the Netherlands). The 4% wheat-based liquid medium (4% WLM) was prepared by suspending ground WK (4 g) in H2O (100 ml) that was sterilised for 15 min at 120°C. Surface growth was performed for determination of the radial extension rate on 1% agar plates of 4% WLM supplemented with 10 mM uridine at 30°C at 98% humidity in the incubator (VEA Instruments).

Extraction of A. oryzae cultures grown on solid substrates and measurements of enzyme activities

Extracts of A. oryzae grown on WK and ground WK were prepared as described by te Biesebeke et al. (2005, personal communication). The α-amylase and glucoamylase activities were measured in the extracts as described by te Biesebeke et al. (2005, personal communication). Protease activities were measured as described (te Biesebeke et al. 2005, personal communication), except that the 2 μl sample+13 μl water was mixed with 75 μl reagent [5 g l−1N,N-dimethylcaseine in 0.1 M NaAc/Hac (pH 5.5), 0.1 M K2HPO4 (pH 7.0) or 0.1 M Na2B4O7·10H2O (pH 8.5)] to measure the protease activities at pH 5.5, 7.0 and 8.5, respectively.

Construction of the pclA and pg/pi-tp disruption vectors

The (KEX2-like) proprotein convertase-encoding gene (pclA) of A. sojae, a strain that is closely related to A. oryzae, was isolated by functional complementation of an Aspergillus niger pclA disruption strain, which was described by Punt et al. (2003), with an A. sojae cosmid library (Heerikhuisen et al. 2004). A genomic clone was identified to complement the hyperbranching phenotype of the A. niger pclA strain, resulting in wild-type growth characteristics. PCR and sequence analysis confirmed the presence of the homologous pclA gene on this genomic clone. A 5.2-kb XbaI/BamHI fragment was sub-cloned, resulting in the vector pAS2-4. Subsequently, the pclA disruption vector pAS2-4pcl was constructed by cloning the repeat-flanked A. niger pyrG gene from pAB4-1rep as a SmaI fragment into the unique EcoRV site of the pclA gene of pAS2-4 (Heerikhuisen et al. 2004).

The pg/pi-tp gene (and a tandemly arranged thiolase gene) was amplified from genomic DNA from A. oryzae ATCC 16968 using primers (5′-CGGTTGTCGTCATAGTGAGC-3′) and (5-′CGATCGGAATCTAGAGAGACGG-3′) (Record et al. 2001). The total DNA sequence of 4,279 bp was cloned in pGEM-T easy vector (Promega) and verified by sequencing. The pGEM-T vector containing the pg/pi-tp gene was digested with AvrII and Eco47III to excise a 723-bp fragment from the pg/pi-tp gene. This fragment was replaced with the 2.8-kb SpeI/SmaI fragment from vector pAB4-1rep carrying the A. niger pyrG selection marker (Heerikhuisen et al. 2004), resulting in vector pG-PLTPdel.

Southern and Northern blot analysis

Southern blot analysis was performed according to the standard procedure (Sambrook et al. 1989) after cutting the chromosomal DNA with EcoRI and BamHI. Southern blots were hybridised with 32P labelled (Random Prime Labeling Kit, Pharmacia) PCR-amplified probes for the pclA and pg/pi-tp genes and hybridised at 68°C in standard hybridisation buffer (Sambrook et al. 1989). Hybridisation signals were visualised after exposure to X-OMAT AR films (Kodak, cat. 1651512) and scanned with the Hewlett-Packard 6200C Scanjet at 600 dpi. A probe (401 nucleotides) for the pclA gene was PCR-amplified in 35 cycles of 1 min at 94°C, 1 min at 55°C, 3 min at 68°C using primers (5′-CCTGATGATAGAGACCGATTTCCCG3-′) and (5′-GCCTTTGCATTATCTGACCGCCGCGC-3′) and vector pAS2-4 as template. A probe (361 nucleotides) for the pg/pi-tp gene was PCR amplified in 40 cycles of 1 min at 94°C, 1 min at 30°C, 2 min at 72°C using primers (5′-ACCCTCTGGAGTATTGTAATG3-′) and (5′-TTCAGGCAAGTGATATGCTC-3′) and chromosomal DNA of A. oryzae as template. In Southern analysis, the probes for the pclA and pg/pi-tp genes both showed single specific hybridising bands in the wild-type A. oryzae strain. Disruption of the pclA and pg/pi-tp genes was confirmed in the respective pclA and pg/pi-tp disrupted strains. Northern blot analysis was performed with isolated total RNA and PCR-amplified probes for the α-amylase, glucoamlyase B, alkaline protease A and neutral protease B genes as described (te Biesebeke et al. 2004, personal communication).

Microscopy and hyphal measurements

To obtain the number of spores per milliliter extract, spores were counted microscopically with an extract (25 ml H2O) of A. oryzae strain grown on WK (5 g). To obtain a constant and reliable value for our measurement of hyphal branching, expressed as the hyphal growth unit length (L hgu) (Prosner 1994), germination and subsequent branching was observed of colonies originating from single spores of A. oryzae grown on 2% WSM. From the moment that hyphal tips appeared, every 10 min the length of the hyphal elements was determined, and the number of tips (N t) on each hyphal element were counted. For determination of L hgu of A. oryzae wild type, pclA and pg/pi-tp disrupted strains were grown on 2%WSM, and pictures were taken of hyphal elements having five to 12 hyphal tips using a Zeiss microscope (AXIOLAB) with a ×10 objective. Pictures were taken with the Sony 3CCD camera and imported and saved as TIFF files using the Paintshop Pro software. The pictures were analysed on a Macintosh (iMAC, Macintosh Incorporated, USA) computer using the IMAGEJ software (http://www.imageJ.com), according to the manufacturer’s protocol. L hgu was determined by dividing the total hyphal length of the growing hyphal element (L tot) by the number of tips (N t) present on the growing hyphal element. L hgu was expressed in hyphal length per hyphal tip (mm tip−1) and is an indication for the frequency at which hyphal branching occurs on a hyphal element (branch frequency). For all experiments, 40–50 hyphal elements were analysed.

Results

Construction of the pclA and pg/pi-tp disrupted strains

A previous study has shown that disruption of the (KEX2-like) proprotein convertase-encoding gene (pclA) in A. niger resulted in a hyperbranching phenotype (Punt et al. 2003). Mycelial morphology is also influenced by the concentration of the cell membrane phospholipid, phosphatidylcholine, apparently by controlling branch initiation (Markham et al. 1993). The phospholipid transfer protein encoded by the pg/pi-tp gene of A. oryzae transferred besides phosphatidylglycerol and phosphatidylinositol, and also phosphatidylcholine (Record et al. 1995). Moreover, a correlation was shown between growth and phospholipid transfer activity of A. oryzae under specified conditions (Record et al. 2001). Therefore, an A. oryzae mutant strain with a disruption in the pg/pi-tp gene may also have a different growth phenotype compared to the wild type.

A transformation system based on the orotidine-5′-phosphate decarboxylase gene (pyrG) as selection marker (Mattern et al. 1987) was developed for the A. oryzae ATCC16868 strain. An ATCC16868pyrG mutant strain was used for the construction of the pclA and pg/pi-tp disruption strains. For the disruption of the pclA gene in ATCC16868pyrG, the pclA disruption vector pAS2-4Δpcl (Heerikhuisen et al. 2004) was used. From this vector, an 8.2-kb NotI/AscI fragment, containing the pclA gene interrupted by the pyrG gene, was used to transform ATCC16868pyrG. Of the Pyr+ transformants, 12 out of 15 showed a compact growth morphology as was observed for the pclA disrupted strain in A. niger (Punt et al. 2003). Southern blot analysis revealed that two of the compact Pyr+ transformants showed a shift in band size with the probe for the pclA gene compared to the wild type as a consequence of the insertion of the 8.2-kb NotI/AscI fragment at the pclA locus (results not shown). To disrupt the pg/pi-tp gene, a 6.3-kb EcoRI fragment from pG-PLTPdel was used to transform ATCC16868pyrG. In two of the three Pyr+ transformants analysed, the pg/pi-tp gene was shown to be disrupted using Southern analysis with a probe for the pg/pi-tp gene (results not shown). Two strains for which Southern analysis confirmed the disruption of the pclA or the pg/pi-tp gene were selected for further analysis.

Growth phenotypes of the pclA and pg/pi-tp disrupted strains

To determine the effects of the disruption of the pclA or pg/pi-tp gene in ATCC16868pyrG, the growth phenotype was compared to ATCC16868pyrG during growth on WK. The amount of biomass that was formed during surface growth on 2% WSM was equal for all strains and equal as previously described for the A. oryzae ATCC16868 strain (te Biesebeke et al. 2004). Although the obtained biomass was the same, the pclA disrupted strain clearly showed a more compact morphological phenotype compared to the ATCC16868pyrG and the pg/pi-tp disrupted strain. The radial extension rate is a reliable method to determine the growth of a fungal colony (Trinci 1974). After spotting 10 μl containing 108 conidia on 4% WSM, only minor differences in radial extension rate were observed between the ATCC16868pyrG and the pg/pi-tp disrupted strain, whereas that of the pclA disrupted strain was about 70% of ATCC16868pyrG (Table 1). After 3 days of growth on 4% WSM, ATCC16868pyrG started to sporulate in the centre of the colony, whereas sporulation of the pclA and pg/pi-tp disrupted strains appeared after 4 days. After 3 days of growth on ground WK, the aerial hyphae of the mycelial layer of these two strains were longer (>3 mm), compared to the pclA disrupted strain (1–2 mm) (Table 1). Colonisation on WK, in terms of the time that it takes to fully cover the substrate with filamentous fungal mycelium, was equal for all strains (4 days). However, sporulation of the pclA or pg/pi-tp disrupted strains occurred 2 days and 1 day later, respectively, compared to ATCC16868pyrG (Table 1). There were no differences observed between ATCC16868 wild type and the ATCC16868pyrG strain under these cultivation conditions (Table 1).

Table 1 Growth phenotype of ATCC16868, ATCC16868pyrG, pclA and pg/pi-tp disruption strains
Full size table

Microscopical phenotype of the pclA and pg/pi-tp disrupted strains

To determine the effect of the disrupted genes on hyphal branching, the pclA or pg/pi-tp disrupted strains and ATCC16868pyrG were grown for up to 24 h at 37°C in a 1- to 2-mm layer of 2% WSM at a density of 20 conidiospores ml−1. The conidiospores were observed microscopically with an amplification of ×100. During germination of the conidiospores, the pclA or pg/pi-tp disrupted strains clearly showed a higher number of hyphal tips compared to ATCC16868pyrG. To define this process in a quantitative way, Lhgu was determined as a measure to describe branching (Dynesen and Nielsen 2003; Christiansen et al. 1999; Prosner 1994; Trinci 1974). For the successful use of this parameter, a similar approach was used as described by Dynesen and Nielsen (2003). First, it was determined for the ATCC16868pyrG strain that at least six branches should be formed on 2% WSM before Lhgu approached a constant value (Fig. 1a). To determine branch frequency of the pclA or pg/pi-tp disrupted strains, only germinating conidiospores with hyphal elements possessing five to 12 hyphal tips were analysed. Figure 1b shows that Lhgu is clearly distinct in the different strains, whereas for all strains, Lhgu is almost constant after the formation of at least seven hyphal tips. Table 1 shows that the pclA or pg/pi-tp disrupted strains show an average value for Lhgu (calculated with the seven to 12 tips Lhgu values shown in Fig. 1b) that is respectively 50 and 74%, compared to that of the wild type. Since more hyphal tips occurred in both pclA and pg/pi-tp disrupted strains during growth on 2% WSM compared to ATCC16868pyrG, more hyphal tips were also assumed for the pclA and pg/pi-tp disrupted strains grown on WK and ground WK compared to ATCC16868pyr.

Fig. 1
figure 1

a The hyphal growth unit length (Lhgu, in mm tip−1) of the ATCC16868pyrG strain grown on 2% WSM. The hyphal length (mm) (filled triangle) and the number of tips were determined from hyphal elements derived from germinating conidiospores. Lhgu (open diamond) was calculated from the hyphal length and the number of tips. bLhgu (in mm tip−1) of the ATCC16868pyrG strain (open diamond) and the pclA (open triangle) and pg/pi-tp (filled circle) disrupted strains grown on 2% WSM. Of every strain, 40–50 hyphal elements with five to 12 hyphal tips were analysed

Full size image

Expression of amylase and protease by the pclA and pg/pi-tp disrupted strains

To study the correlation between the number of hyphal tips and protein secretion, the activity of amylases, glucoamylases and proteases was studied in the extracts of A. oryzae ATCC16868pyrG and the pclA and pg/pi-tp disrupted strains grown on WK. A previous study reported that after 5 and 6 days of growth of A. oryzae on WK and ground WK, considerable enzyme activities were measured (te Biesebeke et al. 2005, personal communication). After 6 days of growth, the α-amylase activities measured in the extracts of the pclA and pg/pi-tp disrupted strains grown on WK and ground WK were all more than 50% higher, compared to those of ATCC16868pyrG and the wild-type ATCC16868 strain (Table 2). The glucoamylase activities measured in the extracts of the pclA and pg/pi-tp disrupted strains grown for 6 days on WK and ground WK were more than 100% higher compared to those of ATCC16868pyrG (Table 2). The protease activities measured at pH 5.5, 7 and 8.5 in the extracts of the pclA and pg/pi-tp disrupted strains grown on WK were all more than 90% higher compared to those of ATCC16868pyrG (Table 2). In contrast, all protease activities measured at pH 5.5, 7 and 8.5 in the extracts of the pclA and pg/pi-tp disrupted strains grown on ground WK were all about equal (see Table 2), compared to those of ATCC16868pyrG.

Table 2 α-Amylase, glucoamlyase activities and protease activities at pH 5.5, 7 and 8.5 in extracts of the wild-type ATCC16868, ATCC16868 pyrG and the pclA and pg/pi-tp disrupted strains grown for 5 and 6 days (d) on wheat kernels (WK) or ground wheat kernels (GWK)
Full size table

Transcriptional analysis of the pclA and pg/pi-tp disrupted strains

Northern analysis was performed to determine the transcription levels of a selection of genes encoding hydrolytic enzymes of ATCC16868pyrG and the pclA and pg/pi-tp disrupted strains. The strains were grown for 6 days on WK, and mRNA samples were isolated and analysed for transcription of the α-amylase, the glucoamylase, the alkaline and the neutral protease genes (Fig. 2). The pclA and pg/pi-tp disrupted strains showed equal transcription levels for the α-amylase, and the alkaline and neutral protease genes compared to ATCC16868pyrG. However, the transcription level of the glucoamylase B gene in the pclA and pg/pi-tp disrupted strains was clearly higher compared to ATCC16868pyrG (Fig 2) at 6-day growth on WK.

Fig. 2
figure 2

Northern analysis of the ATCC16868pyrG strain (left lane of each panel) and the pclA (middle lane of each panel) and pg/pi-tp (right lane of each panel) disrupted strains grown for 6 days on WK with probes for the α-amylase (amyA), glucoamylase B (glaB), alkaline protease (alpA) and neutral protease II (nptB) genes. As a loading control, the ribosomal RNA (rRNA) was used

Full size image

Discussion

Filamentous fungal growth of A. oryzae on WK solid substrate was used to study protein secretion. In solid state cultivation, penetrative hyphae secrete hydrolysing enzymes to degrade the solid substrate and subsequently penetrate the substrate. To study whether secretion of proteins by A. oryzae depends on the number of hyphal tips, two site-directed, gene-disrupted mutants were constructed and analysed during growth on WK.

The pclA disrupted strain of A. oryzae showed a twofold lower L hgu, corresponding to a twofold higher branch frequency during surface growth on 2% WSM compared to ATCC16868pyrG. The A. oryzae pclA gene encodes a homologue of the A. niger endoproteolytic proprotein processing enzyme kexB that processes dibasic cleavage sites (KR) in target proteins that are transported through the secretion pathway (Jalving et al. 2000; Punt et al. 2003). The hyperbranching phenotype of the pclA disrupted strain is apparently the result of at least one unprocessed proprotein. Interestingly, chitin synthase B (GenBank accession no AAK31732), encoded by the chsB gene of A. oryzae, contains putative mono and dibasic processing sites for PclA. As an abnormal branching phenotype has been observed for a strain with a disrupted chsB gene in A. oryzae (Muller et al. 2002), an unprocessed dysfunctional chitin synthase B might contribute to the observed phenotype of the pclA disrupted strain. This is in agreement with a recent study suggesting disordered cell integrity signalling in response to perturbation of the cell wall (Mizutani et al. 2004). Alternatively, PclA might be required for the processing of sensor proteins in the cell integrity pathway (Mizutani et al. 2004).

The pg/pi-tp disrupted strain showed 1.5 times lower L hgu on 2% WSM compared to ATCC16868pyrG. The amino acid sequence of the phosphatidylglycerol/phosphatidylinositol transfer protein (GenBank accession no AAG13652) suggests that it is an outer cell membrane targeted protein (Record et al. 1999). The purified phospholipid transfer protein (AAG13652), a putative homologue of human NPC2 whose deficiency causes Niemann–Pick type C2 disease (Ikonen and Holtta-Vuori 2004), was found to transfer besides phosphatidylglycerol and phosphatidylinositol and also phosphatidylcholine (Record et al. 1995). Therefore, the phospholipid transfer protein may be involved in the integrity of the lipid content of the cell membrane. Choline appears as phosphatidylcholine (lecithin) in the outer cell membrane, and reduced concentrations of choline result in multiple tip-formation of Fusarium graminearum (Markham et al. 1993). The phenotype of the pg/pi-tp strain might therefore be a consequence of a reduced phosphatidylcholine transfer activity. The resulting lipid imbalance in A. oryzae cell membranes is suggested to result in a higher branch frequency.

As shown, both branching mutants showed increased levels of amylase and protease production. The option that increased biomass yields cause increased protein production seems unlikely as biomass yields for the mutant strains were not higher than the wild-type strain during growth on 2% WSM (results not shown). It is only in glucoamlyase that increased production may be due to increased gene expression (Fig. 2). The glaB gene is induced by maltose (Ishida et al. 1998, 2000) that is liberated from the WK (te Biesebeke et al. 2004). Our results could indicate that the maltose induction is increased in the pclA and pg/pi-tp disrupted strains. Compared to the wild type, the growth phenotype of both mutant strains on WK could imply more penetrative hyphae involved in uptake of inducer for glaB gene transcription. Moreover, the retarded sporulation in the pclA and pg/pi-tp disrupted strains may indicate improved nutrient uptake compared to the wild type, as nutrient state is one of the regulatory cues for sporulation (Adams et al. 1990; Skromme et al. 1995). For amylase and protease production, we suggest that a higher number of WK penetrating hyphae per hyphal growth unit may explain the higher secretion of enzyme activities of the mutant strains. Our results show an important contribution of hyphal tips to the production of proteins during growth of A. oryzae on a solid substrate.