Introduction

Expression tools are essential for genetic and application studies. Fungi are convenient cell factories widely used for the production of recombinant proteins, using strong and regulated promoters. One of the most widely used fungal promoters is the Aspergillus oryzae Taka-amylase promoter, which is used to drive the expression of a large number of different enzymes (Brakhage et al. 1999). Other widely used fungal promoters include the promoter of the Aspergillus gpdA, alcR or glaA genes (Punt et al. 1995; Santerre-Henriksen et al. 1999; Mathieu et al. 2000). Aspergillus promoters have also been used with success in many other ascomycetes for genetic studies or to establish transformation procedures. One prominent example of this is the A. nidulans gpdA promoter and plasmid pAN7-1 (Wang et al. 1999).

In zygomycetes, a combination of scarce genetic knowledge and cumbersome sexual cycle has resulted in a limited development of tools for expression studies. However, the current use of this group of fungi in industrial processes (production of carotene, gamma-linolenic acid, enzymes) and their potential use in novel processes make the development of expression tools a challenge. Although promoters have been identified in this fungal class, they have been used mainly to develop transformation protocols (Mackenzie et al. 2000; Wöstemeyer 1987). To our knowledge, only the Rhizopus niveus pgk2 promoter has been studied in detail (Takaya et al. 1995).

Fungal promoters display a number of features, such as a CT-rich stretch adjacent to the transcription start site, a TATA box and CAAT boxes involved in recognition by the RNA polymerase complex (Brakhage et al. 1999). A wide variety of additional signals is present in regulated promoters. In the A. nidulans and A. niger gpdA promoters, a number of consensus regulatory boxes have been identified (Punt et al. 1995), but sequence homology in more distant homologues has not been reported.

Recently, we showed that expression of the Mucor circinelloides gpd1 gene for glyceraldehyde-3-phosphate dehydrogenase (gpd1P) was significantly stronger when growing the cells in glucose, compared with glycerol or ethanol. Furthermore, expression and secretion of active glucose oxidase 1 (GOX) was shown using a gpd1P-expression cassette on an episomal plasmid (Wolff and Arnau 2002).

Using the gox1 gene as a reporter for expression studies in M. circinelloides, it is possible to investigate aspects like promoter strength and regulation. However, for these studies, the expression cassette has to be integrated rather than maintained extrachromosomally, to prevent genetic variation due to genetic instability inherent to episomal plasmids. In M. circinelloides, the available transformation systems that have been developed yield mitotically unstable transformants (Van Heeswijck and Roncero 1984; Anaya and Roncero 1991; Benito et al. 1992). Integrative transformation is not straightforward but rather is an exception, although a few reports have demonstrated integration and targeted gene inactivation in this fungus (Arnau et al. 1991; Arnau and Strøman 1993; Navarro et al. 2001).

M. circinelloides produces a yellow mycelium when exposed to light, due to the accumulation of β-carotene (Navarro et al. 2001). Like other related species, M. circinelloides possesses a light-regulated carotene biosynthetic pathway (Navarro et al. 1995; Fraser et al. 1996). Overexpression of the carotenogenesis regulatory gene (crgA) can abolish the light requirement for carotene production in M. circinelloides, causing an accumulation of carotenes in the dark (Navarro et al. 2000). Remarkably, disruption of crgA leads also to yellow pigmentation, independent of light. CrgA is therefore proposed as a negative regulator of the carotene biosynthetic genes (Navarro et al. 2001). Targeting an expression cassette for integration at the crgA locus could be used as a method to identify homologous recombination at this non-essential locus, since gene disruption would result in yellow colonies in the dark.

In this paper, we report on the characterisation of gpd1P using GOX strains containing integrated expression cassettes with deletion derivatives to aid our understanding of regulated promoters in this model zygomycete and to develop improved promoters for heterologous gene expression.

Materials and methods

Strains and media

M. circinelloides strain R7B (ATCC 90680, a leucine auxotrophic derivative of ATCC 1216b) was used throughout this study as a transformation host (Roncero 1984). Strain ATCC 1216b has been referred to as M. circinelloides, M. racemosus and M. circinelloides syn. racemosus in the past two decades. For clarity, we amend the species name to M. circinelloides. M. circinelloides strains were grown in YPG (complete medium; Bartnicki-Garcia and Nickerson 1962), YNB (minimal medium; Lasker and Borgia 1980), Vogel’s medium supplemented with 1.5 g glutamic acid/l and 5 g casamino acid/l (McIntyre et al. 2002). All media were supplemented with 1 mg niacinamide/l and 1 mg thiamine chloride/l. When indicated, glucose was replaced with a mixture of alternative carbon sources at the following final concentrations: 1% glycerol, 0.4% ethanol and 20 mM gluconolactone (GDL medium).

Escherichia coli strains DH10B and Top10 (Invitrogen Corp.) were used for cloning. These strains were grown in LB medium (Sambrook et al. 1989) with 100 μg ampicillin/μl or 50 μg kanamycin/μl.

Construction of gpd1P deletion derivatives

Two genetic constructions, UP194 and UP186, were devised for integration that incorporated homologous flanking regions derived from the M. circinelloides gpd2 and crgA genes, respectively (Fig. 1B). For UP194, a 1.2-kb fragment of the gpd2 gene was amplified using primers SmaIgpd2fwd (CCCGGGATGGTTACTCAAGTTGGTATTAACG, at positions 351–371 in the M. circinelloides gpd2 sequence, SmaI site in italics; accession AJ293013) and SmaIgpd2rev (CCCGGGTTATTGAGCAGCAGCATCAACC, at positions 1,563–1,542 in the gpd2 sequence, SmaI site in italics). The amplified fragment was cloned in the pCR2.1 vector, resulting in pCRgpd2. Digestion of this clone with PshAI, which has a unique site in the cloned gpd2 sequence (at position 1,011 in accession AJ293013) resulted in a blunt-ended fragment containing approx. 500–600 bp of gpd2 sequence at either end (Fig. 1B). A PvuII-digested pEUKA4-gox1 vector (Wolff and Arnau 2002) was used to isolate a 7.7-kb fragment containing the leuA gene and the gox1 gene under the control of gpd1P. After purification from an agarose gel, the fragment was ligated to the linearised pCRgpd2 vector. The resulting vector, UP194, contained the leuA gene and the gpd1P-gox1 expression cassette flanked by gpd2 sequences (Fig. 1B). After SmaI digestion, a linear fragment of approx. 10 kb and containing the above elements was purified from gel and used to transform M. circinelloides.

Fig. 1A, B
figure 1

Analysis of the Mucor circinelloides full-length gpd1P promoter and genetic constructions for the integration of gox1 expression cassettes. A The full-length 741-bp gpd1P is shown as a black box and the deletion derivatives as grey boxes; and the position of each perfect CATCAC repeat in the gpd1P sequence is shown as a black R box. Other boxes depicted: white boxes heat-shock elements, C CAAT boxes, Ce CAAT enhancer box, CTr CT-rich stretch, D developmental sequence motif, ATGAAAT, T TATA box, G GCN4 box. The transcriptional start site is shown by an arrow (tss). B Genetic constructions used for integration. UP194 contains gpd2-flanking regions, while UP186 has crgA-flanking regions. Derivatives of UP186 were constructed for the three promoter derivatives by replacement of the ApaI-XhoI full-length gpd1P fragment with the corresponding DNA fragment. White arrow Aspergillus niger gox1 gene, striped arrow M. circinelloides leuA gene (selective marker), dotted box and arrow flanking regions derived from the M. circinelloides gpd2 gene (in UP194), grey box and arrow flanking regions derived from the M. circinelloides crgA gene (in UP186), white box transcriptional terminator from the M. circinelloides gpd1 gene. Only relevant restriction sites are shown: A ApaI, E EcoRI, N NotI, M MfeI, P PstI, Ps PshAI, Pv PvuII, S SmaI and X XhoI. The orientation of the expression cassettes in the genetic constructions was confirmed by PCR using the primers depicted as arrows 1–4; see Materials and methods

Full size image

Similarly, a genetic construction for integration into crgA (UP186; Fig. 1B) was also performed, cloning a 1.6-kb fragment of the M. circinelloides crgA gene into pCR2.1, using primers SmaIcrgAfwd and SmaIcrgArev (CCCGGGATGAATTGTATCCCAGATGAA TCCG and CCCGGGTTAAGAAAT ACAACAAGAGACAGGC, positions 626–650 and 2,209–2,233, respectively, SmaI site in italics in each; accession AJ250998). Subsequent cloning was as described above for UP194. For UP186 and derivatives, the expression cassette and selective marker were cloned into the unique MfeI site present in the amplified crgA fragment (position 1,101 in the crgA sequence), providing a 453-bp and 1,155-bp flanking region at the 5′ and 3′ ends, respectively (Fig. 1B). For the analysis of integration at the crgA locus, two primers named crgA-5′fwd and crgA-3′rev (depicted as primers 1 and 4 in Fig. 1B, CTATCGTCATCTAGCAGAATAC and CACGGCTCCATCCTGATATG, at positions 262–283 and 2,652–2,671, respectively, in accession AJ250998) were used together with primer leuA3′fwd (CATCAAGCGTACTGGTCTCGG, primer 2 in Fig. 1B, positions 2,083–2,103 in accession AJ548466) and primer gpd1termfwd (AATCATTTCTAGTCATTGCATTTC, primer 3 in Fig. 1B, positions 1,875–1,898 in accession AJ293012). For the wild-type crgA locus, the use of primers 1 and 4 should result in a 2.4-kb fragment. For the other primer combinations used, primers 1 and 2 should result in a 0.9-kb fragment and primers 3 and 4 should result in a 2.2-kb fragment, if the expression cassette is intact in the transformants.

The three gpd1P deletion derivatives were amplified using Taq polymerase (Invitrogen Corp.) and pEUKA4-crgA as a template (Wolff and Arnau 2002), together with the following 5′ primers: ApaIgpd1Pfwd1 (GCGGGCCCGAACAATTCATCCCTATAATCTC, ApaI site in italics, spanning positions 22–44 of the full-length gpd1P), ApaIgpd1Pfwd4 (GCGGGCCCCTTTATTCACCAAGGAAAGAAGTG, positions 375–398) and ApaIgpd1Pfwd6 (GCGGGCCCGGATCCGAATCCGATTCAAAC, positions 584–604), to generate the 713-, 361- and 151-bp derivatives, respectively. All gpd1P derivatives were amplified with the same 3′ primer: XhoIgpd1Prev1 (GGCTCGAGATTTATAAAATATATAGAGATATAG, XhoI site in italics, spanning positions 711–741 in gpd1P and including the restriction site). The PCR products were cloned in pCR2.1 (Invitrogen Corp.) and confirmed by sequencing. The relevant promoter fragments were subsequently cloned into the ApaI and XhoI sites of the integration cassette (UP194 or UP186, Fig. 1B), resulting in plasmids pGG28, pGG41 and pGG60, containing the 713-, 361-, and 151-bp promoter derivatives, respectively. In all the above constructions, digestion with SmaI resulted in a large linear fragment (approx. 10 kb) that included the flanking DNA regions at the ends, the leuA selective marker and the gox1-expression cassette (Fig. 1B).

DNA sequencing and sequence analysis

Sequence reactions were carried out using cycle sequencing with fluorescent primers on an ALF Express (Pharmacia), according to the manufacturer’s instructions. Sequence analysis of the gpd1P sequence (positions 1–741, accession AJ293012) and derivatives was carried out using the DNASTAR package (Lasergene, Madison, Wis.).

Transformation of M. circinelloides

Protoplast formation and transformation of M. circinelloides were performed as described by Van Heeswijck and Roncero (1984) and Wolff and Arnau (2002). Transformants were allowed to sporulate prior to three subsequent cycles of single-colony isolation on YNB.

Flask experiments

M. circinelloides strains were grown in Vogel’s medium (McIntyre et al. 2002) supplemented with 5.0% glucose, 0.5% glucose or GDL medium for 4 days in 500-ml flasks and shaking at 200 rpm, 28 °C. Samples (4 ml) were taken from the cultures and cells were removed by filtration, using acid-washed filters (Frisenette, Denmark). The cell-free supernatants were stored at −20 °C until analysis. The mycelia from the filtration were stored at −80 °C for DNA isolation.

Analysis of GOX production in the constructed M. circinelloides strains

GOX production was analysed by a 2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) (ABTS) assay. Commercial GOX from A. niger was used to generate a standard curve. Typically, a 10-μl sample or standard was mixed with 900 μl reaction buffer [10 mg ABTS, 1 ml glucose (20%), 5 μl peroxidase (10 units/μl), 16 ml 50 mM sodium phosphate buffer, pH 6.3, up to 20 ml H2O]. The reactions were incubated for 15 min at room temperature and the absorbance measured at 420 nm.

The amount of GOX secreted into the supernatant was also analysed on ABTS plates (YPG medium containing 1 unit peroxidase/ml, 1 mg ABTS/ml). GOX activity resulted in a green colour (Fig. 2).

Fig. 2
figure 2

GOX activity in spores from constructed strains. Spores (approx. 103) were spotted on 2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) indicator plates to reveal glucose oxidase (GOX) activity in strains containing integrated expression cassettes. Pictures were taken within 1 h of application; see Materials and methods

Full size image

Southern blot analysis

Chromosomal DNA for Southern blot analysis was isolated using FastDNA and FastPrep FP120 (BIO101, Calif.). Approximately 10 μg DNA were loaded onto an agarose gel containing ethidium bromide, subjected to electrophoresis and transferred to a GeneScreen membrane (Du Pont). Hybridisation was performed as described by Wolff and Arnau (2002). Probe labelling was performed using Ready-To-Go DNA labelling beads (Amersham Pharmacia Biotech, Sweden). Unincorporated nucleotides were removed using a NICK column (Amersham Pharmacia Biotech, Sweden).

Results

Analysis of the gpd1P sequence for relevant regulatory sequences

The 741-bp gpd1P fragment (accession AJ293012) was successfully used to drive heterologous gene expression in M. circinelloides. The initial sequence characterisation showed standard promoter features in the 741-bp fragment, e.g. a CT-rich stretch just upstream of the transcriptional start site, a TATA box and two CAAT boxes within the 200 bp upstream of the ATG start codon (Wolff and Arnau 2002; Fig. 1A). The expression of gpd1 in M. circinelloides is regulated by the carbon source. In fact, the level of gpd1 expression is directly correlated to the concentration of glucose or galactose in the medium (Wolff and Arnau 2002). Therefore, additional regulatory signals involved in the hexose-dependent expression of the gpd1 gene could be present in the cloned gpd1P fragment. Sequence analysis of the full-length (741-bp) fragment was carried out using MotifFinder (a program based on the TRANSFAC database, available at http://motif.genome.ad.jp). A significant number of putative regulatory sequences with a relevant score (cut off at 85) were identified when searching in the fungal database. Only sequences that are relevant for the regulation of gpd1P are described below.

There are 14 putative heat shock element (HSE) boxes (consensus AGAAN) present throughout the gpd1P sequence (Fig. 1A), suggesting that expression of the M. circinelloides gpd1 gene might be under stress regulation. HSE boxes are present in the promoter region of genes activated by stress via the co-operative binding of heat shock factor (HSF) trimers between adjacent and distantly spaced HSF recognition sequences like HSE. In the budding yeasts Saccharomyces cerevisiae and Kluyveromyces lactis, HSF can bind DNA in two different conformations, one that results in strong transcription induction and another in which induction is not very effective (Erkine et al. 1999).

The involvement of the above stress-related sequences in gpd1P regulation is supported by the fact that spores of producing strains display a high reporter enzyme activity (GOX) when plated on indicator ABTS plates before hyphal growth initiation (Fig. 2). Sporulation occurs upon nutritional stress, suggesting a role for the gpd1 gene during the sporulation process in M. circinelloides. Also, during batch fermentation, GOX activity increases during the stationary phase (J. Breum, personal communication), indicating that a complex growth-dependent regulation is exerted on gpd1P. The functionality of the above sequences and other regulatory sequences identified with MotifFinder remains unclear but their presence in gpd1P suggests a complex regulation.

Since M. circinelloides is a zygomycete and most of the fungal regulatory sequences in the MotifFinder database belong to ascomycetes, novel or zygomycete-specific regulatory sequences would not be revealed by this analysis. Therefore, a search was carried out for sequences implicated in the regulation of gene expression in other zygomycetes and repeated sequences present in gpd1P. In the promoter region of the zygomycete R. niveus pgk1 and pgk2 genes encoding 3-phosphoglycerate kinase, a sequence element (ATGAAAT) was described as implicated in the regulation of expression. A similar motif was also described in other fungal promoters. This motif is related to a transcriptional element involved in development and differentiation in eukaryotes (Takaya et al. 1995). A conserved ATGAAAT sequence is also present in gpd1P (positions 47–53, D box, Fig. 1A), suggesting a developmental regulation of gpd1P. The same motif is present in the promoter region of the M. circinelloides gal1 and glaM genes encoding galactokinase and glucoamylase, respectively (accessions AJ438267, AY168303). Carbon source regulation has also been shown for the M. circinelloides glaM gene (Houghton-Larsen and Pedersen 2003).

Remarkably, four perfect CATCAC repeats are scattered in the 741-bp gpd1P fragment (R boxes, Fig. 1A). This sequence shows no homology to known regulatory signals. These repeats are not present in the promoter region of the M. circinelloides gpd2 and gpd3 genes. No expression of these two homologous gpd genes was detected during vegetative growth and fermentation (Wolff and Arnau 2002). Both gpd2 and gpd3 include consensus TATA and CAAT boxes in the promoter region. Thus, gpd1 might represent the only functional gpd gene in M. circinelloides during vegetative growth and sporulation. Some of the identified regulatory signals in gpd1P may provide this functionality and regulation. In an effort to characterise the regulatory mechanisms involved in gpd1P regulation, a deletion analysis was performed.

Construction of gpd1P deletion derivatives

The original strategy to study the regulatory mechanisms involved in gpd1P-dependent expression was to construct sequential deletions from the 5′ end and place them upstream of the A. niger gox1 gene encoding a secreted GOX. Subsequently, integration of the different expression cassettes would result in strains where measurement of GOX activity in the culture supernatant would correlate with promoter activity. In order to achieve this, homologous recombination is desirable to integrate the expression cassette at the same locus. However, although feasible in M. circinelloides, homologous recombination is a rare event. We chose two different loci for targeting the expression cassettes: crgA and gpd2. Targeted gene disruption of crgA was shown to have no growth phenotype, although crgA strains showed reduced sporulation (Navarro et al. 2001). Similarly, no expression of the gpd2 gene was detected (Wolff and Arnau 2002). Therefore, both genes were considered suitable targets for integration.

Three gpd1P deletion derivatives (151, 361, 713 bp) were constructed with an original 3′ end fused to the gox1 gene, in order to maintain the sequence around the transcriptional start site (Fig. 1A). The shortest derivative lacks the upstream CAAT box and spans positions 584–741. The 361-bp and 713-bp derivatives include the region at positions 375–741 and 22–741, respectively. Both derivatives should include all the consensus promoter features (Fig. 1A; Brakhage et al. 1999).

The full-length gpd1P fragment was used in two sets of genetic constructions, one with gpd2 flanking regions (UPO194, Fig. 1B) and the other using crgA flanking regions (UPO186, Fig. 1B). All the other promoter derivatives (pGG28, pGG58 and pGG60 for 713-, 361- and 151-bp derivatives, respectively) were only used with crgA flanking regions (Fig. 1B; see Materials and methods for details of genetic constructions).

Construction and characterisation of M. circinelloides strains containing integrated gox1 expression cassettes

Transformation of strain R7B was carried out with SmaI-digested linear DNA fragments that included the relevant promoter derivative, the leuA selective marker and flanking regions for gpd2- or crgA-targeting (Fig. 1B). The transformation efficiency was in all cases more than 500-fold lower, compared with transformation using circular plasmids. Typically, between one and six transformants were obtained for each construction.

M. circinelloides produces uninucleated spores (Orlowski 1991). Thus, single-colony isolation and sporulation of the obtained transformants was performed to avoid heterokaryon strains. The obtained transformants were screened for GOX activity on plates. An overview of the strains obtained is shown in Table 1. For the full-length gpd1P, two transformants were obtained with UP194 (the genetic construction containing gpd2 flanking regions, Fig. 1B). One of these, strain UPO1171, showed high GOX activity on plates (Fig. 2) while the other strain, UPO1136, did not show GOX activity. In both cases, PCR analysis of the gpd2 locus showed that integration in these strains did not occur in the gpd2 gene but at other loci. Both strains contained a full-length gox1 gene and gpd1P, as shown by PCR with specific primers (data not shown). The lack of GOX activity in strain UPO1136 remains therefore unclear. Different integration events occurring in strains UPO1171 and UPO1136 may account for the observed differences. UPO1136 was not analysed further.

Table 1 Mucor circinelloides strains used in this study
Full size table

For constructions containing crgA-flanking regions, four strains were selected for further analysis. One of these, strain UPO1299, contains the full-length gpd1P. Strain GG86 includes the 713-bp gpd1P derivative. Strains GG103A and GG102A were obtained with pGG58 and pGG60 that contain the 361-bp and 151-bp derivatives, respectively. For the shortest derivative, neither GG102A nor the additional transformants obtained with pGG60 displayed GOX activity on ABTS plates.

Similar to the situation for the transformants obtained with the gpd2-flanking construction, some of the strains obtained with the crgA-flanking construction did not show GOX activity. Overall, the integration event following transformation with linear DNA in M. circinelloides seems rather variable, as reported by Arnau and Strøman (1993).

For strains constructed with crgA-flanking DNA, homologous integration at the crgA locus should result in a yellow colony phenotype during growth in the dark, if the endogenous crgA gene is disrupted (Navarro et al. 2001). In fact, strains UPO1299 and GG86 formed yellow colonies in the dark. Strains GG102A and GG103A produced colonies with an overall wild-type phenotype. Therefore, integration at the crgA locus might have occurred in UPO1299 and GG86, while other integrative events would be expected in GG102A and GG103A.

PCR was carried out to further characterise the integration events in these strains, using primers located at the junctions between the flanking regions used in the genetic constructions and the neighbouring region in the chromosomal locus. Thus, homologous integration occurred solely in strains UPO1299 and GG86, as demonstrated by the amplification of fragments spanning the junctions (Fig. 3A, lanes 7–8, 11–12). A 2.4-kb wild-type crgA PCR fragment was obtained for the host strain R7B and also for strain GG103A, suggesting that integration in this strain occurred ectopically, as for strain UPO1171 (Fig. 3A, lanes 1–2). It is conceivable that ectopic integration is favoured in M. circinelloides when targeting genes that are expressed at a low level, like crgA, or genes for which no expression is detected, like gpd2. Alternatively, integration might occur by homologous recombination at gpd1 or leuA sequences also present in the transforming DNA. Strain GG102A did not include gox1 sequences, as shown using gox1-specific primers, while positive amplification was obtained from all the remaining strains. Thus, GG102A might result from gene conversion at the leuA locus, since the strain was able to grow in defined medium without leucine and did not contain the reporter gene.

Fig. 3A–D
figure 3

Analysis of the integration site in the constructed strains. Molecular sizes (on the left or right of plates) are given in kilobases. A PCR analysis of the crgA locus using primers 1 and 4 (lanes 1–4), primers 1 and 2 (lanes 5–8), primers 3 and 4 (lanes 9–12), and genomic DNA from strains R7B (lanes 1, 5, 9), GG103A (lanes 2, 6, 10), UPO1299 (lanes 3, 7, 11) and GG86 (lanes 4, 8, 12). B Southern blot analysis of PstI-digested DNA from transformant strains UPO1299 (lane 1), GG86 (lane 2), GG102A (lane 3), GG103A (lane 4), UPO1171 (lane 5) and the host strain R7B (lane 6). A leuA-specific probe (see Fig. 1B) was used. The size of the band corresponding to the endogenous leuA locus is shown on the left. C Southern blot analysis of EcoRI-digested DNA from strains R7B (lane 1), UPO1171 (lane 2) and GG103A (lane 3), using the leuA-specific probe. D Southern blot analysis as in C, using the gox1-probe (see Fig. 1B)

Full size image

Southern blot analysis was carried out to further study the integration events occurring in strains UPO1171, UPO1299, GG86 and GG103A. Using a leuA-specific probe (Fig. 1B), the presence of an additional band was demonstrated in all four strains (Fig. 3B). Strain GG102A contained a 4.4-kb PstI band diagnostic of the endogenous leuA locus and also present in the host strain R7B, establishing that gene conversion or replacement occurred in this strain and adding conclusive evidence for the lack of an integrated expression cassette in this strain.

The relative intensity of the additional leuA band in strains UPO1171, UPO1299, GG86 and GG103A was in all cases similar to the endogenous leuA band, indicating that single-copy integration occurred in these strains (Fig. 3B) and precluding any possible effect or variation due to heterokaryon stages. For GG86 and UPO1299, a similar-size band corresponding to the integrated DNA was observed also demonstrating homologous integration at the crgA locus in these strains (Fig. 3B, lanes 1, 2).

For strains GG103A and UPO1171, a different large band was observed in PstI-digested DNA, confirming integration at different chromosomal loci in these strains (Fig. 3B, lanes 4, 5).

A further investigation was carried out to address the intactness of the integrated DNA. A Southern blot analysis using EcoRI-digested DNA was performed. As shown in Fig. 3C, two additional EcoRI bands were detected with the leuA probe in GG103A and UPO1171, in addition to the endogenous 2.5-kb band also present in the host strain. For GG103A, a 4.3-kb and a 1.4-kb band were detected. The size of the larger band is in agreement with the expected size of the corresponding construction containing at least the leuA gene and the complete gox1 expression cassette (Fig. 1B), while the smaller band represents the upstream region. Furthermore, the presence of leuA adjacent to the gpd1P::gox1 cassette was demonstrated, since the 4.3-kb band was also detected using the gox1-specific probe (Fig. 3D), establishing the presence of the leuA gene immediately upstream of the 361-bp gpd1P derivative in GG103A. Similarly, a 4.6-kb and a 1.6-kb band were detected in strain UPO1171 with the leuA probe; and the larger band was also detected with the gox1 probe. The presence of a 1.5-kb band in both strains confirms that the region downstream of the gox1 gene is also intact in GG103A and UPO1171 (Figs. 1B, 3D).

Analysis of GOX production in the constructed M. circinelloides strains

The obtained strains were further used to study GOX production. The fact that the strains resulted from different integrative events complicated the comparison of production levels between the strains but did not prevent a study of the regulation of expression for the same strain in different media or for strains resulting from a similar integration event at the same locus.

Different overall GOX levels could readily be detected on ABTS plates (Fig. 2). UPO1171 showed the highest level of GOX, while similar levels were observed for UPO1299 and GG103A. Low GOX activity was displayed by GG86, while no activity was detected for GG102A or the host strain R7B, as expected.

A comparison of the levels of GOX produced under different growth conditions for the different strains was performed to investigate whether the regulatory mechanisms involved in gpd1P expression were affected in the constructed strains.

Flask experiments were conducted using different carbon sources (5.0% glucose, 0.5% glucose, GDL medium; see Materials and methods) and culture supernatant samples were taken at different time-points to determine GOX activity. For strain UPO1171 containing a full-length gpd1P, the highest level of GOX was obtained when grown in 5% glucose, reaching about 2 mg/l, confirming previous observations on the glucose-regulated expression of the gpd1 gene (Wolff and Arnau 2002). Lower levels of GOX activity were measured during growth on low glucose and on the non-fermentable GDL medium, both conditions expected to decrease gpd1P promoter strength. Under these conditions, the measured GOX activity remained below 0.5 mg/l throughout growth (Fig. 4).

Fig. 4
figure 4

GOX activity measurements during growth. The activity present in culture supernatants was determined during growth in Vogel’s medium. The strains were inoculated in 5.0% glucose (circles), 0.5% glucose (squares) or GDL medium (triangles; see Materials and methods). The data are the result of three independent experiments. Vertical bars show standard deviations. L Litres

Full size image

Strain UPO1299 contains the same full-length promoter fusion as UPO1171, but the expression cassette is integrated at the crgA locus. As shown in Fig. 4, overall lower levels of GOX were detected for UPO1299 in all media, compared with UPO1171, suggesting that the crgA locus might be located in a chromosomal region of overall lower transcriptional activity. Expression of crgA is light-regulated and expression levels under induced conditions are low (Navarro et al. 2001). As for UPO1171, the highest GOX activity levels were obtained in medium with 5.0% glucose, reaching 0.45 mg/l at the end of growth. Low levels were obtained in 0.5% glucose, confirming that the glucose concentration-dependent regulation of gpd1P was maintained. The lowest levels for UPO1299 were detected in GDL medium (below 0.2 mg/l). The overall pattern of glucose-dependent production of GOX in UPO1299 was therefore comparable with the results obtained for UPO1171.

Together with UPO1299, GOX activity was also determined for the additional three strains that were obtained using crgA-flanking DNA with the 713-, 361- and 151-bp promoter derivatives. For strain GG86, GOX levels were overall 10-fold lower than for UPO1299, suggesting that removal of a small (28-bp) region at the 5′ end of the promoter might be needed for a high level of expression. However, as for UPO1171 and UPO1299, glucose regulation was maintained in GG86, since the highest levels of GOX were detected in 5% glucose. No activity was detected in GDL medium (Fig. 4). For strain GG102A, no GOX activity was detected in any of the conditions used. These results were expected, since this strain resulted from gene conversion at the leuA locus as previously reported in M. circinelloides, and they correlated with the observed lack of GOX activity and gox1 sequences in this strain (Figs. 2, 3).

A 361-bp promoter derivative drives expression of gox1 in a constitutive fashion in strain GG103A

As shown above, strain GG103A contains a 361-bp promoter fusion ectopically integrated and displays GOX activity on ABTS plates (Fig. 2). During growth in flasks, the levels of GOX for GG103A were similar when comparing media with 5.0% or 0.5% glucose. In both media, over 0.6 mg GOX/l were measured after 96 h (Fig. 4). The glucose concentration-dependent expression pattern of the full-length gpd1P was therefore lost in this derivative. The fact that relatively high levels of GOX were detected at high and low glucose suggested that a negative regulatory mechanism involved in glucose concentration-dependent induction was absent in this promoter derivative. Moreover, comparable levels of GOX activity were also measured in GDL medium. Thus, the 361-bp promoter derivative displayed a consistent expression pattern, even in the absence of non-fermentable carbon sources.

Discussion

An analysis of the M. circinelloides gpd1P was carried out using gpd1P deletion derivatives and the A. niger GOX1 as reporter. Strains containing integrated cassettes were constructed and used to study the carbon source regulation of gpd1P. In spite of the inherent difficulty in constructing strains in M. circinelloides by DNA integration, a set of strains carrying integrated expression cassettes was generated. The integrative events varied strongly, although targeting the DNA to the crgA locus resulted in half of the studied events in homologous integration. Homologous integration at this locus was also reported by Navarro et al. (2001).

To facilitate comparisons, the full-length promoter was included in strains carrying the expression cassette at two different locations (strains UPO1171, UPO1299). For all strains used to study GOX levels, the intactness of the expression cassette and the sequences upstream of the promoter was verified by PCR and Southern analysis (Fig. 3). Analysis of GOX levels during growth using different carbon sources or concentrations showed a consistent glucose-dependent expression pattern for the full-length promoter (Fig. 4). In contrast, carbon source-independent expression was observed in strain GG103A, providing the first example of a strong constitutive promoter in M. circinelloides.

Scarce information is available on regulated promoters and the mechanisms of gene expression in zygomycetes. A putative glucose-inducible sequence element was conserved in the promoter of the zygomycete R. niveus pgk1 and pgk2 genes, encoding 3-phosphoglycerate kinase (Takaya et al. 1995). Sequence analysis of the gpd1P sequence showed no homology between the suggested 21-bp glucose-inducible element present in the promoter region of these genes and the CATCAC repeat of the M. circinelloides gpd1P. Galactose induction occurs for gpd1P (Wolff and Arnau 2002) but not for pgk genes, suggesting a different regulatory mechanism for the promoter of fungal pgk genes and gpd1P.

With the data obtained in this work, a tentative model for the regulation of gpd1P can be put forward that incorporates the regulatory sequences identified and the measurement of GOX activity in the constructed strains (Fig. 5). The model assumes that the observed differences in the overall level of GOX activity between the strains used are due to the functionality of the promoter and, as we established, not due to the presence of neighbouring upstream sequences that might affect the expression of the gox1 gene, e.g. via read-through from a strong upstream promoter. Read-through from an upstream promoter is rather unlikely, since a common region that includes the leuA gene (in opposite orientation, with respect to the gpd1P derivative) and the 5′ flanking DNA is present in strains GG103A and UPO1171 (Fig. 3).

Fig. 5
figure 5

Model for the regulation of gpd1P. Two major mechanisms are involved in regulation of gpd1P: glucose induction and stress regulation. First, the model proposes the involvement of the four CATCAC repeats present in gpd1P (black R boxes) in recognition by a negative regulator. During growth in low glucose, a negative regulator (shaded oval) binds or interacts with gpd1P, lowering the transcription levels. In high glucose, the negative regulator is released, allowing high expression. Second, a positive regulator analogue to heat-stress transcription factor (HSF) binds co-operatively to heat-stress elements, resulting in an induction of expression

Full size image

A comparison of GOX levels in strains UPO1299 and GG86, both containing a homologous integrated transforming DNA at the crgA locus (Fig. 3), strongly suggested that the small deletion in the gpd1P of GG86 has a role in gpd1P promoter strength (Fig. 4).

The presence and distribution of numerous HSE boxes [especially the HSE triplets present in the region included in the 361-bp promoter derivative (Fig. 1A), which are often found in the promoter region of heat-shock genes in fungi and are recognised by HSF trimers] correlate with the observed GOX activity present in spores of UPO1171, UPO1299, GG86 and GG103A (Fig. 2). It remains unclear whether the five additional HSE boxes found in the upstream region of gpd1P not included in the 361-bp derivative also contribute to stress-induced expression. HSE boxes are widely found in the promoter region of other M. circinelloides genes, like gal1 and glaM (Houghton-Larsen and Pedersen 2003).

We confirmed by batch fermentation of strain UPO1171 that the level of GOX increases significantly during the stationary phase, after glucose is depleted. This observation adds evidence to the developmental regulation of gpd1P, as suggested by the presence of several regulatory sequences like the ATGAAAT motif, which is also found in the M. circinelloides gal1 promoter.

Carbon source induction and the effect of sugar concentration are lost in the 361-bp derivative (Fig. 4). Since this derivative lacks three out of the four CATCAC repeats, it is conceivable to speculate that these repeats are involved in the regulatory mechanism. The simplest model to explain these data includes a negative regulator that may bind to the CATCAC repeats in conditions of low concentration of hexose, reducing the expression level. An increase in glucose concentration would result in the release of the negative regulator and an increase in the level of expression (Fig. 5). Thus, removal of the recognition signals (e.g. three CATCAC repeats lacking in the 361-bp derivative) would result in a high expression level at low glucose concentration or in GDL medium, since no negative regulation is exerted. To prove this, a different constitutive promoter can be used, to which an upstream region spanning the four CATCAC boxes of gpd1P can be added. Glucose-induced expression of such a promoter will demonstrate the involvement of these sequences in the negative regulation of gpd1P.

The constitutive nature of the expression pattern observed for the 361-bp promoter derivative in strain GG103A was further demonstrated using strains that carried a plasmid construction containing the same gox1 expression cassette with the leuA gene located upstream and placed on a geneticin resistance-based selection system developed recently (data not shown). Positional effects that may influence the gox1 expression pattern in strain GG103A are therefore excluded.

The availability of a strong promoter whose strength is independent of glucose concentration represents a valuable genetic tool for expression studies in M. circinelloides.

Moreover, the use of the 361-bp promoter derivative provides an advantage for use in heterologous gene expression. Growth in high-glucose concentrations to induce suitable gpd1P-driven expression levels normally results in significant ethanol production and growth inhibition, since M. circinelloides is a crab-tree fungus (McIntyre et al. 2002).