The transcription factor Snf 1 p is involved in a Tup 1 p-independent manner in the glucose regulation of the major methanol metabolism genes of Hansenula polymorpha

Hansenula polymorpha is a methylotrophic yeast widely employed in biotechnology as a “protein factory”. Most promoters used for heterologous protein expression, like MOX (methanol oxidase) and DAS (di-hydroxy acetone synthase), are involved in the peroxisomal methanol metabolism (C1 metabolism) and are under strong glucose repression. Interestingly, the MOX promoter is subjected to glucose regulation also in Saccharomyces cerevisiae, a non-methylotrophic yeast in which this phenomenon is well studied. In this species, the transcription factor Tup1p plays an essential role in glucose repression of several genes. This effect is counteracted by the activator Snf1p when glucose is exhausted from medium. Therefore, to test whether this regulatory circuit has been conserved in H. polymorpha, HpTUP1 and HpSNF1 were partially cloned and disrupted. Deletion of HpTUP1 did not affect glucose repression of the major C1 metabolism genes (MOX, DAS). Thus, though conserved, HpTUP1 does not seem to take part in a general glucose repression in H. polymorpha. In contrast, the deletion of HpSNF1 led to significant decreases in the activation of these genes in the absence of glucose. Therefore, the effect of HpSNF1 in transcriptional activation may be through an HpTUP1independent circuit.


Introduction
Hansenula polymorpha (syn.Pichia angusta) belongs to a limited group of yeast capable of using methanol as the sole carbon and energy source (Veenhuis et al., 1983).The expression of the major enzymes accounting for peroxisomal methanol metabolism (C 1 metabolism), namely methanol oxidase (Moxp), catalase (Catp) and di-hydroxi acetone synthase (Dasp), is controlled at the transcriptional level.The genes encoding these three enzymes are fully induced when cells are grown on methanol, and repressed to different extents in the presence of glucose (Genu et al., 2003;Janowicz et al., 1985;Roggenkamp et al., 1984).For instance, Moxp can represent up to 30% of the total soluble cellular proteins in continuous cultures growing on methanol (Giuseppin et al., 1998).In the presence of glucose, however, no methanol oxidase activity can be detected in crude extracts (Veenhuis et al., 1989).In view of this, the MOX promoter has been widely employed for heterologous protein expression, and H. polymorpha has been considered a "versatile protein factory" (Gellissen, 2000;van Dijk et al., 2000).Nevertheless, although it is a very suitable source of carbon for general industrial fermentation, glucose cannot be regularly employed for protein expression in H. polymorpha.It would therefore be a significant technological improvement if the MOX promoter could also function in glucose-containing media.
The MOX promoter has been studied in detail.Promoter mapping has revealed four cis-acting regulatory elements (Godecke et al., 1994), but no H.polymorpha factor directly involved in MOX regulation has been identified to date.Remarkably, the MOX promoter can drive glucose-repressible expression of a reporter gene in Saccharomyces cerevisiae (Pereira and Hollenberg, 1996), a non-methylotrophic yeast.In this species, activation of the MOX promoter is dependent on Adr1p (Pereira and Hollenberg, 1996), a transcription factor that regulates the expression of a number of genes encoding peroxisomal enzymes (Simon et al., 1991).This finding indicates that mo-lecular mechanisms governing regulation by glucose may be conserved in both species.
In S. cerevisiae, the Mig1p-Ssn6p-Tup1p complex represses several genes involved in the utilization of different carbon sources (e.g.sucrose, galactose) when cells are grown in the presence of glucose (Bu and Schmidt, 1998;Rohde et al., 2000).When glucose is exhausted from the growth medium, the same genes need the protein kinase Snf1p to be properly derepressed (Gancedo, 1998;Ostling et al., 1996).Both the Mig1p-Ssn6p-Tup1p and the Snf1 complexes are highly conserved among several species and seem to contribute, in addition to glucose regulation, to many other processes (Cassart et al., 1997;Lo et al., 2001;Yamashiro et al., 1996;Zaragoza et al., 2000).As an example, the deletion of a TUP1 homolog in Candida albicans causes failure in hypha development (Braun and Johnson, 1997).
In this study, H. polymorpha genes highly homologous to TUP1 and SNF1 were cloned, disrupted and analyzed for their possible conserved roles in the regulation of genes involved in C 1 metabolism.The results show that mutants lacking HpTUP1 did not exhibit altered levels of MOX and DAS transcription when grown on methanol or glucose.On the other hand, mutants lacking HpSNF1 presented lower MOX and DAS induction on methanol, when compared to wild-type cells.Taken together, these results indicate that the participation of Snf1p in the regulation of the major H. polymorpha C 1 metabolism genes occurs in a Tup1p-independent manner, therefore implying a minor involvement of these conserved proteins in the mechanism of glucose repression of methanol metabolism.

Yeast strains and growth conditions
The H. polymorpha strain used to disrupt homologs of the TUP1 and SNF1 genes was isogenic to strain CBS 4732 (wild type -WT).YPD (1% yeast extract, 2% peptone, 2% glucose) was used as rich medium.SD (0.17% Yeast Nitrogen Base, 0.5% ammonium sulphate) was used as synthetic medium and was supplemented with 5% glucose or 1% methanol.Cells were grown at 37 °C.

Bacterial strains, plasmids and oligonucleotides
Escherichia coli strain DH10B (Invitrogen) was used for all DNA manipulations (grown at 37 °C in LB medium).The plasmid pF6A-kanmx4 (Wach et al., 1994) was used for construction of the disruption modules (see below).pGEM-T (Promega) and pUC18 were used for routine cloning of PCR products.The oligonucleotides used in this work are listed in Table 1 and were synthesized by Operon.

Cloning of H. polymorpha TUP1 and SNF1 genes
The cloning strategy was based on degenerated oligonucleotides.We designed degenerated primers for conserved regions of the TUP1 and SNF1 gene sequences from S. cerevisiae, Schizosaccharomyces pombe and C. albicans (Table 1).These primers were used to conduct PCRs, using H. polymorpha genomic DNA as template.Fragments containing 600 bp and 800 bp were amplified using primers DTup1 and DSnf1, respectively.These DNA fragments were cloned into pGEM-T (Promega) and sequenced.The DNA sequences obtained were then analyzed by Blast X and found to present high similarities with C. albicans TUP1 (4e-84) and SNF1 (5e-54).The sequences of the 5' and 3' extremities of each fragment were used to design the specific primers shown in Table 1 (Stup1F,  Stup1R; Ssnf1F, Ssnf1R).Each of these primers was used in combination with T7 or SP6 custom primers to amplify cDNA fragments from a H. polymorpha cDNA library, constructed in our laboratory using the vector pSPORT (Invitrogen).By this procedure, the putative HpTUP1 and HpSNF1 gene fragments were expanded.The predicted amino acid sequences encoded by the H. polymorpha TUP1 and SNF1 genes were compared with known homologs of other yeast species (see Figure 1) using the Multalin tool (www.prodes.toulouse.inra.fr/multalin/multalin.html).

Disruption and screening of Hptup1 and Hpsnf1 strains
The HpTUP1 and HpSNF1 DNA sequences obtained as described above were also used to design chimerical disruption oligonucleotides.These oligonucleotides (see Ta- ble 1 and Figure 2) present ~30 bp of the genes to be disrupted in their 5' extremities; the 3' extremities are composed of ~25 bp complementary to the Multiple Cloning Site of the pF6A plasmid in which the kanmx4 disruption module was integrated (Wach et al., 1994).Specific TUP1 and SNF1 kanmx4 modules were constructed with these primers (CDTup1.1+CDTup1.2and CDSnf1.1+CDSnf1.2,respectively) by standard PCR using the pF6A-kanmx4 plasmid as template.Either of these linear disruption modules bears sequences, at both the 5' and the 3' ends, which are homologous to TUP1 and SNF1 genes.These were used to transform H. polymorpha WT electro-competent cells (Cregg and Russell, 1998).The kanmx4 module confers resistance to the drug G418, used for preliminary screening.Positive transformants growing in the presence of G418 (Amersham) were confirmed by colony PCR and also by PCR with genomic DNAs, using the specific primers STup1.1+STup1.2;SSnf1.1+SSnf1.2(Table 1).

PCR conditions
All PCRs were performed in a final volume of 50 µL, using kits (Promega) and following instructions provided by the manufacturer.The reactions were carried out in a 9700 thermocycler (Applied Biosystems) using the following programs: i) for fragment amplification -96 °C for 4 min, followed by 30 cycles at 94 °C for 30 s, 40 °C (degenerated primers) or 50 °C (specific primers) for 30 s, and 72 °C for 1 min; ii) for construction of the disruption modules -96 °C for 4 min, followed by 30 cycles: 94 °C for 30 s, 70 °C for 30 s, and 72 °C for 1.5 min.

Induction experiments and Northern Blotting
Cells were grown overnight in 100 mL YPD shaking at 280 rpm, harvested (centrifuged at 5000 rpm, 10 min) and inoculated in the appropriate medium at an OD 600 = 2.0.After 6 h of growth, cells were collected for RNA extraction.Total RNA was prepared by the hot acid phenol method (Ausubel et al., 1998).Samples were resolved by formaldehyde gel electrophoresis, blotted onto a positively charged nylon membrane (Hybond N + -Amersham Pharmacia) and fixed following standard protocols.The probes used were the following DNA fragments (numbers refer to position relative to each ORF start codon): MOX -a 2.3 kb EcoRV fragment (+21 to +2323), CAT -a 1.7 kb EcoRI/EcoRV fragment (+172 to +1695), and DAS -a 1.0 kb (+197/+1196) PCR fragment obtained with the primers Dasf and Dasr.All probes were labeled by random primer extension with α-32 P[dATP] (Amersham Pharmacia).Hybridization was carried out at 42 °C following standard protocols (Ausubel et al., 1998).rRNA bands in agarose/formaldehyde gels stained with ethidium bromide prior to blotting were used as internal loading control.

Growth characterization
The different strains were grown to log-phase in YPD, harvested and resuspended in the appropriate media (indicated in Results) to an OD 600 = 0.01.Aliquots were collected each hour for 24 h and the A 600 was measured.The results presented are the average of three independent experiments.

Results
Identification of H. polymorpha TUP1 and SNF1 homologs H. polymorpha DNA fragments were obtained by PCR, using degenerated primers for important regions of the TUP1 and SNF1 genes (Table 1 and Figure 1).The putative HpTUP1 fragment encompassed six WD repeats (Figure 1), a typical motif of this protein family, which is involved in protein-protein interaction and signalling (Smith et al., 1999).The putative HpSNF1 fragment bears a region with high similarity to other related proteins, including part of the serine/threonine domain which is present in all SNF1-like kinases (Hanks and Hunter, 1995) (Figure 1).Fragments of both genes presented the highest similarities with homologs from the Candida genera, which is phylogenetically close to Hansenula.Sequence comparison analyses led us to the conclusion that the DNA fragments obtained encompass the most important functional regions of H. polymorpha TUP1 and SNF1 genes.

Construction of TUP1 and SNF1 H. polymorpha mutants by PCR
The HpTUP1 and HpSNF1 fragments were used to construct disruption modules, as described in Materials and Methods.H. polymorpha wild-type cells were transformed and positive colonies growing on G418 were selected.About 150 transformants for each gene were screened for the presence of the disruption module by PCR amplification using specific primers for the HpTUP1 and HpSNF1 gene fragments.Figure 2 depicts the disruption strategy and the PCR products of correct integrations, observed for 7% of the transformants, in comparison to amplification of the wild-type fragments.Growth abilities of tup1 and snf1 cells on plates containing G418 were also re-checked in comparison to wild-type cells (not shown).  .Screening for positive mutants growing in G418 was done by PCR, using specific primers for the HpTUP1 and HpSNF1 genes.These primers (STup1.1,STup1.2;SSnf1.1,SSnf1.2) amplify respectively 500 bp and 300 bp of the wild-type genomic copies of each gene.Mutants in which the genes were efficiently disrupted present a single 1.5 kb band as PCR product.

Expression of C 1 metabolism genes in tup1 and snf 1 mutants
It has been established in several yeast species that Tup1p plays a key role in repressing genes involved in the utilization of a number of carbon sources in response to the availability of glucose in a growth medium (Bu and Schmidt, 1998;Rohde et al., 2000).Conversely, it is also well known that Snf1p accounts for relieving the repression caused by the Mig1p-Ssn6p-Tup1p complex at different promoters (Gancedo, 1998;Ostling et al., 1996).Since the major H. polymorpha C 1 metabolism genes are severely repressed by glucose, it was hypothesized that Tup1p and Snf1p homologous proteins might also be involved in the mechanism of glucose repression/derepression of these genes.To address this issue, the expression of the genes MOX, DAS and CAT was analyzed by Northern Blot in wild-type, tup1 and snf1 cells grown on glucose or methanol.
In tup1 cells grown on either glucose or methanol, the transcription levels of the three genes were similar to those observed in the wild type (Figure 3).This indicates that the repression mediated by glucose on H. polymorpha MOX, DAS and CAT promoters is Tup1p-independent.Interestingly, snf1 cells presented a significant reduction in the levels of MOX and DAS transcription in cells grown on methanol (Figure 3), suggesting that SNF1 might be needed for activated transcription of these genes under this condition.These results indicate that Snf1p action in the derepression of the genes MOX and DAS happens through a different molecular mechanism (e.g., different from relieving the Mig1p-Ssn6p-Tup1p complex repression at promoters).On the other hand, the levels of CAT transcription remained unaffected in snf1 cells under all conditions tested (Figure 3), which indicates that the glucose repression/derepression mechanism acting on this promoter is also independent of Snf1p.

Effects of tup1 and snf1 disruptions on glucose and methanol utilization
In S. cerevisiae, disruption of the genes TUP1 or SNF1 lead also to several pleiotropic phenotypes, in addition to transcriptional regulation mediated by glucose.For instance, a TUP1 disruption caused cell flocculation (Lipke and Hull-Pillsbury, 1984), whereas a SNF1 disruption led to a general slow growth (Schuller and Entian, 1987).Thus, to examine whether H. polymorpha tup1 and snf1 cells presented altered growth abilities on glucose or methanol, or showed any additional phenotype, we monitored the growth rates of cells from each strain in liquid media (Figure 4B).Surprisingly, no difference was observed among the wild-type, tup1 and snf1 cells grown in either glucoseor methanol-containing media.Moreover, we could not detect any tup1 or snf1 pleiotropic phenotype under the conditions tested.To confirm this observation, we carried out patch growth assays (Demasi et al., 2001;Verdoucq et al., 1999) on plates containing glucose or methanol as carbon sources (Figure 4A).Once again, we did not observe differences in either glucose or methanol utilization among the three strains.Taken together, these results indicate that, contrarily to S. cerevisiae, tup1 or snf1 disruptions failed to lead to any direct effect in the carbon utilization pathways of H. polymorpha.

Discussion
The transcriptional regulation of H. polymorpha C 1 metabolism promoters, especially the MOX promoter, has been intensively studied because of their potential in mediating industrial heterologous protein expression.The most important feature of these promoters is the strong glucose repression mechanism that functions when glucose is available in the growth medium, which makes this sugar useless as a carbon source for industrial fermentation of H. polymorpha.Thus, understanding the precise molecular mechanisms governing glucose repression of C 1 metabolism promoters may be useful to expand their expression profiles and allow the utilization of these promoters under conditions where they are presently ineffective.For these reasons, we investigated possible contributions of the conserved H. polymorpha TUP1 and SNF1 genes in the glucose repression mechanism.
TUP1 mutations eliminate or reduce glucose repression of many S. cerevisiae promoters.In addition, they induce several other phenotypes, like flocculation, failure to sporulate, and sterility of MAT-α cells (Williams and Trumbly, 1990).Surprisingly, no apparently abnormal phenotype was observed in H. polymorpha tup1 cells.These cells were indistinguishable from the wild-type by microscopic analysis, and no variations in growth behavior or in expression of the C 1 metabolism genes were observed.It is important to note that, although the deletion achieved does not encompass the complete ORF, the removed region bears the WD40 repeats, which are essential and sufficient for TUP1 repressing function in S. cerevisiae (Sprague et al., 2000;Williams and Trumbly, 1990).These results were unexpected, since clear effects of TUP1 mutations have been demonstrated in other yeast species closer to H. polymorpha than S. cerevisiae.For example, Candida albicans TUP1, which works in a complex similar to that identified is S. cerevisiae, is involved in general regulation of metabolism, cellular morphogenesis and stress response (Murad et al., 2001).Although TUP1 is not essential in this species, its deletion led to female sterility, reduced growth rates and failure to turn on filamentous growth (Braun and Johnson, 1997;Yamashiro et al., 1996).
With regard to proper SNF1 function in S. cerevisiae, it is required for appropriate derepression of several glucose-repressible genes, such as SUC2 (encoding invertase), GAL1 (encoding galactokinase) and ADH2 (encoding alcohol dehydrogenase 2) (Abrams et al., 1986;Denis and Audino, 1991;Flick and Johnston, 1990).The snf1 mutation leads to pleiotropic effects, such as a petite-like phenotype and a generally slow growth, which gives the mutant cells a "sick" appearance (Carlson et al., 1981).One of the main roles of Snf1p in S. cerevisiae is to relieve repression mediated by the Mig1p-Ssn6p-Tup1p complex, by phosphorylating the Mig1p protein and releasing the complex from promoters (Ostling and Ronne, 1998).Additionally, Snf1p is also involved in the operation of transcription factors like Adr1p (Young et al., 2002), and possibly other still unidentified factors (Gancedo, 1998).For example, disruption of the Candida glabrata SNF1 homolog resulted in the loss of the ability to utilize trehalose, indicating that even in an organism with a limited carbon utilization spectrum, the regulatory mechanism governing catabolite repression is preserved (Petter and Kwon-Chung, 1996).
Consistent with the possibility of playing a role in derepression in H. polymorpha, the deletion of SNF1 led to a significant reduction of the MOX and DAS expression in cells growing in methanol.This effect seems to be specific for some genes, since CAT expression remained unaffected by the deletion.These results reinforce the view that MOX and DAS are tightly co-regulated, and that possibly Snf1p plays a role in their regulatory circuits.However, this role may not be due to the inactivation of the Mig1p-Ssn6p-Tup1p complex, as has been reported for several S. cerevisiae genes (Gancedo, 1998).As described above, HpTUP1 does not seem to repress either MOX or DAS expression.Probably, HpSNF1 exerts its effect by interacting with a putative Adr1p-like factor that is believed to regulate MOX (Pereira and Hollenberg, 1996).In S. cerevisiae, several genes, including some that encode peroxisomal proteins, are regulated by both factors (Denis and Audino, 1991;Navarro and Igual, 1994;Simon et al., 1992).Moreover, ScSnf1p has been recently demonstrated to promote binding of Adr1p to chromatin (Young et al., 2002).Taken together, these facts indicate that both factors (Snf1p and Adr1p) act in conjunction in the regulation of some genes, a situation that might be conserved in other yeast species.
However, despite the reduction in MOX and DAS expression, deletion of HpSNF1 did not lead to any obvious external phenotype, in contrast to what has been observed in C. albicans (Petter et al., 1997) and C. tropicalis (Kanai et al., 1999), species which are phylogenetically close to H. polymorpha.Another interesting observation was that the snf1 growth rate on methanol-containing media was indistinguishable from wild-type cells, indicating that the reduction in MOX expression was not sufficient to create a bottleneck in peroxisomal methanol metabolism.and swi3 cells in agar plates containing 5% glucose or 1% methanol (Verdoucq et al., 1999;Demasi et al., 2001).Plates were photographed after 48 h of incubation at 37 °C.B: Growth rates of WT, tup1 and snf1 strains on liquid media containing 5% glucose or 1% methanol were determined by monitoring growth for up to 24 h.The results are the average of three independent experiments.
In summary, our work indicates that, in contrast to the observations made for other yeast species, TUP1 and SNF1 do not play essential roles in carbon source regulation in H. polymorpha.Nonetheless, Snf1p may participate in the activation of genes repressed by glucose, but at least in the case of the genes MOX and DAS, its mechanism of action differs from counteracting the repression mediated by the Tup1p-containing complex.

Figure 2 -
Figure2-Disruption strategy and screening of H. polymorpha tup1 and snf1 mutants.The specific kanmx4 disruption modules were constructed by PCR using chimeric oligonucleotides (see Materials and Methods for details).Screening for positive mutants growing in G418 was done by PCR, using specific primers for the HpTUP1 and HpSNF1 genes.These primers (STup1.1,STup1.2;SSnf1.1,SSnf1.2) amplify respectively 500 bp and 300 bp of the wild-type genomic copies of each gene.Mutants in which the genes were efficiently disrupted present a single 1.5 kb band as PCR product.

TUP1Figure 3 -
Figure 3 -Expression of MOX, CAT and DAS in WT, tup1 and snf1 cells.Northern blotting of samples collected after shifting cells from repressing (cells grown overnight in glucose 5% -time 0) to derepressing conditions (methanol 1%) for 6 h: G -glucose; M -methanol.rRNA bands of samples used in each filter are shown in agarose-formaldehyde gels stained with ethidium bromide prior to blotting.The probes used are described in Materials and Methods.

Figure 4 -
Figure4-Effects of tup1 and snf1 deletions on growth on glucose and methanol.A: Patch growth assays were performed with wild-type, swi2 and swi3 cells in agar plates containing 5% glucose or 1% methanol(Verdoucq et al., 1999;Demasi et al., 2001).Plates were photographed after 48 h of incubation at 37 °C.B: Growth rates of WT, tup1 and snf1 strains on liquid media containing 5% glucose or 1% methanol were determined by monitoring growth for up to 24 h.The results are the average of three independent experiments.