Trm1p, a Zn(II)₂Cys₆-Type Transcription Factor, Is a Master Regulator of Methanol-Specific Gene Activation in the Methylotrophic Yeast Candida boidinii

Abstract

The methylotrophic yeasts are commonly used as hosts for heterologous gene expression. In this study, we describe a novel gene, TRM1, in Candida boidinii, responsible for the transcriptional activation of several methanol-inducible promoters. The encoded protein, Trm1p, is a Zn(II)₂Cys₆-type zinc cluster protein. Deletion of TRM1 completely inhibits growth on methanol but causes no growth defect on glucose or other nonfermentative carbon sources, glycerol, ethanol, or oleate. Trm1p is responsible for transcriptional activation of five methanol-inducible promoters tested, but not for peroxisome assembly or peroxisomal protein transport. Expression of the TRM1 gene was constitutive, and Trm1p localizes to the nuclei regardless of the carbon source. Two cis-acting methanol response elements (MREs), MRE1 and MRE2 are present in the promoter of the dihydroxyacetone synthase gene. Trm1p is shown to be required for MRE1-dependent methanol-inducible gene expression. Chromatin immunoprecipitation assays reveal that Trm1p binds to five methanol-inducible promoters upon methanol induction but does not bind in glucose-grown cells. Thus, the TRM1 gene encodes a master transcriptional regulator responsible for methanol-specific gene activation in the methylotrophic yeasts.

Methylotrophic yeasts, which can utilize methanol as the sole carbon and energy source, have been used as the hosts for production of a wide variety of heterologous proteins and also as a model organism for studies of peroxisome biogenesis and degradation (4, 14, 19). To date, heterologous expression systems have been established in Candida boidinii, Hansenula polymorpha (Pichia angusta), Pichia methanolica, and Pichia pastoris (2, 4, 15, 22). With these expression systems, a large number of useful proteins have been produced, e.g., enzymes, antibodies, cytokines, plasma proteins, and hormones (4). The unique feature of these systems is that heterologous gene expression can be driven by methanol-inducible promoters, allowing production of the target protein to be tightly regulated; this is particularly important when the desired protein is toxic to the host cell. The molecular mechanism of methanol-inducible gene expression, however, has yet to be elucidated.

When methylotrophic yeasts grow on methanol as the sole carbon source, peroxisomes massively proliferate and can occupy up to 80% of the intracellular volume. Methanol-induced peroxisomes contain methanol-metabolizing enzymes, alcohol oxidase (AOD), dihydroxyacetone synthase (DAS), and catalase. AOD catalyzes the oxidation of methanol to form formaldehyde and hydrogen peroxide. Formaldehyde is situated at the branching point of the assimilation and dissimilation pathways. DAS catalyzes the first reaction of the assimilation pathway by fixing formaldehyde with xylulose-5-phosphate. Alternatively, formaldehyde can be oxidized to carbon dioxide by the glutathione-dependent formaldehyde oxidation pathway, catalyzed by formaldehyde dehydrogenase, S-formylglutathione hydrolase, and formate dehydrogenase (FDH) in the cytosol. More than 10 enzymes involved in the methanol metabolism are induced, along with peroxisome proliferation, during growth on methanol (32, 33, 35).

The regulatory profile of methanol-inducible gene expression has been studied mainly by focusing on the promoters of AOD-encoding genes. When grown on glycerol, C. boidinii and H. polymorpha exhibited ∼10% and 80% of the methanol-induced maximum AOD expression level, respectively (Fig. 1). Glucose-limited chemostat culture experiments (3) also showed that the levels of AOD in H. polymorpha gradually increased with decreasing dilution, whereas the derepression of AOD was lower in C. boidinii than in H. polymorpha. Therefore, the extent and mode of derepression differ among the methylotrophic yeast species. Furthermore, in C. boidinii, the maximum expression of the AOD1 gene required not only derepression (gene activation without methanol) but also methanol-specific induction (activation by methanol) (Fig. 1). Thus, compared to other methylotrophic yeasts, C. boidinii has several distinct features with respect to regulation of methanol-inducible gene expression. These features enabled us to determine the genetic factors specific to methanol induction.

FIG. 1.

Glucose derepression and methanol-specific induction. The expression levels of the H. polymorpha MOX gene, C. boidinii AOD1 gene, and C. boidinii DAS1 gene during growth on various carbon sources are compared with those during growth on methanol and shown as relative expression levels. On glucose-containing medium, expression is completely repressed. When glucose is exhausted or the cells are shifted to glycerol-containing medium, the expression of MOX and AOD1 is induced (glucose derepression). The extent of derepression of AOD genes differs between H. polymorpha and C. boidinii. When the cells are grown on methanol, the maximum level of expression of AOD genes is achieved not only by glucose derepression but also by methanol-specific induction. The contribution of methanol-specific induction to the expression of the AOD gene in C. boidinii is much higher than in H. polymorpha. The induction of DAS1 on methanol-containing medium is achieved only by methanol-specific induction. Through this study, Trm1p is shown to be required for methanol-specific induction.

In previous work, we isolated the promoter regions of five methanol-inducible genes from C. boidinii, evaluated their strength, and studied their regulation in detail (34). Among the five promoters studied, the promoter of the gene encoding DAS (P_DAS1) was the strongest and attained transcription rates several times higher than that of the commonly used AOD1 gene promoter (P_AOD1). P_AOD1 showed a maximum level of expression in cells grown on methanol, a derepressed level of expression in cells grown on glycerol or oleate, and was repressed in cells grown on glucose or ethanol (Fig. 1). In contrast, P_DAS1 did not show a derepressed level of expression in any of the alternate carbon sources. Similar results have been reported for the DAS-encoding gene in H. polymorpha and P. pastoris (16, 30).

In H. polymorpha, Mpp1p regulates the levels of peroxisomal matrix proteins and peroxins (8). Also, in H. polymorpha, the SWI/SNF complex (Swi1p and Snf2p) plays a role in transcriptional control of methanol-inducible gene expression (13), suggesting that chromatin remodeling participates in transcriptional regulation of methanol-inducible genes. In P. pastoris, Mxr1p (a homologue of the C₂H₂-type transcriptional factor Adr1p in Saccharomyces cerevisiae) was shown to control the transcriptional level of methanol-metabolizing genes, especially AOD1, as well as the PEX genes (9). All of these transcriptional regulators seem to be involved in derepression of gene expression, yet the mechanism of methanol-specific induction remains completely unknown.

Here, we aim to clarify the genes responsible for the regulation of methanol-inducible promoters, especially in regard to methanol-specific induction as opposed to derepression. C. boidinii has several advantages as a model system for the study of this mechanism. (i) Its methanol-inducible promoters exhibit both the derepressed levels and the induced levels of expression. (ii) The difference between the derepressed and the methanol-induced transcription levels is larger than in other species of methylotrophic yeasts (Fig. 1). Previously, we have developed a new gene-tagging mutagenesis method that utilizes random integration of linear DNA fragments in the C. boidinii genome (24). Using this method, we isolated mutants with defects in transcriptional activation of methanol-inducible promoters (P_AOD1 and P_DAS1); we refer to the mutants obtained from this screen as TRM genes (transcriptional regulation of methanol induction). In this study, we describe a novel gene, designated TRM1, which encodes a Zn(II)₂Cys₆-type protein, and demonstrate that Trm1p positively regulates the methanol-specific induction of multiple methanol-inducible promoters.

MATERIALS AND METHODS

Strains, media, and cultivation conditions.

The haploid strain C. boidinii S2 was used as the wild-type strain. C. boidinii strain TK62 (ura3 [17]) was used as the host for transformation. Strain AP and strain DP, which express the acid phosphatase (APase) gene derived from S. cerevisiae (S. cerevisiae PHO5 [ScPHO5]) under the control of P_AOD1 and P_DAS1, respectively, were used as host strains for gene-tagging mutagenesis (34). Escherichia coli DH10B and SA116 (17) were used for plasmid propagation.

The yeast strains were grown on either YPD medium (2% glucose, 2% Bacto peptone, 1% Bacto yeast extract) or YNB medium (0.67% yeast nitrogen base without amino acids). One or more of the following were used as the carbon source in YNB medium: 2% (wt/vol) glucose (YND), 2% (vol/vol) glycerol (YNG), 0.7% (vol/vol) methanol (YNM), and 0.5% (vol/vol) oleate (YNO). Tween 80 was added to the medium containing oleate at a concentration of 0.05% (vol/vol). For peroxisome observation, YPM (0.7% methanol, 2% peptone, 1% yeast extract) or YPO (0.5% oleate, 2% peptone, 1% yeast extract) medium was used. The initial pH of the medium was adjusted to 6.0. Cultivation was performed at 28°C under aerobic conditions with reciprocal shaking, and the growth of the yeast was monitored by measuring the optical density at 610 nm (OD₆₁₀).

E. coli was grown at 37°C in LB medium (1% tryptone, 0.5% yeast extract, and 0.5% NaCl) supplemented, when required, with ampicillin (50 μg/ml) or zeocin (Invitrogen, Carlsbad, CA) (50 μg/ml).

Enzyme assays.

APase activity was measured as described previously (29, 34) with one modification: at the beginning of induction, the initial OD₆₁₀ of the medium was adjusted to 1.0.

Cloning and sequencing of the TRM1 gene.

The genomic DNA of mutant strain AP9 was digested with either EcoRI or PstI, self-ligated, and introduced into E. coli DH10B cells. The resulting plasmids were sequenced. The 4.2-kb disrupted gene of mutant strain AP9 (24) was named TRM1. An approximately 6.6-kb DNA fragment harboring the TRM1-encoding region was sequenced.

Disruption of the TRM1 gene.

The upstream region of the TRM1 gene was amplified by PCR with the primers TRM1-up-KpnI and TRM1-up-SacI, using genomic DNA as a template. The 0.8-kb PCR-amplified fragment was inserted into pGEM-T Easy (Promega, Madison, WI), yielding pGEM-TRM1-up. The downstream region of the TRM1 gene was amplified with the primers TRM1-down-XhoI and TRM1-down-BglII using the genomic DNA as a template. The 0.7-kb DNA fragment was inserted into pGEM-T Easy, yielding pGEM-TRM1-down. Four DNA fragments [the 3.0-kb NotI-EcoRI fragment from pBluescript II SK(+), the 4.3-kb SacI-XhoI fragment from pSPR harboring the C. boidinii URA3 gene (21), the 0.8-kb NotI-SacI fragment from pGEM-TRM1-up, and the 0.7-kb XhoI-EcoRI fragment from pGEM-TRM1-down] were ligated, yielding the TRM1 disruption vector pTRM1-dis. After pTRM1-dis was digested with KpnI and BglII, the 5.8-kb fragment was gel purified and used to transform C. boidinii TK62 to uracil prototrophy using a modified version of the lithium acetate method (17). The disruption of the TRM1 gene was confirmed by Southern blot analysis using EcoRI-digested genomic DNA of transformants and 0.7-kb XhoI-EcoRI fragment from pGEM-TRM1-down as the probe. The trm1Δ strain was converted to uracil auxotrophy by 5-fluoroorotidic acid selection, yielding the trm1Δ ura3 strain.

Observation of peroxisomes and fluorescence microscopy.

The plasmids harboring the green fluorescent protein-peroxisome targeting signal 1 (GFP-PTS1) gene expression cassette (pGFP-PTS1) (23) were integrated into C. boidinii TK62, C. boidinii trm1Δ ura3, and C. boidinii aod1Δ das1Δ ura3 (18) strains; the transformants of these three strains were named GFP-PTS1/wt, GFP-PTS1/trm1Δ, and GFP-PTS1/aod1Δdas1Δ, respectively. Observation of GFP fluorescence was done with an IX70 fluorescence microscope (Olympus, Tokyo, Japan). Images were acquired with a Sensys charge-coupled device camera (Photometrics, Tucson, AZ) and analyzed on MetaMorph imaging software (Universal Imaging, West Chester, PA). To observe the C. boidinii cells expressing GFP-PTS1, cells were incubated for 6 h with a starting OD₆₁₀ of 0.5 in YPM or YPO medium.

Construction of trm1Δ strains expressing the ScPHO5 gene under the control of various methanol-inducible promoters.

The plasmids harboring the ScPHO5 gene under the control of various methanol-inducible promoters (pAPU, pDPU, pFPU, p20PU, and p47PU) (34), as well as the ACT1 promoter (P_ACT1) as a negative control (pActPU), were transformed into the trm1Δ ura3 strain. The integrative events of the transformants were analyzed by Southern blot analysis with EcoRI-digested chromosomal DNA, using the BamHI-PstI fragment of the URA3 gene as a probe (data not shown). Transformants that showed a single integration event of pAPU, pDPU, pFPU, p20PU, p47PU, and pActPU at the ura3 locus of the chromosomal DNA in C. boidinii trm1Δ ura3 strain were isolated and named APtrm1Δ, DPtrm1Δ, FPtrm1Δ, 20Ptrm1Δ, 47Ptrm1Δ, and ActPtrm1Δ strain, respectively.

Construction of a strain expressing Trm1p protein fused to YFP.

To visualize the localization of Trm1p, a strain expressing a yellow fluorescent protein (YFP)-Trm1p fusion protein was constructed as follows. First, the TRM1 coding region was amplified by PCR with the primers SalI-TRM1-Fw and XbaI-TRM1-Rv (Fw indicating forward, and Rv indicating reverse), using genomic DNA as a template. Next, the TRM1 promoter (P_TRM1) region was amplified by PCR with the primers SacI-PTRM1-Fw and BamHI-PTRM1-Rv, using genomic DNA as a template. Then, PCR was performed with the primers BamHI-YFP-Fw and SalI-YFP-Rv, using a chemically synthesized fragment harboring the coding sequence of YFP protein optimized for the preferred codon usage of C. boidinii as a template (unpublished). The 5.4-kb SacI-SpeI fragment of pAPU, the 0.8-kb BamHI-SalI fragment of amplified P_TRM1 region, the 0.7-kb BamHI-SalI fragment of amplified YFP coding region, and the 4.2-kb SalI-XbaI fragment of the amplified TRM1 coding region were ligated to yield pTYT. The pTYT was linearized with EcoT22I and transformed into a trm1Δ ura3 strain. The resulting strain was named YFP-Trm1p strain.

Fluorescence microscopy and nuclear staining were performed as follows. YFP-Trm1p cells grown to mid-log phase in YND or YNM medium were harvested, washed once, and fixed with 1 ml of 70% ethanol for 30 min at room temperature. Fixed cells were then washed twice, resuspended in 150 μl sterilized water, and stained with 150 μl of 0.125 μg/ml DAPI (4′,6′-diamidino-2-phenylindole) solution. After 10 min of incubation, fluorescence was observed.

Western blot analysis.

Yeast cells grown in YNM medium at an OD₆₁₀ of 1.0 were collected and resuspended in 100 mM potassium phosphate buffer (pH 7.5) containing 1 mM PMSF (phenylmethylsulfonyl fluoride). Cells were disrupted by zirconia beads with a diameter of 0.5 mm. The crude extract (100 μg) was separated by 7% sodium dodecyl sulfate-polyacrylamide gel electrophoresis and blotted onto a nitrocellulose membrane. YFP-Trm1p was detected by anti-GFP polyclonal antibody (Molecular Probes, Eugene, OR) and horseradish peroxidase-linked anti-rabbit antibody.

Deletion analysis of the P_DAS1 promoter.

The deletion series of P_DAS1 was constructed as follows. The plasmids pDPUΔ1480, pDPUΔ878, and pDPUΔ524 were obtained as gifts from Kirin Brewery Co., Ltd. The PCR was performed by primers 5S1603 and 3′-DAS1 for pDPUΔ1603 and by primers 5S1006 and 3′-DAS1 for pDPUΔ1006, using pDPU as a template. Each of the 5′ deleted P_DAS1 fragments were digested with SacI and BamHI, and subcloned into pBluescript II SK (+). The SacI-BamHI-digested fragments were replaced by the native P_DAS1 region of pDPU to produce pDPUΔ1603 and pDPUΔ1006. The plasmids pDPUΔ484-684, pDPUΔ484-984, and pDPUΔ484-1084, containing the ScPHO5-FDH1 terminator cassette under internal deletion derivatives of the P_DAS1 promoter, were obtained as gifts from Kirin Brewery Co., Ltd.

pM1ActPU and pM2ActPU, containing chimeric promoters in which the P_DAS1 fragment connects upstream of the P_ACT1 fragment, were constructed as follows. The P_ACT1 fragment (bp −585 to −1) was amplified by PCR with pACT1 (23) harboring P_ACT1 using the primer pair ACTPD1 and ACTPD2 or the pair ACTPD1-5S and ACTPD2, yielding HBACT1 or SBACT1, respectively. SBACT1 was digested with SacI and BamHI and replaced by P_AOD1 of pAPU to produce pActPU. PCR was performed with the primers MRE1-5-a and MRE1-3-c, using pDPU as a template. The SacI and HindIII fragment of the amplified product, the HindIII-BamHI fragment of HBACT1, and the 6.9-kb SacI-BamHI fragment of pAPU were ligated to produce pM1ActPU. Similarly, PCR was performed with the primers MRE2-5-a and MRE2-3-c, using pDPU as a template. The SacI-HindIII fragment of the amplified product was inserted into the SacI-HindIII fragment of pM1ActPU to produce pM2ActPU. The constructed plasmids were linearized by EcoT22I and transformed into C. boidinii aod1Δ das1Δ ura3 strain and aod1Δ das1Δ trm1Δ ura3 strain. Single-copy transformants were isolated by Southern blot analysis using the 1.9-kb PstI- and BamHI-digested fragment of pSPR as a probe. APase activity was measured after incubation in the presence of 0.7% methanol for 24 h.

ChIP assay.

The chromatin immunoprecipitation (ChIP) assay was done as follows. YFP-Trm1p cells grown in YNM medium or YND medium at mid-log phase were cross-linked for 10 min. Immunoprecipitation was performed by an anti-GFP antibody at a dilution of 1:400. The primers used for amplification of the promoter region are listed in Table 1.

TABLE 1.

Oligonucleotide primers used in this study

Primer	Sequence (5′→3′)a
TRM1-up-KpnI	GGTACCATCAGCCTAGAAATTGTCTCA
TRM1-up-SacI	GAGCTCGACTATTAGTAATGGAAATTATGT
TRM1-down-XhoI	CTCGAGTTTTACTAAACATTCCATACATAC
TRM1-down-BglII	AGATCTCAAATCAAAACTGGTCTTGATTTA
SalI-TRM1-Fw	ACGCGTCGACATGGTTTCTTCAAAATCACA
XbaI-TRM1-Rv	GCTCTAGATTATTTAGAATTATGTGAATTATAATC
SacI-PTRM1-Fw	AAACGAGCTCAATTACGGCACACGAGCATT
BamHI-PTRM1-Rv	CGGGATCCGACTATTAGTAATGGAAATTATG
BamHI-YFP-Fw	CGGGATCCATGGTTTCTAAAGGTGAAGAATTATTC
SalI-YFP-Rv	ACGCGTCGACTTTATATAATTCATCCATAC
5S1603	CGAGCTCGTGGAAAGCAGAATAATTAAAAT
3′-DAS1	CGGGATCCTTTTGTAATTTTTTTTAATAATTTATTAT
5S1006	CGAGCTCTAGGCTCCCTATTCCAAAAAG
ACTPD1	CCCAAGCTTCAGGCAGGCATTAAGGCAT
ACTPD2	CGGGATCCTTTTGTAATATATATTAAATTAAATTA
ACTPD1-5S	CGAGCTCCAGGCAGGCATTAAGGCAT
MRE1-5-a	AAACGAGCTCGGACCTCTCAGTTGCATTTC
MRE1-3-c	CCCAAGCTTCCCTTTTTGGAATAGGGAGCC
MRE2-5-a	AAACGAGCTCTCCGTGGAAAGCAGAATAAT
MRE2-3-c	CCCAAGCTTGATTACATAACAGACAATAGG
Zn-Fw	ACTTATGATCATCCACCTTCAAGAACTGTT
Zn-Rv	AGCTCTAGTGACTCTTCTTCTCTTTTGTAA
Linker-Fw	GAATTATTACAAAGAATGACAAATAGATTA
Linker-Rv	TCTTGAAGGTGGATGATCATAAGTACATTC
CC1-Fw	GATGATTCAAATAATAGTACTAATACAAAT
CC1-Rv	AGGATTAGTACTTGTTATTTGTGGAGGATT
MHR-Fw	AGTGAAGATGATTTTGATCAAGATATGCCA
MHR-Rv	TATACCATAAGTATCTCGGGTATCTGTAAT
CC2-Fw	TCAACTAGTACTACTACTACTACTTCTACA
CC2-Rv	TTCTGGATCTAAATCACCATTTCTATTTAA
Q-rich-Fw	GACAACCTGTCAGACTCAGTCTACAACAAT
Q-rich-Rv	AGTCAAATTACCCAATTCGTTATTTCTATT
Acidic-Fw	GGTGATTATAATTCACATAATTCTAAATAA
Acidic-Rv	TTTATTAAGTTGTTCTTGACGAGCTGGTAA
Zeo cassette Hind-Fw	CCCAAGCTTTGACTGACTGACATGGTCATA
Zeo cassette Hind-Rv	CCCAAGCTTTCCAGCTTGCAAATTAAAGCC
DAS1-ChIP-Fw	TGATTCTGTTCCGTGGAAAGCA
DAS1-ChIP-Rv	CAGACAATAGGTTAAATCTGTCATCATG
AOD1-ChIP-Fw	CCCAGCTTTTCAATTTAATAAAATAGCC
AOD1-ChIP-Rv	GATAGTAAATATAGTAAAATGTGATATGGG
FDH1-ChIP-Fw	TGTTACAATTGTCACAATTCTTGGATATAC
FDH1-ChIP-Rv	AACATCTGACTAGTATTACCATAAATGTAC
PMP20-ChIP-Fw	CAAGAGAATGAACCTGATTTTAACTGTC
PMP20-ChIP-Rv	TTCCACAGGTTCATCATACCAAAC
PMP47-ChIP-Fw	GATCTCACTCTCACAGATTCTTCAG
PMP47-ChIP-Rv	AGTATGAACAAGACCCAGATCCAG
TRM1-ChIP-Fw	TCATCTCCACCGGACTTTCC
TRM1-ChIP-Rv	TGAGCTGGATGTTTGCTTTTTGTCG
ACT1-ChIP-Fw	AGTTTACCCACTTGTTTTCATTCTCG
ACT1-ChIP-Rv	TAAGTGAATGAATGATGAATGAAATGGC

^aThe underlined nucleotide sequences are additional restriction enzyme recognition sequences.

Construction and characterization of domain deletion mutants.

Domain deletion mutants of YFP-Trm1p were constructed as follows. PCR was performed with the primer pairs Zn-Fw and Zn-Rv, Linker-Fw and Linker-Rv, CC1-Fw and CC1-Rv, MHQ-Fw and MHQ-Rv, CC2-Fw and CC2-Rv, Q-rich-Fw and Q-rich-Rv, and Acidic-Fw and Acidic-Rv using pTYT as a template; amplified fragments were self-ligated to produce pTYTZnΔ, pTYTLinkerΔ, pTYTCC1Δ, pTYTMHRΔ, pTYTCC2Δ, pTYTQ-richΔ, and pTYTAcidicΔ, respectively. The constructed plasmids were linearized by EcoT22I and transformed into C. boidinii trm1Δ ura3 strain. The domain deletion mutants generated from plasmids pTYTZnΔ, pTYTLinkerΔ, pTYTCC1Δ, pTYTMHRΔ, pTYTCC2Δ, pTYTQ-richΔ, and pTYTAcidicΔ were named ZnΔ, LinkerΔ, CC1Δ, MHRΔ, CC2Δ, Q-richΔ, and AcidicΔ, respectively. The DNA fragment harboring the zeocin resistance gene expression cassette was amplified by PCR with the primers Zeo cassette Hind-Fw and Zeo cassette Hind-Rv using pREMI-Zc (24) as a template. The amplified fragment was digested with HindIII and inserted into the HindIII site of pDPU. The resulting plasmid was named pDPU-Zeo. pDPU-Zeo was linearized by EcoT22I and transformed into each domain deletion strain. APase activity was measured in the presence of 0.7% methanol for 8 h.

Nucleotide sequence accession number.

The sequence of the TRM1 gene has been submitted to DDBJ and assigned accession number AB365355.

RESULTS

Screening for the trm mutants defective in transcriptional activity of methanol-inducible promoters.

We used the C. boidinii strains AP and DP, which express the acid phosphatase gene derived from S. cerevisiae (ScPHO5) under the control of P_AOD1 and P_DAS1, respectively, as host strains for gene-tagging mutagenesis. Mutants exhibiting lower APase activity were obtained (41 mutants from strain AP and 42 mutants from strain DP) by the procedure described previously (24), and we named these the trm mutants. Sequence analysis of the region flanking the pREMI-Zc inserted locus (24) revealed that three mutants, the trm1-1, trm1-2, and trm1-3 mutants, were disrupted in the same open reading frame. The trm1-1 mutant was derived from strain AP (AP9 in our previous report [24]), and the mutants trm1-2 and trm1-3 were derived from strain DP. We named the disrupted gene TRM1.

Characterization of the TRM1 gene.

We determined the sequence of the TRM1 gene. The TRM1 gene consists of a 4,230-bp open reading frame corresponding to 1,410 amino acid residues. The previously known gene most similar to TRM1 is H. polymorpha MUT3 (GenBank accession number AY046886; identity of 26.3% and similarity of 43.7%). The H. polymorpha mut3 mutant is not able to grow on methanol and has reduced activity of alcohol oxidase promoter (31). However, its molecular function has not been studied in detail. A motif search performed on the deduced amino acid sequence of the Trm1p protein revealed that Trm1p has a fungus-specific Zn(II)₂Cys₆ (Zn cluster)-type zinc finger domain close to the amino terminus (Fig. 2A). Most proteins belonging to the Zn(II)₂Cys₆ family are known to regulate either carbon metabolism, nitrogen metabolism, or amino acid biosynthesis in fungi (11). A primary sequence comparison of the region comprising the Zn(II)₂Cys₆ domain of H. polymorpha Mut3, P. pastoris RPPA05209, and the deduced amino acid sequence of Trm1p is shown in Fig. 2B. The primary sequences of the Zn(II)₂Cys₆ domains are well conserved among the three methylotrophic yeasts.

FIG. 2.

Functional domains of Trm1p. (A) Schematic drawing of the annotated domains within Trm1p. Zn, zinc cluster; Linker, linker region between the zinc cluster and coiled-coil 1 (CC1); MHR, middle homology region; Q-rich, glutamine rich domain; Acidic, acidic domain; a.a., amino acids. (B) Alignment of Trm1p and its homologous proteins of other methylotrophic yeasts. The N-terminal amino acid sequences of C. boidinii Trm1p (CbTrm1p) (GenBank accession number AB365355) including the Zn(II)₂Cys₆ zinc cluster, H. polymorpha Mut3p (HpMut3p) (GenBank accession number AY046886), and P. pastoris RPPA05209 (PpRPPA05209). Residues that are identical in all three proteins are shown on a black background. Residues that match in two of the three proteins are shown on a light gray background. Solid black circles represent cysteine residues.

Disruption of the TRM1 gene caused growth defect on methanol but not on other carbon sources.

The TRM1 gene was disrupted by replacing the open reading frame with the C. boidinii URA3 gene as a selective marker (Fig. 3A). Disruption of the TRM1 gene was confirmed by Southern blot analysis with EcoRI-digested genomic DNA from each transformant, using the downstream region of the TRM1 gene as the probe (Fig. 3B). In order to restore the selective marker, we isolated the trm1Δ ura3 strains as described in Materials and Methods.

FIG. 3.

One-step disruption of the TRM1 gene in C. boidinii. (A) Restriction map of the cloned fragment and disruption strategy. The boxes with diagonal lines represent repeated sequences for homologous recombination to remove the URA3 gene after the gene disruption. (B) Genomic Southern blot analysis of EcoRI-digested total DNAs (5 μg of each) from the wild-type strain (lane 1), the trm1Δ strain (lane 2), and the trm1Δ ura3 strain (lane 3). The probe used for the analysis is indicated by a thick black line.

The trm1Δ strain showed a severe growth defect on methanol but no growth defect on glucose or other derepressing carbon sources, such as glycerol and oleate (Fig. 4). The trm1Δ strain also showed normal growth on ethanol (data not shown). The Trm1p homologue in P. pastoris obtained by gene-tagging mutagenesis also showed a growth defect on methanol (data not shown). These results suggested that Trm1p is more involved in the regulation of methanol-specific induction rather than derepression.

FIG. 4.

Growth of the wild-type and trm1Δ strains on various carbon sources. Closed circles and open circles represent the wild-type strain and trm1Δ strain, respectively. The open triangles in panel B represent the YFP-Trm1p strain.

The TRM1 gene positively regulates the transcriptional activity of methanol-inducible promoters.

In order to investigate the influence of the TRM1 gene deletion on the activity of methanol-inducible promoters (P_AOD1, P_DAS1, P_FDH1, P_PMP20, and P_PMP47), we constructed reporter strains expressing the ScPHO5 gene under the control of these five promoters in the trm1Δ strain and assayed the enzyme activity of the reporter APase (Table 2). In all methanol-inducible promoters, the transcriptional activities in medium containing methanol drastically decreased in the trm1Δ strain (Table 2). This result indicated that Trm1p is a master regulator for methanol-specific induction. With respect to P_AOD1, the reporter activity in the trm1Δ strain still showed 6.55% of that in the wild-type strain. This means that the activation of P_AOD1 is regulated not only by methanol-specific induction but also by some other mechanisms.

TABLE 2.

Level of APase activity on glucose and methanol in the trm1Δ strain

Carbon source	Strain	APase activitya
		PAOD1	PDAS1	PFDH1	PPMP20	PPMP47	PACT1
Glucose	Wild type	0.37 ± 0.10	0.12 ± 0.06	0.35 ± 0.02	0.17 ± 0.03	0.21 ± 0.04	1.29 ± 0.09
	trm1Δ	0.28 ± 0.08	0.15 ± 0.03	0.37 ± 0.06	0.14 ± 0.01	0.28 ± 0.00	1.19 ± 0.12
Methanol	Wild type	74.5 ± 6.59	507 ± 18.4	83.1 ± 8.53	44.3 ± 3.87	1.75 ± 0.36	3.53 ± 0.09
	trm1Δ	4.88 ± 0.61	1.49 ± 0.28	0.15 ± 0.02	0.23 ± 0.01	0.17 ± 0.02	3.63 ± 0.28

^aThe level of APase activity of the strain on the respective carbon source is expressed as U/OD₆₆₀. The means ± standard deviations from three independent experiments are shown. The wild-type and trm1Δ strains were incubated in the presence of 2% glucose or 0.7% methanol for 8 h.

Trm1p does not participate in peroxisome assembly and transport of peroxisomal proteins.

In order to investigate whether Trm1p is involved in peroxisomal assembly or import of peroxisomal proteins, we expressed GFP tagged to a typical PTS1, AKL, under the control of P_ACT1, and observed the wild-type and trm1Δ strains with a fluorescence microscope (Fig. 5).

FIG. 5.

Trm1p is not involved in transport of peroxisomal protein. GFP-AKL fluorescence of wild-type, trm1Δ, and aod1Δ das1Δ cells in methanol- or oleate-containing medium. GFP fluorescence localizes to peroxisomes. (A) GFP-AKL/wt, GFP-AKL/trm1Δ, and GFP-AKL/aod1Δdas1Δ cells grown in YPM medium. (B) GFP-AKL/wt and GFP-AKL/trm1Δ cells grown in YPO medium.

When cells were induced in medium containing methanol, green fluorescence in the trm1Δ cells (GFP-PTS1/trm1Δ) was punctate, but its size was relatively small compared with the wild-type cells (GFP-PTS1/wt). We considered that such small peroxisomes in the methanol-induced trm1Δ cells were due to the lack of peroxisomal matrix proteins, such as AOD and DAS. To verify this, we observed GFP-PTS1 fluorescence in aod1Δ das1Δ cells (GFP-PTS1/aod1Δdas1Δ). As expected, the aod1Δ das1Δ cells in methanol-containing medium contained small peroxisomes, similar to those in the trm1Δ cells.

When cells were induced in another peroxisome-inducing carbon source, oleate, the trm1Δ cells exhibited a peroxisomal morphology similar to that of the wild-type cells. No differences concerning the number and size of peroxisomes between the wild-type cells and trm1Δ cells were observed. These observations indicate that Trm1p is not directly involved in peroxisomal assembly, peroxisomal proliferation, or transport of peroxisomal proteins.

Trm1p is constitutively expressed and localizes to the nucleus.

In order to investigate the expression of the TRM1 gene and the intracellular localization of Trm1p, we expressed a YFP-tagged Trm1p under the control of P_TRM1. The YFP-tagged Trm1p complemented the growth defect on methanol in the trm1Δ strain (Fig. 4B).

Western blot analysis was performed for protein extracts prepared from glucose- or methanol-grown cells of YFP-Trm1p strain using anti-GFP antibody. YFP-Trm1p was expressed in both glucose- and methanol-containing medium (Fig. 6A). In addition, YFP-Trm1p was expressed on other carbon sources, such as glycerol and oleate (data not shown). These data indicated that Trm1p is expressed constitutively.

FIG. 6.

YFP-Trm1p is constitutively expressed and localized in the nuclei regardless of carbon source. YFP-Trm1p was expressed under the control of P_TRM1. (A) YFP-Trm1p cells grown in YND medium or YNM medium was harvested. Cell extracts were prepared and analyzed by Western blot analysis using anti-GFP antibody. (B) YFP-Trm1p cells were grown in YND medium or YNM medium.

Localization of YFP-Trm1p was observed with a fluorescence microscope. YFP-Trm1p was localized in the nucleus (Fig. 6B), when the cells were grown on glucose or methanol.

Domain deletion analysis of Trm1p.

Trm1p has a Zn(II)₂Cys₆-type zinc cluster. In addition to this motif, two coiled-coil regions are predicted by COILS (http://www.ch.embnet.org/software/COILS_form.html) (10) (Fig. 2A). One coiled-coil region (coiled-coil 1 [CC1]) is positioned C terminal to the zinc cluster. The other (CC2) is positioned in the middle of Trm1p. Most Zn(II)₂Cys₆-type zinc cluster proteins have coiled-coil structure, which is most likely responsible for dimerization and protein-protein interactions. The region between the zinc cluster and coiled-coil is a linker region, which is responsible for DNA-binding specificity. The middle homology region and acidic region are also characteristic structures among Zn(II)₂Cys₆ zinc cluster proteins. The role of the middle homology region is not yet known. The acidic region acts as an activation domain. The glutamine-rich domain (Q-rich domain) is a unique sequence in Trm1p. There are some reports that glutamine-rich domains may act as transcriptional activation domains (27).

In order to characterize the functional domains of Trm1p, we constructed each domain deletion mutant of Trm1p. We analyzed growth on methanol, subcellular localization of Trm1p, and transcriptional activity of P_DAS1 for each mutant (Table 3). Growth on methanol was completely impaired in ZnΔ, LinkerΔ, MHRΔ, and CC2Δ strains. In CC2Δ and AcidicΔ strains, growth was partially impaired. In Q-richΔ strain, the growth on methanol was almost the same as that for full-length YFP-Trm1p strain. Interestingly, MHRΔ strain showed cytosolic localization. This suggested that the nuclear localization signal exists in the middle homology region (MHR) or that MHR masks nuclear export signal to regulate the nuclear localization of Trm1p. We measured transcriptional activity of P_DAS1 in each domain deletion mutant. In LinkerΔ, MHRΔ, and CC2Δ strains, P_DAS1 activity was completely abolished. This reduction of transcriptional activity may cause the growth defect on methanol. In ZnΔ and CC1Δ strains, P_DAS1 still retained detectable activity even though it was below 10% compared to full-length YFP-Trm1p strain. In AcidicΔ strain, P_DAS1 activity was ∼20% of the full-length YFP-Trm1p strain. Binding of domain-deleted Trm1ps to P_DAS1 was analyzed by ChIP assay (cf. Fig. 8), revealing that the zinc cluster, linker region, MHR, and CC2 are required for binding to P_DAS1. On the other hand, the CC1, Q-rich domain, and acidic regions were not required.

TABLE 3.

Characterization of domains in Trm1p

Strain	Growth on methanola	Localization of YFP-Trm1pb	APase activityc of PDAS1	Relative activityd	Binding to PDAS1 by ChIPe
YFP-Trm1p	+++	N	359 ± 52.8	100	+
ZnΔ	−	N	22.4 ± 2.22	6.24	−
LinkerΔ	−	N	2.70 ± 0.14	0.75	−
CC1Δ	+	N	21.1 ± 9.92	5.88	+
MHRΔ	−	C	2.61 ± 0.58	0.73	−
CC2Δ	−	N	2.69 ± 0.36	0.75	−
Q-richΔ	+++	N	285 ± 37.5	79.4	+
AcidicΔ	+	N	70.5 ± 4.75	19.6	+

^aGrowth on methanol as the sole carbon source. +++, same growth as YFP-Trm1p; +, weak growth; −, no growth.

^bN, nucleus; C, cytosol.

^cAPase activity is expressed as U/OD₆₆₀. The means ± standard deviations from three independent experiments are shown.

^dActivity is shown as a percentage of the activity of YFP-Trm1p.

^eBinding to P_DAS1 was assayed by the ChIP assay as described in the legend to Fig. 8. +, binding detected; −, binding not detected.

FIG. 8.

Trm1p specifically binds the promoter regions of methanol-inducible genes. ChIP assay was performed with cells grown on methanol (A) or glucose (B). YFP-tagged Trm1p and native Trm1p are immunoprecipitated with (+) or without (−) anti-GFP antibody. IP, immunoprecipitation; WCE, whole-cell extract.

Identification of cis-acting elements in P_DAS1.

We showed that among all methanol-inducible promoters tested, P_DAS1 is the most highly inducible (34). In addition, P_DAS1 shows a more rapid response to methanol than P_AOD1 (20). Furthermore, the transcriptional activity of P_DAS1 in the trm1Δ strain was reduced below 1% compared with that of the wild-type strain (Table 2). These features prompted us to determine the DNA sequences that are responsible for methanol-specific induction in P_DAS1. We constructed deletion series of P_DAS1 fused to the ScPHO5 reporter gene. The plasmid carrying these deletion derivatives of the P_DAS1 were integrated into the ura3 locus of C. boidinii aod1Δ das1Δ ura3 strain. Single-copy transformants were selected by Southern blot analysis (data not shown) and incubated in the presence of 0.7% methanol for 24 h. The expression of ScPHO5 under the deletion mutants of P_DAS1 was monitored by measuring the APase activity.

Figure 7 shows the series of 5′ deletions (Fig. 7A) and internal deletions (Fig. 7B), and their corresponding activity on methanol compared to that with the native P_DAS1. The 5′-endpoint nucleotide in each deleted promoter is numbered relative to the translation initiation codon of the DAS1 gene. Deletions extending from −1084 to −984 and from −1603 to −1480 showed drastic loss of the reporter activity; we named these two sequences methanol response elements (MREs), MRE1 and MRE2, respectively. In addition, it appears that there would be upstream repression sequences between −1480 and −1006 and between −984 and −684.

FIG. 7.

Deletion analysis of P_DAS1. ScPHO5 expression constructs with 5′ deletions (A) or internal deletions (B) are shown. The deletion breakpoints are numbered relative to the initiation codon of the DAS1 gene. The C. boidinii aod1Δ das1Δ strain carrying these chimera promoters fused to the ScPHO5 reporter was incubated in the presence of 0.7% methanol for 24 h. The enzyme activity is presented as the mean value ± standard deviation of the APase activity (U/OD₆₆₀) from three independent experiments and as a percentage of the activity of the strain carrying the control plasmid pDPU. The positions of methanol response elements MRE1 and MRE2 are shown.

In order to investigate whether these sequences are sufficient for methanol induction, we constructed artificial promoters. The MRE1 and MRE2 or the region from −1603 to −984 containing both MRE1 and MRE2 were fused upstream of P_ACT1 (Table 4). P_ACT1 is a constitutively expressed promoter, not induced by methanol. The plasmids expressing ScPHO5 under these artificial promoters (pM1ActPU, pM2ActPU, and pM1,2ActPU) were constructed and integrated into the ura3 locus in the genome of C. boidinii aod1Δ das1Δ ura3 strain. The level of APase activity was induced approximately 10 times in the presence of methanol by the addition of MRE1 to P_ACT1 (Table 4). This result suggests that the MRE1 is sufficient for methanol-specific induction. The MRE2-fused P_ACT1 was able to induce expression in the presence of methanol, albeit a minor effect compared to that of MRE1. Addition of MRE2 to MRE1 on P_ACT1 also enhanced the expression, indicating that MRE1 and MRE2 have distinguishable roles in methanol induction.

TABLE 4.

TRM1 is required for MRE1-dependent methanol induction

^aThe C. boidinii aod1Δ das1Δ strain and the aod1Δ das1Δ trm1Δ strain expressing the ScPHO5 gene under the control of P_ACT1 (−585 to −1) were incubated in the presence of 0.7% methanol or 2% glucose for 24 h or 8 h, respectively. The enzyme activity is presented as the mean value ± standard deviation of APase activity (U/OD₆₆₀) from three independent experiments, and APase activity compared to that in a strain carrying the control plasmid pActPU is shown as a percentage in parentheses.

Trm1p is involved in MRE1-dependent methanol induction.

In order to investigate whether Trm1p is involved in MRE-dependent methanol induction, pM1ActPU, pM2ActPU, and pM1,2ActPU were integrated into the ura3 locus in the genome of C. boidinii aod1Δ das1Δ trm1Δ ura3 strain. The APase activity was measured in the same way as for the aod1Δ das1Δ strain in glucose and methanol medium (Table 4). In the absence of Trm1p, MRE1-dependent methanol induction was fully abolished, indicating that Trm1p is required for MRE1-dependent methanol induction.

Trm1p specifically bound to multiple methanol-inducible promoters in methanol-grown condition.

In order to investigate whether Trm1p binds to the promoter region of methanol-inducible genes in vivo, ChIP assay was performed with methanol-grown cells and glucose-grown cells expressing YFP-Trm1p (Fig. 8). After immunoprecipitation, the promoter regions of five methanol-inducible genes (P_AOD1, P_DAS1, P_FDH1, P_PMP47, and P_PMP20) and P_TRM1 could be amplified from the template DNA from methanol-grown cells, whereas the P_ACT1 region could not. In contrast, only the P_TRM1 region could be amplified using the template DNA purified from glucose-grown cells. This indicated the specific association of Trm1p to the promoter regions of these five methanol-inducible genes on methanol, but no association on glucose. Trm1p also bound to the P_TRM1 region, which suggests a self-regulation mechanism.

DISCUSSION

In C. boidinii, full expression under methanol induction is exerted through two activation steps, glucose derepression and methanol-specific induction. Furthermore, both activation steps are repressed by the presence of glucose. In S. cerevisiae, Adr1p has been well characterized as a major regulator of glucose derepression. In the methylotrophic yeasts, homologues of Adr1p, e.g., Mxr1p in P. pastoris, play an important role in glucose derepression under methanol-induced conditions. However, the protein involved in methanol-specific induction has not been identified so far.

In this study, using the gene-tagging mutagenesis in C. boidinii (24), we identified a novel gene, TRM1, as a putative regulator of methanol-specific induction. Trm1p belongs to a Zn(II)₂Cys₆-type zinc cluster protein family, known as a transcriptional regulator in many fungi. The trm1Δ strain lost the ability to grow on methanol but showed normal growth on glucose or other carbon sources, together with decreased levels of transcriptional activity in several methanol-inducible genes. Furthermore, by ChIP assay, we showed that Trm1p bound to five methanol-inducible promoters when cells are grown on methanol but did not bind when cells are grown on glucose. Judging from our results, Trm1p plays a major role in transcriptional activation of methanol-specific induction but not in glucose derepression.

In the trm1Δ strain, the transcriptional activities of methanol-inducible promoters were drastically decreased (Table 2). Particularly, the transcriptional activities of P_DAS1, P_FDH1, P_PMP20, and P_PMP47 were reduced almost to their basal levels in methanol growth conditions. The transcriptional activity of P_DAS1 in the trm1Δ strain grown on methanol was retained. This may be derived from a minor derepression effect. P_AOD1 activity was retained at approximately 6.55% of the wild-type strain. This suggested that in addition to Trm1p, other activators that regulate the P_AOD1 transcriptional activity also exist. The regulation of P_AOD1 is distinct from other methanol-inducible promoters in C. boidinii in that the AOD1 gene showed a derepressed level of gene induction in glycerol-containing medium, ∼20%, compared to the level of gene induction in methanol-containing medium. However, derepression of other promoters is almost negligible (34). This indicates that P_AOD1 is regulated not only by methanol-specific induction but also by glucose derepression.

In H. polymorpha, cells lacking the MPP1 gene generally contain a single peroxisome per cell. This phenotype is different from the C. boidinii trm1Δ strain. The trm1Δ strain showed the normal size and number of peroxisomes in oleate-containing medium. In methanol-containing medium, the number of peroxisomes is normal, although the size of peroxisomes is smaller than the that of the wild-type strain because of a reduced level of peroxisomal protein (Fig. 5). These observations strongly indicate that the TRM1 is not involved in peroxisome assembly and transport of peroxisomal proteins. The fact that the growth of the trm1Δ strain was normal on medium containing oleate also supports this conclusion. On the other hand, the P. pastoris mxr1Δ strain could not grow on either methanol or oleate. Thus, our results indicate that Trm1p is a transcriptional regulator which is specific to the methanol-specific induction.

Deletion analysis of P_DAS1 revealed that two methanol response elements exist. Of the two MREs, MRE1 exhibited strong response to methanol compared to MRE2. To date, several consensus sequences have been reported to be conserved in promoter regions of methanol-metabolizing enzyme genes from P. pastoris, H. polymorpha, and C. boidinii (7, 12). Recently, the sequence responsible for methanol or both methanol and formate in C. boidnii P_FDH1 were identified (6). However, both MRE1 and MRE2 are not similar to these previously reported sequences. Furthermore, MRE1 does not have any similar sequences in the five methanol-inducible promoters we studied in C. boidinii. In contrast, sequences similar to MRE2 [TGGCTNTTT(T/G)A] exist in the other four methanol-inducible promoters we examined (data not shown). Although Trm1p was shown to associate with promoter regions of methanol-inducible genes (Fig. 8) and Trm1p is required for MRE1-dependent methanol induction (Table 4), the fact that conserved sequences similar to MRE1 are not present among these promoters remains a mystery.

The deduced amino acid sequence of Trm1p contained a Zn(II)₂Cys₆-type zinc cluster. The Zn(II)₂Cys₆-type zinc cluster proteins often contain an acidic domain near the C terminus. The acidic domain acts as a transcriptional activation domain (25). In Trm1p, deletion of the acidic region reduced transcriptional activity of P_DAS1 but did not affect the binding ability. This fact implies that the acidic region of Trm1p acts as a transcriptional activation domain. In many cases, fungal Zn(II)₂Cys₆ proteins are known to act as homodimers or heterodimers. In S. cerevisiae, the consensus sequence that Zn(II)₂Cys₆-type zinc cluster proteins bind in their target promoters is a CGG or CCG triplet (11, 25, 28). The MRE1 contains such sequences; therefore, we tried to determine whether Trm1p binds to the MRE1 sequence. His₆-tagged Trm1p was overexpressed under the control of P_DAS1 in medium containing methanol and was purified by passage through a Ni²⁺ Sepharose affinity column. However, the migration of purified His₆-Trm1p was not retarded in a band shift assay using MRE1 as a probe (data not shown). This indicated that Trm1p by itself does not bind to the MRE1 sequence in vitro. We speculate that another partner exists and that it is required for binding to MRE1, while other possibilities e.g., activation by protein modification, can be considered. Most Zn(II)₂Cys₆ zinc cluster proteins form homodimers or heterodimers through their coiled-coil regions. For instance, Oaf1p and Pip2p, two Zn(II)₂Cys₆-type zinc cluster transcriptional activators that regulate oleate-inducible genes in S. cerevisiae, are not able to bind the oleate response element by themselves; Oaf1p-Pip2p heterodimer formation is required for oleate response element binding and oleate induction (5, 26). The heterodimeric binding partners of Zn(II)₂Cys₆-type zinc cluster are often other proteins of the same type, although some exceptions are reported (1). In this context, a homologue of H. polymorpha Mpp1p is one possible partner of Trm1p. In H. polymorpha, Mpp1p is known to be a Zn(II)₂Cys₆-type zinc cluster protein, which was shown to regulate the synthesis of peroxisomal proteins and peroxins, such as DAS, AOD, Pex3p, Pex5p, and Pex10p (8). However, since Mpp1p homologues in C. boidinii have not yet been found, it is unclear whether Trm1p and Mpp1p form a heterodimer. However, judging from our results, Trm1p should be localized in a transcriptional regulatory complex. We are now searching other components of this gene activation complex.