Protein Arginine Methylation in Candida albicans: Role in Nuclear Transport

The yeast strains used in this study are listed in Table 1. Oligonucleotides used in plasmid and strain construction were synthesized at Integrated DNA Technologies, Inc., and are shown in Table 2. All S. cerevisiae strains were grown and genetic manipulations performed as previously described (26, 27, 34). C. albicans strains were grown in YPD medium or in synthetic dropout (SD) media lacking appropriate supplements to select for integrated markers (34); for Uri⁻ strains, YPD and SD media were supplemented with 80 μg/ml uridine. Generation times for C. albicans strains were determined by diluting overnight cultures to an optical density at 600 nm (OD₆₀₀) of 0.1 in YPD and monitoring the OD₆₀₀ every 30 min. For each strain, the log of normalized OD₆₀₀ values was graphed as a function of time, and the linear part of the graph was used to calculate generation times.

All plasmids used in this study are listed in Table 3. Plasmids expressing the C. albicans HMT1 and Homo sapiens PRMT1 genes under the control of S. cerevisiae HMT1 regulatory regions were constructed as follows. C. albicans HMT1 was amplified by PCR from CAI-1 genomic DNA (a gift of R. Wheeler) using oligonucleotides AM80 and AM81. H. sapiens PRMT1 was amplified from pPS1302 using oligonucleotides AM79 and AM82. PCR products were digested with NdeI and NsiI and inserted into NdeI-NsiI-digested pPS1872 to produce pAM160 (CaHMT1) and pAM161 (HsPRMT1), which express arginine methyltransferases with an additional MH dipeptide from S. cerevisiae Hmt1 at the C terminus.

The protein A (PrA)-ScHMT1 expression plasmid pAM91 was constructed by inserting an NdeI-MscI fragment from pPS1872 into NdeI-SmaI-digested pNOPPATA. The URA3 CaHMT1 plasmid pAM390 was constructed by inserting an XbaI-HindIII fragment from pAM160 into XbaI-HindIII-digested pRS316. The C. albicans HMT1 gene and surrounding regions were cloned by PCR amplification of BWP17 genomic DNA with oligonucleotides AM113 and AM114, XhoI and SpeI digestion, and insertion into XhoI-SpeI-digested pRS315, resulting in pAM322. The C. albicans HMT1 reconstitution plasmid (pAM385) resulted from the insertion of an XhoI-BsrBI fragment of pAM322 into XhoI-SmaI-digested pDS10 (30) followed by the elimination of an internal NcoI site by QuikChange mutagenesis (Stratagene) using oligonucleotides AM203 and AM204, resulting in a silent C-to-A mutation at the Ala196 codon.

The C. albicans NPL3 open reading frame (ORF), as defined in assembly 19 of the C. albicans genome sequence, was amplified from BWP17 genomic DNA with oligonucleotides AM143 and AM144, digested with NdeI and BamHI, and inserted into NdeI-BamHI-digested pNOPPATA to create the PrA-CaNPL3 plasmid pAM361. Homologous recombination in S. cerevisiae was used to create a gene fusion between the first three domains of S. cerevisiae NPL3 and the C. albicans NPL3 RGG domain. First, an ApaI site was introduced into the PrA-ScNPL3 plasmid pPS2389 through a silent G-to-C mutation in the Gly312 codon using QuikChange mutagenesis and oligonucleotides AM135 and AM136, resulting in pAM463. A PCR fragment containing the C. albicans NPL3 RGG domain flanked by sequences upstream and downstream of the S. cerevisiae NPL3 RGG domain was amplified using oligonucleotides AM141 and AM142. This fragment was cotransformed into FY23 with ApaI-linearized pAM463, and plasmids were rescued from cells that grew on medium lacking leucine. After transformation into Escherichia coli and purification, plasmids were sequenced at the University of Maine, Orono, DNA sequencing facility to verify proper fusion (ScNPL3 codons 1 to 278-CaNPL3 codons 231 to 336-CCC [proline codon]-ScNPL3 codons 404 to 414).

A PCR-based homologous recombination strategy was used to create C. albicans strains lacking methyltransferase genes according to established protocols (45). PCR amplification of the HIS1 (from pGEM-HIS1) and ARG4 (from pRSARG4ΔSpeI) genes was performed with oligonucleotides AM145 and AM146 (HMT1) or AM147 and AM148 (RMT2). Sequential transformation of BWP17 with HIS1 and ARG4 PCR products to create homozygous deletion mutants was followed by reintroduction of the URA3 gene at its endogenous locus by transformation of PstI-NotI-digested pBSK-URA (30), resulting in AMC11 and AMC30, respectively. HMT1 was reintroduced into the hmt1Δ/hmt1Δ strain AMC35 by transformation of NcoI-linearized pAM385 and selection on medium lacking uridine. The URA3 marker was removed by selection on 5-fluoroorotic acid (5-FOA) and replaced at its endogenous locus as described above, to create AMC14. The RMT2 3′ sequences available in assembly 19 of the C. albicans genome sequence at the time of strain construction did not include the stop codon. To extend these sequences by inverse PCR, BWP17 genomic DNA was digested with TaqI, ligated, and digested with NsiI, and RMT2 3′ sequences were amplified using oligonucleotides AM117 and AM118. Sequencing with AM117 added 111 nucleotides to the assembly 19 RMT2 sequence but did not reveal the stop codon. Although RMT2 shows “complex sequence changes in assembly 20” due to the use of a different reference strain to fill in gaps (1), these 111 nucleotides are identical in our inverse PCR data, generated using BWP17 (a third-generation derivative of the original reference strain SC5314 [45]) and in those of assembly 20. Given these considerations, only the first 369 of 412 known codons were deleted from RMT2, and URA3, but not RMT2, was reintroduced into the rmt2Δ/rmt2Δ strain (AMC30). The heterozygous strain AMC28 serves as a positive control strain that should express Rmt2.

At each step, integration was tested by PCR using the following primer combinations: AM124/AM125 (HMT1), AM124/AM154 (hmt1Δ::HIS1), AM113/AM131 (hmt1Δ::ARG4), AM167/AM170 (RMT2), AM154/AM167 (rmt2Δ::HIS1), AM208/AM131 (rmt2Δ::ARG4), and AM213/AM214 (URA3). Final strains were also tested by Southern blotting: HMT1 alleles were detected by probing DraIII-digested genomic DNA with nucleotides −622 to +36 to the HMT1 ORF; RMT2 alleles were detected by probing HindIII/XbaI-digested genomic DNA with a 1-kb fragment PCR fragment generated from BWP17 genomic DNA using primers AM170 and AM199.

NPL3 was tagged with the green fluorescent protein (GFP) gene in HMT1/HMT1 and hmt1Δ/hmt1Δ C. albicans strains using established protocols (19). The GFP-URA3 cassette was amplified from pMG1602 with oligonucleotides AM173 and AM247 and products transformed into BWP17 and AMC36. Strains that grew on medium lacking uridine were tested for Npl3-GFP expression by anti-GFP immunoblotting and fluorescence microscopy. To introduce the S340A mutation and GFP tag into NPL3 simultaneously, oligonucleotides AM173 and AM244 were used to amplify the GFP-URA3 cassette from pMG1602 prior to transformation. The presence of the mutation and the sequence of the NPL3-GFP junction were confirmed by PCR amplification with oligonucleotides AM143 and AM258 and sequencing.

Protein expression in S. cerevisiae and C. albicans was tested by lysing mid-log-phase cells with glass beads in radioimmunoprecipitation buffer supplemented with protease inhibitors, using a FastPrep cell disruptor (Bio101), followed by immunoblot analysis as described previously (27). CaNpl3-GFP methylation was tested by similarly lysing mid-log-phase cells in lysis buffer (150 mM KCl, 5 mM MgCl₂, 20 mM Tris-HCl, pH 8.0) containing protease inhibitors and 0.2% Triton X-100. CaNpl3-GFP was precipitated from lysates (3 mg total protein) by incubation with anti-GFP (Roche; 5 μl) followed by incubation with protein G-agarose (Santa Cruz, sc-2002; 5 μl). After washing, CaNpl3-GFP was eluted by boiling in Laemmli buffer and analyzed by immunoblotting with anti-GFP and anti-ADMA (ASYM24; Upstate) antibodies.

PrA pulldown assays were performed essentially as described previously (26). Briefly, mid-log-phase cells expressing PrA fusion proteins were lysed in the lysis buffer described above with either 1% (Npl3) or 0.1% (Hmt1) Triton X-100. PrA fusion proteins and interacting proteins were precipitated with immunoglobulin G (IgG)-Sepharose (Pharmacia), washed, eluted with 3 M MgCl₂, precipitated with trichloroacetic acid, and analyzed by immunoblotting.

For immunoblot analyses, antibodies were used at the following dilutions: polyclonal anti-ScNpl3, 1:5,000 (7); monoclonal 1E4 (anti-methyl-ScNpl3) (38, 46), 1:2,500; polyclonal anti-ScHmt1 (27), 1:2,000; polyclonal anti-myc, 1:2,000 (sc-789; Santa Cruz Biotechnology); monoclonal anti-GFP (Roche), 1:1,000; and secondary anti-mouse-horseradish peroxidase and anti-rabbit-horseradish peroxidase, 1:5,000 (Amersham). Polyclonal anti-Hmt1 and anti-myc antisera also detect PrA fusion proteins. All proteins were visualized by enhanced chemiluminescence (Amersham), and migration was compared to that of Benchmark standards (Invitrogen).

C. albicans strains were grown at 30°C in YPD medium as previously described by Porras-Yakushi et al. (33), with the following changes. The strains were harvested by centrifugation at 1,000 × g for 3 min at 4°C, and the pellets were washed a total of three times by the addition of 1 ml of sterile water followed by centrifugation for 5 min at 10,600 × g at room temperature.

Fractionation by sodium dodecyl sulfate (SDS) gel electrophoresis was carried out as previously described (33). For fluorography, the gels were treated with En³Hance (Perkin-Elmer Life Sciences) for 1 h and then washed for 20 min in water. Gels were dried in vacuo for a total of 3 h (2 h at 70°C and 1 h at room temperature).

Analysis was performed as previously described by Niewmierzycka and Clarke (29) with the following modifications. After precipitation with trichloroacetic acid, lysates were centrifuged at 1,000 × g for 20 min at 25°C. The pellets were washed once with 50 μl of cold acetone, centrifuged at 1,000 × g for 20 min at 25°C, and allowed to dry prior to being acid hydrolyzed. Hydrolysates were mixed with nonradiolabeled standards (50 μl of 2 mM SDMA and 5 μl each of 0.2 M ADMA and 0.2 M ω-MMA [Sigma Chemical Co., St. Louis, MO]) prior to being loaded on a Beckman AA-15 sulfonated polystyrene cation-exchange column (29).

Cells expressing GFP-tagged CaNpl3 proteins were grown to mid-log phase in synthetic dropout medium lacking uridine, washed with phosphate-buffered saline (PBS), and incubated with 10 μg/ml DAPI (4′,6′-diamidino-2-phenylindole) in PBS for 30 min to visualize nuclei. Cells were washed with PBS and visualized by fluorescence microscopy (Olympus BX51; GFP and DAPI filters). Images were captured with an EvolutionVF color digital camera (noncooled, 12-bit; MediaCybernetics) and QCapture Pro 5.0 software. For each filter, exposure times were equivalent for all strains.

A BLAST search of nonredundant databases with the major type I arginine methyltransferase from S. cerevisiae reveals a clear C. albicans ortholog. These proteins not only are very similar among numerous fungal species (Fig. 1A and data not shown), but also share a high degree of similarity with the predominant human PRMT, PRMT1 (Fig. 1A) (35, 41). To test whether PRMT function is conserved between these two fungal species, C. albicans HMT1 (CaHMT1) was subcloned into a plasmid under the control of the endogenous S. cerevisiae HMT1 (ScHMT1) promoter (official gene names consist of three letters and a number, and therefore the Ca and Sc prefixes are used to distinguish orthologous genes from the two organisms). Although HMT1 is not an essential gene in S. cerevisiae, Hmt1 is required in a strain that contains a point mutation in one of its substrates, the major mRNA-binding protein Npl3 (npl3-1) (24), or that lacks the 80-kDa mRNA cap-binding protein Cbp80 (cbp80Δ) (37). Plasmids that express S. cerevisiae, C. albicans, and human PRMTs were transformed into these S. cerevisiae strains. Cells expressing CaHmt1 grow as robustly as cells expressing ScHmt1 (Fig. 1B). Thus, C. albicans Hmt1 supports all essential Hmt1 functions in these S. cerevisiae strains. In contrast, expression of human PRMT1 allowed partial growth in the absence of Cbp80 but did not support growth of the npl3-1 strain (Fig. 1B).

To test methyltransferase activity, these PRMT expression plasmids were transformed into S. cerevisiae cells lacking endogenous HMT1 (hmt1Δ), and levels of Hmt1, Npl3, and methylated Npl3 were determined by immunoblotting (Fig. 1C). The antiserum raised against ScHmt1 (27) also recognizes CaHmt1, which migrates slightly faster, but not human PRMT1 (Fig. 1C, top panel); relative expression levels of the three PRMTs cannot be determined due to epitope differences, but all three PRMTs are expressed from the ScHMT1 promoter. Although Npl3 levels are similar in all strains (Fig. 1C, middle panel), levels of methylated Npl3 vary depending on which methyltransferase is expressed. CaHmt1 methylates S. cerevisiae Npl3 almost as well as ScHmt1, whereas Npl3 methylation by the human enzyme is significantly reduced in comparison (Fig. 1C, bottom panel). Thus, C. albicans Hmt1 demonstrates significant methyltransferase activity for a known Hmt1 substrate in S. cerevisiae.

The crystal structures of S. cerevisiae Hmt1 and human PRMT1 reveal PRMT dimerization, which is important for enzyme function (44, 51). To determine whether C. albicans Hmt1 showed similar protein-protein interactions, heterodimerization of CaHmt1 and ScHmt1 proteins was tested. CaHmt1 was coexpressed with a PrA-tagged form of ScHmt1 in hmt1Δ S. cerevisiae cells. After PrA-ScHmt1 isolation, untagged Hmt1 proteins associated with PrA-ScHmt1 were detected by immunoblotting (Fig. 1D). Both ScHmt1 and CaHmt1 copurify with PrA-ScHmt1, supporting cross-species Hmt1 hetereodimerization (lane 3, top panel) as well as homodimerization of ScHmt1 (lane 2). The presence of a ∼40-kDa copurifying protein depends on the presence of untagged Hmt1 (lanes 1 and 2), the presence of the Hmt1 dimerization domain (lanes 2 and 4) (44) and the expression of PrA-ScHmt1 (lanes 2 and 5). The lower band in the C. albicans lane is unlikely to represent a degradation product of PrA-ScHmt1, since the protein in this lane migrates slightly faster than the minor band seen in the absence of untagged Hmt1 expression (lane 1). The ability of C. albicans Hmt1 to bind to S. cerevisiae Hmt1, combined with the conservation of regions involved in ScHmt1 dimerization (Fig. 1A) suggests that the C. albicans enzyme may dimerize in vivo.

To determine whether HMT1 and RMT2 genes encode functional PRMTs in C. albicans, strains with homozygous deletions of each gene were created by homologous recombination. The entire HMT1 gene was removed in hmt1Δ/hmt1Δ strains, and over 89% of RMT2 was removed in rmt2Δ/rmt2Δ strains, including all methionine codons and the entire AdoMet-binding motif. Immunoblotting with the antibody raised against ScHmt1 revealed the presence of a ∼39-kDa protein in the reconstituted strain that was absent in hmt1Δ/hmt1Δ cells (Fig. 2A). All hmt1Δ/hmt1Δ and rmt2Δ/rmt2Δ strains grew as robustly as reconstituted or heterozygous strains in rich media. The generation times of two hmt1Δ/hmt1Δ strains were 1.35 ± 0.08 h and 1.36 ± 0.02 h, compared to 1.36 ± 0.07 h and 1.32 ± 0.05 h for two reconstituted hmt1Δ/hmt1Δ+HMT1 strains. Generation times of two rmt2Δ/rmt2Δ strains were 1.37 ± 0.09 h and 1.34 ± 0.06 h, compared to 1.32 ± 0.07 h for the rmt2Δ/RMT2 heterozygous strain. Therefore, similar to the case for the S. cerevisiae orthologs (18, 24, 29), neither HMT1 nor RMT2 is essential for C. albicans growth.

To test PRMT activity in vivo, C. albicans cells with and without each methyltransferase were grown in the presence of ³H-AdoMet. Radiolabeled proteins were analyzed by SDS-polyacrylamide gel electrophoresis and fluorography and acid hydrolyzed for amino acid analysis to identify methylarginine species. Radioactive labeling of multiple proteins was lower in lysates of hmt1Δ/hmt1 cells than in those of hmt1Δ/hmt1Δ+HMT1 cells (Fig. 2B, right panel), suggesting lower levels of arginine methylation of these substrates in the absence of Hmt1. The radioactive ADMA peak visible in acid hydrolysates of proteins from the hmt1Δ/hmt1Δ+HMT1 strain (Fig. 2C, upper panel) is absent in hmt1Δ/hmt1Δ acid hydrolysates; the ω-MMA peak is also significantly reduced in hmt1Δ/hmt1Δ cells (Fig. 2C, lower panel). These results indicate that, as in S. cerevisiae (18), Hmt1 is the major type I arginine methyltransferase in C. albicans.

In contrast, levels of radioactively labeled proteins are similar in rmt2Δ/RMT2 and rmt2Δ/rmt2Δ cells (Fig. 3A, right panel). This result is consistent with the presence of only one known Rmt2 substrate in S. cerevisiae: Rmt2 monomethylates ribosomal protein L12 (Rpl12) in S. cerevisiae at a single site (10), and CaRpl12 may comigrate with strongly labeled proteins between the 14.4- and 21.5-kDa markers. In amino acid analysis chromatography, δ-MMA migrates similarly to SDMA (29), and, as predicted, a small peak that elutes close to the SDMA standard in the rmt2Δ/RMT2 and HMT1 samples is reduced in rmt2Δ/rmt2Δ hydrolysates (Fig. 3B). These results are consistent with Rmt2 catalyzing δ-MMA formation in C. albicans. The presence of a small amount of radioactivity in the δ-MMA position suggests that an additional methyltransferase activity may be present that forms symmetric dimethylarginine residues, although no gene that may encode this activity has been identified.

The first cellular function demonstrated for arginine methylation in vivo was a role in nuclear export of two Hmt1 substrates in S. cerevisiae, including the mRNA-binding protein Npl3 (37). Subsequent studies demonstrated that methylation facilitates nuclear import of human RNA helicase A (40) and that PRMT1 deletion increased cytoplasmic localization of mammalian RNA-binding protein Sam68 (12), also suggesting a role for methylation in import. We therefore wished to test whether arginine methylation has a role in either nuclear export or import of methylated RNA-binding proteins in C. albicans. A putative Npl3 ortholog was identified in the C. albicans genome (Fig. 4A). Npl3 is not as highly conserved between fungal species as is Hmt1, but S. cerevisiae Npl3 and its C. albicans ortholog both contain two RNA recognition motifs and a C-terminal domain with many arginine-glycine-glycine (RGG) tripeptides that are extensively methylated by Hmt1 in S. cerevisiae (26). The two proteins also share a heptapeptide at the C terminus that includes a serine targeted by the SR protein kinase in S. cerevisiae, Sky1 (20, 50).

To test whether Npl3 was functionally conserved between fungal species, we subcloned full-length C. albicans NPL3 (CaNPL3) into an S. cerevisiae PrA expression vector. A second plasmid expresses a PrA-tagged chimeric protein containing the first three domains of ScNpl3, the RGG domain of CaNpl3, and the last 11 amino acids of ScNpl3. These plasmids and control plasmids were transformed into npl3Δ S. cerevisiae cells bearing a URA3 plasmid for ScNpl3 expression. After growth on 5-FOA medium, cells expressing the chimeric protein grew as well as those expressing ScNpl3, but cells expressing full-length CaNpl3 grew only slightly better than cells bearing the expression vector (Fig. 4B). Therefore, although all proteins are expressed at similar levels (Fig. 4C), the RGG domain of CaNpl3 can replace essential functions of the ScNpl3 RGG domain, but the full-length C. albicans protein cannot perform functions essential for growth of the npl3Δ cells.

Methylation of ScNpl3 disrupts protein-protein interactions, including Npl3 self-association (49). The RGG domain helps direct ScNpl3 self-association, since PrA-ScNpl3 can interact with the RGG domain alone in hmt1Δ cells (A. McBride, unpublished data). Therefore, we tested whether the CaNpl3 RGG domain could mediate methylation-sensitive Npl3-Npl3 interaction in S. cerevisiae (Fig. 4C). The PrA-Npl3 plasmids were transformed into Δhmt1 and HMT1 cells expressing myc-tagged ScNpl3 and PrA-Npl3, and associated proteins were isolated with IgG-Sepharose. Whereas the chimeric Npl3 showed slightly less interaction than ScNpl3 with ScNpl3-myc, binding of full-length CaNpl3 to ScNpl3-myc was severely reduced, suggesting that domains other than the RGG domain influence ScNpl3 self-association. In all cases, however, Npl3-Npl3 binding is decreased in the presence of HMT1 (Fig. 4C).

The methylation sensitivity of Npl3 dimerization in S. cerevisiae has been linked to the role of arginine methylation in ScNpl3 export (26). Given the increased interaction of Npl3 with proteins containing the CaNpl3 RGG domain in hmt1Δ S. cerevisiae cells, we tested whether Npl3 is methylated in C. albicans and whether Hmt1 influences its nuclear transport. To detect Npl3 methylation in C. albicans, a GFP tag was inserted at the C terminus of one of the endogenous NPL3 alleles. Npl3-GFP immunoprecipitated from cells with or without Hmt1 was detected with anti-GFP antibodies to detect relative protein levels and with an antidimethylarginine antiserum to determine relative methylation levels (Fig. 5A). The methylation-specific antiserum only detected Npl3-GFP proteins in cells expressing HMT1. Immunoprecipitates from cells lacking the GFP tag or expressing the GFP tag alone do not contain any proteins recognized by the dimethylarginine antiserum. Therefore, Npl3 is an Hmt1 substrate in C. albicans.

Fluorescence microscopy revealed that wild-type Npl3-GFP colocalizes with DAPI staining and is therefore nuclear in C. albicans in the presence or absence of Hmt1 (Fig. 5B). Although this result indicates that Hmt1 is unlikely to have a role in expediting nuclear import, inhibitory effects of HMT1 deletion on nuclear export cannot be determined in these cells. In S. cerevisiae, nuclear export defects have been probed using a temperature-sensitive nucleoporin mutation that slows Npl3 import (25, 37). The diploidy of C. albicans and paucity of selectable markers led us to devise another strategy to slow Npl3 nuclear import to test the effect of Hmt1 on Npl3 export in C. albicans. An S-to-A mutation in the serine that is phosphorylated by Sky1 decreases Npl3 import in S. cerevisiae (20, 50). The amino acids surrounding this serine in the C terminus of Npl3 are conserved in C. albicans (Fig. 4A), which also contains orthologs of Sky1 and Mtr10, the Npl3 import receptor (2, 36). Hmt1 methylates npl3-S340A in C. albicans (Fig. 5A), and npl3-S340A localizes throughout HMT1 cells at steady state (Fig. 5B), indicating that the rate of Npl3 import relative to export is decreased by this mutation. When this mutation was introduced into hmt1Δ/hmt1Δ cells, however, npl3-SA-GFP showed nuclear localization similar to that of wild-type Npl3-GFP (Fig. 5b). This result demonstrates that Hmt1 deletion lowers the rate of npl3-SA-GFP export relative to that of import, suggesting that Hmt1 facilitates nuclear export of Npl3 in C. albicans.