ABSTRACT
The methylotrophic yeast Pichia pastoris has been utilized for heterologous protein expression for over 30 years. Because P. pastoris secretes few of its own proteins, the exported recombinant protein is the major polypeptide in the extracellular medium, making purification relatively easy. Unfortunately, some recombinant proteins intended for secretion are retained within the cell. A mutant strain isolated in our laboratory, containing a disruption of the BGS13 gene, displayed elevated levels of secretion for a variety of reporter proteins. The Bgs13 peptide (Bgs13p) is similar to the Saccharomyces cerevisiae protein kinase C 1 protein (Pkc1p), but its specific mode of action is currently unclear. To illuminate differences in the secretion mechanism between the wild-type (wt) strain and the bgs13 strain, we determined that the disrupted bgs13 gene expressed a truncated protein that had reduced protein kinase C activity and a different location in the cell, compared to the wt protein. Because the Pkc1p of baker’s yeast plays a significant role in cell wall integrity, we investigated the sensitivity of the mutant strain’s cell wall to growth antagonists and extraction by dithiothreitol, determining that the bgs13 strain cell wall suffered from inherent structural problems although its porosity was normal. A proteomic investigation of the bgs13 strain secretome and cell wall-extracted peptides demonstrated that, compared to its wt parent, the bgs13 strain also displayed increased release of an array of normally secreted, endogenous proteins, as well as endoplasmic reticulum-resident chaperone proteins, suggesting that Bgs13p helps regulate the unfolded protein response and protein sorting on a global scale.
IMPORTANCE The yeast Pichia pastoris is used as a host system for the expression of recombinant proteins. Many of these products, including antibodies, vaccine antigens, and therapeutic proteins such as insulin, are currently on the market or in late stages of development. However, one major weakness is that sometimes these proteins are not secreted from the yeast cell efficiently, which impedes and raises the cost of purification of these vital proteins. Our laboratory has isolated a mutant strain of Pichia pastoris that shows enhanced secretion of many proteins. The mutant produces a modified version of Bgs13p. Our goal is to understand how the change in the Bgs13p function leads to improved secretion. Once the Bgs13p mechanism is illuminated, we should be able to apply this understanding to engineer new P. pastoris strains that efficiently produce and secrete life-saving recombinant proteins, providing medical and economic benefits.
INTRODUCTION
The methylotrophic yeast Pichia pastoris (also called Komagataella pastoris) is a popular host for recombinant protein expression. The yeast has been genetically engineered to express over 5,000 heterologous proteins and has been valued for industrial, pharmaceutical, and basic research purposes for the past 25 years (1, 2). In fact, more than 70 P. pastoris products, ranging from anticancer therapeutics to vaccine antigens, are commercially available or in late-stage development (phase 2 or phase 3 clinical trials) (3, 4). Because P. pastoris secretes few of its own proteins, the expressed recombinant protein is usually the major polypeptide species found in the extracellular medium (ECM). Thus, programmed export acts as a valuable step in the purification of the heterologous protein and is considered a strong asset of this microbial host. Unfortunately, one of the greatest weaknesses of the system is that some recombinant proteins that are engineered to be secreted are retained or degraded inside the cell (5, 6). Furthermore, these heterologous proteins may lack the correct posttranslational modifications and folding of the native proteins, which may cause problems, including reduced activity and triggering of an immune response if injected into the human body.
To illuminate the mechanism of P. pastoris secretion and to help alleviate the problem of inefficient protein export, we identified 13 β-galactosidase supersecretion (BGS) genes through a selection of genomic disruption mutants generated by using a restriction enzyme-mediated integration (REMI) plasmid (7). In these selected mutants, the pREMI plasmid was inserted into a chromosomal gene whose modified function led to an increase in β-galactosidase secretion. The enhanced secretion of this enzyme enabled the mutated cells to metabolize lactose, allowing the mutants to be selected on lactose-containing medium. The affected genes appear to have functions in intracellular signaling or vesicle transport. One particular mutant strain, the bgs13 strain, showed enhanced secretion of the majority of recombinant proteins tested. Because most supersecreting mutant strains identified by other laboratories demonstrated enhanced secretion of only a small subset of peptides (8), the bgs13 mutant raised the intriguing possibility of being a universal supersecreter.
The Bgs13 polypeptide (Bgs13p) shares similarities with protein kinase C (PKC) 1 protein (Pkc1p) from the yeast Saccharomyces cerevisiae. The predicted amino acid sequence of wild-type (wt) P. pastoris Bgs13p is 50% identical and 68% similar to that of S. cerevisiae Pkc1p in the N-terminal region. In baker’s yeast, Pkc1p plays a critical role in the cell wall integrity (CWI) pathway, which is responsible for detecting and responding to cell wall stress that arises under normal growth conditions or through environmental challenges (9, 10). The CWI pathway in S. cerevisiae relies on a family of cell surface sensors coupled to a small G protein, Rho1p, which is considered to be the master regulator of CWI signaling not only because it senses signals from the cell surface but also because it influences a variety of mechanisms involved in cell wall biogenesis, actin organization, and polarized secretion (10). The transcriptional output of the CWI pathway is under the control of a mitogen-activated protein kinase (MAPK) cascade dependent on Pkc1p, which interacts with Rho1p. Pkc1p induces the activity of the MAPK cascade by phosphorylating Bck1, which stimulates a redundant pair of kinases, Mkk1 and Mkk2, that finally regulate Slt2/Mpk1. Ultimately, Slt2/Mpk1 interacts with Rlm1p to regulate transcriptional responses of cell wall genes on an intricate, global scale; some responses are positive, while others are negative (11). Like the Δpkc1 strain of baker’s yeast, the bgs13 strain of P. pastoris released more alkaline phosphatase (a vacuolar enzyme) than its wt parent, suggesting a more permeable cell wall than normal (7). For both species, alkaline phosphatase export was inhibited by the addition of sorbitol to the medium, strongly suggesting that CWI was compromised by the variant bgs13p function. Interestingly, in S. cerevisiae, the loss of PKC1 resulted in a more severe growth defect than that produced by deletion of any of the members of the MAPK cascade under the control of this kinase, which suggests that Pkc1p regulates additional steps along the CWI pathway (10).
Although several lines of evidence point to modification of the cell wall as the cause of enhanced secretion in the bgs13 strain, some previous results suggest that other factors may be involved (7). For example, unlike the S. cerevisiae Δpkc1 strain, bgs13 strain cells did not require the presence of 1 M sorbitol in the medium for growth. This result raised the question of whether pREMI insertion into the BGS13 locus caused only a partial loss of function. In addition, other experimental results implicated Bgs13p in the protein-sorting process. For instance, studies using the α-mating factor secretion signal sequence fused to an N-terminal maltose-binding protein (MBP) and a C-terminal enhanced green fluorescent protein (EGFP) indicated that the hybrid protein was proteolyzed at the linker region between the two domains prior to secretion. Compared to the wt strain, the bgs13 mutant increased the secretion of MBP-EGFP while decreasing its proteolysis in the linker region between the two domains (12). Furthermore, fluorescence microscopy indicated that, in the wt background, MBP-EGFP gave rise to green fluorescence mainly in the vacuole, while the bgs13 allele caused MBP-EGFP to be shunted from the vacuole and localized to other parts of the secretory network prior to its export into the ECM. The suppression of vacuolar targeting has been linked to increased secretion in baker’s yeast and P. pastoris (13, 14). Thus, it is not clear whether changes in cell wall structure are (i) the cause of increased secretion of proteins or (ii) an effect of bgs13 mutant protein activity that is not directly responsible for enhanced peptide export.
To help resolve this question, we have focused on characterizing the genetic nature of the chromosomal bgs13 mutation, determining whether it yields complete or partial loss of function, and illuminating how this mutation affects the localization and kinase activity of its gene product. In addition, we have examined how the mutation affects the population of secreted proteins, as well as the constituents of the cell wall. Taken together, these results provide clues for comprehending how the mutation in the bgs13 strain triggers enhanced protein secretion.
RESULTS AND DISCUSSION
Characterization of mutant bgs13 mRNA.The bgs13 mutant strain was generated by the insertion of the 2.8-kb pREMI plasmid into the locus (7). The bgs13 strain had no obvious phenotypic difference in growth, compared to its wt parent, except that its doubling time was about 50% longer. We initially hypothesized that the mutant P. pastoris bgs13 strain was not a null strain, since, unlike the Δpkc1 strain of baker’s yeast (15), it could grow without osmotic stabilization and appeared to have normal architecture inside its cells, based on transmission electron microscopy analysis (7). Therefore, it was postulated that the insertion of pREMI into the BGS13 locus did not cause a gene knockout but allowed for the expression of a Bgs13p variant. To determine whether BGS13 mRNA was present inside the bgs13 strain cells, we first examined the genomic DNA from the bgs13 stain obtained by a plasmid rescue strategy (7, 16). Genomic DNA 5′ and 3′ of the pREMI plasmid insertion indicated that the vector had integrated within the codon 147 of the BGS13 open reading frame (ORF) (between nucleotide 441 and nucleotide 442) (Fig. 1A and B). We reasoned that either or both of two hybrid mRNAs could be produced, one driven by the BGS13 promoter, containing only the 5′ end of the BGS13 mRNA, with the first 147 codons, and the 3′ sequence from the pREMI plasmid (Fig. 1B, left), and/or one driven by a promoter within the pREMI plasmid, containing 5′ plasmid sequence fused to codons 148 through 1035 of the BGS13 gene at the 3′ end (Fig. 1B, right). We hypothesized that either mRNA could potentially encode a protein with some form of residual Bgs13p activity that would promote supersecretion.
PCR analysis used to determine the structure of the wt and bgs13 mRNA in mutant cells. (A and B) Schematics show the PCR used to determine the regions of the BGS13 sequence present in the wt (A) and bgs13 (B) strain mRNAs. Primers annealing to the 5′ end of the BGS13 cDNA should amplify a 500-bp product, while primers annealing to the 3′ end of the BGS13 cDNA should amplify a 650-bp product. CDS, coding sequence. (C and D) cDNA reactions were performed on total RNA isolated from wt and bgs13 strain cells. +, reaction mixture containing reverse transcriptase; −, reaction mixture lacking reverse transcriptase. Primers annealed to either the 3′ end (C) or the 5′ end (D) of the BGS13 gene. Plasmids containing the entire BGS13 gene or no recombinant gene served as the templates for the positive- and negative-control reactions, respectively. For clarity, duplicate samples and some extraneous controls were spliced out of the image shown in panel C.
To distinguish between these two possibilities, total mRNA from wt and bgs13 strain cells (grown under either glucose or methanol conditions) was reverse transcribed, and the resulting cDNA was subjected to PCR amplification with primer combinations annealing to either the 5′ region or the 3′ region of the BGS13 ORF (Fig. 1A and B). A PCR product of the expected size (approximately 650 bp) was synthesized with primers annealing to the 3′ end of the BGS13 coding sequence using cDNA produced from the wt strain and the bgs13 strain, indicating that mRNA containing the 3′ region of BGS13 was found at significant levels inside the bgs13 strain cells (Fig. 1B, right, and Fig. 1C). A PCR product of approximately 500 bp was synthesized with primers annealing to the 5′ end of the BGS13 coding sequence from cDNA produced by the wt strain but not the bgs13 strain, indicating that the bgs13 strain cells did not contain mRNA containing the 5′ region of BGS13 (Fig. 1B, right, and Fig. 1D). mRNA containing the 5′ BGS13 sequence might have been transcribed in the bgs13 strain cells but most likely was degraded due to an early termination codon created in the fusion of the BGS13 and pREMI sequences, which triggered the up frameshift (UPF) mechanism (17). As negative controls, cDNA reactions that lacked reverse transcriptase did not yield significant amounts of either 5′ or 3′ PCR products, as expected (Fig. 1C and D); the weak bands were most likely amplified from undigested genomic DNA. Thus, the bgs13 strain cells contained a stable mRNA transcribed from the 3′ region of the BGS13 gene.
Cloning of Remi-bgs13 cDNA.Our next step was to determine the sequence of the hybrid mRNA, which was hypothesized to contain the 5′ pREMI sequence fused to the BGS13 sequence 3′ of codon 147. Using a modified 5′-rapid amplification of cDNA ends (5′-RACE) strategy with a commercially available kit, we first reverse transcribed total RNA using a 5′ random primer mix. The reverse transcriptase added several nontemplated residues to the 3′ end of the cDNA products. An oligonucleotide annealing to these nontemplated residues was used by the reverse transcriptase to synthesize the complementary second cDNA strand. A 5′-RACE reaction was then performed with a primer that annealed near nucleotide 1200 of the BGS13 coding sequence and a kit-provided universal primer mix (UPM) that annealed to the nontemplated bases. A smeared 800- to 1,500-bp PCR product, which was seen only in reactions with both primers and not in those with the individual primers, was produced in this first 5′-RACE reaction (Fig. 2, lane 1).
Cloning of bgs13 strain cDNA. (A) Lane 1, product of 5′ RACE reaction; lane 2, 800-bp product of nested PCR after gel isolation; lane 3, 1,100-bp product of nested PCR after gel isolation. (B) Sequence of the Remi-bgs13 cDNA, which contains 12 codons from pREMI (nonitalicized) fused in frame to codon 148 and the remainder of the BGS13 coding sequence (italicized). The start codon is underlined. The arrow indicates the border between pREMI and BGS13 DNA sequence. The predicted protein sequence is shown below the DNA sequence. In the protein sequence, the amino acids encoded by pREMI are nonitalicized, while the residues resulting from BGS13 coding sequence are italicized.
A second nested PCR was then performed on this reaction product, using a primer that annealed near nucleotide 1000 of the BGS13 coding sequence and the UPM. The two PCR products (approximately 800 bp and 1,100 bp) resulting from this amplification reaction (Fig. 2, lanes 2 and 3) were then gel purified and cloned into a sequencing vector. While the 1,100-bp product contained no BGS13 or pREMI sequence, the 800-bp sequence, which represented the Remi-bgs13 hybrid transcript, contained a 5′ untranslated region (UTR) and 12 codons (initiated by an ATG) originating from the pREMI plasmid, fused in frame to codon 148 to the stop codon of the BGS13 coding sequence (Fig. 2). The resultant protein, designated Remi-bgs13p, was presumed to have some Bgs13p activity and to be responsible for the supersecreter phenotype. Similarly to Saccharomyces cerevisiae Pkc1p, P. pastoris Remi-bgs13p, due to the lack of its first 147 amino acids, does not contain a homologous region 1 (HR1) domain (15). In baker’s yeast, the HR1 domain of Pkc1p is associated with (i) binding to Rho1p, the upstream activator of Pkc1p, and (ii) the regulation of oligosaccharyltransferase activity and actin depolarization. However, Remi-bgs13p contains domains such as the C1 domain, which is involved in the control of the transcription of cell wall genes and Rho1p binding, as well as its kinase domain (18).
Creation of a null bgs13 strain.To reveal the effect of inhibiting all Bgs13p activity on secretion, we then endeavored to create a null bgs13 strain. Attempts to knock out the BGS13 gene, using two different strategies, were unsuccessful. First, using the traditional technique, a knockout construct was created with approximately 500 nucleotides of BGS13 regions upstream and downstream of its coding sequence (19, 20). These regions flanked a central selectable marker, Saccharomyces cerevisiae HIS4. His+ transformants of the yDT39 (met2 his4) strain were generated, but PCR analysis of genomic DNA indicated that none of them contained the expected disruption/deletion of the chromosomal BGS13 (data not shown). Recombination at sites lacking homology is dependent on the targeted gene in P. pastoris (21). As a second strategy, a CRISPR technique optimized for P. pastoris (22) was used. Six CRISPR vectors, with a Zeocin- or G418-selectable marker, contained guide RNA (gRNA) sequences that targeted three different stretches of the coding region (positions +816 to +836, +1049 to +1068, and +1069 to +1088) responsible for the N terminus of Bgs13p. Zeocin- and G418-resistant transformants were selected on medium supplemented with 1 M sorbitol in case the Bgs13 knockout resulted in cell wall dysfunction, which is true of the baker’s yeast Δpkc1 strain (15). Genomic DNA from the targeted BGS13 region in 60 randomly selected transformants was amplified by PCR and sequenced. Approximately one-third of the colonies contained genomic mutations within the BGS13 sequence, but all were either 3- or 6-bp deletions in the BGS13 coding region (data not shown). Because none of these deletions resulted in frameshifts, the DNA could encode Bgs13p lacking 1 or 2 amino acids at most, most likely retaining some Bgs13p activity. This inability to create a null mutant strongly suggests that the BGS13 gene is essential for viability. Therefore, we had to proceed in our studies without the use of a BGS13 knockout strain.
Examining the PKC activity of wt Bgs13p and mutant Bgs13p.Our next step was to compare the effects of wt Bgs13p and Remi-bgs13p, to characterize the relationship between their functional activities and their effects on secretion efficiency. Due to its homology to S. cerevisiae PKC1, we hypothesized that the BGS13 gene product would have PKC activity. While a deletion in its N terminus would remove the HR1 domain, the mutant peptide should contain its putative kinase domain. We sought to compare the protein kinase activities of the Bgs13p and Remi-bgs13p.
A commercially available kit detected only very low PKC activity in intracellular extracts from the wt strain, which could not be distinguished from the levels in bgs13 strain cells, helping to establish a baseline of activity. Both wt Bgs13-c-myc-His6 (pKdSC5) and Remi-bgs13-c-myc-His6 (pAH1) were expressed from an AOX1 promoter in wt cells that first were grown on repressing glucose medium and then were induced on methanol growth medium for 24 h. Spot Western blot analysis of extracts demonstrated the presence of the expected c-myc-tagged proteins in each extract (Fig. 3). Overexpression of either Bgs13-c-myc-His6 (pKdSC5) or Remi-bgs13-c-myc-His6 (pAH1) did not affect the growth rate of either strain, compared to a negative-control strain that harbored an empty plasmid (data not shown). For reasons that were not clear, however, we were unable to isolate either wt Bgs13-myc-His6 or Remi-bgs13-c-myc-His6 from these extracts, using both nickel and cobalt affinity columns. Thus, to try to compare specific activity, spot Western blot analysis was used to determine volumes of extracts with equal levels of the wt Bgs13-c-myc-His6 and Remi-bgs13-c-myc-His6 peptides. These volumes were then assayed for PKC activity. Our results indicated that the Remi-bgs13-c-myc-His6 protein had much lower PKC specific activity than did the wt Bgs13-c-myc-His6 peptide (Table 1).
Production of Bgs13p and Remi-bgs13p. Strains expressing the wt Bgs13p, Remi-bgs13p, or no recombinant protein were grown on glucose, induced with methanol, and harvested. Both proteins were expressed with a c-myc epitope. Intracellular extracts were isolated, and equal amounts of protein were used in a spot Western blot analysis. Protein extracts and controls were probed with an anti-c-myc antibody. Spot Western blot images captured with a ChemiImager 5500 were quantified with integrated density value (IDV) analysis. −, wt strain carrying an empty pPICZB plasmid; +, 20 ng of a commercially available, c-myc-tagged, control protein. For clarity, duplicate samples were spliced out of the image.
PKC activity of extractsa
This result led to the initial hypothesis that improved secretion of the bgs13 strain resulted from a decrease in PKC activity. We reasoned that if this were true, then elevated levels of PKC activity would decrease the secretion of a reporter. We chose human secretory leukocyte protease inhibitor protein (SLPI) as the recombinant reporter because it is a secreted human protein, containing eight disulfide bonds, that had been shown previously by our laboratory to be secreted at higher levels by our bgs13 strain than by the wt parent strain (7, 23). Therefore, a strain that expressed SLPI under the control of the AOX1 promoter was transformed with a plasmid that overexpressed the wt Bgs13-myc-His6 peptide (pKdSC5) or an empty plasmid. As shown in Table 2, overexpression of Bgs13p reduced secretion of SLPI by approximately 50%, compared to a strain with normal levels of endogenous Bgs13p. It could be argued that the reduction in SLPI secretion was due to the secretory stress created by the addition of an AOX1 promoter-BGS13 expression construct to a strain that already contained an AOX1 promoter-SLPI expression cassette. However, a previously published report (23) from our laboratory demonstrated that a strain carrying two copies of an AOX1p-SLPI expression cassette secreted approximately twice as much SLPI peptide as a strain harboring a single copy of this expression cassette. This finding supported our hypothesis that the secretion efficiency of a strain may be inversely proportional to its PKC activity.
Secreted levels of SLPI protein in strains producing normal (yAM1-pPICZB) or elevated (yAM1-pKdSC5) levels of Bgs13p
Determining the localization of Bgs13p.Although these experiments suggested that secretion levels were regulated by the PKC activity of Remi-bgs13p, we also wanted to determine whether the localization of Remi-bgs13p played a role in modulating the levels of recombinant protein secretion. Because deletion of the HR1 domains in S. cerevisiae Pkc1p affected its cellular location (15), we attempted to determine whether there was a difference in the cellular locations of Bgs13p and Remi-bgs13p. To pursue this idea, an N-terminal wt Bgs13p or Remi-Bgs13p sequence was fused to a C-terminal EGFP sequence and expressed in wt cells. Fluorescence microscopy was subsequently employed to determine the location of the fusion proteins, using FM 4-64 dye [N-(3-triethylammoniumpropyl)-4-(6-[4-(diethylamino)phenyl]hexatrienyl)pyridinium dibromide] to stain the vacuole as a reference point in the P. pastoris cells (12). While the wt Bgs13-EGFP peptide was concentrated near the neck and bud of dividing cells, the mutant protein-EGFP fusion was found on the periphery of the cell (Fig. 4). The observation that the two Bgs13p forms were localized to different regions of the cell suggests that, potentially, the peptides were able to interact with different molecular partners in the cell. Interestingly, wt Bgs13p showed localization similar to that of S. cerevisiae wt Pkc1p; however, deletion of the N-terminal region of Pkc1p caused it to be relocated to the nucleus in baker’s yeast, not to the cell periphery as seen in P. pastoris (15). Furthermore, the kinase activity of Pkc1p was found to be necessary for its localization to the bud neck region; a 1-amino-acid substitution in the kinase domain that reduced activity caused mislocalization. For Remi-bgs13p, the different localization might have changed its molecular partners, which in turn reduced activity, or its reduced activity interfered with its normal localization. Overall, the substitution of the first 147 amino acids of the Bgs13p with 12 amino acids in Remi-bgs13p caused changes in its cellular location and PKC activity, either or both of which may ultimately result in enhanced secretion.
Confocal fluorescence images of Bgs13-EGFP and Remi-bgs13-EGFP fusions. Strains expressing Bgs13-EGFP and Remi-bgs13 fusions from the AOX1 promoter were grown on glycerol and then induced on methanol-containing medium. These cells were then subjected to confocal fluorescence microscopy to detect EGFP, and the vacuolar membranes were stained with FM 4-64. The images were subsequently merged.
Characterization of the structural defects of the bgs13 strain cell wall.Having focused on the properties of the wt strain and the mutant Bgs13ps strain, we next turned our attention to the downstream effects of Remi-bgs13p, compared to its wt counterpart. Because of S. cerevisiae Pkc1p involvement in CWI, we first focused on examining the influence of Remi-bgs13p on the cell wall. In our first test, Congo red (Cr) and Calcofluor white (Cfw) were used to detect any cell wall defects in the wt and bgs13 strains. Both compounds are thought to interfere with cell wall assembly by binding to chitin in S. cerevisiae (24). The more chitin that is present, the greater is the binding of Cr and Cfw and thus the more severe is the weakening of the cell wall structure. In response to cell wall stress, yeast cell wall chitin levels can increase to as much as 20% of total wall polymer (10). Most cell wall mutants, such as those with disturbances in the synthesis of 1,3-β-glucan or the mannosylation of proteins, have more chitin in their cell walls than wt strain cells and thus are more sensitive to both Cfw and Cr (24). Before determining the sensitivity of our bgs13 strain, we confirmed that the FWK1 strain, a P. pastoris Δoch1 strain with a deletion in a mannosyltransferase gene and a well characterized cell wall defect (25), showed decreased colony density with increasing levels of both Cr and Cfw, compared to its wt parent Ku70 (Fig. 5). In the same manner, the bgs13 strain displayed cell wall defects demonstrated by greater sensitivity to these compounds, compared to its parent strain. Furthermore, transmission electron microscopy showed that bgs13 strain cells contained a thicker cell wall than wt parent cells (Fig. 6), another indication that a cell wall defect is present (26). These results were consistent with our previous finding that the bgs13 strain released higher levels of alkaline phosphatase than the wt strain, which was also interpreted as a sign of a cell wall weakness (7). Furthermore, in the previous work, the addition of the osmotic stabilizer sorbitol decreased alkaline phosphatase release into the ECM, strongly suggesting that CWI is compromised by the Remi-bgs13p function.
Sensitivity of bgs13 cell walls to Cr and Cfw. The indicated strains were grown to mid-log phase (OD600 values of 5 to 10) on YPD medium. Cell densities were normalized to an OD600 of 1.0 and diluted four times in a 10-fold series (100, 10−1, 10−2, 10−3, and 10−4). Equal volumes of each dilution were spotted on YPD plates containing the indicated concentrations of either Cr or Cfw. The plates were incubated at 30°C for 3 to 4 days and photographed.
Transmission electron microscopy of wt strain and bgs13 strain cells. Both wt strain and bgs13 strain cells were grown in YPD medium, and electron micrographs were taken at ×12,000 magnification. Arrows indicate the cell walls.
Determining the effect of DTT on cell wall proteins.To delve further into this question and to examine the structure of the cell walls of the wt and mutant strain cells, we performed a dithiothreitol (DTT) extraction analysis. DTT liberates noncovalently and dithiol-linked cell wall proteins, yielding increased amounts of extractable proteins when the cell wall is defective (27). It is expected that defects in the cell wall composition or structure would lead to increases in DTT-extractable proteins. Treatment of equivalent numbers of wt and bgs13 strain cells with DTT demonstrated that the mutant strain had higher levels (approximately 2- to 3-fold higher) of DTT-extractable proteins, as detected by SDS-PAGE with silver staining (Fig. 7). Novel protein bands appeared to be released from the bgs13 strain cell walls, while other protein bands seemed to become more prominent, especially those at 40 to 45 kDa; this was true for the strain grown on both simple and enriched methanol-based media.
SDS-PAGE analysis of proteins extracted from the cell wall. Two samples (each) of wild-type and bgs13 strain cells were grown on glucose, induced on BMM, and harvested. After washing, the cells were subjected to DTT treatment to extract proteins from the cell wall. A silver stain analysis of proteins separated by SDS-PAGE was then performed.
Measurement of cell wall porosity.However, two pieces of evidence seemed to contradict the hypothesis that a weakened cell wall was the cause of increased secretion (28). First, if a weakened cell wall were responsible for allowing greater protein secretion, then sorbitol addition should decrease secretion of a reporter such as SLPI. However, the addition of sorbitol to the bgs13 strain, which used a GAP promoter to constitutively secrete a SLPI reporter, did not reduce secretion as expected but actually increased export into the ECM (data not shown). Second, hypersensitivities to Cfw and Cr are associated with a more porous cell wall (25). A more porous cell wall could be expected to be more permeable to proteins delivered outside the plasma membrane by the secretory system, allowing entry into the ECM. To measure cell wall porosity, both wt and bgs13 strain cells were exposed to DEAE-dextran, which has a large hydrodynamic radius, and poly-l-lysine, which has a small hydrodynamic radius. While DEAE-dextran can damage the plasma membrane and release nucleic acids only if sufficiently large pores exist in the cell wall, poly-l-lysine constitutively causes cell leakage and nucleic acid release (25, 29). Therefore, by measuring the UV-absorbing compounds released from the cells by DEAE-dextran versus poly-l-lysine, the relative cell wall porosity can be determined. To ensure that the assay was being performed correctly, we measured the relative cell wall porosity of a Δoch1 strain of P. pastoris and its wt parent. The Δoch1 strain of Saccharomyces cerevisiae, which lacks a Golgi mannosyltransferase, has been demonstrated to have increased cell wall porosity, compared to its parent (25). After confirming that the P. pastoris Δoch1 strain showed the same phenotype, we found that the bgs13 mutant actually had a cell wall with slightly less porosity than its wt parent (Table 3). Thus, the bgs13 strain contains a cell wall with structural defects that make it more sensitive to Cfw and Cr but with less porosity than the wt cell wall, in contrast to findings observed for the P. pastoris Δoch1 strain, which showed structural anomalies and greater porosity. Moreover, a previous study demonstrated that the P. pastoris Δoch1 strain actually displayed a 30% reduction in the export of horseradish peroxidase (30). This finding suggests that any impact of cell wall porosity on secretion of a specific polypeptide depends on the unique properties of that protein. Thus, the combination of these results complicates the simple conclusion that a defective cell wall led to greater permeability, which in turn increased protein export.
Relative cell wall porosity, as determined by a polycationic assaya
ESI-MS/MS analyses of proteins in the ECM of wt and bgs13 strains.Because the bgs13 mutation resulted in different proteins being deposited in the cell wall and these proteins travelled through the secretory organelles of the yeast before reaching this destination (26), we wished to extend our analysis to the proteins delivered by the secretory pathway (i) into the ECM and (ii) to the cell wall of both wt and mutant strains. Previous experiments increased the secretion of four of five reporter proteins in the bgs13 strain (7). Thus, our first approach was to use electrospray ionization (ESI)-mass spectrometry to compare the entire populations of proteins secreted by the wt and mutant strains. Through a label-free quantitative approach, we analyzed equal amounts of proteins in the strains’ ECM and spiked the samples with equivalent amounts of bovine serum albumin (BSA) and myoglobin. To identify the proteins present and to quantitate their relative abundance in the shake flask culture medium from both mutant bgs13 and wt strains, the strains were induced with methanol, and the peptides were collected, solubilized in 3 M urea, and subjected to overnight digestion with trypsin. The tryptic peptides were then separated via nano-liquid chromatography, eluted, and ionized by ESI for tandem mass spectrometry (MS/MS) analysis. Within both wt strain ECM and mutant bgs13 strain ECM, a total of 337 proteins were identified with high levels of confidence (>2 unique peptides mapping back to each identified protein across all six technical replicates). The relative abundance of the 335 identified proteins found in the postinduction ECM was calculated as the ratio of the protein’s grouped (i.e., average) abundance in the bgs13 strain ECM to the protein’s grouped abundance in the wt strain ECM. Fold changes in the abundance of proteins were calculated with the log2(abundance ratio), which was subsequently displayed in a volcano plot (Fig. 8). The volcano plot displays fold changes versus significance (on the x and y axes, respectively), resulting in data points with low P values (highly significant) appearing toward the top of the plot. Therefore, proteins with the greatest increase in relative abundance in the bgs13 strain secreted fraction and with high significance are found in the top right portion of the graph.
Differential protein populations found in the ECM of shake flask cultures grown at 28°C. A volcano plot illustrates the differential abundance of proteins found in the ECM of wt and bgs13 strain cells expressing the MBP-EGFP reporter. Differentially abundant proteins of statistical significance (P ≤ 0.05) are shown in red (wt strain) and green (bgs13 mutant); x axis, log2(fold change), i.e., fold change in abundance between the two populations (wt strain and bgs13 mutant); y axis, −log10(P), i.e., statistical significance of the fold change in abundance. Fold changes were calculated as log2(abundance ratio) (bgs13 strain/wt strain); for the three technical replicates for each strain, the abundance of each identified protein was grouped (averaged) for use in the abundance ratio and fold change analyses. “A” represents MBP, while “B” represents EGFP.
Both the wt and bgs13 strains depicted in the volcano plot were engineered to secrete a fusion of MBP and EGFP under methanol growth conditions, the export of which was studied previously (12). The fusion was proteolyzed into its two domains prior to secretion. As demonstrated in a Western blot analysis (Fig. 9), the bgs13 strain secreted about 20 times as much MBP-EGFP as its parent strain, according to densitometry analysis. The MBP from MBP-EGFP is slightly larger than the positive-control MBP purified from bacteria because it contains a few extra amino acids from a linker region located between the MBP and EGFP domains. Consistent with the Western blot data, the volcano plot indicated that MBP (indicated by an A) and EGFP (indicated by a B) were found to have abundance ratios of 17.52 and 24.04, respectively, validating the relevance of our analysis (Fig. 8).
Western blot analysis of extracellular fractions from strains expressing MBP-EGFP. Wild-type and bgs13 strains containing an MBP-EGFP fusion and a wt strain containing an empty plasmid, pPICZαB (negative control), were grown on glycerol-containing medium to accumulate biomass, induced on methanol-containing medium to express recombinant protein, and pelleted. Volumes of extracellular supernatants, corresponding to equivalent numbers of cells (based on OD600 values), were probed with anti-MBP antibody. Fifty nanograms of commercially available MBP was used as a positive control. Images were captured with a ChemiImager 5500 and quantified with IDV analysis.
Overall, the secretome analysis demonstrated that, of the proteins released by both strains, most were released at significantly higher levels by the bgs13 strain. Through manual annotation of the protein hits provided by Proteome Discoverer 2.2, nine previously characterized proteins that are secreted or contain a signal peptide sequence were identified (Table 4). These proteins included cell wall-associated proteins such as 1,3-β-glucanosyltransferase (ATY40_BA7501626) and exo-1,3-β-glucanase (EXG), as well as other native proteins such as vacuolar proteases (BA75_01312T0 [PRB1] and BA75_03408T0 [PEP4]). Extracellular protein X1 (Epx1) (BA75_03408TO [PRY2]) is one of the most abundantly secreted proteins in chemostat cultures of wt P. pastoris and is considered a major contaminant during the purification of recombinant proteins (57). If Epx1 export was increased by the bgs13 mutation, this would have worsened an already difficult problem. However, Epx1 was found at equal levels (abundance ratio of 1.014) in the ECM of both bgs13 and wt strains. Thus, the bgs13 mutation does not increase the secretion of a major contaminant, which is an attractive aspect of the bgs13 mutant strain.
Summary of information on identified representative proteins in the ECM that contained a signal peptide sequence
Furthermore, of the 335 proteins that were identified in the ECM of both strains, many intracellular proteins were found, most likely due to cell lysis of older cells, which was noted in a previous proteomic analysis of a wt P. pastoris secretome (31). A defective cell wall would make the bgs13 strain cells more susceptible to cell lysis. Cell wall defects have been demonstrated to arise from perturbations in the CWI pathway, which has been shown to be coordinated regulated with the unfolded protein response (UPR) (10, 32). Sensitivity to Cr and Cfw, increased levels of DTT-extractable cell wall proteins, and enhanced secretion of some recombinant proteins are phenotypes of cells with a disrupted UPR. The UPR is an endoplasmic reticulum (ER)-to-nucleus signaling pathway initiated by ER stress, such as the buildup of unfolded proteins in the ER that can result from protein overexpression (33–35). The UPR attempts to reduce this ER stress by activating a large number of target genes, whose products function in a range of activities, such as enhanced protein folding to improve secretion efficiency and protein degradation to remove unfolded aggregate peptides. The combined actions of these UPR mechanisms should ultimately reduce the buildup of unfolded proteins in the secretory mechanism. Thus, it could be hypothesized that Remi-bgs13p triggers abnormal UPR activity, which leads to greater secretion in addition to a weaker cell wall.
The differentially abundant intracellular proteins also included those associated with the UPR mechanism (the ER chaperones BiP [KAR2] and protein disulfide isomerase [PDI]) and others involved in protein folding and aggregation (BA75_00603T0 [CPR6], BA75_00961T0 [SGT2], BA75_01760T0 [HSP104], BA75_02652T0 [STI1], BA75_00236T0, and ATY40_BA7500236) (see Fig. S1 in the supplemental material). Three pieces of evidence connect the bgs13 strain to a change in the UPR. First, the overexpression of the Hac1 protein, a master transcriptional activator of UPR genes, causes the secretion of Kar2/BiP, protein disulfide isomerase (PDI), and other chaperone-type proteins (containing the HDEL retention sequence) into the ECM, possibly because the capacity of the ER retrieval mechanism is exceeded when the UPR is induced (36). Second, Kar2p/BiP and PDI were found in the ECM when recombinant secretory insulin precursor secretion triggered the UPR in another study (37). Lastly, Yu et al., when comparing the transcriptomes of recombinant strains containing either 12 copies or single copies of a bacterial phospholipase gene, identified the most significant, differentially expressed genes found at higher levels in the multicopy strain, which were involved in ER protein processing and heat shock responses (38). Of the eight top genes, six were found in greater abundance in the bgs13 strain than in the wt strain (Fig. S1). These similarities suggest that Bgs13p may be involved in the UPR mechanism and the lack of the first 147 amino acids in Remi-bgs13p may affect normal UPR activity. Thus, the bgs13 mutation has a global effect on the secretory mechanism, influencing the export of a diverse array of proteins, with links to the CWI and UPR.
ESI-MS/MS analyses of proteins in the cell wall of wt and bgs13 strains.The increased amounts of ECM proteins from the bgs13 strain cells justified the use of the DTT-extracted cell wall proteins as the next protein population for mass spectrometry analysis. After removal of all traces of ECM, the protein populations were extracted from the cell walls of both the wt strain and the mutant bgs13 strain (Fig. 7) and underwent sample preparation, tryptic digestion, and subsequent data analysis. A total of 359 proteins were identified with high confidence in the cell wall extracts from both strains, with 23 proteins being classified as either uncharacterized or hypothetical proteins. The fold changes in the differential abundance of the 359 proteins are graphically represented in a volcano plot in Fig. 10. Similar to those found in the ECM, the majority of the proteins were more abundant in the mutant bgs13 strain cells than in the wt strain cells (Fig. S2). The recombinant reporter proteins EGFP and MBP were found to be more abundant in the bgs13 strain (abundance ratios of 2.262 and 1.977, respectively). These proteins may be present either because of being retained inside the cell wall during transit to the ECM or because of being released by cell lysis.
Differential protein populations identified from DTT extraction of the cell wall from cells grown in shake flask cultures at 28°C. A volcano plot illustrates the differential abundance of proteins extracted from the cell walls of wt and bgs13 strain cells after methanol induction, as shown in Fig. 8. Differentially abundant proteins of statistical significance (P ≤ 0.05) are shown in red (wt strain) and green (bgs13 mutant); x axis, log2(fold change), i.e., fold change in abundance between the two populations (wt strain and bgs13 mutant); y axis, −log10(P), i.e., statistical significance of the fold change in abundance. Fold changes were calculated by taking the log2(abundance ratio) (bgs13 strain/wt strain); for the three technical replicates for each strain, the abundance of each identified protein was grouped (averaged) for use in the abundance ratio and fold change analyses.
Through manual annotation of the protein hits provided by Proteome Discoverer 2.2, eight previously characterized cell wall-associated proteins were identified, including multiple isoforms of 1,3-β-glucanosyltransferase, an endo-1,3-β-glucanase, and a putative aspartic protease (Table 5). Two of the 1,3-β-glucanosyltransferase isoforms, encoded by ATY40_BA7501626 and GAS1, were the only characterized cell wall-associated proteins that displayed greater abundance in the bgs13 strain extract. Surprisingly, most of the cell wall-associated peptides were found in greater abundance in the wt strain; the reason for this is unclear. In S. cerevisiae, the Pkc1 kinase in the CWI pathway represses some enzymes involved in the biosynthesis and assembly of the cell wall, including β-glucanase encoded by BGL2 (39). In S. cerevisiae Δpkc1 strain cells, overproduction of β-glucanase was partially responsible for a weakened cell wall and a growth defect. It is possible that a similar scenario is occurring in the bgs13 strain of P. pastoris; the hybrid Remi-bgs13p may be affecting the regulation of specific cell wall-associated proteins.
Summary of information on identified representative cell wall-associated proteins from DTT extraction of the cell wall
Furthermore, many of the 359 cell wall-extracted proteins were of intracellular origin, such as key proteins involved in the UPR and protein folding, including PDI (abundance ratio of 1.70) and BiP/Kar2p (abundance ratio of 2.292), just as in the ECM. Their presence in the cell wall extract could suggest that Remi-bgs13p leads to either (i) inefficient retention of these ER-resident proteins, leading to mistargeting and incorporation into the cell wall, with liberation by DTT, or (ii) lysis due to a weakened cell wall, leading to the release of chaperones into the extract.
Quantitation of intracellular PDI expression.To help differentiate between these two possibilities, we quantified the intracellular levels of PDI. We hypothesized that, if Remi-bgs13p caused PDI to be overexpressed, not retrieved efficiently, and allowed to enter the cell wall and ECM, then the intracellular levels of PDI should be roughly the same inside wt strain cells and mutant cells. However, if PDI was overexpressed, accumulated inside the ER, and was released only after cell lysis, then the levels of PDI would be much higher inside bgs13 strain cells, relative to wt strain cells. To quantitate intracellular PDI levels, proteins were extracted from equal numbers of mutant and wt cells and separated into soluble/cytoplasmic and insoluble/membrane-associated fractions. PDI is usually found in an insoluble membrane fraction of P. pastoris cells (40, 41). Equal amounts of protein from each strain’s soluble and insoluble fractions were subjected to Western blot analysis with a primary antibody against PDI (Fig. 11A). In addition, a strain overexpressing PDI from the AOX1 promoter was utilized as a positive control (23). In our immunoblots, the PDI overexpression strain contained about 20% more detectable PDI than these strains (Fig. 11B), which was consistent with previous results (23). Furthermore, our Western blot analysis indicated that, while PDI was undetectable in the soluble fraction of each strain, approximately equal amounts of the chaperone were detected in the insoluble membrane fractions of both mutant and wt cells. Taken together, these results suggest that, if PDI is overproduced due to the bgs13 mutation, then it saturates the capacity of the ER and proceeds through the secretory system, arriving at the cell wall and ECM. Thus, in the context of previous findings regarding the bgs13 strain, our work demonstrates that Remi-bgs13p changes the protein-sorting mechanism and leads to changes in protein localization in P. pastoris cells.
Intracellular PDI expression levels in P. pastoris strains after methanol-induced recombinant protein expression. (A) Immunoblot analysis of intracellular PDI levels extracted from insoluble membrane-associated fractions. The strains, which expressed the human SLPI protein, were grown on glucose, induced with methanol, and harvested. Insoluble membrane-associated fractions were isolated from the cell pellets, and equal amounts of protein were probed with an anti-PDI antibody. The positive control overexpressed PDI from the AOX1 promoter. (B) Quantitation of the intensity of each strain’s detected PDI with IDV analysis.
Conclusions.The insertion of the pREMI plasmid into the BGS13 locus created a gene that expressed a Bgs13p lacking its first 147 amino acids. This mutant version had lower PKC activity and was localized to a different cellular region, compared to its wt counterpart. Because overexpression of wt Bgs13p decreased the secretion of a SLPI reporter, we think that the decreased PKC activity found for Remi-bgs13p was at least partially responsible for increases in the secretion levels of multiple reporters (including MBP-EGFP), as well as native proteins, as revealed by mass spectrometry analysis. The reduction in kinase activity would justify an examination of the members of the MAPK cascade in the CWI pathway. However, certain proteins were secreted at the same levels in the wt and bgs13 strains, indicating that the mutation did not have a universal effect.
One salient effect of this enhanced protein export was a defect in cell wall construction, since the bgs13 strain cell wall not only became more sensitive to Cr and Cfw but also released greater amounts of protein when subjected to DTT extraction. These effects are also seen in S. cerevisiae cells with a decreased UPR (27) but, unlike that strain, the P. pastoris bgs13 mutant did not display increased porosity of the cell wall or a thinner cell wall. However, the bgs13 strain demonstrated overexpression of BiP and PDI, an effect seen in P. pastoris strains with elevated Hac1p levels and increased UPR (36). In the bgs13 strain, PDI was present in the cell wall and ECM, suggesting a difference in protein sorting. Coupled with this result is the previous finding (12) that the bgs13 allele caused MBP-EGFP to be shunted from the vacuole (its normal destination in the wt strain) and localized to other parts of the secretory network prior to its export into the ECM. Taken together, these results suggest that the bgs13 strain is a supersecreter not because it has a more porous cell wall but because its novel protein-sorting mechanism causes a large increase in the secretion of proteins, which results in an altered composition and structure of its cell wall.
The bgs13 strain has several similarities and differences with respect to cells with decreased UPR. Thus, Remi-bgs13p does not simply activate or deactivate the UPR. The UPR has two major mechanisms to ameliorate the ER stress brought on by protein buildup in secretory organelles associated with high levels of recombinant peptide expression, namely, the degradation of misfolded proteins and the enhanced folding of nascent peptides (33). In our current model, Remi-bgs13p may be influencing these two mechanisms by either (i) downregulating the UPR, which causes misfolded and/or incorrectly modified proteins to circumvent vacuolar targeting and allows them to enter the cell wall or ECM (it has been suggested that defects in cell wall structure may stem from the incorporation of misprocessed and/or misfolded proteins in the cell wall [32, 42]), or (ii) upregulating the UPR, which leads to an increase in folding with high fidelity and allows more correctly proteins to be successfully secreted. Our future experiments will determine which model is correct by using mass spectrometry to examine the folding of proteins that show enhanced secretion. Using matrix-assisted laser desorption–time of flight mass spectrometry, the differences in disulfide linkages of the reporter protein SLPI produced in the wt strain versus the bgs13 strain can be elucidated. Thus, differences in disulfide linkages will give information about the protein-folding capabilities of the two strains. Understanding this critical step in how Bgs13p modification leads to supersecretion not only will shed light on the basic principles of protein export but also could lead to new approaches to make P. pastoris a better recombinant host system.
MATERIALS AND METHODS
Strains and growth conditions.Escherichia coli One Shot TOP10 chemically competent cells (Life Technologies, Carlsbad, CA) were used for transformation and plasmid DNA amplification unless stated otherwise. The TOP10 transformants were grown at 37°C in Lennox broth (LB) with the addition of 100 μg/ml ampicillin or 25 μg/ml Zeocin, in a New Brunswick Scientific C25 incubating shaker (New Brunswick Scientific Edison, NJ) at 225 rpm. The P. pastoris wt strain, yJC100, was derived from the original strain NRRL Y11430 (North Regional Research Laboratories, US Department of Agriculture, Peoria, IL). The yDT39 strain (his4 met2) (43) and the bgs13 mutant (7) derived from it were generated in our laboratory. These yeast strains were cultured in YPD medium (1% yeast extract, 2% peptone, 2% glucose) supplemented with 100 μg/ml Zeocin or 0.5 mg/ml G418 for antibiotic resistance selection, YND medium (0.17% yeast nitrogen base [YNB] with 0.5% ammonium sulfate and 0.4% glucose), basic medium with glucose and YNB (BMGY) (1% glycerol, 2% peptone, 2% glucose, 1.34% YNB, 0.00004% biotin, 100 mM potassium phosphate [pH 6.0]), or basic medium with methanol and YNB (BMMY) (0.5% methanol, 2% peptone, 1% yeast extract, 1.34% YNB, 0.00004% biotin, 100 mM potassium phosphate [pH 6.0]) (44). P. pastoris cells were grown in liquid culture at 28°C in a shaking incubator (model 1585; VWR, Batavia, IL) set to 325 rpm, while cells that were grown on agar plates were incubated at 30°C in a Fisher Scientific Isotemp incubator (Fisher Scientific, Pittsburgh, PA). Optical density at 600 nm (OD600) values were measured using a Spectronic Genesys 2 spectrophotometer (Spectronic Instruments Inc., Rochester, NY).
Recombinant DNA procedures, including bacterial transformation, were performed essentially as described previously (45). Plasmid DNA was purified from E. coli cultures using a QIAprep Spin miniprep kit (Qiagen, Chatsworth, CA). Linear plasmid DNA and PCR products were cleaned and concentrated using the Zyppy DNA Clean and Concentrator kit (Zymo Research, Irvine, CA). Restriction enzymes were purchased from MBI Fermentas (Hanover, MD). Precast 1.2% agarose Flash gels (Lonza, Rockland, ME) were used to analyze the restriction digestion products. Oligonucleotides were synthesized by Sigma Genosys (Plano, TX). All mutated sites and ligation junctions in newly synthesized vectors were confirmed by DNA sequencing (Quintara Biosciences, South San Francisco, CA). The CRISPR BGS13 knockout plasmids were kindly provided by Anton Glieder (Technical University of Graz). The MBP used as a positive control in Western blot analyses was purchased from New England Biolabs (Ipswich, MA). DNA concentrations were determined using a NanoDrop 2000c spectrophotometer (Thermo Fisher Scientific) set to 260 nm. Sequence analyses were performed with SnapGene software (GSL Biotech, Chicago, IL).
PCR analysis of cDNA from the bgs13 strain.Wild-type and bgs13 strain cells were grown as described below (in “Large-Scale Expression”) but the cell pellets were harvested after a 24-h methanol induction, washed with diethyl pyrocarbonate (DEPC)-treated water, and frozen at −80°C. Total RNA was isolated from both cell types by a standard procedure adapted for total yeast RNA isolation (46). After incubation with RNase-free DNase, the RNA was reverse transcribed from approximately 1 μg of total RNA using the ProtoScript first-strand cDNA synthesis kit (New England BioLabs, Ipswich, MA), according to the manufacturer’s instructions. Two microliters of the cDNA product was then subjected to standard PCR using the primers TTGGATCCCATACTGGGCAATTGTGTGC and TTGAGCTCGGTTCGTCGGTAAGAATGGA, to determine whether the 3′ end of BGS13 was present in the wt and Remi-bgs13 transcripts (expected 650-bp product). The same PCR procedure was performed with cDNA product template primers GAGGTACCTGTCAGACACAAGCAATGATGA and AAAGCCATCTTCATCAAATG, to determine whether the 5′ end of BGS13 was present in the wt and Remi-bgs13 transcripts (expected 500-bp product). A plasmid containing the wt BGS13 gene and reverse transcription reaction mixtures lacking transcriptase were used as the templates for positive- and negative-control reactions, respectively. The PCR products were then analyzed by agarose gel electrophoresis.
Cloning of hybrid Remi-bgs13 cDNA.The hybrid Remi-bgs13 mRNA was reverse transcribed, amplified, and cloned using the SMARTer RACE 5′/3′ kit (Clontech Laboratories, Mountain View, CA) according to the manufacturer’s instructions, with the following specifications. Briefly, total RNA was isolated from bgs13 strain cells grown overnight in YPD medium by a standard procedure adapted for total yeast RNA isolation (46). Approximately 2 μg of total RNA was used in the cDNA synthesis reaction, after which 90 μl of Tris-EDTA (TE) buffer was added to dilute the first-strand cDNA synthesis reaction mixture. The initial 5′-RACE reaction mixture contained 2 μl of a 5 mM solution of the Bgs13antisense1201 primer (GATTACGCCAAGCTTGTAAGCAGAATCTGCCCGGCAGGTGCTA), using 25 cycles of the program 94°C for 30 s, 68°C for 30 s, and 72°C for 3 min. After gel electrophoresis indicated a 1,000- to 1,300-bp product from this first 5′-RACE reaction, a nested PCR (optional, according to the kit directions) containing 5 μl of a 1:50 dilution of this reaction mixture, 25 μl of 2×SeqAmp buffer (from the kit), 1 μl of SeqAmp DNA polymerase (from the kit), 17 μl of distilled water, 1 μl of 10 μM Bgs13antisense1001 primer (GATTACGCCAAGCTTGCGGCTGTTGCTGTTCTGGTCTCCCTGCC), and 1 μl of 10 μM universal primer short (from the kit) was performed using the program described above. The PCR products were gel purified using the NucleoSpin gel and PCR clean-up kit (included with the SMARTer RACE 5′/3′ kit), according to the manufacturer’s instructions. These purified fragments were ligated into linearized pRACE vectors using the In-Fusion HD cloning kit and were transformed by heat shock into Stellar competent E. coli cells (included with the SMARTer RACE 5′/3′ kit), according to the manufacturer’s instructions. Transformants were plated on LB agar plates supplemented with ampicillin. After colonies were grown overnight in 3 ml of LB containing antibiotic, their plasmids were isolated, and the inserts were characterized by restriction enzyme digestion followed by sequencing.
BGS13 knockout strain construction.The plasmid used to perform a conventional knockout of BGS13, named pKdSC2, was constructed as follows. First, the S. cerevisiae HIS4 gene (47) was inserted into pBluescript SK II to create pMS4. Then, a fragment containing approximately 500 bp of 3′ BGS13 genomic sequence was generated with PCR using the primers 5′ LL1BACKBamH1 (TTGGATCCCATACTGGGCAATTGTGTGC) and 3′ LL1BACKSac1 (TTGAGCTCGGTTCGTCGGTAAGAATGGA). The PCR product was digested with BamHI and SacI and inserted into the same sites of pMS4 to construct pKdSC1. A fragment containing the first approximately 500 bp of 5′ BGS13 coding sequence was then amplified by using the primers 5′ LL1FRONTKpn1 (GAGGTACCTGTCAGACACAAGCAATGATGA) and 3′ LL1FRONTPst1 (GACTGCAGCGTGGTAATTCTTCAAAGCCA). The PCR product was digested with KpnI and PstI and inserted into the same sites of pKdSC1 to construct pKdSC2. pKdSC2 contains S. cerevisiae HIS4 flanked by 5′ and 3′ BGS13 sequences to promote homologous recombination. The knockout sequence was liberated by digestion with KpnI and SacI, gel purified, and then transformed into yGS115 (his4). His+ transformants were selected on YND medium (both with and without 1 M sorbitol) and subjected to colony PCR using a forward primer located in the BGS13 promoter region and a reverse primer located within P. pastoris HIS4. A chromosomal BGS13 knockout would be expected to produce a 1.2-kb PCR product.
The creation of BGS13 knockout strains was also attempted using a CRISPR/Cas9 system developed for P. pastoris (22). Briefly, six plasmids (kind gifts from Anton Glieder, Technical University of Graz) were utilized, one containing gRNA sequences of BGS13 coding sequence positions 1069 to 1088 (pKO13_gRNA1), a second containing coding sequence positions 1049 to 1068 (pKO13_gRNA2), and a third containing coding sequence positions 816 to 836 (pKO13_gRNA3) (position 1 represents the A of the initiation codon ATG) in sets of vectors with either Zeocin or G418 resistance genes. As a control, a plasmid harboring no BGS13 gRNA sequence was also utilized (pKO13_gRNA0). The undigested plasmids were transformed into competent yJC100 or ku70 cells (19), and the cells were selected for antibiotic resistance. All groups of transformants were purified by single-streak isolation and were subjected to colony PCR using the primers BGS13656-680for (CCAAGAAATTAGAAACGTTGGTTTC) and BGS131189-1165rev (GCTGACGAATAGCTCCATGACGACC), which amplified approximately 530 nucleotides of the chromosomal BGS13 coding sequence, containing all three of the CRISPR/Cas9 target sites. The PCR products were purified and sequenced to confirm the predicted mutations. Alignments between CRISPR transformant and wt BGS13 DNA were performed with SnapGene (GSL Biotech, Chicago, IL).
Plasmid constructions.The Bgs13p overexpression construct, pKdSC5, was produced by amplifying the entire BGS13 coding sequence without the stop codon with the primers 5′LL1CDSXho1 (GCCTCGAGATGTCAGACACAAGC) and 3′LL1CDSNot1 (AAGCGGCCGCAGTGTGATCATCG). The resultant PCR product was inserted into the pPICZαB vector with the restriction enzymes XhoI and NotI. The pKdSC5 vector has a Zeocin-selectable resistance marker and BGS13 coding sequence, fused in frame with C-terminal c-myc and His6 tags, under the control of the AOX1 promoter.
pAH1 was constructed to express the Remi-bgs13 hybrid transcript under the control of the AOX1 promoter. It contained 12 codons (initiated by an ATG) originating from the pREMI plasmid fused in frame to codon 148 to the stop codon of the BGS13 coding sequence, with C-terminal c-myc and His6 tags. It was constructed by amplifying the complete Remi-bgs13 coding sequence from the chromosomal DNA of the bgs13 strain using colony PCR (7) with the primers AH5′Xho1hulk (GGCTCGAGATGAGATCAGATCGTAAAACG) and 3′ LL1CDSNot1 (AAGCGGCCGCAGTGTGATCATCG). The 3-kb PCR product was digested with restriction enzymes XhoI and NotI and inserted into the corresponding sites of pPICZB to create pAH1.
pBGS13-EGFP, which contains the wt Bgs13 coding sequence fused to a C-terminal EGFP under the control of the AOX1 promoter, was constructed by combining the 6-kb NotI-BamHI fragment of pKdSC5 with the 1.4-kb NotI-BamHI fragment pJGBG-EGFP (7). pREMI-BGS13-EGFP was constructed by digesting pBGS13-EGFP with HindIII and SphI and replacing the 2-kb fragment containing a portion of the BGS13 coding sequence with a 1.9-kb HindIII-SphI fragment from pAH1 containing the corresponding portion of the Remi-bgs13 coding sequence. Thus, pREMI-BGS13-EGFP expresses the Remi-bgs13 hybrid fused to a C-terminal EGFP under the control of the AOX1 promoter.
pAMI, which contains the AOX1 promoter driving the expression of the MATα-SLPI-c-myc-His6 fusion, causes SLPI-c-myc-His6 to be secreted into the ECM; its construction was described previously (23). pKanJV4, which contains the AOX1 promoter driving the expression of the MATα-MBP-EGFP fusion in a pKanαB vector, was described previously (12).
P. pastoris transformation.Transformations of P. pastoris were performed by the electrotransformation method, as described (48). Plasmids were linearized at unique sites in the AOXI promoter with restriction enzymes SacI or MssI, purified with the DNA Clean and Concentrator kit (Zymo Research), and electroporated into competent yJC100 cells, yDT39 cells, or bgs13 strain cells. Transformed cells were allowed to recover for 1 h at 30°C in 1 ml of a 50% 1 M sorbitol/50% YPD solution and then were plated on selective medium. For selection of P. pastoris transformants, Zeocin was added to a final concentration of 100 μg/ml and G418 was added to a final concentration of 500 μg/ml (49). Transformed colonies were then purified by streaking for isolated colonies on selective medium.
Small-scale induction.Cells were grown in a sterile 96-deep-well-plate (DWP) format for enzymatic assays (50, 51). Five hundred microliters of sterile water was added to the perimeter of the plate to prevent evaporation. Next, 250 μl of liquid buffered minimal medium with dextrose (BMD), along with any amino acid supplements needed, was added to the wells that were designated to hold the cell cultures. Isolated colonies were selected from fresh plates and inoculated into their designated wells containing 1× BMD. Parafilm was wrapped around the edges of the DWP and the corners were taped to block evaporation. The DWP was placed in a 28°C shaking incubator at 325 rpm. After 48 h, 5-μl aliquots of cell culture from randomly selected wells were added to glucose test strips to confirm that all of the glucose had been exhausted. Subsequently, the cells were induced with 250 μl of 2× BMMY, yielding a final methanol concentration of 0.5%. The plate was sealed with Parafilm and placed back in the shaking incubator with the same settings. During the next two consecutive 24-h periods, 50 μl of 10× buffered minimal medium with methanol (BMM) was added to each well. On the sixth day, after a total of 72 h of methanol induction, the OD600 of each sample was measured at a 1:20 dilution. The samples were transferred from the wells into 1.5-ml microcentrifuge tubes, centrifuged at maximum speed for 1 min, and either stored in the –80°C freezer or used immediately for enzymatic assays.
Large-scale expression.Cultures were first grown overnight in YPD medium to stationary phase. On the second day, the OD600 was measured, and 5.0 OD600 units of each culture were pelleted for 5 min at 2,000 × g at room temperature. The cells were suspended in 5 ml of BMGY in a 50-ml conical centrifuge tube. On the third day, the OD600 was measured, and 10 OD600 units of each culture were pelleted for 5 min at 2,000 × g at room temperature. The cells were suspended in 10 ml of BMMY in a 50-ml conical centrifuge tube. The cultures were induced for 48 and 72 h at 28°C with shaking (225 rpm), with the addition of methanol to 0.5% (vol/vol) every 24 h to compensate for losses resulting from evaporation and metabolism. At each time point, the OD600 of each culture was measured, and 1.0 ml was centrifuged at 10,000 × g for 1 min to separate cells from extracellular supernatant. The supernatants were transferred to new microcentrifuge tubes, and Protease Arrest protease inhibitor (G Biosciences, St. Louis, MO) was added. Pellets and supernatants were immediately frozen and stored at –80°C. This protocol was also scaled up 10-fold (100 ml of BMMY) for larger preparations of MBP fusion peptides.
PKC assays.P. pastoris strains were grown overnight in BMMY to approximately 10 OD600 and were harvested by centrifugation at 13,000 × g for 1 min. After the ECM was removed, 500 μl of chilled phosphate-buffered saline (PBS) and 1 μl of 1 M phenylmethylsulfonyl fluoride (PMSF) were added to each pellet. After 150 μl of chilled glass beads (500- to 600-μm) was added to each sample, the lips of the tubes were cleaned with cotton swabs so that the tubes could seal well. The tubes were closed and vortex-mixed for a total of 3 min, consisting of three cycles of 1 min of vortex-mixing followed by 1 min on ice. The tubes were then centrifuged at 13,000 × g for 5 min at 4°C. The supernatant was transferred to a fresh tube and kept on ice. PKC activity was measured using a PKC activity kit (Enzo Life Sciences, Farmingdale, NY), according to the manufacturer’s directions.
Spot Western blotting.Spot Western blots were used to detect the presence of antigen in the ECM or intracellular extracts of recombinant P. pastoris cells (50). The vacuum attached to the 96-well Spot Blot apparatus (Topac, Boston, MA) was turned on, and the amount of supernatant that was added to the wells on the prewet nitrocellulose membrane was measured, to equal the OD600 of the sample with the lowest OD600 measurement. The spotted nitrocellulose membrane was allowed to dry in a 60°C oven for 5 min. The membrane was then soaked in 1× PBS while the SNAP ID protein detection system blot holder (Millipore, Billerica, MA) was wet with water. The nitrocellulose membrane was inserted into the blot holder, followed by a spacer sheet. The closed blot holder was placed in the SNAP ID apparatus, attached to a vacuum. With the vacuum turned on, 30 ml of I-Block (Thermo Fisher Scientific) solution (1× PBS with 0.2% I-Block and 0.1% Tween 20) was added to the membrane. The vacuum was turned off before incubation of the membrane with 3 ml of I-Block solution containing 10 μl of the primary antibody, i.e., mouse anti-GFP, rabbit anti-MBP, or mouse anti-c-myc (Santa Cruz Biotechnology, Santa Cruz, CA). After 10 min of incubation with the primary antibody, the vacuum was turned on again, and the membrane was washed three times with a total of 100 ml of wash buffer (1× PBS with 0.1% Tween 20). The membrane was then incubated for 10 min in 3 ml of I-Block containing a goat anti-mouse IgG or goat anti-rabbit IgG secondary antibody, conjugated to alkaline phosphatase (Applied Biosystems, Foster City, CA), with the vacuum off. The membrane was then washed three times with wash buffer, as described previously, before it was incubated for 5 min in a petri dish containing 20 ml of femtoLUCENT PLUS-AP 1× Tris-buffered saline with Tween (G Biosciences). The blot was laid flat on a plastic-wrap surface, and 2 ml of the femtoLUCENT PLUS-AP detection reagent was added dropwise onto the membrane. After 5 min of incubation at room temperature, the detection reagent was drained off. The membrane was placed in a plastic envelope or wrap and developed using the Bio-Rad ChemiDoc XRS+ imaging system (Bio-Rad, Hercules, CA), with exposure times of 1 to 2 min. Signals were quantified with Bio-Rad Image Lab software.
Polyacrylamide gel electrophoresis and Western blot analysis.Protein concentrations were determined using the Pierce (Rockford, IL) bicinchoninic acid (BCA) protein assay kit with BSA as the standard. Either equivalent amounts of protein or volumes of ECM from equivalent numbers of cells were electrophoresed on SDS-PAGE gels. The proteins were either stained for total protein visualization using GelCode Blue (Pierce, Rockford, IL) or Silver Snap II (Pierce) stain or analyzed by Western blotting. For immunoblotting, proteins were transferred onto nitrocellulose membranes using an iBlot apparatus (Life Technologies), according to the manufacturer’s instructions. Immunoblots were processed as described above, using the SNAP i.d. system and femtoLUCENT reagents. The anti-PDI antibody was a generous gift from Carl Batt (Cornell University, Ithaca, NY).
Transmission electron microscopy.Transmission electron microscopy was performed as described previously (52).
Fluorescence microscopy.(i) Growth and induction. Fresh colonies were inoculated into 5 ml of BMGY in 50 ml conical tubes. These cultures were incubated overnight at 28°C, at 325 rpm. Next, 2 OD600 units of the cells were centrifuged and resuspended in 2 ml of BMMY, in new 50-ml conical tubes, to start the induction with a concentration of 1 OD600 unit/ml. The cell cultures were incubated again in a 28°C shaking incubator for 5 to 6 h.
(ii) Vacuolar staining with the FM 4-64 dye. Some strains were stained with the FM 4-64 dye (Life Technologies). This dye stains the vacuolar membranes in P. pastoris, which serves as a point of reference for visualizing the localization of fluorescent proteins inside the cells (53, 54). When the cells were switched to BMMY, 2 μl of FM 4-64 (1 mg/ml) was added to the BMMY-cell mixture. Cell cultures were incubated for 5 to 6 h in a 28°C shaking incubator before the cells were visualized with fluorescence microscopy.
(iii) Visualization with fluorescence microscopy. After induction for 5 to 6 h, the cells were centrifuged in 50-ml conical tubes at 3,000 × g for 5 min and resuspended in 2 ml of fresh 1 M sorbitol or sterile water. The samples were centrifuged briefly for pelleting. The sorbitol or water was discarded and 500 μl of fresh 1 M sorbitol or water was added to resuspend the cells. Then, 5 μl of each sample was added to dried 1% agarose in water on microscope slides (Fisher Scientific, Pittsburgh, PA), and a coverslip (Corning Life Sciences, Lowell, MA) was placed over each sample. Confocal fluorescence images were acquired with a Leica DMIRE2 inverted fluorescence microscope using MetaMorph software (Molecular Devices) and a Yokogawa CSU-X1 spinning disc confocal system with a QuantEM:5125C camera. Excitation and emission wavelengths were as follows: excitation at 491 nm and emission at 525 ± 25 nm for visualization of EGFP and excitation at 561 nm and emission at 605 ± 25.5 nm for visualization of FM 4-64 dye.
CWI assays.CWI assays were performed using both Cr and Cfw, as described previously (24). Briefly, YPD plates were made with different concentrations of Cr (0, 10, 20, and 40 μg/ml) and Cfw (0, 1, 2, and 5 μg/ml). Serial dilutions (1, 10−1, 10−2, 10−3, and 10−4) of 1 OD600 unit were of the bgs13 strain and its parent strain yDT39, as well as FWK1 and its parent strain KU70. FWK1 is a Δoch1 mutant that served as a positive control. Five microliters of each dilution was spotted onto all plates, which were incubated at 30°C for 3 days, and cells were analyzed for viability.
Cell wall porosity assay.The permeability of the wt strain (yDT39-pAM1) and bgs13 strain (ybgs13-pAM1) cell walls was measured using a protocol adapted for several different fungal species (55). In addition, the assay was performed with a mutant strain (FWK1) with a known cell wall defect in the OCH1 gene and its parent strain (Ku70), for comparison (30). Briefly, overnight cultures of cells grown in YPD medium were diluted in fresh medium to an OD600 of 0.1 at 28°C and then were allowed to grow to an OD600 of approximately 1. For each strain, approximately 8 OD600 units of cells were centrifuged at 3,500 × g for 7 min at room temperature. The supernatant was discarded, and the cell pellet was washed three times with 0.22-μm sterile-filtered Milli-Q water. The cells were then resuspended in 10 mM Tris (pH 7.4) to a final OD600 of approximately 2.0, and the resuspension was divided into three 1.5-ml centrifuge tubes with an approximate volume of 0.9 ml in each. To each 0.9 ml of cell suspension was added either (i) 100 μl of 10× poly-l-lysine (30 to 70 kDa; Sigma-Aldrich, St. Louis, MO) (100 μg/ml in 10 mM Tris [pH 7.4]), (ii) 100 μl of 10× DEAE-dextran (500 kDa; Sigma-Aldrich) (50 μg/ml in 10 mM Tris [pH 7.4]), or (iii) 100 μl of 10 mM Tris (pH 7.4). The cell suspensions were then incubated at 28°C for 30 min in shaking incubator, at 200 rpm. The cells were centrifuged two times for 2 min at 10,000 × g to isolate the supernatant. The A260 reading for each supernatant was taken to measure the UV-absorbing compounds, using the supernatant from cells incubated in only 10 mM Tris buffer (pH 7.4) as a blank. The assay was performed on bioreplicate samples in duplicate. The relative cell wall porosity was defined as follows: porosity (%) = [(A260 of DEAE-dextran)/(A260 of poly-l-lysine)] × 100%.
Mass spectrometry and proteomic analysis of proteins found in the ECM.yDT39-pKanJV4 (wt) and ybgs13-pKanJV4 cultures were grown overnight in YPD medium to stationary phase. OD600 measurements were obtained, and 5.0 OD600 units of each culture were pelleted and resuspended in 10 ml of BMD with histidine and methionine. Cultures were grown for 24 h, and then 50 OD600 units of wt strain culture and 100 OD600 units of bgs13 strain culture (to compensate for the lower growth rate) were pelleted and resuspended in 50 ml of BMM with histidine and methionine. Cultures were induced for 48 h at 28°C, with shaking at 225 rpm; methanol was added to 0.5% at 24 h postinduction, to compensate for losses resulting from evaporation and metabolism. At harvest, the OD600 of each culture was measured, cells were centrifuged, and the supernatant was filtered (0.22 μm) to remove remaining cell particulates. Cell pellets and supernatants were immediately frozen and stored at –80°C.
Supernatant total protein concentrations were measured at 280 nm for both strains, using a NanoDrop spectrophotometer (Thermo Fisher Scientific). Triplicate aliquots containing 2 mg of protein were prepared, and each replicate was spiked (1 part spiked reference proteins/500 parts proteins) with protein standards, including BSA (Thermo Fisher Scientific) and horse myoglobin (Sigma-Aldrich), for subsequent label-free proteomics quantitation analyses. Aliquots were precipitated using acetone (–20°C for 3 h). Proteins were dried, pelleted, and then resuspend in 200 μl of 3 M urea-100 mM Tris (pH 7.8). Precipitated protein concentrations were determined using the same methods as above, and then all samples were diluted to standardize concentrations (0.5 μg/μl; 50 μg total protein) prior to sample preparation for proteomics. Proteins were reduced with 5 mM DTT (Gold Biotechnology, St. Louis, MO) for 30 min at room temperature and were alkylated with 15 mM iodoacetamide [IAA] (Sigma-Aldrich) for 30 min at room temperature in the dark. Unreacted IAA was quenched by addition of 20 μl of 200 mM DTT and incubation for an additional 30 min. Each reaction was diluted with 3 volumes of sterile water (∼750 μl), to reduce the urea concentration to <2 M, and then digested overnight at 37°C using trypsin (Promega, San Louis Obispo, CA) (1 part trypsin/10 parts proteins). Digestions were halted with the addition of trifluoroacetic acid (TFA) to a final volume of 5% and digested peptides were purified using OMIX C18 spin columns, according to the manufacturer’s instructions (Agilent Technologies, Santa Clara, CA). Samples were then diluted, lyophilized, and resuspended in 40 μl of 0.1% formic acid in high-performance liquid chromatography (HPLC)-grade water; peptide samples were diluted to 150 ng/μl. All samples were stored at −80°C until mass spectrometry analysis.
HPLC-MS/MS.For each sample, 5 μl was loop injected by a Dionex Ultimate 3000 autosampler onto an EASY-Spray C18 liquid chromatography column (75 μm i.d. by 15 cm, 100 Å; Thermo Fisher Scientific) held at 35°C for HPLC. Flow rates were kept at 300 nl/min, set to 140 min. Solvents A and B were 0.1% formic acid in water and 0.1% formic acid in acetonitrile, respectively. Solvent B was used as follows: 3% for 5 min, 3% to 28% in 75 min, 28% to 45% in 25 min, 45% to 95% in 5 min, 95% for 5 min, return to 3% in 5 min, and 2% for 25 min.
Mass spectrometry analysis was performed using an Orbitrap Fusion Tribrid mass spectrometer equipped with nanospray HPLC (Thermo Fisher Scientific), operated in a data-dependent acquisition (DDA) manner via Xcalibur 4.0 software (Thermo Fisher Scientific). Briefly MS1 spectra were collected by using the Orbitrap mass analyzer; precursor ions were selected via DDA using a quadrupole mass filter and then were fragmented using higher energy collisional dissociation in the collision cell. Instrument and data acquisition settings were the same as reported previously (56), with the exception of the scan range (200 to 1,400 Da), MS1 Orbitrap resolution (120,000), scan time (10 to 130 min), MS/MS Orbitrap resolution (30,000), and MS/MS maximum injection time (150 ms).
Protein identification.Analysis of MS/MS data was performed via Proteome Discoverer 2.2.0.388 (Thermo Fisher Scientific). Peptide spectra were searched in the complete UniProt/Swiss-Prot database (downloaded on 13 February 2018) using SEQUEST. To identify any associated contaminants that might have arisen during sample preparation, spectra were searched in the common Repository of Adventitious Proteins (cRAP) (https://www.thegpm.org/crap/index.html). Search parameters were the same as reported previously (56). False-discovery rates for peptide spectral matches and peptides were estimated by searching reversed decoy databases generated from the UniProt/Swiss-Prot and cRAP databases. Results were filtered to remove identified contaminants. Peptides with target false-discovery rates of <1% were retained. For proteins to be categorized as unique, ≥2 unique peptides that mapped back to the spectra were required.
Protein quantitation.Protein quantitation was performed using Proteome Discoverer 2.2.0.388 (Thermo Fisher Scientific). To align chromatographic runs for each biological condition, the feature mapper node was used with a maximum retention time shift of 10 min, a mass tolerance of 10 ppm, and a minimum signal/noise threshold of 5 for feature linking mapping. MS/MS data were filtered to retain only proteins that had ≥2 unique peptide hits and were identified in all samples with high levels of confidence. Biological replicates were grouped by condition. Quantitative abundances were calculated, normalized to those of reference proteins (BSA and horse myoglobin), and scaled with a label-free method using the precursor ion quantifier node in Proteome Discoverer 2.2. Abundance ratio calculations were made by using summed abundance values. An analysis of variance for individual proteins was performed to determine whether protein abundances differed significantly (P ≤ 0.05) between samples. Abundance ratio and log2(fold change) values were calculated for the sample comparisons between wt and mutant strains.
DTT extraction of cell wall proteins.DTT extracts of wt strain (yDT39-pKanJV4) and bgs13 strain (ybgs13-pKanJV4) cells were prepared according to a described previously protocol (27). pKanJV4 expresses an MBP-EGFP fusion protein under the control of the AOX1 promoter. Briefly, 40 OD600 units of cells were grown to early stationary phase (OD600 of ∼10) overnight in either BMM or BMMY, harvested by centrifugation (5 min at 300 × g), washed twice with 10 ml of water, and resuspended (0.5 OD600 unit equivalents/μl) in extraction buffer (50 mM Tris [pH 7.5], 5 mM DTT). As negative controls, both strains were resuspended in 50 mM Tris (pH 7.5) alone. The cell suspensions were shaken in a multivortex apparatus for 2 h at 4°C, and the supernatants were used for further analysis. An aliquot of the supernatant was removed, mixed with 2× Laemmli sample buffer, boiled for 5 min at 100°C, and fractionated by SDS-PAGE (Bio-Rad Mini-Protean TGX 4 to 20% gradient gels). Fractionated proteins were visualized by silver staining according to the manufacturer’s protocol (Pierce silver stain kit; Thermo Fisher Scientific).
Mass spectrometry and proteomics of proteins extracted from the cell wall.From the DTT extraction of cell wall proteins, supernatant total protein concentrations for both strains were measured at 280 nm using a Nanodrop spectrophotometer (Thermo Fisher Scientific). Duplicate aliquots containing 120 μg of protein were prepared. Aliquots were precipitated using acetone (at –20°C for 3 h). Proteins were dried, pelleted, and then resuspended in 200 μl of 3 M urea-100 mM Tris (pH 7.8). Precipitated protein concentrations were determined using the same methods as described above, and then all samples were diluted to standardized concentrations (approximately 0.21 μg/μl; 25 μg total protein) prior to proteomics sample preparation. Prior to sample preparation for tryptic digestion, each replicate was spiked (1 part spiked reference proteins/500 parts proteins) with protein standards, including BSA and horse myoglobin, for subsequent label-free proteomics quantitation analyses. Samples were reduced in 5 mM DTT (Gold Biotechnology) for 30 min at room temperature and alkylated using 15 mM IAA (Sigma-Aldrich) for 30 min in the dark at room temperature. Unreacted IAA was quenched by addition of 20 μl of 200 mM DTT and incubation for an additional 30 min. Each reaction mixture was then diluted with 3 volumes of sterile water (∼750 μl), to reduce the urea concentration to <2 M, and digested overnight at 37°C using trypsin (Promega) (1:10 trypsin/protein). The digestion was halted with the addition of TFA, and digested peptides were purified using OMIX C18 spin columns, according to the manufacturer’s instructions (Agilent Technologies). Samples were then diluted and lyophilized. Lyophilized samples were resuspended in 40 μl of 0.1% formic acid in HPLC-grade water; peptide samples were diluted to 150 ng/μl. All samples were stored at −80°C until mass spectrometry analysis. Subsequent HPLC-MS/MS, protein identification, and protein quantitation procedures and parameters were the same as those used for the ECM.
Cytoplasmic and membrane-associated protein extraction.Protein extractions from soluble cytoplasmic and insoluble membrane-associated fractions were performed according to a previous extraction procedure (6). Cells (OD600 of 4) stored at −80°C were thawed, washed in PBS (pH 7.4), and then resuspended in yeast lysis buffer (50 mM sodium phosphate [pH 7.4], 1 mM PMSF, 1 mM EDTA, 5% [vol/vol] glycerol). An equal amount of acid-washed glass beads was added, and cells were lysed by vortex-mixing at maximum speed 10 times for 1 min, with 1-min periods on ice. The lysate was centrifuged at 10,000 × g for 30 min at 4°C. The supernatant, containing mainly cytoplasmic proteins, was collected and designated the soluble fraction. The remaining pellet with cell debris and glass beads was resuspended in 100 μl of lysis buffer containing 2% (wt/vol) SDS and was centrifuged at 4,000 × g for 5 min at 4°C. The supernatant was collected as designated the membrane-associated fraction. Extraction and detection of PDI was also performed on yJC100-pAM1-pPICZPDI cells, as a positive control (23).
ACKNOWLEDGMENTS
This work was supported by NIH AREA grants GM65882-03 and GM129758-01, as well as Scholarly and Artistic Activity grants from University of the Pacific to J.L.-C. and G.P.L.-C. This research was also funded by a National Science Foundation grant (grant DBI 1531417) to C.A.V.
FOOTNOTES
- Received 20 July 2019.
- Accepted 22 September 2019.
- Accepted manuscript posted online 4 October 2019.
Supplemental material for this article may be found at https://doi.org/10.1128/AEM.01615-19.
- Copyright © 2019 American Society for Microbiology.
