Adaptation of Candida albicans to growth on sorbose via monosomy of chromosome 5 accompanied by duplication of another chromosome carrying a gene responsible for sorbose utilization

Abstract

Candida albicans, a fungus that normally inhabits the digestive tract and other mucosal surfaces, can become a pathogen in immunocompromised individuals, causing severe or even fatal infection. Mechanisms by which C. albicans can evade commonly used antifungal agents are not fully understood. We are studying a model system involving growth of C. albicans on toxic sugar sorbose, which represses synthesis of cell wall glucan and, as a result, kills fungi in a manner similar to drugs from the echinocandins class. Adaptation to sorbose occurs predominantly due to reversible loss of one homolog of chromosome 5 (Ch5), which results in up regulation of the metabolic gene SOU1 (SOrbose Utilization) on Ch4. Here, we show that growth on sorbose due to Ch5 monosomy can involve a facultative trisomy of a hybrid Ch4/7 that serves to increase copy number of the SOU1 gene. This shows that control of expression of SOU1 can involve multiple mechanisms; in this case, negative regulation and increase of gene copy number operating simultaneously in cell.

Introduction

Candida albicans is a unicellular fungus that is a benign inhabitant of the mucosal surfaces of the gastrointestinal tract in approximately two thirds of the human population. In healthy individuals, C. albicans can cause superficial mucosal infections, but in immunocompromised patients C. albicans is associated with significant morbidity and mortality, and is therefore an opportunistic pathogen.

C. albicans is usually a diploid organism with a genome organized into 8 pairs of chromosomes. However, aneuploid chromosomes, as well as changes in DNA sequence, are frequently generated in vivo and in vitro and aneuploidies are well tolerated, similar to other fungi (Ahmad et al., 2008; Arbour et al., 2009; Yang et al., 2011; Rustchenko, 2007; Selmecki et al., 2010). Furthermore, reversible aneuploidy of a specific chromosome or of a large portion of a chromosome confers adaptation and survival in environments that otherwise kill cells or inhibit their proliferation. We have previously demonstrated survival of C. albicans via specific aneuploidy upon culture in media in which glucose was replaced by alternative sugars such as L-sorbose or D-arabinose, or which were supplemented with the antimetabolite 5-fluoro-orotic acid or the antifungal agent fluconazole (reviewed in Rustchenko, 2007; 2008). Another study demonstrated that in multiple fluconazole-resistant clinical isolates, triploidy of various chromosomes is a common condition with frequent formation of an iso-chromosome 5 containing two left arms (Selmecki et al., 2006).

In the well-studied case of survival on medium containing toxic sorbose that kills fungi similarly to antifungal agents of the echinocandins class (reviewed in Yang et al., 2013), a major mechanism for survival is the reversible loss of one homolog of chromosome 5 (Ch5), as demonstrated with various strains (Ahmad et al., 2008; Rustchenko, 2007; 2008; Janbon et al., 1998; Lephart et al., 2005; Andaluz et al., 2007). We reported that Ch5 monosomy reduces gene dose and downregulates multiple CSU (Control of Sorbose Utilization) genes on Ch5 that are negative regulators of the metabolic gene SOU1 (SOrbose Utilization) (orf19.2896) on Ch4 (Ahmad et al., 2012; Kabir et al., 2005). SOU1 encodes sorbose reductase catalyzing the first step of sorbose catabolism pathway, and is responsible for sorbose utilization, the Sou⁺ phenotype (Janbon et al., 1998; Greenberg et al., 2005). Ch5 monosomy upregulates SOU1, while restoration of Ch5 disomy by spontaneous duplication results in normal copy number of CSUs, down regulation of SOU1 and reversion to the normal Sou⁻ phenotype of no growth on sorbose (Janbon et al., 1998; Rustchenko, 2003). Ch5 loss is often accompanied by change in ploidy or length of ChR (Janbon et al., 1999). However, such ChR instability also occurs often in laboratory strains, as well as in various derivatives containing aneuploid chromosomes (Ahmad et al., 2008; Rustchenko, 2007). The changes of chromosomes other than ChR are not as frequent in the Sou⁺ strains and were not systematically studied (Janbon et al., 1999; Rustchenko, 2007).

Here, we report a duplicated chromosome, which increases the SOU1 copy number to a total of three, in the Sou⁺ strain Sor55. We previously reported Sor55 as a derivative of the popular laboratory strain 3153A upon culture on medium containing toxic sorbose as the sole source of carbon, and showed that it was monosomic for Ch5 (Janbon et al., 1998; Kravets et al., 2010). Here, we show that the 3153A strain harbors a reciprocal translocation between chromosomes 4 and 7. Thus, 3153A contains one normal Ch4 or Ch7 and two chimeric chromosomes 4/7a and 4/7b that combine portions of Ch4 or Ch7. In addition, we show that in Sor55 the hybrid Ch4/7b harboring SOU1 has been duplicated, resulting in trisomy for the SOU1 gene.

Materials and Methods

Strains, genes and primers

The Sou⁻ laboratory strain 3153A and its Sou⁺ derivatives Sor55 (monosomic for Ch5b) or Sor60 (monosomic for Ch5a) were previously reported (Janbon et al., 1998; Kravets et al., 2010). The strain SC5314 is the C. albicans sequencing strain [see Candida Genome Database (CGD) (http://www.candidagenome.org/]. Genes and primers used in this work are listed in Table S1 in Supporting Information.

Comparative genome hybridization microarrays (aCGH)

Roche NimbleGen Inc., (Madison, WI) designed and provided us with custom tiling DNA microarrays for the genome supercontigs from genomic assembly 19 of the C. albicans reference strain SC5314, as found in CGD (http://www.candidagenome.org/). Each array contained a total of 710,907 50-mer probes, tiling at every 35 nucleotides. Data were obtained from two-channel experiments with three independent batches of total DNA prepared for each strain, 3153A or Sor55. Raw data are available at Gene Expression Omnibus (GEO) (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=) with the accession number GSE21616. Changes of DNA copy number by aCGH were calculated from two-channel experiments, as averaged ratios Sor55/3153A using either the probes tiling across the entire genome sequence (Kravets et al., 2010) or the probes tiling across each ORF. See Kravets et al. (2010) for more details about the microarrays, the DNA hybridization to microarrays as well as handling of chips.

For graphic presentation of DNA profiling of Sor55, a scatter plot was prepared using Microsoft Excel. ORF annotations and chromosome positions were updated using genome assembly 21 of the reference strain SC5314, as found in CGD. Averaged ratios for probes tiling across each ORF were plotted according to the ORF chromosomal positions.

Pulse-field gel electrophoresis (PFGE) analysis

We used various PFGE running conditions to optimize separation of chromosomes with different sizes as previously published by Ahmad et al. (2008). Precisely separated portions of each electrophoretic karyotype were presented as composite schematics, reproducing the complete PFGE chromosome banding pattern.

Various procedures

Southern blot analysis, preparation of primers or probes, and procedures that diminish undesirable chromosome instability were carried out, as previously reported (Ahmad et al., 2008).

Results and Discussion

Identification of chimeric chromosomes 4/7a and 4/7b in the parental strain 3153A

In prior studies, the chromosome banding pattern of the laboratory strain 3153A was assigned to eight chromosomes using PFGE and Southern blot hybridization techniques (reviewed in Rustchenko & Sherman, 2002). Here, we determined the sizes of some chromosomes of 3153A from multiple PFGEs according to (Ahmad et al., 2008) (Fig. 1A). 3153A was separated side by side with the C. albicans reference strain SC5314. Chromosome sizes of SC5314 due to PFGE were in good agreement with chromosome sizes due to sequencing that are available from the genomic assembly 21 in CGD (Fig. 1A).

Fig. 1. Schematic chromosome patterns of strains 3153A, its aneuploid derivative Sor55, and reference strain SC5314.

A. Horizontal bars represent the chromosomes of the strains, as indicated. Chromosomes are designated from 1 to 7 and R, on the left, as their sizes decrease from top to bottom. ChR refers to the chromosome containing a cluster of tandemly repeated rDNA units. Homologous chromosomes are indicated with “a” and “b” and are shown, as a single bar if their sizes are identical. However, individual bars conveniently present equally sized Ch6b and Ch4/7b of 3153A and Sor55, as well as the duplicated Ch4/7b of Sor55. Single size for each chromosome of SC5314 was retrieved from C. albicans genome assembly 21 of CGD. Sizes of some, but not all chromosomes of SC5314 or 3153A, were also determined using PFGE (Ahmad et al., 2008). The sizes from assembly 21 are always smaller than the sizes from PFGE, presumably due to an incomplete representation of various repetitive sequences. SOU1 is indicated by a black dot. B. Positions of seven markers of Ch4 and four markers of Ch7 are indicated in SC5314, 3153A, and Sor55 as follows: 1, RIS1; 2, DAO2; 3, MET4; 4, ECM1; 5, LYS1; 6, BRN1; 7, SOU1; 8, SAC7; 9, RPA135; 10, GPH1; and 11, LOS1. Note that a reciprocal translocation between Ch4 and Ch7 in 3153A resulted in two chimeric chromosomes, Ch4/7a and Ch4/7b (Fig. 1A), of which Ch4/7b duplicated in Sor55 (Figs. 1A and B). Arrows indicate break points and junctions. ♦ - indicates centromeres.

Previous limited Southern blot analysis indicated that one copy of the SOU1 gene in 3153A was translocated from Ch4 to Ch7 (Janbon et al., 1998) (Fig. 1A). We have now extended this analysis by using seven Ch4 markers, RIS1 (orf19.5675); DAO2 (orf19.3365); orf19.5312; ECM1 (orf19.5299); LYS1 (orf19.1789.1); BRN1 (orf19.1251); and SOU1; as well as four Ch7 markers, SAC7 (orf19.7115); RPA135 (orf19.7062); GPH1 (orf19.7021); and LOS1 (orf19.7153). These markers are denoted, respectively, from 1 to 7, Ch4, or from 8 to 11, Ch7, and are presented according to chromosome positions in the reference sequencing strain SC5314 (Fig. 1B). The markers were used to prepare probes and to hybridize them to chromosomal blots from PFGE (Fig. 2).

Fig. 2. Southern blot analyses of chromosomes of the parental strain 3153A and its derivatives Sor55 or Sor60.

Various PFGE running conditions were used to separate chromosomes of 3153A and its mutants Sor55 or Sor60 that are monosomic for Ch5b or Ch5a, respectively. The gels on the left and on the right contain well separated Ch7, Ch6, and Ch5, as indicated; the remaining chromosomes are compressed on the top of each gel. Ch7, Ch6, Ch5, and Ch4 banding pattern is shown schematically at the extreme right. Dashed or continuous lines represent single or co-migrating chromosomes, respectively. The low-resolution gel in the middle shows separation of all chromosomes. Also shown are hybridization signals that were obtained from blots of the corresponding shown gels with the indicated probes. See Fig. 1B and its legend, as well as Results and Discussion for the explanation of the chromosomal positions of the probes.

This approach revealed a reciprocal exchange between one Ch4 and one Ch7 in the strain 3153A. As a result, two chimeric chromosomes Ch4/7a and Ch4/7b were formed leaving one intact Ch4a or Ch7a (Figs. 1A and B). Two chromosomal bands of 3153A that are denoted here Ch4/7a and Ch4/7b, hybridized with markers from both Ch4 and Ch7 (Fig. 1B). A total of five Ch4 markers, orf19.5312 (3); ECM (4); LYS1 (5); BRN1 (6); and SOU1 (7); showed, as expected, one signal with an intact Ch4a, but another signal from a distinct band, which we designate Ch4/7b, as exemplified by SOU1 (7) in Fig. 2 (see also schematics in Fig. 1B). The bands for Ch4/7b and Ch6b co-migrate in the PFGE, hence these two chromosomes are of approximately the same size. The remaining two markers of Ch4, RIS1 (1) and DAO2 (2), showed one signal, as expected, from Ch4a and another from the band which we now designate Ch4/7a (not presented, but see schematics in Fig. 1B). Also, the marker of Ch7 LOS1 (11) showed one signal, as expected, from an intact Ch7a, but another signal from the Ch4/7a (Fig. 2).

The remaining three markers of Ch7, SAC7 (8); RPA135 (9); and GPH1 (10); showed one signal, as expected, from Ch7a and another from Ch4/7b, as exemplified by RPA135 in Fig. 2. We do not know why the translocation occurred and which phenotypes it controls. The translocation may be not relevant to sorbose utilization. Interestingly, reciprocal exchange between Ch4 and Ch7 is similar to that previously reported in the strain NUM1000, as compared to the strain 1006 (Chibana et al., 2000), with roughly the same portion of Ch7 and a slightly larger portion of Ch4 involved in 3153A.

Chimeric Ch4/7b of 3153A is duplicated in the Sou⁺ strain Sor55 monosomic for Ch5

We have previously shown that the Sou⁺ strain Sor55, which arose from 3153A on solid medium containing sorbose as the sole carbon source has lost Ch5a while retaining Ch5b (Janbon et al., 1998; Kravets et al., 2010). Here, we continued analyzing Sor55 by Southern blot and found that, in addition, this strain has acquired trisomy of the chimeric Ch4/7b. This is shown by all analyzed markers residing on Ch4/7b (Fig. 1B), as exemplified with SOU1 (7) and RPA135 (9) in Fig. 2. In order to address generality of Ch4/7b trisomy, we analyzed another representative Sou⁺ strain Sor60, also derived from 3153A but retaining Ch5a while losing Ch5b (Kravets et al., 2010). We found no duplication of Ch4/7b using the same probes that were applied to Sor55 (see examples of hybridization signals in Fig. 2).

In order to clarify the condition of Ch6, we generated two signals with Ch6 markers, NAG1 (orf19.2156) and ALS2 (orf19.1097) that hybridized, as expected, with Ch6a and b producing signals with equal brightness (not shown). This result confirmed that Ch6b co-migrates in PFGE with the duplicated Ch4/7b of the mutant Sor55 (see the high resolution separation, as well as schematics of the banding pattern of the small Ch7, Ch6, and Ch5 in Fig. 2).

Confirmation of duplication of Ch4/7b by aCGH

To confirm and quantify Ch4/7b duplication indicated by Southern blot analysis, and to more thoroughly examine the ploidy of the remainder of the genome in the Sor55 strain, we hybridized genomic DNA of Sor55 and parental 3153A to tiling arrays covering the entire C. albicans genome and we determined changes of DNA levels in Sor55 as compared to the parental strain 3153A (Materials and Methods). Fig. 3 presents the graphic image of DNA profiling data.

Fig. 3. DNA profiling of the mutant Sor55 by tiling aCGH.

The mean aCGH ratio Sor55/3153A for each ORF is plotted according to the ORF chromosome position in the C. albicans genome assembly 21 in CGD. X axis indicates the position of the probes on each chromosome in the reference strain SC5314, as annotated in genomic assembly 21 in CGD. Y axis is the average ratio of each probe. Note a massive duplication of DNA across the 749 kb segment of Ch4 and the 236 kb segment of Ch7. These segments correspond to the chimeric Ch4/7b in 3153A that is duplicated in Sor55 (Figs. 1A and B; Results and Discussion). Ch5 monosomy in Sor55 is also shown, as previously reported by Kravets et al. (2010).

Consistent with the above Southern blot analysis, Sor55 showed a clear 1.5 fold increase of DNA, which is presented in Fig. 3 as two segmental duplications, one across 749 kb of Ch4, from 845 kb to the right telomere, and another across 236 kb of Ch7, from the left telomere to 236 kb. Combined, the 985 kb size approximately corresponds to the length of Ch4/7b (1,051 kb) in 3153A or Sor55, as determined by PFGE (Fig. 1A). This is because Ch4/7b does not exist as an independent entity in reference SC5314, which was used to assign aCGH probes to chromosomal positions. In SC5314, sequences that comprise Ch4/7b are part of Ch4 and Ch7 (see Fig. 1B). Apparently, in 3153A, the breaks occurred on the left arms of one Ch4 at 854 kb and one Ch7 at 236 kb followed by a reciprocal exchange and formation of the chimeric Ch4/7a and Ch4/7b of different sizes (Fig. 1B). The larger chimeric Ch4/7a, which is comparable in size to Ch4a, combined the 854 kb Ch4 portion and approximately the 715 kb Ch7 portion including the Ch7 centromere. The smaller chimeric Ch4/7b, which is comparable in size to Ch7a, combined the 236 kb Ch7 portion and the 749 kb Ch4 portion including the Ch4 centromere. As a result, the SOU1 gene remains on the right arm of Ch4a and appears on the smaller chimeric Ch4/7b (Figs. 1A and B).

Approximately 91% of Ch4/7b genes, 321 out of a total of 353, exhibited DNA ratios in the range of 1.4 to 1.6, which is strongly indicative of 1.5-fold increase of DNA amount or trisomy, as also determined with the Southern blot analysis. The remaining genes, 18 within the range 1.2–1.3 and 13 within the range 1.7–1.8, may represent variability in hybridization signal intensities. A single gene, orf19.3074, showed no increase of DNA, 1.0. This gene could reside on a chromosome other than Ch4/7b in 3153A and its derivatives.

Of special note, because the high throughput aCGH method does not show the genome position of the segmental aneuploidy, aCGH needs to be combined with Southern blot analysis of PFGE separated chromosomes, as we have done.

The remaining chromosomes of Sor55, except the previously reported monosomic Ch5 (Kravets et al., 2010), showed no change in the DNA content, except for a few scattered genes (Fig. 3). Identical results were obtained when all 710,907 probes tiling across the genome, not only probes tiling across ORFs, were prepared as scatter plots and analyzed (data not shown).

We have previously produced evidence that Ch5 monosomy alone is sufficient to render the Sou⁺ phenotype, i.e., resistance to sorbose and sorbose utilization as a carbon source. Our study included Ch5 monosomic derivatives from two different strains, 3153A or CAF4-2; the latter being an auxotrophic descendent of the reference strain SC5314 (reviewed in Rustchenko, 2007). This study supports this generalization, as we found an additional duplication of Ch4/7b rendering an extra-copy of SOU1 in the Ch5 monosomic derivative Sor55, but not in a representative derivative Sor60. We regard Ch4/7b trisomy in Sor55 as a facultative alteration facilitating the Sou⁺ phenotype due to the Ch5 monosomy, which is the major mechanism for growth on sorbose.

We have previously studied the properties of SOU1, which encodes sorbose reductase catalyzing the first step in the sorbose catabolic pathway (Greenberg et al., 2005) and is responsible for the Sou⁺ phenotype (Janbon et al., 1998). Introduction of an extra copy of SOU1 into a normal diploid strain with two copies of Ch5 does not result in growth on sorbose, Sou⁺. Such growth, however, does occur in the presence of multiple copies of SOU1 that are either integrated in genome or overexpressed from plasmids (Janbon et al., 1998; Greenberg et al., 2005; reviewed in Rustchenko, 2003). Most importantly, SOU1 is reversibly up regulated in cells with the monosomic Ch5, as shown by Northern analysis using two independent Sou⁺ strains that are monosomic for Ch5 and are different from Sor55 (Janbon et al., 1998), as well as their derivatives with duplicated Ch5. Expression profiling of Sor55 consistently showed 2.4-fold up regulation of SOU1, as determined with method described by Kravets et al. (2010). Raw data for the expression profiling of Sor55 are available at GEO with the accession numbers GSM455127–GSM455132.

As demonstrated here with the mutant Sor55, C. albicans possesses multiple strategies to control the metabolic gene SOU1 that is important for survival on and utilization of toxic sorbose. In addition to the well-documented up regulation of normal two copies of SOU1 by loss of one Ch5, an additional copy of SOU1 can be furnished by duplicating the Ch4/7b. Most probably, an extra-copy of SOU1 is relevant to sorbose resistance; however, the relative contributions of each copy of SOU1 or Ch5 monosomy and Ch4/7b duplication to sorbose resistance remain elusive. We believe that the survival strategies involving a concomitant aneuploidy of different chromosomes could inform about mechanisms of resistance due to caspofungin and other echinocandins.