The role of aneuploidy in the emergence of echinocandin resistance in human fungal pathogen Candida albicans

Fungal diseases largely affect human and animal health and dramatically diminish food crop yields [1]. Among fungi, systemic Candida infections are the second or third most common pathogens isolated from blood cultures in the USA [2]. Candida albicans is still the predominant Candida species, causing up to 50% of candidemia despite an increase in diversity of Candida species isolated from clinical samples [3]. In healthy individuals, C. albicans is a harmless inhabitant of mucosal surfaces throughout the body. However, in immune-compromised individuals, C. albicans can become a dangerous pathogen, causing severe or even fatal infections. In this review, we summarize recent data linking the reduced susceptibility of C. albicans cells to mainline echinocandin (ECN) drugs to aneuploidies of chromosomes 5 (Ch5) and Ch2.

resistance in a series of isogenic clinical isolates of the related species Candida glabrata, they demonstrated that mutation to resistance was preceded by a mutation in a different gene that conferred a relatively small increase in MIC [14]. Indeed, quickly arising mutations conferring decreased susceptibility to ECNs are currently viewed as a means to provide a window of opportunity for temporary survival and subsequent formation of resistant FKS1 mutations. Based on such observations, Healey and Perlin [15] proposed a multistep model in which spontaneously acquired mutations lead to some decrease of drug susceptibility prior to acquisition of FKS1 resistance mutations and thus play an important role in the evolution of ECN resistance (Fig 1).
Subsequently, given their importance, some of the genes that confer decreased susceptibility have been identified by screening C. albicans deletion libraries against the ECN caspofungin [16]. In addition, such genes were also identified in mutants that were caspofungin-generated in vitro [17]. Clearly, understanding mechanisms that promote ECN resistance by influencing ECN susceptibility is of high importance. Although currently the incidence of clinical resistance to ECN drugs is relatively low, it is persistent, and the number of resistant cases is growing, primarily due to the increased use of ECNs for disease and prophylactic treatment.

Spontaneous aneuploidies of C. albicans control vital functions
C. albicans has a diploid genome, which is organized into eight pairs of chromosomes that are known for their instability (reviewed in [18]). Aneuploidy, defined as a change in the number of chromosome(s) or large portion of a chromosome, is well tolerated in fungi, including C. albicans. Spontaneous aneuploidy can be found in populations of C. albicans cells at high frequencies, between c. 10 −4 and 10 −2 , with a clear tendency to increase under external stresses, and seems to be a basic property of this microbe [19]. It has been demonstrated that any chromosome of C. albicans can become aneuploid [18]. Various aneuploidies introduce diversity  in a population of cells by controlling vital physiological functions, such as, for example, utilization of different carbon and nitrogen sources [20].

Specific aneuploidies of C. albicans Ch5 or Ch2 control adaptation to ECNs
In earlier work, the Rustchenko group demonstrated that C. albicans employs reversible alterations of specific chromosomes to adapt for growth in the presence of toxic agents that kill cells or prevent cell propagation, including fluconazole, 5-fluoro-orotic acid, and the toxic sugar L-sorbose (Table 1) (reviewed in [18]). Interestingly, adaptation to utilize the secondary sugar D-arabinose also relied on specific alterations of two specific chromosomes [18,21]. Recently, similar experiments demonstrated that reversible aneuploidies of Ch5 or Ch2 can control adaptation to the ECN caspofungin (Table 1) [13,22].
Loss of one Ch5 results in reduced susceptibility to caspofungin (denoted Cas + ). In the absence of caspofungin, spontaneous duplication of the remaining Ch5 reverts Cas + cells to the original (parental strain) susceptibility [13] (Table 1). Loss and reduplication of Ch5 also controls resistance to the toxic sugar sorbose (denoted Sou + ) [24], which kills fungi via a mechanism similar to ECN drugs by inhibiting glucan synthase [25,26]. Interestingly, Ch5 monosomy also confers adaptation to the antifungal pyrimidine analog 5-fluorocytosin (denoted 5-Fl + ), as well as increased susceptibility to the azole antifungal fluconazole and the polyene antifungal amphotericin B (denoted Flu S and AmB S , respectively) [23] (Table 1). Another aneuploid state conferring a Cas + phenotype involves duplication of the right arm of Ch5 to create an isochromosome with two right arms (iso-Ch5R) (Fig 2). Note that these cells carry one iso-Ch5R and one normal Ch5, resulting in three right arms and one left arm of Ch5. Importantly, spontaneous duplication of the remaining normal Ch5 reverts cells to caspofungin susceptibility [13]. In contrast, duplication of the left arm of Ch5 resulting in iso-Ch5L with two left arms confers decreased susceptibility to fluconazole, Flu R (Fig 2), and is found in resistant isolates, as well as in vitro generated mutants [27]. Also, trisomy of Ch2 confers Cas + combined with adaptation to hydroxyurea, Hu + [22]. Of note, genes residing on Ch2 responsible for Cas + phenotype have not yet been identified. However, it is clear that the reversible duplication of Ch2, in contrast to reversible loss of one Ch5, implies that genes for positive regulation of the Cas + phenotype exist on Ch2. It is of interest to this review that exposure of C. albicans cells to chemotherapeutic drug hydroxyurea results in an Hu + phenotype due to Ch2 trisomy, similarly to exposure to caspofungin [22], the effect, which might contribute to development of Candida infection in chemotherapy-treated patient [28]. It is worth mentioning that occasionally a mutant adapts to sorbose via a large deletion within the right arm of Ch5,  Fig 2).

Complexity of regulation by aneuploidy
Long-term studies of reversible aneuploidy in response to challenge by caspofungin or the toxic sugar sorbose as a model system for C. albicans adaptation to ECNs have unraveled an astonishingly complex and multilayered regulation of the genes responsible for these phenotypes. These studies show that the rate of formation of (predominantly Ch5 monosomic) Sou + mutants per viable cell per day increased from 10 −6 at the initial time of detection to 10 −2 after four days of incubation on selective media [24]. These data suggest that nondisjunction of a Ch5 homologue can occur by different mechanisms and that preexisting and adaptive mutants occur by different processes; the latter ones possibly involving a nonmitotic mechanism. At least nine spatially separated, functionally redundant regions that negatively control resistance to sorbose were identified on Ch5 (Fig 2). These regions, which fall into two functionally redundant pathways I and II (Fig 2), bear no sequence similarity among themselves and four of them bear no similarity to any known sequence. The regions are thought to encompass CSU (Control of Sorbose Utilization) genes for negative control of sorbose resistance, three of which CSU51, CSU53, and CSU57 were identified in three corresponding regions (Fig 2) [29,31,32]. Most importantly, CSU51, a putative transcription factor, negatively controls both ECN susceptibilities and sorbose resistance [17]. It remains to be determined whether the two other known CSUs are also implicated with ECNs. The fact that CSU51 performs this dual function is remarkable, as it opens the possibility that many of the unique CSU regulators predicted to reside in different regions of Ch5 also have dual function. Furthermore, two genes, PGA4 and CHT2, which participate in cell wall construction and encode negative regulators of only ECN susceptibilities, have also been identified (Fig 2) [17]. However, the final number of either CSUs or genes involved only in ECN susceptibility remains elusive. It is tempting to explain the Cas + phenotype resulting from Ch5 monosomy as due to diminished gene dose of multiple negative regulators scattered across this chromosome.  length (1,295 kb); centromere (C); and telomeres (T). Also shown is the approximately 395 kb truncation of the right arm, and PGA4, CHT2, CNB1, and MID1 on the right arm, as well as TAC1 and ERG11 on the left arm. Note that duplication of the right arm resulting in iso-Ch5R with two right arms confers adaptation to caspofungin, Cas + [13]. While, duplication of left arm resulting in iso-Ch5L with two left arms confers adaptation to fluconazole, Flu R [30]; (B) various regions of Ch5, 133-135, and 137-142 that were cloned from Ch5 library and regions A, B, and C that were identified by their function. Also shown are CSU51, CSU53, CSU57 identified in three of the regions and antisense elements ASU51 and ASU53 [29,31,32]. Note that genes in red encode negative regulators of ECN susceptibility [17], while genes in blue encode positive regulators of fluconazole susceptibility [33]. Ch5, chromosome 5; ECN, echinocandin. https://doi.org/10.1371/journal.ppat.1009564.g002 However, the fact that the formation of iso-Ch5R, resulting in three right arms versus one left arm of Ch5, also results in diminished ECN susceptibility in corresponding cells implies that an additional or more complicated scenario can be in play. The right arm of Ch5 carries two genes, MID1 and CNB1, which encode positive regulators of ECN susceptibility [34,35]. Therefore, it is possible that the effects of various Ch5 ploidies are due to a balance between negative and positive factors expressed from this chromosome. From this point of view, the loss of one Ch5 diminishes the action of both negative and positive regulators expressed from this chromosome; negative regulators overriding positive regulators. Whereas, in the condition of iso-Ch5R, the amplified positive regulators on the right arm of Ch5 override the action of negative regulators on this arm.
As an example, Ch5 contains two key genes of interest, TAC1, a positive regulator of CDR1 and CDR2 genes encoding fluconazole efflux pumps on Ch3, and ERG11, a target gene for fluconazole (Fig 2A) [27]. Thus, loss of one Ch5, diminishing the copy number of TAC1 and ERG11, likely leads to increased sensitivity to fluconazole via loss of activity of these efflux pumps and the fluconazole target. In addition, as loss of one Ch5 also increases sensitivity to amphotericin B (see Table 1), it may be expected that Ch5 also carries genes for positive regulation of susceptibility to this drug. On a contrary, negative regulator(s) of antifungal 5-fluorocytosin could be also expected to reside on Ch5, as cells with monosomic Ch5 acquire 5-Fl + phenotype (see Table 1).
An additional level of regulation includes regulatory elements denoted ASU (Antisense regulators of Sorbose Utilization) that are embedded within CSUs in an antisense configuration (Fig 2). In respect to the CSU transcripts, the ASU long noncoding transcripts are completely overlapped by CSU transcripts, are in lesser amounts, and are inversely related. Presumably, ASU transcripts modulate CSU transcripts [31]. Some genes residing on aneuploid chromosomes are also controlled by transcriptional compensation for gene dosage, which keeps expression of select genes at or near the diploid level, irrespectively of chromosome ploidy [36,37]; MID1 and CNB1 exemplifying such genes (Fig 2) [13]. Indeed, widespread dosage compensation occurs across monosomic Ch5 and correlates with increased chromosome-wide acetylation of histone H4 [38]. This epigenetic feature involves the histone acetyltransferase complex NuA4, which could be a novel drug target to reduce the viability of resistant cells. On the other hand, decreased expression of some genes to the diploid level has been shown to occur within the trisomic Ch4/7 and correlates with increased acetylation of histone H3, but the histone acetyltransferases involved have not been identified [38].
Last, but not least, the C. albicans genome contains two genes, FKS2 and FKS3, which have considerable sequence identity with the key ECN resistance gene FKS1 but act as negative regulators of FKS1 [39]. While heterozygous deletion of the essential FKS1 gene results in increased ECN susceptibility, complete removal of either FKS2 or FKS3 results, in contrast, in decreased ECN susceptibility due to resultant overexpression of FKS1, leading to an increase in cell wall glucan. Other indications of the involvement of FKS2 and FKS3 in ECN susceptibility include wide variations in the expression levels of these genes relative to FKS1 in clinical resistant isolates, down-regulation in spontaneous laboratory mutants harboring FKS1 resistance mutations [40], as well as down-regulation in model mutants bearing monosomic Ch5 or iso-Ch5R but lacking FKS1 resistance mutations [13].
ECN susceptibility in C. albicans can be diminished by a limited number of distinct aneuploidies of Ch5 and Ch2, which is consistent with an earlier assumption that genes of C. albicans are distributed over chromosomes nonrandomly [29,41]. A number of genes and processes relevant to this control already have been elucidated and inform about much-needed potential drug targets. However, we are clearly at the beginning of an exciting journey into understanding the regulation of ECN drug susceptibility by chromosome aneuploidy, which involves a complex interplay between ratio of negative and positive regulators on Ch5 and Ch2, the regulation of FKS genes residing outside Ch5 and Ch2, additional factors involved in the complex regulation of genes on aneuploid chromosomes, and still unidentified genes and features. A better understanding of the control exerted by aneuploidies will help to better understand evolution of ECN drug resistance and will facilitate the identification of new drug targets.