Widespread occurrence of chromosomal aneuploidy following the routine production of Candida albicans mutants

Abstract

It has come to our attention that approximately 35% of >100 published microarray datasets, where transcript levels were compared between two different strains, exhibit some form of chromosome-specific bias. While some of these arose from the use of strains whose aneuploidies were not known at the time, in a worrisome number of cases the recombinant strains have acquired additional aneuploidies that were not initially present in the parental strain. The aneuploidies often affected a different chromosome than the one harboring the insertion site. The affected strains originated from either CAI-4, RM1000, BWP17 or SN95 and were produced through a variety of strategies. These observations suggest that aneuploidies frequently occur during the production of recombinant strains and have an effect on global transcript profiles outside of the afflicted chromosome(s), thus raising the possibility of unintended phenotypic consequences. Thus, we propose that all Candida albicans mutants and strains should be tested for aneuploidy before being used in further studies. To this end, we describe a new rapid testing method, based on a multiplex quantitative PCR assay, that produces eight bands of distinct sizes from either the left or right arms of each C. albicans chromosome.

Fig. 1

Example of chromosomal bias in published transcriptional profiles. Each spot represents fluorescence ratio data (log 2) from genes that were ordered according to their position on the eight Candida albicans chromosomes. The top panel represents a comparison between a Δmkc1 strain and its CAI4 parental strain (Oberholzer et al., 2006). The bottom panel shows the profile obtained from a comparison between a ‘white’ morphology strain expressing the α1, α2, a1 and a2 MTL transcriptional regulators and its CAI4 parental strain at 23°C (Tsong et al., 2003).

Fig. 3

Aneuploidies can affect transcriptional profiles outside of the afflicted chromosomes. (a) Fluorescence intensities in a transcriptional profiling experiment of one of three Δfun31 mutants compared with a DAY185 control strain. Downregulation of Chr R genes is apparent. (b and c) Scatter plots showing the similarity of transcriptional profiles between the genes outside of Chr R in a comparison of two individual Δfun31 mutants lacking a copy of Chr R (b), or a Δfun31 mutant lacking a Chr R, when compared with a control strain without the Δfun31 mutation and an equal number of chromosomes. Spots represent the fluorescence ratios of 486 genes in Chr 1–7 that had a 1.5-fold change or more in at least one experiment. R² values represent the similarities in the profiles between the two strains and indicate that the aneuploid strains produce profiles that are more similar to each other.

Fig. 2

Example of different aneuploidies from two distinct colonies. These graphs represent the fluorescence ratios (log 2) from individual probes in a CGH comparing one of two colonies expressing the HA-Rfg1p transcription factor with a CAI4-pCaEXP empty-vector control strain that had previously been confirmed to have two copies of each chromosome. While the fold change in CGH should be expected to be at least 1.5-fold for a triploid vs. diploid comparison, we note that the quantarray software used to quantify our microarrays tends to underestimate fluorescence ratios.

Fig. 4

Aneuploidy detection with a multiplex PCR assay. (a) Bioanalyzer profiles of multiplex PCR reactions using primer set A (left panel) or primer set B (right panel). We used as templates genomic DNA from either a validated SC5314 control strain, a Δnrg1 strain (with extra copies of Chr 2 and 4) or the BWP17 strain carrying a heterozygous deletion on the right arm of Chr 5. Images of the profiles were scaled to similar sizes, thus allowing the identification of amplicons with a different abundance (arrowheads). In (b), the multiplex PCR assay was conducted with primer set A on genomic DNA from strains SC5314, DkCa169 (Legrand et al., 2008) and SZY20 (Znaidi et al., 2007). Graphs represent the mutant/SC5314 ratio of the median normalized peak heights on the X-axis and each chromosome on the Y-axis. A log₂ ratio above 0.2 (in red) was considered to be significant and indicative of aneuploidy for these chromosomes.