ALS51, a Newly Discovered Gene in the Candida albicans ALS Family, Created by Intergenic Recombination: Analysis of the Gene and Protein, and Implications for Evolution of Microbial Gene Families

Abstract

The Candida albicans ALS family has eight genetic loci, each encoding a large glycoprotein. Als protein function is discussed most frequently in terms of adhesion to host and abiotic surfaces. Analyses of C. albicans strain WO-1 indicated variation within the ALS1 locus compared to other isolates such as SC5314. Investigation revealed recombination between the contiguous ALS5 and ALS1 loci to generate a new coding region, named ALS51, because it encodes the 5' domain of ALS5 fused in-frame to the tandem repeat region and 3' domain of ALS1. ALS51 was detected in 11 isolates (4.6%) from a collection of 239 C. albicans strains of diverse origin and clade assignment. The 12 ALS51-positive strains identified in this study represented three different ALS family genotypes with respect to the presence and copy number of ALS51, ALS5 and ALS1. ALS51 transcription was detected by real-time RT-PCR in WO-1. Although the cell-surface abundance of Als51 on WO-1 and Als5 on SC5314 was too low to visualize by indirect immunofluorescence using an anti-Als5 monoclonal antibody, both proteins were observed on Western blots of beta-1,6-glucanase-digested C. albicans cell walls. Characterization of ALS51 illustrates one of the recombination mechanisms that generate diversity within C. albicans gene families.

Introduction

Candida albicans is an opportunistic fungal pathogen that causes oral and vaginal mucosal infections as well as systemic disease. C. albicans has several gene families that encode proteins involved in host-pathogen interactions (Jones et al., 2004). Among these is the ALS (Agglutinin-Like Sequence) family that encodes large cell-surface glycoproteins most frequently considered for their role in adhesion to host and abiotic surfaces (reviewed in Hoyer et al., 2008). ALS genes share a similar basic organization consisting minimally of a relatively conserved 5’ domain, a central domain of tandemly repeated sequence units, and a 3’ domain of relatively variable length and sequence. ALS genes are located on three of the eight C. albicans chromosomes (Hoyer et al., 2008). Analysis of C. albicans strain collections, as well as completion of the genome sequence of strain SC5314 (Jones et al., 2004), suggested that there are eight different ALS genes (Hoyer et al., 2008). Within these eight distinct genetic loci is a considerable degree of sequence variation, most commonly observed in the number of copies of the tandemly repeated sequence units included in the central domain. Sequence variations within the 5’ and 3’ domains of ALS alleles have also been described (Hoyer et al., 2008).

A current experimental priority is development of a monoclonal antibody (mAb) specific for each Als protein to investigate cell-surface localization patterns (Coleman et al., 2009). Development of an anti-Als1 mAb and immunolabeling of C. albicans strains of diverse clade and origin showed that while Als1 is obvious on the surface of yeast forms after inoculation into fresh culture medium, no labeling was observed on strain WO-1 (Coleman et al., in press). WO-1 is the original white-opaque phenotypic switching strain described by Slutsky et al. (1987) and a strain frequently used in experiments that explore the molecular biology of C. albicans mating (reviewed in Soll, 2009). Historic observations suggested differences in ALS1 in strain WO-1 compared to other isolates. For example, ALS1 transcript could not be detected in strain WO-1 grown under conditions that induced ALS1 expression in other C. albicans isolates (Hoyer et al. 1998). Southern hybridizations demonstrated that the sequences immediately 5’ of ALS1 were absent in strain WO-1 despite the presence of genomic sequences from the ALS1 3’ domain (Hoyer et al. 1998). At the time of those observations, it was suggested that ALS1 in WO-1 might be under control of different regulatory mechanisms than in the other C. albicans isolates. Genome sequence data provided further insights into the ALS1 locus in different isolates. In the strain SC5314 genome sequence, the coding regions for ALS5, ALS1 and ALS9 are adjacent to each other on chromosome 6, and all transcribed in the same direction (Zhao et al., 2003). Genome sequence assembly for strain WO-1 failed in this region (http://www.broadinstitute.org/annotation/genome/candida-albicans/MultiHome.html). The sum of previous observations made at both the DNA and protein level suggested that strain WO-1 is different from most other strains at the ALS1 locus. The goal of this work was to determine the differences between strains SC5314 and WO-1 in this chromosomal region and explain the previous experimental results that were obtained for strain WO-1. The work led to identification of a new Als protein, Als51, and to insights regarding methodology for detection of Als proteins on the C. albicans cell surface and evolution of the ALS gene family.

Materials and methods

Fungal strains and culture conditions

C. albicans strains SC5314 (Gillum et al., 1984) and WO-1 (Slutsky et al., 1987) were described previously and were used for the majority of the studies. Strain 163 is an oral isolate from a normally healthy human. A collection of 239 C. albicans isolates was assembled from the collections described by Zhao et al. (2007b) (Collection A) and Wrobel et al. (2008) (Collection B). Strains in Collection A were from three populations previously analyzed by Ca3 fingerprinting (Wrobel et al., 2008; Pujol et al., 1997; Pujol et al., 2002). The geographic origin and clade distribution of Collection A was described previously (Zhao et al., 2007b). Clade status of most of the strains was confirmed by multilocus sequence typing (MLST) and included clades 1, 2, 3, 4 and 11 (Odds et al., 2007). Collection B was comprised of human oral isolates collected from residents of the Midwestern United States, as well as isolates from wildlife species from the same geographic area (Wrobel et al., 2008). Isolates, including some that were not described in the published analysis, were assigned to clades by MLST. Clades represented were 1, 2, 3, 4, 5, 7, 8, 9, 10, 11 and 12. All C. albicans strains were stored as glycerol stocks at −80°C and, when needed, streaked to YPD agar plates (per liter: 10 g yeast extract, 20 g peptone, 20 g glucose, 20 g Bacto agar). Plates were incubated for 24 h at 37°C and then stored at 4°C for no longer than one week. Liquid cultures routinely were grown in YPD medium. The opaque form of strain WO-1 was selected from colonies grown on modified Lee’s medium agar plates (Bedell & Soll, 1979). Opaque colonies were identified by their pink color on plates containing 5 µg ml⁻¹ of Phloxine B. A single opaque colony was inoculated into 20 ml of modified Lee’s medium (pH 4.5) and the culture grown at 25°C for 48 h. Cells were observed microscopically to verify the elongated shape of opaque cells prior to fixation in paraformaldehyde and immunolabeling (see below).

To demonstrate the capabilities of an anti-Als5 monoclonal antibody in immunolabeling the fungal cell surface, two strains that were constructed over 10 years ago but never published, were utilized from the Hoyer lab culture collection. Strain 884 is Saccharomyces cerevisiae that displays an Als5-Agα1 fusion protein on its surface. All PCR amplifications for strain construction used Pfu polymerase (Stratagene) according to the manufacturer’s instructions. The 5’ domain of ALS5 (approximately 1.3 kb) was amplified from cloned ALS5 from strain 1161 (GenBank accession number AF068866) using primers ALS5Xho (5’ CCC CTC GAG ATG ATT CAA CAA TTT ACA TTG TTA TTC C 3’) and ALSCla3 (5’ TTG TAC CAC CAC TGT ATC GAT TGA ATC AGT 3’). The 3’ half of AGα1 (approximately 1 kb) was amplified from S. cerevisiae genomic DNA using primers ALSCla5 (5’ TCA ATC GAT ACA GTG GTG GTA CAA GGT ACA GCT AGC GCC AAA AGC TCT 3’) and AGA1Bgl (5’ CCC AGA TCT TTA GAA TAG CAG GTA CGA CAA AAG CAG 3’). These primers also encoded a portion of ALS5 that overlapped with the first PCR product (above). PCR products from both reactions were extracted with phenol/chloroform and precipitated with ethanol. Resuspended DNA from each PCR was combined into a single amplification reaction with primers ALS5Xho and AGA1Bgl. The resulting PCR product had a 5’ XhoI site, sequences encoding the N-terminal 430 amino acids of Als5 (with an in-frame ClaI site at the sequence encoding amino acids 424 and 425), sequences encoding the C-terminal 323 amino acids of Agα1 (amino acids 328 to 650), and a 3’ BglII site. XhoI-BglII-digested fragment was ligated into identically cut plasmid p138NB (Hoyer et al., 1998). The resulting construct produced the Als5-Agα1 fusion protein under control of the copper sulfate-inducible CUP1 promoter. The accuracy of the construct was verified by DNA sequencing. Methods for transformation of the plasmid into S. cerevisiae YPH274, selection on SC-Trp agar plates, growth and induction of protein production in SC-Trp medium using CuSO₄ were detailed previously (Hoyer et al., 1998). Yeast cells were harvested from the culture, and fixed in paraformaldehyde as described previously (Coleman et al., 2009) for use in indirect immunofluorescence analysis with the anti-Als5 mAb (see below).

Strain 1002 is C. albicans that produces full-length Als5 under control of the high-level constitutive TPI1 promoter. A plasmid construct for high-level constitutive gene expression, based on pRC2312 (Cannon et al., 1992), was described previously (Green et al., 2005a). This plasmid, pRC2312/TPI encodes the TPI1 promoter and terminator sequences separated by a XhoI-SmaI-NotI-BglII polylinker. ALS5 was amplified from cloned strain 1161 DNA using primers ALS5Xho and ALS5Bgl (5’ CCC AGA TCT TCA TAG AAA GAA GAA TAA TGC AAC GCT AAT AA 3’). The PCR product was digested with XhoI-BglII and cloned into the identically digested pRC2312/TPI vector. The resulting plasmid was transformed into strain CAI4 (Fonzi & Irwin, 1993) and selected for uridine prototrophy. Consistent production of cell-surface Als5, even after growth in non-selective conditions, suggested integration of the plasmid into the C. albicans genome. Strain 2373 (Zhao et al., 2007a), from which both ALS5 alleles are deleted, was also used in this work.

Amplification and subcloning of WO-1 genomic DNA

PCR primers based on genomic sequence from strain SC5314 were used to attempt amplification of the strain WO-1 region of chromosome 6 that encodes ALS5 and ALS1. Pfu Turbo polymerase (Stratagene) was used according to the manufacturer’s instructions to ensure high fidelity of the resulting product. Primers ALS5upF (5’ CCC CCT AGG ACC AGC ATT GTC AAT CGA ACC A 3’) and ALS1dnR (5’ CCC GCC GGC TTG GTT TAA CTT TCT TGA CTG G 3’) amplified a 4040 bp product (Fig. 1). The product was purified on an agarose gel using GeneClean III (MP Biomedical), cloned into the pJet1.2/blunt vector (Fermentas) and transformed into E. coli TOP10 (Invitrogen). The resulting strain was named 3052. The cloned fragment was digested with BglII to release it from the pJet1.2/blunt vector, and then with NdeI to create three smaller fragments for DNA sequencing. Blunt ends were created on the restriction fragments using Pfu polymerase and dNTPs. Blunted fragments were subcloned into pJet1.2/blunt as described above. DNA sequences of the resulting subcloned fragments were derived using commercial primers pJetF and pJetR at the W.M. Keck Center for Comparative and Functional Genomics, University of Illinois. DNA sequences from the fragments were assembled according to the orientation of sequences in the strain SC5314 genome sequence (http://www.candidagenome.org). Custom DNA sequencing primers were designed from this preliminary assembly and used to verify the sequence of the entire fragment in clone 3052. The completed sequence of clone 3052 encoded two NdeI sites (nt 1110, within the 5’ domain of the gene, and nt 2097, near the end of the tandem repeat domain), consistent with production of three subcloned fragments that were used to create the preliminary sequence assembly. Overlapping double-stranded DNA sequencing of clone 3052 provided a high degree of confidence that the completed sequence was correct. The completed sequence was deposited into GenBank under accession number GQ149128, designated as ALS51-2, the smaller allele (fewer tandem repeat copies) from strain WO-1.

Fig. 1.

(A) Line drawing of the region of chromosome 6 that encodes ALS5 and ALS1. The size of this region is variable between C. albicans strains and often between alleles in the same strain (Hoyer et al., 2008). The example shown here is adapted from drawings of the ALS alleles on one copy of chromosome 6 in strain SC5314 (Zhao et al., 2003). Recombination between the ALS5 and ALS1 coding regions resulted in formation of a new gene, ALS51. Numerals in parentheses indicate the position of PCR primers used to amplify the 4040-bp genomic fragment encoding the 3240-bp ALS51-2 (1 = ALS5upF, 2 = ALS1dnR). Primer (3) is ALS5repF and primer (4) is ALS1GenoR; these primers were used in the analysis in (D) below. (B) Alignment of ALS51 allelic sequences from strain WO-1 with the sequence of an ALS5 allele from strain SC5314 (GenBank accession number AY227439) and ALS1 from strain 1161 (GenBank accession number L25902). The region in which recombination between ALS5 and ALS1 occurred to form ALS51 is featured. Asterisks are used to mark positions where all four nucleotide sequences are identical. Nucleotides shown in gray in the ALS1 sequence are those that do not match ALS51 or ALS5. Underlined nucleotides in bold type at position 912 indicate where one ALS51 allele matches ALS5 and the other matches ALS1. Underlined bold nucleotides at position 945 indicate the point where ALS51 alleles match ALS1, and not ALS5. (C) Graphical depiction of ALS5 and ALS1 sequence identity within the 5’ domain of each gene. The sequences of the 5’ domains from ALS5 and ALS1 were aligned. A window size of 10 nucleotides was used to calculate percent sequence identity at each position. The resulting graph demonstrates areas of sequence conservation and divergence within the 5’ domain and shows that recombination to form ALS51 (5’ of nucleotide 945) occurred immediately 5’ of the long region of 100% identity. (D) Ethidium bromide-stained agarose gel of PCR products resulting from amplification of genomic DNA from strain WO-1 or SC5314 with primers ALS5repF and ALS1GenoR. The location of these primers is shown in (A). The primers anneal to the 5’ domain of ALS5 and the 3’ domain of ALS1, respectively, and amplify the tandem repeat domain. Two fragments were amplified from WO-1 indicating the presence of two ALS51 alleles that differed in the number of copies of the tandemly repeated 108-bp sequence in the central domain of the gene. No products were amplified from SC5314 because, in this strain, the ALS5 and ALS1 sequences were too far apart for successful amplification using a standard protocol (see Materials and methods). Molecular size markers (in bp) are indicated to the left of the image.

To provide comparisons between ALS51 alleles from strain WO-1, the larger ALS51 allele was amplified and cloned in a similar manner. Amplification used primers ALS51cdF (5’ ATG ATT CAA CAA TTT ACA TTG TTA TTC 3’) and ALS51cdR (5’ CTA AAT GAA CAA GGA CAA TAA TGT GAT 3’). The DNA sequence of the cloned fragment was determined using the custom primers designed for ALS51-2. The DNA sequence was completed on both strands, with the exception of the tandem repeat domain, which was too long to sequence without subcloning. Subcloning was accomplished by ScaI digestion, which cut 5’ of the tandem repeat domain and within the tandem repeat domain to create a fragment of 1198 bp that was cloned into pJet1.2/blunt for DNA sequencing. The ALS51-1 sequence was deposited into GenBank under accession number HM164053.

Detection of ALS51 transcription

C. albicans strains SC5314 and WO-1 were streaked from −80°C glycerol stocks to YPD plates and incubated for 24 h at 37°C. One representative colony from each plate was used to inoculate a flask of 20 ml YPD medium that was incubated at 30°C and 200 rpm shaking for 16 h. Yeast cells from each culture were collected by filtration over a 0.45 micron pore-size membrane. Each membrane was transferred to a 50 ml conical Falcon tube, flash frozen in a dry ice-ethanol bath, and stored at −80°C until RNA was extracted using a hot phenol method (Hoyer et al., 1995). RNA was treated with DNaseI and the samples purified using an RNeasy Mini kit (Qiagen). PCR with primers PIR1upF (5’ CCC AAG CTT CTA TAA TCT GCA TCG ATT AAG 3’) and PIR1upR (5’ CCC CTC GAG AGT TGA TGT TAT TAT AGT TGT 3’) was used to verify that all detectable DNA was removed from the RNA preparation. The quality and integrity of the RNA was checked on a formaldehyde agarose gel (Hoyer et al., 1995). Primers QRTALS5F and QRTALS5R were used for the analysis (Green et al., 2005b). The method for real-time reverse transcription polymerase chain reaction (RT-PCR) and data analysis was published previously (Zhao et al., 2005).

PCR screening of C. albicans strain collections

A PCR assay was developed to screen other C. albicans isolates to determine how common ALS51 was in the larger population of C. albicans strains and to determine if its presence was clade-specific. The ALS5repF forward PCR primer (5’ TTT CTC CCT CAG ATA ATA ACC AGT AT 3’) was located within the ALS5-derived sequences in the 5’ domain of ALS51 and the ALS1GenoR reverse primer (5’ CTG TTG ACA TAA TGA GGA CGG G 3’) was located in the 3’ domain of ALS51, derived from ALS1. PCR used Taq polymerase (Invitrogen) with the manufacturer’s supplied buffer containing 1.5 µm MgCl₂, 1 µm of each primer, 200 µm dNTPs and 100 ng DNA template. One cycle of 95°C for 5 min was followed by 35 cycles of 95°C (30 sec), 55°C (30 sec) and 72°C (2 min). Following a final extension step (72°C for 7 min), the reaction was stored at 4°C until analyzed by agarose gel electrophoresis. DNA quality was monitored by amplification of a portion of the TEF1 gene using primers TEF1F (5’ CAC GTT ACC GTC ATT GAT GC 3’) and TEF1R (5’ GCA GAG ATT TGA CCT GGA TGG 3’). A positive reaction with the TEF1 primers and a negative reaction with the ALS51 primers were interpreted to mean that the isolate did not experience a recombination event leading to the presence of ALS51.

Primers ALS51F (5’ TTA CTT GGG AAG CGA GTG GGT CTT 3’) and ALS51R (5’ ACC TAC GGA ATA GAA TGG AGG CGA 3’) were designed to assess the loss of the ALS5/ALS1 intergenic region during recombination to create ALS51. The primers amplify a 500-bp region between the ALS5 and ALS1 coding regions that is deleted during the recombination event that creates ALS51. Therefore, ALS51-positive homozygous strains should not produce a product in this PCR assay while ALS51-positive heterozygous and ALS51-negative strains should produce the 500-bp fragment. Twenty-five cycles of PCR with Taq polymerase were used, with a 30 sec annealing step at 58°C and a 1 min extension step at 72°C.

Primers ALS5repF (5’ TTT CTC CCT CAG ATA ATA ACC AGT AT 3’) and ALS5repR (5’ AAG ACA GTT CTT CCA ATG GAT CA 3’) were used to assay for the presence of the ALS5 coding region. These primers anneal 5’ and 3’ of the tandem repeat domain in ALS5. A PCR product will not be observed in strains that are homozygous for ALS51, but will be observed for strains with at least one ALS5 copy.

Strains that were ALS51-positive were typed by MLST analysis as described by Bougnoux et al. (2003) and assigned to a clade as detailed by Wrobel et al. (2008). Mating type (MAT) was also determined for ALS51-positive strains (Tavanti et al., 2003).

A PCR assay was developed to screen the C. albicans strain collections for the direct repeat-mediated recombination event in the 3’ end of ALS1 or ALS51. Primers ALS1DRF (5’ TCA GTG ACA TCA TTG ACT CAG TTG T 3’) and ALS1DRR (5’ GAT TGA TCA TTT GAA GCA CTG GCA A 3’), which flank the direct repeat-containing region, were used. Amplification used Taq polymerase as described above except each PCR cycle included a 1 min extension step at 72°C.

Serial passage of C. albicans strains

Strain WO-1 and strain 163 were passaged serially to assess the stability of ALS genes over generations of culture growth. Strains were taken from −80°C stocks and streaked onto YPD plates. Plates were incubated for 24 h at 37°C, producing the ‘zero’ generation samples of cells. From this initial plate, a single colony was selected and streaked to a fresh YPD plate. The plate was incubated for 24 h at 37°C and the process continued for 20 days. From the last plate, 20 colonies were selected for analysis. A culture was also grown from a sample taken from the primary streak of the final plate; these cells represented the overall final population, rather than an individual colony following serial passage. Genomic DNA was extracted from each sample and used in PCR analysis with primers ALS5repF and ALS1GenoR that amplify the tandem repeat domain of ALS51, as well as primers ALS1DRF and ALS1DRR that amplify the direct repeat in the 3’ end of ALS51. Primers ALS3GenoF and ALS3GenoR (Oh et al., 2005) were used to amplify the tandem repeat domain of ALS3.

Analyses with anti-Als5 monoclonal antibody (mAb)

Methods for production and validation of the anti-Als mAbs and soluble Als N-terminal domain fragments were described previously (Coleman et al., 2009; Coleman et al., in press). Methods for BIAcore analysis, Western blotting, flow cytometry analysis, and indirect immunofluorescent labeling of C. albicans were also described in those sources. Immunofluorescent labeling signal was amplified using the Tyramide Signal Amplification (TSA) fluorescence system (Perkin Elmer NEL 701A) according to the manufacturer’s instructions. Briefly, paraformaldehyde-fixed C. albicans cells were incubated with blocking solution (Dulbecco’s phosphate buffered saline (DPBS) containing 0.5% blocking reagent) for 30 min at room temperature. Cells were washed with DPBS and incubated with anti-Als5 mAb diluted in blocking solution. After 1 h, cells were washed with DPBS three times and then incubated with fluorescein isothiocyanate (FITC)-conjugated goat anti-mouse antibody for 1 h at 4°C. Cells were washed with DPBS three times and incubated with anti-fluorescein horse radish peroxidase conjugate (Perkin Elmer NEF 710) diluted 1:1000 in blocking solution for 1 h at 4°C. Cells were washed with DPBS three times and incubated with 100 µl of TSA solution diluted 1:50 in the kit’s amplification diluent solution for 10 min at room temperature. Cells were washed three times with DPBS and examined as wet mounts using an Olympus BX50 Fluoview confocal microscope.

Analysis of C. albicans cell wall proteins

A 10 ml YPD starter culture was grown from a single C. albicans colony overnight at 30°C and 200 rpm shaking. The starter culture was used to inoculate 1 L flasks containing 400 ml fresh YPD to OD₆₀₀ = 0.1. Four flasks were grown at 30°C and 200 rpm shaking to OD₆₀₀ = 1.0. Cells were collected by centrifugation, washed once in ice-cold MilliQ water and four times in ice-cold lysis buffer (10 mM Tris-HCl, pH 7.5; 1 mM PMSF). The cell pellet, an equal volume of glass beads (0.425 – 0.6 mm; Sigma), and an equal volume of lysis buffer (containing Complete Mini protease inhibitor cocktail (Roche) and 1 mM pepstatin) were combined in a 50 ml conical tube (Falcon) and vortexed until cells were lysed. The tube was alternately vortexed and rested on ice. Cell breakage was monitored microscopically. Glass beads were allowed to settle by gravity and the supernatant removed. This step was repeated until fresh lysis buffer added to the beads was clear. The pooled washings were centrifuged at 3000 × g for 10 min at 4°C and the pellet was washed in succession with MilliQ water, 5% sodium chloride, 2% sodium chloride, 1% sodium chloride and then with MilliQ water. All wash solutions were ice-cold and contained 1 mM PMSF. The mass of the purified cell wall pellet was recorded.

Beta-1,6-glucanase from Trichoderma harzianum was prepared from the Pichia pastoris overexpression clone described by Bom et al. (1998) according to the authors’ method. The clone was a kind gift from Unilever Research and Development (Vlaardingen, The Netherlands). Enzymatic activity was determined using the 3,5-dinitrosalicylic acid (DNS) reagent described by Bernfield (1955). Due to the lack of commercial availability of pustulan (beta-1,6-glucan), S. cerevisiae cell wall beta glucan (a mixture of beta-1,3-glucan and beta-1,6-glucan) for which a typical analysis was available, was used (Southeastern Pharmaceutical). Lack of detectable beta-1,3-glucanase activity in the purified enzyme preparation was verified using the DNS assay with laminarin (beta-1,3-glucan; Sigma) as the substrate.

Between 100 and 150 mg of purified cell walls (wet weight) were placed into a 1.5 ml microfuge tube and washed twice with beta-1,6-glucanase digestion buffer (50 mM potassium acetate, pH 5.0; 10 mM DTT; 1 mM PMSF). The pellet was resuspended in digestion buffer with protease inhibitors (plus pepstatin) and approximately 2 units of enzyme added. Identical reactions, without enzyme, were included as negative controls. Reaction mixtures were incubated overnight at 37°C and then centrifuged at 3000 × g for 10 min. Supernatant volumes corresponding to the equivalent of 10⁹ cells were loaded onto a 3–8% Tris-acetate polyacrylamide gel (Invitrogen). Western blotting and anti-Als5 mAb detection followed previously published protocols (Coleman et al., 2009; Coleman et al., in press).

Results

Deducing the ALS51 sequence

The DNA sequence from the ALS5-ALS1-encoding region of chromosome 6 in strain WO-1 was PCR amplified and cloned as described in Materials and methods, yielding clone 3052. Assembly of the DNA sequence for clone 3052 revealed an in-frame fusion between ALS5 and ALS1 (Fig. 1A). Alignment of the nucleotide sequence of the WO-1-derived fragment with ALS5 and ALS1 sequences from strain SC5314 suggested that the recombination event that generated the WO-1 sequence occurred in the region shown in Fig. 1B. This location is within the 5’ domain of the ALS gene and close to the start of the sequences that encode the Thr-rich region that immediately precedes the tandem repeat domain (Hoyer et al., 2008). Overall, the 5’ domains of ALS5 and ALS1 are 89% identical at the nucleotide level, with regions of greater and lesser sequence identity (Fig. 1C). The resulting open reading frame was named ALS51 because it consisted of the 5’ domain of ALS5 and the tandem repeats and 3’ domain of ALS1. Estimating the amount of DNA that was deleted as a result of the recombination event was complicated by the fact that the number of copies of the 108-bp tandemly repeated sequence in an ALS coding region is often variable (Hoyer et al., 2008). In the example shown in Fig. 1, approximately 8.8 kb of DNA was lost in the creation of ALS51, however this number will differ for each isolate in which the recombination occurs. Examination of the C. albicans SC5314 genome sequence suggested that no other open reading frames were lost in the recombination event that created ALS51. Approximately 500 bp of sequence upstream of ALS51 and 700 bp of sequence downstream were included in the original fragment from clone 3052. Only 1 bp upstream and 5 bp downstream did not match those from the corresponding region in the SC5314 genome database. PCR with primers that amplify ALS9 sequences (immediately downstream of ALS1 in strain SC5314) suggested that the ALS9 coding region was intact in strain WO-1 (data not shown).

ALS51 alleles in strain WO-1

Initial PCR experiments to amplify the ALS5-ALS1 region from strain WO-1 suggested the presence of two different-sized products. Differences in the number of copies of the 108-bp tandem repeat sequence in the central domain of ALS genes are often responsible for allelic size variation (Hoyer et al., 2008). Primers ALS5repF and ALS1GenoR were used to amplify this region of ALS51, and produced two different-sized fragments (Fig. 1D). These fragments indicated the number of copies of the tandem repeat sequence in each ALS51 allele and confirmed that the smaller of the two alleles (ALS51-2; eight tandem repeat copies) had been sequenced in its entirety.

The two alleles of ALS51 were also evident on the initial amplification of WO-1 genomic DNA with primers ALSupF and ALSdnR (primers (1) and (2) in Fig. 1A; data not shown). The larger product (ALS51-1) was cloned and its DNA sequence determined as described in Materials and methods. The 5’ domains of the ALS51 alleles had 3 nucleotide polymorphisms, none of which translated into a different amino acid sequence. ALS51-1 encoded 15 copies of the tandemly repeated 108-nt sequence in the central domain, compared to the eight copies encoded by ALS51-2. Alignment of the 3’ domains of the ALS51 alleles showed that ALS51-2 was shorter than ALS51-1 (Fig. 2A). The missing nucleotides comprised one copy of a direct repeat sequence that was noted in the initial description of ALS1 (Hoyer et al., 1995). Allelic variation in the 3’ domain was detected by PCR using primers ALS1DRF and ALS1DRR. As expected, strain WO-1 was heterozygous in this region, showing an 807-bp and a 431-bp fragment (Fig. 2B). Information from DNA sequencing and PCR analysis was combined to produce a schematic of the two ALS51 alleles from strain WO-1 (Fig. 2C).

Fig. 2.

(A) Amino acid sequence alignment of a region from the C-terminal domain of Als1, Als51-1 and Als51-2. Alignment of the sequences indicated that Als51-2 was missing 109 amino acids compared to Als1 and Als51-1. Amino acid coordinates were calculated from the translation of GenBank sequence L25902 (ALS1) and from the ALS51-1 and ALS51-2 sequences described here. Underlining shows the direct repeat sequence. Dashes designate amino acids that were missing in the predicted Als51-2 sequence. (B) Ethidium bromide-stained agarose gel of genomic DNA from C. albicans strain WO-1 amplified with primers ALS1DRF and ALS1DRR that flank the direct repeat sequences (see (C) below). The 807-bp fragment resulted from the ALS51-1 allele while the shorter, 431-bp, fragment resulted from amplification of ALS51-2. Molecular size markers (in bp) are indicated to the left of the image. (C) Line drawings of the two ALS51 alleles from strain WO-1. Asterisks specify the site of recombination between ALS5 and ALS1. All sequences 5’ of the asterisk were derived from ALS5 while all sequences 3’ of the asterisk were from ALS1. Head-to-tail arrows denote copies of the 108-bp tandem repeat sequence. Eight copies were present in ALS51-2 while 15 copies were found in ALS51-1. Block arrows in the 3’ domain of each gene indicate the single-copy direct repeat sequence in ALS51-2 and two direct repeat copies in ALS51-1. The location of PCR primers is shown by numerals in parentheses (3 = ALS5repF; 4 = ALS1GenoR; 5 = ALS1DRF; 6 = ALS1DRR).

Determining the frequency of ALS51-related recombination events in a larger population of C. albicans isolates

Finding ALS51 in strain WO-1 prompted the question of how many other C. albicans isolates might have undergone a similar recombination event between ALS5 and ALS1. Genomic DNA from a collection of 239 C. albicans isolates was screened to detect strains that encode ALS51. PCR primers ALS5repF and ALS1GenoR amplify a product in C. albicans strains, which is only possible when the sequences from the 5’ end of ALS5 are brought into close enough proximity to sequences in the 3’ end of ALS1. In isolates where a recombination event has not occurred, the PCR product would be larger than approximately 9 kb (Fig. 1A). Eleven of the 239 isolates screened (4.6%) had a PCR product, suggesting a recombination event consistent with creation of ALS51. PCR-positive strains were from different genetic clades including 1, 2, 3, 8, 9 and 12. Positive-control PCRs using primers to the TEF1 gene confirmed that each genomic DNA sample could be amplified, and reduced the chance that negative reactions were due to DNA preparation quality.

For each of the 11 ALS51-positive strains, a single ALS5repF/ALS1GenoR PCR product was observed, indicating that either i). the strains were heterozygous with respect to ALS51, or ii). that the ALS51 alleles had the same number of tandem repeat copies. PCR strategies were developed to distinguish between these two possibilities. ALS51F and ALS51R were designed to amplify a fragment between the ALS5 and ALS1 coding regions that would be deleted in an ALS51-positive strain, but present on a chromosome with intact ALS5 and ALS1 sequences. All 11 ALS51-positive isolates were also positive for ALS51F/ALS51R suggesting that each strain was heterozygous, with only one chromosome encoding ALS51. In 10 of the 11 strains, the other chromosome encoded ALS5; the 11^th strain was negative in a PCR using primers ALS5repF and ALS5repR that amplify the tandem repeat domain of ALS5, suggesting ALS5 deletion on the homologous chromosome. This isolate was from clade 2. Previous studies showed that 67% of clade 2 isolates are missing one ALS5 allele (Zhao et al., 2007b).

The collection of 239 C. albicans isolates was also screened to determine the frequency of recombination within the 3’ end of ALS1 (or ALS51) that affected the copy number of the direct repeat sequence. Only 2 other isolates showed the heterozygous pattern that was observed in strain WO-1, suggesting that this event was rare (0.8%). These recombination events presumably affected ALS1 since neither isolate had a PCR product suggesting the presence of ALS51.

Stability of ALS51 was tested over serial passage of two C. albicans strains: strain WO-1 (MTLα/MTLα, clade 6) and strain 163, an oral isolate from a normally healthy human (MTLa/MTLα; clade 1). PCR products derived from primers flanking the central tandem repeat domain in ALS51, the central tandem repeat domain in ALS3, and the direct repeat-encoding region in the 3’ domain of ALS1 or ALS51 did not show any differences between the initial culture plate of each strain and the 20^th plate following serial passage. These results suggest that the tandem repeats within the ALS51 and ALS3 coding regions were stable despite many generations of growth in culture.

ALS51 transcription

Recombination to create ALS51 deleted the tandem repeats and 3’ domain of ALS5, the ALS5/ALS1 intergenic region and the 5’ domain of ALS1. This event left the new ALS51 coding region downstream of the ALS5 promoter sequences. Little is known about ALS5 expression patterns. Since the 5’ domain of ALS5 is 89% identical to the 5’ domain of ALS1 and the 3’ domain of ALS5 is nearly identical to the 3’ domain of ALS6, specific detection of full-length ALS5 transcript requires an oligonucleotide probe (Hoyer & Hecht, 2001); to date, detection of ALS5 transcript on a Northern blot of C. albicans RNA has not been reported. ALS5 expression was detected by RT-PCR of RNA extracted from cultured C. albicans cells, as well as C. albicans present in disease models and clinical specimens (reviewed in Hoyer et al., 2008). However, expression of genes such as ALS1 was detected far more readily, suggesting an overall lower relative level of ALS5 expression.

A pilot experiment was conducted to define a simple growth condition to detect ALS5 and ALS51 expression. Real-time RT-PCR primers for ALS5 anneal within the 5’ domain of the gene, making them equally useful for detecting ALS51 transcript (Green et al., 2005b). C. albicans strains SC5314 and WO-1 were grown in YPD medium and samples collected each hour. Real-time RT-PCR analysis of RNA isolated from the various time points showed similar expression levels, making any of the time points equally valuable to demonstrate ALS5 and ALS51 expression (data not shown). Yeast cells from a 16 h culture at 30°C were used for a more-detailed analysis. The TEF1 gene was used as a control. Duplicate RT-PCR analysis of two independently grown WO-1 cultures showed a mean cycle threshold (C_t) value for TEF1 (control) of 19.6 ± 1.1, while the C_t for ALS51 was 21.9 ± 0.3. C_t values for the negative control reactions (water blank and reverse transcription-negative) were in the range of 33.4 ± 5.1 to 39.8 ± 0.4. These results demonstrated that ALS51 was transcribed in strain WO-1. Results from strain SC5314 showed a TEF1 C_t value of 18.7 ± 1.0, while the C_t value for ALS5 primers was 23.3 ± 0.6. ΔΔC_t analysis suggested that ALS51 expression in WO-1 was 5-fold higher than ALS5 expression in strain SC5314, although the biological meaning of this result must be interpreted carefully. ALS51 transcription was also detected in WO-1 opaque cells (ALS51 C_t = 20.7 ± 0.3; TEF1 C_t = 18.6 ± 0.4).

In search of the Als51 protein

Another method that could be used to demonstrate ALS51 expression is to visualize Als51 protein on the surface of WO-1 cells. This experiment exploited the recent development of an anti-Als5 mAb. Because the mAb was raised against the 5’ domain of Als5 (amino acids 18 to 329), it will also recognize Als51 (Coleman et al., 2009). Measurements of anti-Als5 mAb binding kinetics using BIAcore yielded a dissociation constant (K_D) of 0.28 nM (mean of two independent observations of 0.12 nM and 0.45 nM). Specificity of anti-Als5 was demonstrated initially by ELISA and then with methods such as Western blotting (Fig. 3). The N-terminal domains of Als1 and Als3 are 82% and 79% identical to Als5, respectively, and therefore are the most likely proteins to cross-react with anti-Als5. The blot in Fig. 3 showed that anti-Als5 specifically recognized Als5, and validated the utility of the mAb for Western blotting applications.

Fig. 3.

Demonstration of the specificity of the anti-Als5 mAb and its utility for Western blotting. (A) N-terminal domain protein fragments for Als1, Als3 and Als5 (0.5 µg each) were electrophoresed on a polyacrylamide gel and silver stained using a previously described method (Coleman et al., in press). (B) The same proteins (100 ng of each) were electrophoresed on another polyacrylamide gel, which was transferred to Hybond-P PVDF membrane and Western blotted with anti-Als5 at a concentration of approximately 5 µg ml⁻¹. Recognition of only the Als5 immunogen with the anti-Als5 mAb supported the conclusion that the mAb is specific for Als5, and demonstrated the utility of the mAb for Western blotting. Molecular size markers (in kDa) area shown at the left of each image.

The anti-Als5 mAb clearly recognized Als5 on the surface of C. albicans strain 1002 that overexpressed ALS5 under control of the TPI1 promoter (Fig. 4A). Real-time RT-PCR analysis of these cells showed an ALS5 C_t of 17.2 ± 0.1 and a TEF1 C_t of 18.6 ± 0.4. However, wild-type yeast cells of strains SC5314 and WO-1 (both white and opaque) failed to produce a visible signal with anti-Als5, most likely due to the lower abundance of ALS5/ALS51 transcriptional activity. An immunofluorescence signal amplification kit (Pierce) was used to attempt to raise the signal to the point of detection (Fig. 4A), however this method also did not provide a visible signal for either C. albicans isolate (data not shown). S. cerevisiae strain 884, that produced an Als5-Agα1 fusion protein under control of the copper sulfate-inducible CUP1 promoter, was used as a control to demonstrate utility of the signal amplification kit. Cell-surface immunofluorescence was visible for strain 884 without signal amplification, and more pronounced when subjected to the signal amplification protocol (Fig. 4B). The combination of real-time and immunolabeling data suggest that Als51 should be present on the C. albicans cell surface, but at an abundance below the detection limit for the indirect immunofluorescence assay.

Fig. 4.

Photomicrographs of bright field and anti-Als5 immunolabeled yeast cells demonstrate the cell-surface localization of Als5 and the low-level abundance of Als51 on the surface of strain WO-1. (A) Anti-Als5 immunolabeling of C. albicans strains 1002, which overexpressed ALS5 under control of the constitutive TPI1 promoter, SC5314 and WO-1 (white and opaque cells). Bright-field microscopy was used to create the upper image in each pair. The lower image shows cells immunolabeled with the anti-Als5 mAb and a fluorescein isothiocyanate-conjugated secondary antibody, illuminated with 488 nm light. Real-time RT-PCR threshold values (C_t) are shown below each image pair for cells grown under the same conditions used to produce the micrographs. Constitutive, high-level expression of ALS5 in strain 1002 resulted in abundant cell-surface protein and a lower C_t value than for the wild-type strains. Cell-surface Als5 was not visible in strain SC5314, nor was Als51 visible in white or opaque cells of strain WO-1. (B) Immunofluorescent signal amplification with the TSA system was sufficient to enhance the cell-surface signal strength for a S. cerevisiae strain that produced an Als5-Agα1 fusion protein, but was not sufficient to visualize Als5 or Als51 on the wild-type C. albicans strains (data not shown). These results more closely defined the limit of detection for immunolabeling visualization of cell-surface Als proteins. The scale bar in each figure corresponds to 10 µm.

Als5 on strain SC5314 and Als51 on WO-1 were detected on a Western blot of beta-1,6-glucanase-digested cell walls (Fig. 5). The marked abundance of anti-Als5-reacting protein released from the ALS5 overexpression strain 1002, and absence of a similar-sized protein in the als5Δ/als5Δ strain 2373 (Zhao et al., 2007a) aided identification of Als5, and therefore Als51, on the Western blot. The results suggest that Als5 and Als51 are linked to beta-1,6-glucan in the C. albicans cell wall.

Fig. 5.

Western blot with anti-Als5 to demonstrate beta-1,6-glucanase-dependent release of Als5 and Als51 from the C. albicans cell wall. Cell wall preparations were treated with beta-1,6-glucanase, separated by acrylamide gel electrophoresis and transferred to Hybond PVDF membrane (Amersham) for Western blotting with the anti-Als5 mAb. Sample loading was normalized to cell number. One-fifth as much sample was loaded from strain 1002 (ALS5 overexpression) because of the large relative abundance of Als5 in this strain. Samples without beta-1,6-glucanase were included as negative controls. Additional assurance of the correct identification of Als5 was provided by lack of signal in the negative control lanes, and for the als5Δ/als5Δ strain 2373. Molecular size (in kDa) is indicated at the left of the blot.

Discussion

Determining that the ALS5 and ALS1 open reading frames recombined to form a novel gene, ALS51, explains the previous puzzling observations about ALS1 in strain WO-1 (Hoyer et al., 1998; Coleman et al., 2009). Recombination to form ALS51 deleted the ALS5/ALS1 intergenic region and the 5’ end of ALS1, leaving none of the ALS1 regulatory sequences that are present in most other C. albicans strains, nor the coding region for the N-terminal domain of the Als1 protein. Without these DNA sequences, ALS1 transcript is not produced under growth conditions that produce it in other C. albicans isolates, and the epitope for recognition by the anti-Als1 mAb is missing. However, since the most commonly used ALS1-specific DNA probe is derived from the ALS1 3’ end and those sequences are present in ALS51, strain WO-1 seemed ALS1-positive in early Southern blot-based screens (Hoyer et al., 1995; 1998). PCR screening of a larger set of C. albicans isolates demonstrated that the ALS51 recombination event occurs in approximately 5% of isolates; ALS51-positive strains are from different clades, suggesting that they are not all descendents of a single recombination event. As such, the site of recombination may vary between the isolates, producing sequence variability in the 5’ domain of ALS51.

The search for the Als51 protein provided additional insight into the association between real-time RT-PCR measurements of transcript abundance, the limits for immunofluorescent detection of protein on the C. albicans surface, and visualization of cell wall proteins on Western blots. Both Als5 on strain SC5314 and Als51 on strain WO-1 were released from the cell wall using beta-1,6-glucanase, verifying their biochemical localization. Western data, however, do not allow detection of the spatial localization of Als5 or Als51 on the C. albicans surface (diffuse vs. focal localization, for example). Immunolabeling has provided spatial localization information for Als proteins that are translated from more abundant RNA pools. For example, Als3 is localized diffusely over the surface of germ tubes and hyphae (Coleman et al., 2009) and Als1 is found diffusely over the yeast cell surface except in bud scars (Coleman et al., in press).

The twelve ALS51-positive strains identified in this study (WO-1 plus 11 more) represent three different ALS family genotypes that vary from the wild-type example of strain SC5314 (ALS5, ALS1/ALS5, ALS1; reviewed in Hoyer et al., 2008). Ten of the ALS51-positive strains are heterozygous for ALS51, ALS1 and ALS5, (ALS51/ALS5, ALS1) suggesting that these strains produce functional protein from one allele for each of the genes. One of the ALS51-positive strains is heterozygous for ALS51 and ALS1, with ALS5 deleted by the ALS51 recombination on one copy of chromosome 6 and by direct-repeat-mediated recombination on the other copy (ALS51/Δals5, ALS1). WO-1 is the only strain that is homozygous for ALS51, with loss of both copies of ALS1 and ALS5 (ALS51/ALS51). Observations from earlier studies allow speculation regarding the phenotypic effect of the allelic combinations.

Ten of the 12 ALS51-positive strains produce Als51, Als5 and Als1 from one allele each, adding Als51 function to the complement of 'standard' Als proteins. Als51 is comprised of the N-terminal domain of Als5, and the tandem repeat and C-terminal domain of Als1. The cell-surface abundance of Als51 reflects the abundance of Als5, rather than Als1 (Coleman et al., in press). Function of the various Als domains has been assessed mostly by overexpression of ALS genes in S. cerevisiae. Adhesive function resides within the N-terminal domain of Als1 because mutagenesis of sequences encoding the N-terminal domain lead to decreased adhesive activity (Loza et al., 2004). Using a heterologous expression approach, the Thr-rich region (approximately amino acids 330 to 433) and the 36-amino acid tandemly repeated sequences that immediately follow, enhance S. cerevisiae binding to fibronectin (Rauceo et al., 2006; Frank et al., 2010). Little functional attention has been paid to the C-terminal domain, perhaps due to the prevailing idea that this region is heavily glycosylated and serves as a stalk to project the remainder of the protein outward on the fungal cell surface. Despite extensive sequence differences in the C-terminal domain and in the putative GPI anchor addition site for Als1 (Kapteyn et al., 2000) and Als5, both proteins (and Als51) are linked to beta-1,6-glucan in the C. albicans cell wall. The Als1 C-terminal domain sequences in Als51 may lead to exclusion of Als51 from bud scars (Coleman et al., in press). The ten strains that produce Als51, Als5 and Als1 from one allele each also may exhibit phenotypic effects of haploinsufficiency, although the effect on phenotype will depend on which Als51, Als5 and Als1 alleles are maintained in each strain. The potential for allelic variation in function was demonstrated for Als3, where a C. albicans strain encoding an allele with 9 copies of the tandem repeat sequence exhibited lower adhesion to cultured monolayers of vascular endothelial and pharyngeal epithelial cells than a strain encoding an allele with 12 copies (Oh et al., 2005).

One strain identified in this study encodes Als51 and Als1, but not Als5. Deletion of ALS5 increases adhesion of C. albicans to vascular endothelial and buccal epithelial cells using in vitro assays (Zhao et al., 2007a), suggesting a potential phenotypic difference in the newly identified strain. WO-1 is unique among the strains studied because it has two ALS51 alleles and neither ALS5 nor ALS1. Deletion of ALS1 reduces C. albicans adhesion to cultured vascular endothelial cells (Fu et al., 2002; Zhao et al., 2004) and the balance of this effect against the increased adhesion observed in als5Δ/als5Δ strains remains to be determined. Such interactions likely account for the complex adhesion phenotype among wild-type isolates of C. albicans. Deletion of ALS1 also decreases C. albicans cell size (Zhao et al., manuscript in preparation). The smaller overall size of WO-1 compared to SC5314 is obvious in Fig. 4. Whether this cell size difference is due solely to deletion of ALS1 must be tested experimentally. Since WO-1 is one of the prototype strains for studies of C. albicans mating, neither Als5 nor Als1, per se, are required for this process.

Cell surface adhesion proteins containing tandemly repeated sequences are present in several organisms. Study of the phenotypic effects of variation in repeat copy number in S. cerevisiae FLO1 has gained considerable attention (Verstrepen et al., 2005). Integration of URA3 into the FLO1 tandem repeats and plating under conditions selective for URA3 loss generated new FLO1 alleles with varying numbers of tandem repeat copies. Variation in repeat copy number was associated positively with adhesion phenotype leading to the conclusion that tandemly repeated sequences provide functional diversity of cell surface antigens that allow rapid environmental adaptation and elusion of the host immune system.

ALS alleles commonly vary in the number of copies of the tandemly repeated nucleotide sequence in the central domain (reviewed in Hoyer, 2001). The mere existence of these sequence variations is evidence of the plasticity of this region. However, these changes have been described as ‘rapid’ in frequency and as similar in character to those observed for the well-characterized surface protein antigenic variation mechanisms found in trypanosomes or Plasmodium species (Verstrepen & Fink, 2009). These claims require additional examination to gain an appropriate mental image of Als protein presentation on the C. albicans cell surface and the frequency of allelic change. The relative stability of ALS allele tandem repeat copy number is suggested by lack of detectable differences in strains that were passaged serially for 3000 generations in vitro (Zhao et al., 2007b) and extensively in the present study. The relative stability of ALS alleles is also reflected in the observation that genetic clades of C. albicans strains are enriched for certain ALS alleles and allelic combinations (Oh et al., 2005; Zhao et al., 2007b). Some ALS genes have a very narrow range of tandem repeat copy number, suggesting a mechanism to limit allele size (Zhao et al., 2007b). Serial passage of C. albicans strains in vivo (successively inoculating eight mice per strain) also failed to yield detectable changes in tandem repeat copy number of the surveyed alleles (D. Sachen, X. Zhao & L.L. Hoyer, unpublished data). It is common to detect expression of all eight ALS genes at the same time under many varied growth conditions (including human clinical specimens), suggesting that the Als protein profile on C. albicans is mixed, rather than featuring one protein at a time (reviewed in Hoyer et al., 2008). All of these observations present a stark contrast between the Als family and antigenic variation mechanisms in trypanosomes and Plasmodium. Rando & Verstrepen (2007) lent more global insight to the use of the word ‘rapid’ to describe timescales of genetic inheritance in yeast repeated sequences. They differentiated between environmental change that can occur over an organism’s lifetime and that which occurs over thousands or millions of generations. Repeat-mediated genetic changes in yeast fell into the timescale of 10² to 10⁶ generations, a difference of many orders of magnitude. The extremely plastic genome of C. albicans has yielded some examples of changes that occur in a few hundred generations or less including variations in the major repeat sequence (MRS; Scherer & Stevens, 1988) and changes in the rDNA (Rustchenko et al., 1993). ALS genetic changes may fit into the timescale of repeat-mediated genetic changes, but at the level of tens of thousands or millions of generations.

Comparisons between genome sequences of closely related species also lend insight into evolution of the ALS family. Such a comparison was provided by completion of the genome sequence of the closest relative of C. albicans: C. dubliniensis (Jackson et al., 2009). Genomic comparisons showed that, based on genomic position only, both species have evolved novel ALS genes. However, many conclusions in the paper are drawn from analysis of the ALS 5' domain sequences, without regard to the 3' domains to which they are attached. Some of these conclusions are relevant to the discussion of ALS5 and ALS1 recombination. For example, the authors note that ALS5 is missing in C. dubliniensis without considering that, depending on the strain collection analyzed, ALS5 is missing from 1.6% to 8% of C. albicans isolates (Hoyer & Hecht, 2001; Zhao et al., 2007b). ALS5 is absent from approximately 15% of clade 3 isolates and 23% of clade SA isolates (Blignaut et al., 2002; clade 4 in the MLST phylogeny of Odds et al., 2007). Perhaps analysis of other C. dubliniensis strains would identify one in which ALS5 is present. Also, using only 5' domain sequence data, and the lack of ALS5 in C. dubliniensis, the authors conclude that a tandem duplication of ALS1 must have occurred in C. albicans. The 5' domain sequences only tell a portion of the story, since the 3' domain of ALS5 is nearly 100% identical to that of ALS6, which is localized on C. albicans chromosome 3 (Hoyer & Hecht, 2000). The presence of ALS5 in C. albicans suggests a potential recombination between chromosomes 6 and 3, an event that quite possibly could occur. C. albicans has a great degree of genomic plasticity that may contribute to the generation of novel ALS genes. It is not clear whether C. dubliniensis has a similar level of genomic plasticity, and this variable undoubtedly also affects the conclusions regarding ALS family evolution between the two species.

The observations in this paper explain previous mysteries regarding the nature of the ALS1 locus in C. albicans strain WO-1, which led to additional examples of strain variation for the ALS family. It is very likely that other examples of novel ALS genes remain to be detected and characterized. Such examples will provide further understanding of the extent of genetic diversity in the ALS family.