Tca1, the Retrotransposon-Like Element of Candida albicans, Is a Degenerate and Inactive Element

Abstract

Candida albicans is an asexual fungus and as such must rely on mechanisms other than sexual recombination to generate genetic diversity. Retrotransposons are ubiquitous genetic elements known to generate multiple types of genomic alterations. We have further investigated the nature of the retrotransposon-like element Tca1 in C. albicans. Tca1 is present at two loci in strain SC5314. Both loci have now been cloned, and one element was sequenced in its entirety. This element was flanked by α elements, or long terminal repeats (LTRs), and contained an intervening region of 5,614 bp. The intervening region was highly degenerate and contained no extended open reading frames, indicating that Tca1 is not a functional element. Partial sequence determination demonstrated that the elements from the two loci were nearly identical. Genetic manipulation of the elements showed that both loci were heterozygous for Tca1, that both were transcriptionally active, and that deletion of both had no effect on growth rate or germ tube formation. Thus, it is unclear why this nonfunctional, highly degenerate element has been maintained in many clinical isolates.

Candida albicans is the agent of a large percentage of opportunistic fungal infections. The clinical significance of this organism has spurred efforts to understand its basic biology and genetics. Clinical isolates of C. albicans have been shown to vary in a number of phenotypic properties relevant to the infectious process and virulence (9, 18–20, 26, 27, 29). This variability presumably reflects genetic diversity in these strains. C. albicans is an imperfect fungus with a diploid genome (21). Lacking a sexual cycle, the organism must rely on alternate mechanisms to generate genomic diversity. Chromosomal rearrangements have been shown to occur readily in C. albicans, and this provides one mechanism of potential significance (23, 24). Presumably, additional mechanisms of mutagenesis and genomic change are functional in C. albicans.

Retrotransposons are known to effect several types of genetic alterations in fungal cells, either directly, by virtue of their mobilization and integration at a locus, or indirectly, via recombination between ectopically located copies of the element (2, 5). These alterations include gene inactivation, altered transcriptional control of gene expression, and genomic deletions and inversions (2, 5). It is not known if such retrotransposon-mediated changes occur in C. albicans. However, several groups have reported the presence of retrotransposons or retrotransposon-like elements in this fungus. We previously identified a repetitive element in C. albicans, which we designated alpha (6). This element was 388 bp in length and bound by a 6-bp inverted repeat similar in sequence to the inverted repeats found in the long terminal repeats (LTRs) of retrotransposons. This isolated α element was flanked by a 5-bp direct repeat, suggestive of the target site duplication that occurs with integration of mobile elements. In addition to this solo repeat, we isolated a genomic clone that contained direct repeats of the α element separated by an intervening region of approximately 5.5 kb. This structure had the hallmarks of a bona fide retrotransposon. In addition to its size and the flanking LTRs, the presence of potential primer binding sites for replication were noted, and the entire structure was transcribed into an approximately unit-length RNA (6). Interestingly, transcription of this putative retrotransposon, Tca1, was strongly temperature dependent (6).

More recently, another repetitive element of C. albicans, termed beta, was found to have sequence features suggesting that it was a retrotransposon-derived solo LTR (22). Interestingly, the characterization of several beta elements demonstrated that each was located adjacent to a tRNA gene analogous to the Ty3 retrotransposon of Saccharomyces cerevisiae (11), and a limited sequence similarity with the LTRs of Ty3 was also noted (22). While this finding suggests the presence of a Ty3-like retrotransposon, such an element has not been identified. The only full-length retrotransposon identified in C. albicans is pCAL1 (17). This element is related to the Ty1/copia family of retrotransposons and is unusual in that 50 to 100 copies of unintegrated, linear double-stranded DNA form are present (17).

Because of the potential significance of retrotransposons in the genetics of C. albicans, we have further investigated the nature of Tca1. Here we report the cloning of a second locus containing this element and report its complete nucleotide sequence. Analysis of the sequence indicated that both loci contained degenerate and nonfunctional elements. Deletion analysis demonstrated that both loci were hemizygous for Tca1 and that loss of either or both copies had no gross consequence.

MATERIALS AND METHODS

Strains and culture conditions.

The C. albicans strains used are listed in Table 1. These strains were routinely grown on YEPD medium (25) at 30°C. YNB medium (25) was used for the selection of prototrophic strains. 5-Fluoro-orotic acid (5-FOA)-containing medium (4) was used in the selection of Urd⁻ strains. Germ tube induction was tested in the medium described by Lee et al. (15) at 37°C. The media were supplemented with uridine (25 μg/ml) as needed.

TABLE 1.

C. albicans strains used in this study

Strain	Parent	Genotype	Reference
SC5314		Clinical isolate	10
CAF3-1	SC5314	tca1-1 tca1-2 ura3::imm434/ura3:: imm434	8
CAC-1	CAF3-1	tca1-1::URA3 tca1-2 ura3::imm434/ ura3::imm434	This work
CAC-2	CAC-1	tca1-2 ura3::imm434/ura3::imm434	This work
CAC-4	CAF3-1	tca1-1 ura3::imm434/ura3::imm434	This work
CAC-5	CAC-2	tca1-2::URA3 ura3::imm434/ura3:: imm434	This work
CAC-6	CAC-5	ura3::imm434/ura3::imm434	This work

Hybridization screening; Southern and Northern analysis.

A 1.8-kb HindIII-EcoRI fragment from the internal region of Tca1-1 (6) was used as a hybridization probe to screen a genomic library of strain SC5314 DNA carried in λGEM12 (Promega). The library and screening methods were described previously (6). Methods for DNA isolation, RNA isolation, Southern blot hybridizations, and Northern blot hybridizations were also previously described (6). The hybridization probe for Southern and Northern blots contained equal amounts of the 1.8-kb HindIII-EcoRI fragment and 2.2-kb EcoRV fragment from the internal region of Tca1-1.

DNA sequence determination and analysis.

Nucleotide sequences were determined by the dideoxy-chain termination method using Sequenase (U.S. Biochemical) and [³⁵S]dATP (Amersham). Nucleotide sequence comparisons were conducted by using the BLAST algorithm of Altschul et al. (1).

Plasmid and strain construction.

Plasmids pTca1-1 and pTca1-2 contained the entire insert from lambda clones CJY-3 and CJY-4, respectively. The inserts were released by digestion with BamHI and cloned into the BamHI site of plasmid pBSK(+) (Stratagene). Plasmids pBSKTR1 and pBSKTR2 were constructed by cloning the 3.4- and 3.7-kb EcoRI fragments, respectively, from pTca1-1 into the EcoRI site of pBSK(+).

Plasmid pBSKTca-UR3 containing a deletion and disruption of Tca1-1, in which the internal 0.4-kb EcoRI-XbaI region was replaced with a 1.4-kb ScaI-XbaI fragment of URA3, was constructed in several steps. First, the 1.4-kb ScaI-XbaI fragment of URA3 was cloned into the SmaI-BamHI sites of the polylinker region of plasmid pBSKTR1, which contains the 5′ end of Tca1-1, creating plasmid pBSKTR1-UR3. Next, a 3.2-kb XbaI fragment was isolated from pBSKTR2, which contains the 3′ end of Tca1-1. One XbaI site lies within Tca1-1, and the other lies within the polylinker. This was cloned into the XbaI site of pBSKTR1-UR3, creating pBSKTca-UR3.

Strains containing a partial deletion and disruption of Tca1 were constructed by transformation of the Urd⁻ strain CAF3-1 (8) with the 5.0-kb HindIII-SalI fragment from plasmid pBSKTca-UR3. This fragment contains the URA3 marker nested within the internal region of Tca1-1, with approximately 2.0 kb of 5′-flanking sequence and 1.7 kb of 3′-flanking sequence. Transformation was performed as described by Kelly et al. (13), and Urd⁺ transformants were selected on YNB medium. Integration into Tca1-1 resulted in strain CAC-1. Strain CAC-2, from which Tca1-1 was completely deleted, was isolated by selection of spontaneous Urd⁻ derivatives of CAC-1 on 5-FOA-containing medium (4). Strain CAC-4, from which Tca1-2 was deleted, was obtained in the same fashion. To obtain a strain lacking both copies of Tca1, strain CAC-2 was subjected to a second round of transformation to disrupt Tca1-2. 5-FOA selection of one of these Urd⁺ transformants, CAC-5, resulted in a Urd⁻ strain, CAC-6, which lacked both Tca1-1 and Tca1-2. All recombination events were verified by Southern blot analysis.

Nucleotide sequence accession number.

The complete nucleotide sequence of Tca1-2 has been submitted to GenBank under accession no. AF043301.

RESULTS

Cloning of the second locus containing Tca1.

Previous work established that strain SC5314 contained Tca1 elements at two loci. One of these loci was obtained as a λ clone, CJY-3 (6), and will be referred to as Tca1-1. Preliminary sequence analysis, as discussed later, demonstrated that Tca1-1 lacked any open reading frames characteristic of retrotransposons. This motivated the isolation of Tca1 from the second locus. Hybridization screening of a genomic library in λGEM12 yielded clone CJY-4. Southern blot hybridization of an EcoRI digest demonstrated that this clone contained two fragments, 3.25 and 4.8 kb in length, that hybridized with the internal region of Tca1-1 (Fig. 1). Based on previous results (6), these were of the expected size and were assigned to the second locus. This copy of the element was designated Tca1-2. Restriction endonuclease mapping of the genomic insert carried by CJY-4 demonstrated a region of approximately 6 kb that was largely colinear with the insert of CJY-3, except for a few restriction site polymorphisms (Fig. 2). The restriction maps differed significantly outside this common region, indicating unique flanking sequences (Fig. 2).

FIG. 1.

Southern blot analysis of DNA from lambda clones CJY-3 and CJY-4. Genomic DNA from strain SC5314 or purified DNA from lambda clones CJY-3 and CJY-4 was digested with EcoRI and hybridized with an α-element probe.

FIG. 2.

Restriction site maps of the genomic inserts from lambda clones CJY-3 and CJY-4 containing Tca1-1 and Tca1-2, respectively. The locations of α elements (LTRs) are indicated by the boxed regions. Also indicated are the plus- and minus-strand primer binding sites (+PBS and −PBS), the 5-bp direct repeats flanking the elements, the EcoRI fragments associated with each locus, and the 1.8-kb HindIII-EcoRI and 2.2-kb EcoRV fragments that were used together as a hybridization probe of the internal region. The direction of transcription of Tca1 is left to right, as shown.

Nucleotide sequence analysis of Tca1-1 and Tca1-2.

Nucleotide sequence analysis of the insert in CJY-4 identified flanking direct repeats of identical sequence, 388 bp in length. This sequence was 99% identical to the LTRs of Tca1-1, differing only by the presence of three nucleotide transitions and one transversion (Fig. 3). As previously observed for the LTRs, or α elements, of Tca1-1, the ends of each LTR of Tca1-2 were defined by a 6-bp inverted repeat characteristic of the LTRs of other retrotransposable elements (6). The sequence of this inverted repeat, TGTTCG, resembles those of S. cerevisiae sigma and delta elements and Drosophila copia elements, TGTTGTAT, TGTTGGAA, and TGTTGAATA, respectively (7). A similar sequence, TGTTGG, is also found in the LTRs of the C. albicans retrotransposon pCal (17). The 5′ LTR of Tca1-2 was preceded by the sequence ATTGC. A direct repeat of this 5-bp sequence was found 3′ of the downstream LTR, suggesting that this sequence represents a duplication of the integration target site typically observed with retrotransposons (2, 5). This target site duplication differed in sequence from that of Tca1-1, TTGGT, as expected for integration events at two distinct loci.

FIG. 3.

Comparison of α elements and flanking sequences of Tca1-1 and Tca1-2. The sequence shown is that derived from Tca1-1. The differences in Tca1-2 are indicated below the sequence. The inverted repeats flanking the LTR sequences are indicated by the arrows below the sequence. The 5-bp direct repeats flanking the insertion site of each element are underlined.

The sequence of the entire region between the LTRs was determined and found to be 4,838 bp in length. Thus, the entire Tca1-2 element, including the LTRs, was 5,614 bp, similar in size to other fungal retrotransposons (5). However, examination of the sequence (16) did not reveal any extended open reading frames (Fig. 4), nor did a search for potential splice sites (28) succeed in uniting the short open reading frames that were present. The sequence was translated in all six theoretical translational reading frames, and the putative peptides were examined for conserved motifs characteristic of the proteases, reverse transcriptases, and integrases of retrotransposons (5). None of these motifs were identified, and comparisons to entries in the nonredundant sequence database at the National Center for Biotechnology Information by using the BLAST algorithm (1) failed to reveal homology to known retrotransposons or retroviruses. These results suggested that Tca1-2 is a highly degenerate vestigial element. A comparison of this sequence with approximately 1,400 nucleotides of the intervening region of Tca1-1 demonstrated that Tca1-1 was more than 99% identical (data not shown).

FIG. 4.

Open reading frame analysis of Tca1-2. Each box represents one of the six theoretically possible open reading frames. Short bars indicate the locations of ATG codons, and full-length bars indicate the positions of stop codons.

Disruption and deletion of Tca1.

The extreme degeneracy of Tca1 suggested that these elements have persisted in the genome for an extensive period of time. Several hypotheses could account for their persistence. One or both elements could be essential by virtue of the location of the insertion or their influence on transcriptional activity at this locus. Alternatively, there may exist constraints preventing recombination between the LTRs and excision of the intervening region. These constraints may be due to recombinational silencing of these chromosomal regions or inherent restrictions on intrachromosomal recombination in C. albicans. To test these possibilities, we examined the genetic behavior and phenotypic consequences of targeted disruptions of both loci containing Tca1. By using URA3 as the selectable marker of the disruption, 5-FOA could be used to counterselect for recombination between the LTRs, which would result in loss of the marker (4). The disruption-deletion strategy is depicted in Fig. 5A. Transformation of strain CAF3-1 with the URA3-disrupted HindIII-SalI fragment from Tca1-1 resulted in Urd⁺ transformants in which the DNA had integrated into either Tca1-1 or Tca1-2 (Fig. 5B). Southern blot analysis of the representative strain CAC-1 showed the expected shift in size of the 3.7-kb EcoRI parental fragment associated with Tca1-1. Under 5-FOA selection, strain CAC-1 gave rise to Urd⁻ segregants at a median frequency of 1 per 10³ cells. Southern blot analysis of a representative Urd⁻ segregant, strain CAC-2, demonstrated that the 3.4- and 4.4-kb EcoRI fragments were gone. Parallel manipulations of a Tca1-2 disruptant resulted in strain CAC-4, in which the 3.25- and 4.8-kb EcoRI fragments characteristic of this locus were absent from the genome. These results indicated that disruption of Tca1-1 and Tca1-2 was not lethal and that recombination between the LTRs occurred readily. In addition, both loci were apparently hemizygous for Tca1, since a single disruption and deletion resulted in complete loss of the element at that locus.

FIG. 5.

Construction of the Tca1-1 and Tca1-2 deletion mutants. (A) Restriction map of the disrupted Tca1-1 and Tca1-2 loci and sizes of the expected EcoRI fragments. (B) Results of Southern blot analysis of genomic DNA from the indicated strains. The DNA was digested with EcoRI and hybridized with the internal region probes of Tca1-1 as indicated in Fig. 2.

To determine if simultaneous deletion of Tca1 from both loci affected viability, strain CAC-2 was again transformed to disrupt the remaining element. One of the transformants, CAC-5, had integrated the transforming DNA into Tca1-2, as indicated by the shift in the position of the parental 4.8-kb EcoRI fragment. 5-FOA selection of Urd⁻ segregants resulted in strain CAC-6, in which both Tca1-1 and Tca1-2 were absent. No gross phenotypic consequences of the deletions were apparent. CAC-6 was indistinguishable from the parental strain CAF3-1 with respect to growth rate at 30°C and germ tube-forming ability (data not shown).

Transcription of Tca1 at both loci.

In previous work, it was observed that Tca1-1 hybridized to an approximately unit-length transcript and expression of this mRNA was strongly regulated in response to the growth temperature (6). Construction of the deletion mutants provided an opportunity to determine whether this transcript was generated from one or both loci. Northern blot analysis of RNA derived from these strains demonstrated that the temperature-regulated transcript was present in both strains CAC-2 and CAC-4, lacking Tca1-1 and Tca1-2, respectively (Fig. 6). The transcript was not present in strain CAC-6, which lacked both copies of Tca1. Thus, Tca1-1 and Tca1-2 are both transcribed in a temperature-regulated manner, and both give rise to apparently full-length transcripts.

FIG. 6.

Northern blot analysis of Tca1 deletion mutants. Total RNA from the indicated strains was hybridized with the internal region probe of Tca1-1 (upper panel). The lower panel shows the ethidium bromide-stained gel prior to blotting.

Population distribution of Tca1.

Since deletion of Tca1-1 and Tca1-2 occurred at normal frequency and had no effect on growth or germ tube formation, the selective advantage of maintaining these degenerate elements was unclear. If they do provide some advantage to the organism, they might be expected to have a high frequency of occurrence among natural populations of C. albicans. Consequently, we examined a large number of clinical isolates from several sources and geographical regions. Southern blot hybridization using a probe from the intervening region of Tca1 demonstrated that about 40% of the strains lacked Tca1 elements, suggesting that there is no strong advantage to maintenance of the element within natural populations (Fig. 7A). Although these strains lacked Tca1, it apparently existed in these strains at one time, as evidenced by the presence of solo α elements in the genome (Fig. 7B). A number of the strains contained restriction fragments identical to those of Tca1-1 or Tca1-2, suggesting their presence in the same loci. Most of the strains had one or more solo α elements in common. Interestingly, we observed no strain with more than three DNA fragments that hybridized with Tca1 or with an excessive accumulation of solo α elements, suggesting that a transpositionally active element no longer exists or is only minimally active in these strains.

FIG. 7.

Southern blot hybridization patterns of clinical isolates of C. albicans. Genomic DNA was digested with EcoRI and hybridized with the combined internal region probes of Tca1-1 (A) or with an α-element probe (B). Isolates 1 to 30 are from the United States; isolates 31 to 37 are from Nanjing, People’s Republic of China.

DISCUSSION

The results of these studies clarified several aspects of the genomic organization and expression of the retrotransposon-like element Tca1. Previous work had established that Tca1 was present at a minimum of two loci in strain SC5314 (6). In support of this conclusion, a second locus containing a copy of Tca1 was cloned. This copy was flanked by a 5-bp duplication of the integration target site which was clearly different from that of the locus previously cloned. In addition to the target site duplication, the clones also differed in both the 5′- and 3′-flanking sequences. These observations demonstrate that Tca1-2 was derived from a distinct chromosomal locus. Genetic analysis demonstrated that these are the only two loci containing Tca1. Sequential deletion of Tca1-1 and Tca1-2 resulted in a strain devoid of the element. Furthermore, only a single deletion step was required to remove the element from each locus, indicating that both loci were hemizygous. Both loci are transcribed, and Northern blot analysis of deletion mutants established that their transcription is modulated in response to the growth temperature. A full-length temperature-regulated transcript was present in strains with deletions of either locus alone. Deletion of both loci resulted in the absence of the transcript providing genetic confirmation of the origin of the transcript.

The Tca1 elements from the two loci were nearly identical in structure, differing by less than 1% in nucleotide sequence. While the LTRs associated with these elements contained features expected of a retrotransposon, the transcribed region between the LTRs was completely degenerate and unrecognizable as a retrotransposable element. The coding elements in retrotransposons can differ in organization, but they typically encode the proteins essential for transposition including a processing protease, reverse transcriptase, and integrase (5). Neither copy of Tca1 contained an extended open reading frame, and the open reading frames that were present did not encode peptides with demonstrable homology to other retrotransposable elements. However, there is considerable sequence variation in retrotransposable elements, and only short functional motifs appear to be conserved (5, 14). None of these motifs could be found in Tca1. Thus, both Tca1-1 and Tca1-2 appear to be defective, vestigial elements, in agreement with a partial sequence analysis of Tca1 reported by Matthews et al. (17). It is likely that the cross-hybridizing elements in other strains are also defective, since hybridization was done at high stringency and the sequence is highly divergent from a functional element.

Since Tca1 is defective in structure, it is unclear how nearly identical copies were duplicated at two loci. If the duplication represented an earlier transposition event, prior to degeneration of the coding regions, then the two loci should have diverged considerably in sequence. However, ectopic gene conversion may have maintained sequence identity between the two loci as they degenerated or could have effected a recent duplication of the degenerated form at the site of a solo α element. Another possibility is that a functional retrotransposon in the genome supplied the necessary functions for retrotransposition. Matthews et al. proposed that pCa1 might play such a role, noting the conservation between Tca1 and pCa1 of plus- and minus-strand primer binding sites, as well as the sequence bordering the LTRs (17).

It is not unusual to find defective copies of retrotransposable elements within a genome. Defective elements are common in other organisms (3, 12, 30). However, these are rarely degenerated to the extent of Tca1. This extreme degeneracy implies that Tca1 has been maintained for an extremely long time and raises the question as to why the element has continued to reside in the genome. The Ty5 retrotransposon of S. cerevisiae exhibits preferential integration into regions of silent chromatin (31–33). If Tca1 were trapped in a region of silent chromatin, this might restrict recombination between the flanking LTRs and prevent the consequent loss of the intervening sequences. However, when URA3 was inserted into Tca1 to allow the selection of such recombinational events, they were isolated at typical frequencies. This result also made clear that neither copy of Tca1 is essential since deletion of either or both had no effect on growth of the deletion mutant. This conclusion is also supported by the observation that a number of natural isolates lack Tca1, suggesting that there is little advantage to maintenance of Tca1.