Gene 272 (2001) 157±164
www.elsevier.com/locate/gene
A DNA polymorphism speci®c to Candida albicans strains exceptionally successful as human pathogens L. Giblin a,1, A. Edelmann b, Ningxin Zhang a, N.B. von Maltzahn a, S.B. Cleland a, P.A. Sullivan a, J. Schmid a,* a
Institute of Molecular BioSciences, College of Sciences, Massey University, Palmerston North, New Zealand b Institute of Biochemistry, University of Leipzig, Leipzig, Germany Received 5 March 2001; received in revised form 11 May 2001; accepted 29 May 2001 Received by J.L. Slightom
Abstract A large proportion of infection-causing isolates of the yeast Candida albicans belong to a general-purpose genotype, identi®able by ®ngerprinting with the moderately repetitive sequence Ca3. The high prevalence of this group ± up to 70% in some patient categories ± suggests that its members possess genetic determinants, which enhance their success as pathogens compared to other strains. To ®nd such determinants we are comparing the genomes of representatives of the general-purpose genotype cluster with the genomes of other strains. In this paper we describe the identi®cation of a 985 bp HpaII fragment (MU13-4) speci®c to general-purpose genotype strains. The fragment was present in 90% of these strains, but only in 10% of other strains. The fragment did not hybridize with probe Ca3, used to de®ne the generalpurpose cluster. It contains elevated levels of repetitive DNA. Sequences homologous to MU13-4 are dispersed throughout the chromosomes of general-purpose strains but are rarer or absent in other strains, as judged by Southern hybridization. Using the Stanford C. albicans genome database, we have placed the MU13-4 fragment next to a CARE-1 element. We also found 79 signi®cant homologies between parts of MU13-4 and 19 other contigs. Attempts to amplify the region surrounding the polymorphic fragment in non-general-purpose genotype strains suggest, as do the hybridization data, that the polymorphism is created by a deletion in non-cluster strains. These results show that it is possible to identify polymorphisms speci®c to general-purpose genotype strains. Primers against the fragment will allow PCR-based discrimination between general-purpose genotype strains and other strains, facilitating investigations aimed at determining morbidity and mortality caused by generalpurpose genotype strains compared to other strains. q 2001 Elsevier Science B.V. All rights reserved. Keywords: Evolution; Pathogenicity; Repetitive elements
1. Introduction Candida albicans is a diploid yeast, which is an important human pathogen (Odds, 1988). Using DNA ®ngerprinting with the sequence Ca3, we have recently provided evidence for the existence of a group of genetically similar strains of C. albicans, which cause infections 10±100 times more often than other groups of strains. The ubiquitous nature of these strains suggests that they have a general-purpose genotype, allowing them to out compete other strains in a wide variety of hosts (Schmid et al., 1999). In a local subset of isolates, we showed that general-purpose genotype strains were more
resilient than other strains and adhered more strongly to saliva-coated surfaces (Schmid et al., 1995). We are now attempting to identify DNA polymorphisms characteristic of general-purpose strains to elucidate the genetic basis of their success as pathogens. Such markers will also allow rapid identi®cation of clinical isolates as general-purpose genotype strains for investigations aimed at determining if these strains cause greater morbidity and mortality.
2. Materials and methods 2.1. Strains, DNA ®ngerprinting and tree construction
Abbreviations: bp, base pair; kb, kilobase; Mb, megabase; ORF, open reading frame; PCR, polymerase chain reaction * Corresponding author. Tel.: 164-6-350-4018; fax: 164-6-350-5688. E-mail address:
[email protected] (J. Schmid). 1 Present address: Xanthon Inc., Research Triangle Park, NC, USA.
Sources of most C. albicans strains were described in earlier studies (Schmid et al., 1995, 1999). Additional strains included were SC5314, used to construct the C. albicans genome map (http://www-sequence.stanford.edu/group/
0378-1119/01/$ - see front matter q 2001 Elsevier Science B.V. All rights reserved. PII: S 0378-111 9(01)00548-0
158
L. Giblin et al. / Gene 272 (2001) 157±164
candida/), its derivative CaI4 (Fonzi and Irwin, 1993), and strain ATCC10261. DNA ®ngerprinting by Southern hybridization with the moderately repetitive sequence Ca3, generation of simple genetic distances from ®ngerprints and construction of neighbour-joining trees from distance matrices, using PAUP*Version 4.0b6 (Sinauer Associates), was carried out as described earlier (Schmid et al., 1999). 2.2. In-gel competitive reassociation, Southern hybridization, PCR ampli®cation, sequencing, and pulsed ®eld electrophoresis DNA was extracted (Scherer and Stevens, 1987) from six strains belonging to the general-purpose genotype cluster and ®ve strains not belonging to the cluster plus one unusual cluster strain, HMHc4 (strains used are marked in Fig. 1). The DNA concentration was determined ¯uorometrically using the dye H33258 (Ausubel et al., 1994). Equal amounts of DNA from the six typical cluster strains were mixed. This mixture was digested with HpaII to generate target DNA for in-gel competitive reassociation (Yokota et al., 1995), using as reference a phosphatase-treated HpaII digest of a mixture of DNA from the six other strains. Recovered DNA was ligated into pGEM-7Zf(1) (Promega Corporation) and used to transform Escherichia coli. Plasmid DNA was isolated from ampicillin-resistant clones. Southern hybridization was carried out using standard methods (Sambrook et al., 1989). Ampli®cations with primers ORF-F (5 0 -GGGGGATGCCGACGACA-3 0 ) and CARE-1R (5 0 -TCCCCTGATTTGATTTCTCTTTC-3 0 ) were carried out using 25 cycles of 948C for 30 s, 56.68C for 30 s and 728C for 90 s. Ampli®cations with primers MU134F (5 0 -GATGATGGGTGGTGGGCAGATTG-3 0 ) and MU134R (5 0 -GCCACGATTTCGGCAACTACGA-3 0 ) were carried out using an annealing temperature of 568C (30 cycles). Ampli®cations with primers MU134NPF (5 0 -CTCATCTCGCTTTGCTTGTC-3 0 ) and MU134NPR (5 0 -ACGCTTTATGCGCACGAAGC-3 0 ) were carried out using 30 cycles of 1 min each at 94, 54 and 728C. All ampli®cations were preceded by 5 min at 948C and included a ®nal incubation at 728C for 5 min for extension. Sequencing was performed on an ABI 100 Automated Sequencer Model 377. Universal forward and reverse primers were used for cloned genomic fragments. Puri®ed ampli®cation products were sequenced directly with the primers used for their ampli®cation. Pulsed-®eld gel electrophoresis was carried out as described by Chindamporn et al. (1998) on a CHEF-DRII system (Bio-Rad Laboratories). Sequences reported in this article appear in GenBank under the Accession numbers AF140351 and AF334763. 3. Results and discussion 3.1. A novel DNA fragment speci®c to general-purpose genotype strains To identify a DNA polymorphism speci®c to general-
purpose genotype strains we attempted a genomic subtraction between a HpaII digest of DNA prepared from generalpurpose strains and a HpaII digest of DNA from other strains; see Fig. 1 for separation of strains into generalpurpose genotype cluster strains and other strains, as de®ned by DNA ®ngerprinting. We recovered 16 different HpaII fragments. We then used each fragment to probe genomic HpaII digests of strains belonging to the cluster and other strains. Fifteen of the fragments were not speci®c to the general-purpose genotype (nine produced bands equal to their size in both general-purpose strains and other strains, two produced such bands only in a small percentage of strains and four gave no signal with any of the strains) and were therefore not investigated further. One fragment, MU13-4, hybridized to a band equal to its size (985 bp) in genomic digests of 23 out of 26 general-purpose strains, but only two out of 20 strains not belonging to the generalpurpose genotype possessed this fragment (see Fig. 1 for summary and Fig. 2 for examples of blots). A possible explanation for the presence of fragment MU13-4 in some strains not belonging to the generalpurpose genotype cluster (and the absence in some cluster strains) could be occasional recombination between strains, as has been suggested for other loci (GraÈser et al., 1996). Further, neighbour-joining tree construction involves averaging of distance values and will not always place all strains consistently (Swofford et al., 1996). We next checked if MU13-4 was simply part of the sequences hybridizing with the Ca3 probe, generating the ®ngerprints, which were used to differentiate generalpurpose strains from other strains. A Southern hybridization was carried out with probe Ca3 of a blot containing a lane with 10 ng MU13-4, ¯anked by lanes containing 2.5 mg of EcoRI digests of C. albicans total DNA. The amount of MU13-4 loaded was roughly equivalent to 70 copies per haploid genome in 2.5 mg total C. albicans DNA (Magee and Scherer, 1998). The C. albicans EcoRI digests produced the expected ®ngerprint patterns, but no hybridization was seen in the lane containing MU13-4 (data not shown). Thus, MU13-4 is not part of Ca3 or related sequences that generate the Ca3 ®ngerprint patterns. 3.2. Sequences homologous to the polymorphic fragment MU13-4 are present in multiple copies on different chromosomes The intensity of the 985 bp band hybridizing with MU13-4 in HpaII digests of C. albicans DNA differed between cluster strains (Fig. 2), indicating differences in copy number. HpaII digests of 80% of general-purpose strains also had three or more additional bands hybridizing with MU13-4; the size of these bands varied between strains (Fig. 1). Indeed, all general-purpose strains tested had at least one such band (Figs. 1 and 2). Slightly more than half of the HpaII digests of strains not belonging to the general-purpose genotype strains also had at least one
L. Giblin et al. / Gene 272 (2001) 157±164
band hybridizing with MU13-4, although only 25% of these strains had three or more of such bands (Figs. 1 and 2). When the chromosomes of four general-purpose
159
C. albicans strains, RIHO1, hp10bt, ATCC10261 and SC5314, and two other strains, hol c and hp6ch, were separated by pulsed-®eld gel electrophoresis (Fig. 3),
Fig. 1. Neighbour-joining tree showing the relationships, based upon Ca3 ®ngerprinting, between stains used in this study. The bar represents a genetic distance of 0.05. The short-branched region above the dashed line contains strains belonging to the general-purpose genotype cluster. Two strains, ATCC10261 and HMHc4, are grouped with the cluster but the long branches separating them from the remaining cluster strains suggest that they are unusual (borderline) cluster strains. Strains possessing MU13-4 (as judged by Southern hybridization of a genomic HpaII digest with MU13-4 at 985 bp; see Section 3) are marked with a ®lled dot; strains possessing one to two fragments which hybridize with MU13-4 at a different molecular weight are marked with a (1); strains with more than two of such fragments are marked with (1 1 ). Strains used for in-gel competitive reassociation are marked with an asterisk in an open circle. One unusual cluster strain (HMHc4) was part of the reference set.
160
L. Giblin et al. / Gene 272 (2001) 157±164
3.3. Sequence of MU13-4 and the surrounding area
Fig. 2. Examples of Southern hybridization of HpaII digests of C. albicans strains with MU13-4. Equal amounts of DNA were loaded in all lanes, and this, as well as complete digestion of the DNAs, was con®rmed by visual analysis of the ethidium-bromide stained gels. Prehybridization and hybridization were performed at 658C using Rapid-Hyb Buffer (Amersham Pharmacia Biotech). Blots were washed with 0.2 £ SSC, 0.1% SDS at 658C. Symbols next to strain names are the same as used in Fig. 1 to categorize hybridization with MU13-4. Triangles and a dashed line mark the position of MU13-4 (985 bp).
MU13-4 hybridized to six chromosomes in all generalpurpose strains which possessed the 985 bp band. It also hybridized to six chromosomes in the one non-generalpurpose genotype strain, hp6ch, which lacked the 985 bp band but had other HpaII bands hybridizing to MU13-4. Strain hol c, which digests showed no hybridization with Mu13-4 in Southern blots, showed no hybridization. Likewise, no hybridization was detected with Sacharomyces cervisiae chromosomes, included as molecular weight markers. The presence of sequences homologous to MU13-4 in multiple locations of the genome of generalpurpose strains was further con®rmed when we sequenced the fragment (the sequence is shown in Fig. 6 and its characteristics will be discussed in Section 3.3) and compared its sequence with the Stanford C. albicans genome database, generated from strain SC5314 (Fig. 4; SC5314 falls within the general-purpose genotype cluster; see Fig. 1). Two contigs, 6-2456 and 6-2515, showed 99% homology with the entire fragment. Both contigs are .99% identical to each other for several thousand base pairs surrounding the region containing MU13-4. In addition, we found 79 other signi®cant homologies between MU13-4 and 19 other contigs. The homologies, with 76 to 56% sequence identities, involved parts of MU13-4 of 980 to 93 bp in length (Fig. 4).
Fig. 5A shows regions homologous to MU13-4 in various strains aligned with base pairs 122451±127543 of contig 62515 of the Stanford C. albicans genome database (http:// www-sequence.stanford.edu/group/candida/). Mu13-4 lies in a region of the contig that contains an elevated number of direct repeats (Fig. 5B); the largest repeated sequences are marked in Fig. 5A. Some 800 bp upstream lie two overlapping small putative ORFs with inverse orientations. Putative ORF1 shows homology to two plant cell wall glycoproteins (GenBank Accession numbers: pir:S54157 and pir:S33309; identities of 40/161 (24%) and 31/128 (24%), respectively), mainly due to a PHHHH 1 LHR motif. Putative ORF2 has less homology to sequences in the protein database (the best match had an identity score of 25/97 (25%)). Both are preceded by AT-rich sequences containing TATA boxes (and a CAT box in the case of putative ORF1) within 200 bp from the start codon. In both, the stop codon is followed by an AT-rich region containing sequences resembling the consensus for polyadenylation and transcription termination in yeast (Guo and Sherman, 1996). Five hundred base pairs downstream of
Fig. 3. Hybridization of C. albicans chromosomes with MU13-4. Molecular weights shown are derived using S. cerevisiae chromosomes as markers. Chr1 signi®es Chromosome 1, identi®ed using an actin gene as a probe (C. albicans ACT1; nucleotides 1851±2354; GenBank Accession number: X16377). Prehybridization and hybridization were performed at 658C using Rapid-Hyb Buffer (Amersham Pharmacia Biotech). Blots were washed with 0.1 £ SSC, 0.1% SDS at 658C (high stringency). Symbols next to strain names are the same as used in Fig. 1 to categorize hybridization of HpaII digests with MU13-4.
L. Giblin et al. / Gene 272 (2001) 157±164
161
Fig. 4. Summary of local alignments of MU13-4 with contigs in the Stanford database using BLAST (http://www-sequence.stanford.edu:8080/bncontigs.html). Every alignment is shown as a separate vertical bar, placed on the X-axis in descending order of percent homology. The length and position of the bars indicates the part of MU13-4 showing signi®cant homology to contigs.
Fig. 5. (A) Overview of MU13-4 and surrounding area. Sequences of PCR products obtained from cluster strains hp10bt (primers ORF-F5 0 and CARE-1R5 0 ), ATCC10261 (primers MU134F and MU134R) and CARE-1 (GenBank Accession number: AF140351) have been aligned with a 5000 bp region of contig 62515 from strain SC5314. The original MU13-4 fragment cloned by us is not shown separately; it is identical to its homologous region in hp10bt except for four point mutations, marked with an asterisk above the bar representing the hp10bt sequence. Grey lines, black dashed lines and plain black lines indicate 10 bp regions in the hp10bt, ATCC10261 and CARE-1 sequences that differ from the SC5314 sequence by 10±20, 30±60 and 70±100%, respectively. Where these differences involve deletions of $5 bp a small black triangle is placed next to the bar. Sequences were directly aligned with each other using Clustal, except in the case of CARE-1 where the alignment was improved by splitting off a short piece and aligning it with a different region of the contig. The SC5314 bar also shows the three putative ORFs as numbered arrows (No. 3 is only present in ATCC10261). Black rectangles connected by black lines within the SC5314 sequence bar represent major repeats deduced for a matrix plot (B), showing the density of repetitive DNA in contig 6-2515 in SC5314.
162
L. Giblin et al. / Gene 272 (2001) 157±164
MU13-4 lies a region homologous to parts of the C. albicans repetitive element, CARE-1 (Lasker et al., 1991). Ampli®cation of a 2.3 kb region surrounding MU13-4, including the area homologous to CARE-1, from strain hp10bt, a typical
cluster strain according to its Ca3 ®ngerprint, yielded a sequence 99.3% identical with that of SC5314 (Figs. 5 and 6). To determine the extent of polymorphism between
Fig. 6. Alignment of the 2.3 kb hp10bt region and corresponding regions in contig 6-2515 and ATCC10261. Base pair numbers refer to those of the sequenced 2.3 kb product from hp10bt.
L. Giblin et al. / Gene 272 (2001) 157±164
general-purpose stains and other strains we ®rst ampli®ed and sequenced the region corresponding to MU13-4 in strain ATCC10261. This strain is an untypical generalpurpose genotype cluster strain (Fig. 1). The 5 0 half of the ATCC10261 product was almost identical to the SC5314 and hp10bt sequences, but the remainder diverged considerably (Figs. 5 and 6). The ATCC10261 product also contained a very short putative ORF (No. 3 in Fig. 5) not present in the other strains. This putative ORF was surrounded by AT-rich sequences but the homology to promoter and polyadenylation regions was less pronounced than in putative ORF1 and ORF2. An attempt to obtain a PCR product from non-generalpurpose genotype strains using primers ORF5 0 and CARE1R5 0 , homologous to ¯anking sequences 800 bp removed from the edges of MU13-4, failed, but all general-purpose genotype strains tested produced a product of the expected size. Since one of the primers is located in CARE-1 this suggests that the linkage between CARE-1 and MU13-4 is conserved in cluster strains. Primers MU134NPR and MU134NPF, which would amplify the putative ORF1 and ORF2, plus ¯anking regions which might be involved in their expression and polyadenylation, yielded products not only in cluster strains (the tested strains were Sc5314, CLB42, jam 2c, and RIHO 10) but also in four out of seven non-cluster strains tested (strains gaymc c, svobo c, W13 and HMHc12 yielded a product, whereas sw 17c, OTG19 and hp15bt did not). All products obtained were of identical size as judged by gel electrophoresis (approximately 1300 bp, as expected based on the sequence data). These data suggest, as do the Southern hybridization data, that (i) MU13-4 marks a deletion in non-general-purpose genotype strains (or an insertion in general-purpose genotype strains), (ii) that the size of the DNA missing differs between different non-general-purpose genotype strains, and (iii) that the missing DNA does not include the putative ORF in all non-general-purpose genotype strains. Our data show that it is possible to identify DNA polymorphisms speci®c to the general-purpose genotype strains, which are not associated with the Ca3 ®ngerprinting polymorphisms used to identify these strains. The existence of general-purpose genotype-speci®c polymorphisms con®rms that the cluster is an entity with speci®c genetic characteristics. The polymorphism detected can be used to identify general-purpose genotype strains by simple PCR assays for investigations into differences of morbidity and mortality caused by general-purpose cluster strains and other strains. Our data suggest that with comparative genomic approaches it should be possible to identify additional polymorphisms which explain the success of the general-purpose genotype in causing infection (Schmid et al., 1999), and the enhanced resilience and drug resistance of this cluster (Schmid et al., 1995), i.e. identify loci which determine C. albicans' ability to cause disease. The most likely role in this context of the locus identi®ed by MU13-4 would seem to be that its repetitive nature assists strains in generating genetic variability
163
by mitotic recombination, thus making the genomes of general-purpose genotype strains more ¯exible and thus more likely to evade host defences (Matthews, 1994). We note in this context that although general-purpose cluster strains are genetically less diverse than other strains, we have evidence suggesting that subpopulations of generalpurpose strains in different geographical regions diverge from each other faster than is the case for other groups of C. albicans strains (Schmid et al., 1999). These observations would suggest that general-purpose strains are of fairly recent origin, compared to other groups of strains (and therefore as a group less diverse), but show higher genome plasticity (explaining the faster divergence between geographically separated subpopulations). 3.4. Conclusions 1. Our data verify that the general-purpose genotype strains identi®ed by ®ngerprinting with the sequence Ca3 have genetic characteristics, not linked to Ca3, which set them apart from other strains, and that these can be identi®ed by genomic comparisons. 2. Sequences similar to the locus marked by MU13-4 are dispersed throughout the genomes of general-purpose genotype strains, presumably enhancing their genetic plasticity. Such sequences are rarer or absent in other strains.
Acknowledgements We thank John Tweedie for critical reading of the manuscript. This work was supported by grants from the Massey University Research Fund and the Royal Society of New Zealand. References Ausubel, F.M., Brent, R., Kingston, R.E., Moore, D.D., Smith, J.A., Seidman, J.G., Struhl, K., 1994. Current Protocols in Molecular Biology. Wiley, New York. Chindamporn, A., Nakagawa, Y., Mizuguchi, I., Chibana, H., Doi, M., Tanaka, K., 1998. Repetitive sequences (RPSs) in the chromosomes of Candida albicans are sandwiched between two novel stretches, HOK and RB2, common to each chromosome. Microbiology 144, 849±857. Fonzi, W.A., Irwin, M.Y., 1993. Isogenic strain construction and gene mapping in Candida albicans. Genetics 134, 717±728. GraÈser, Y., Volovsek, M., Arrington, J., SchoÈnian, G., Presber, W., Mitchell, T.G., Vilgalys, R., 1996. Molecular markers reveal that population structure of the human pathogen Candida albicans exhibits both clonality and recombination. Proc. Natl. Acad. Sci. USA 93, 12473±12477. Guo, Z., Sherman, F., 1996. 3 0 -end-forming signals of yeast mRNA. Trends Biochem. Sci. 21, 477±481. Lasker, B.A., Page, L.S., Lott, T.J., Kobayashi, G.S., Medoff, G., 1991. Characterization of CARE-1: Candida albicans repetitive element-1. Gene 102, 45±50.
164
L. Giblin et al. / Gene 272 (2001) 157±164
Magee, P.T., Scherer, S., 1998. Genome mapping and gene discovery in Candida albicans. ASM News 64, 505±511. Matthews, R.C., 1994. Pathogenicity determinants of Candida albicans: potential targets for immunotherapy? Microbiology 140, 1505±1511. Odds, F.C., 1988. Candida and Candidosis, 2nd Edition. Bailliere Tindall, London. Sambrook, J., Fritsch, E.F., Maniatis, T., 1989. Molecular Cloning: A Laboratory Manual, 2nd Edition. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Scherer, S., Stevens, D.A., 1987. Application of DNA typing methods to epidemiology and taxonomy of Candida species. J. Clin. Microbiol. 25, 675±679. Schmid, J., Hunter, P.R., White, G.C., Nand, A.K., Cannon, R.D., 1995. Physiological traits associated with success of Candida albicans strains
as commensal colonisers and pathogens. J. Clin. Microbiol. 33, 2920± 2926. Schmid, J., Herd, S., Hunter, P.R., Cannon, R.D., Yasin, M.S.M., Samad, S., Carr, M., Parr, D., McKinney, W., Schousboe, M., Harris, B., Ikram, R., Harris, M., Restrepo, A., Hoyos, G., Singh, K.P., 1999. Evidence for a general-purpose genotype in Candida albicans, highly prevalent in multiple geographic regions, patient types and types of infection. Microbiology 145, 2405±2414. Swofford, D.L., Olsen, G.J., Waddell, P.J., Hillis, D.M., 1996. Phylogenetic inference. In: Hillis, D.M., Moritz, C., Mable, B.K. (Eds.), Molecular Systematics. Sinauer Associates, Sunderland, MA, pp. 407±514. Yokota, H., Amano, S., Yamane, T., Ataka, K., Kikuya, E., Oishi, M., 1995. Differential cloning using in-gel competitive reassociation. Electrophoresis 16, 286±290.
