Telomere length control and transcriptional regulation of subtelomeric adhesins in Candida glabrata

RIF1 controls telomere length in C. glabrata

We sequenced the RIF1 gene (GenBank AY646924) and found that overall it was 24% identical and 41% similar to the S. cerevisiae RIF1 gene. Upstream of RIF1, we found an orthologue of PPS1 which is also located upstream of the same locus in S. cerevisiae. To determine if rif1Δ mutants in C. glabrata have defects in telomere length regulation, as expected based on the rif1Δ phenotype of S. cerevisiae, we carried out Southern analysis of genomic DNA from four strains: the wild-type strain, and three hyperadherent strains –BG509 (rif1Δ), BG542 (sir3::Tn7) and BG550, which carries the large insertion downstream of EPA7. We digested the DNA with either of two restriction enzymes and probed the blot with an oligonucleotide corresponding to the telomeric repeat. The resulting pattern shows a collection of fragments corresponding to the distal fragments of each chromosome (Fig. 1). Telomeres are not of uniform length in a population of yeast cells and so, in wild type cells the terminal telomeric fragments run as a smear, which is particularly obvious among the smaller bands on the gel. In the rif1Δ background, the average molecular weight of the fragments increases by 300–400 bp, corresponding to an increase in the number of telomere repeats and a concomitant increase in molecular weight of the telomeric fragments. This same effect is not seen in the genomic DNA from the other two hyperadherent mutants demonstrating that an increase in telomere length is not a common feature of hyperadherent mutants but is specific to the rif1Δ mutant. We conclude that in C. glabrata, as in S. cerevisiae, RIF1 is important for correct regulation of telomere length.

Which EPA genes are responsible for the hyperadherent phenotypes of sir3Δ and rif1Δ strains?

Epa1 is able to mediate adherence to a variety of epithelial cells. We tested whether EPA1 overexpression in the sir3Δ or rif1Δ background might be responsible for the hyperadherent phenotype in those two backgrounds. rif1Δepa1Δ and sir3Δepa1Δ strains were still hyperadherent (see Table 3), suggesting that additional adhesins were induced in these two backgrounds. We also tested whether the four additional previously identified EPA genes (De Las Peñas et al., 2003) might be responsible for the sir3Δ or rif1Δ hyperadherent phenotypes by determining if they are even capable of mediating adherence to cultured epithelial cells. While EPA1 expressed in S. cerevisiae allows S. cerevisiae to bind to the Lec2 epithelial cell line (Table 2 and Cormack et al., 1999), we found that expression of EPA2, 3, 4 or 5 in S. cerevisiae did not result in adherence to Lec2 cells (Table 2). This could have been resulting from a failure of expression of EPA2−5 in S. cerevisiae. We do not have specific antibodies to any of the Epa proteins, and therefore could not directly monitor protein levels in these experiments. To partially address this issue, we made chimeric genes fusing the putative N-terminal ligand-binding domains of Epa1–5 with an epitope tagged C-terminal region of Epa1p. All five chimeric proteins were expressed at the S. cerevisiae cell surface as measured by labelling with antibody to the common epitope tag (data not shown), but only constructs containing the N-terminal ligand-binding domain of Epa1 were able to meditate adherence to epithelial cells (data not shown). This was consistent with the data in Table 2 and suggests that Epa2–5 do not mediate adherence to Lec2 cells.

These data suggested strongly that the hyperadherent phenotype of the sir3Δ and rif1Δ strains was because of misregulation of genes (possibly EPA genes) other than EPA1–5. As a first step in identifying the adhesins responsible for the hyperadherent phenotype of the sir3Δ and rif1Δ strains, we first characterized in more detail the complement of EPA genes present in BG2, our standard lab strain. We had previously characterized two loci containing a total of five EPA genes (De Las Peñas et al., 2003). Sequence of the C. glabrata genome (strain CBS138) revealed that there were a number of sequences with homology to the N-terminal ligand-binding domain of EPA1. Using PCR amplification of genomic DNA from strain BG2, our lab strain, we found that the majority of the EPA-related sequences from strain CBS138 were also present in BG2. We made of list of 16 distinct putative EPA-related sequences that were demonstrably present in strain BG2. For each gene, we designed specific oligonucleotides to assess transcript levels by RT-PCR or S1 nuclease protection (see Experimental procedures). We first assessed, using RT-PCR, whether EPA family gene expression was altered in a rif1Δ background. We found that expression for the majority of EPA genes was not significantly increased in the rif1Δ background. Expression of EPA2,3,4/5,6,7,12 and 15 were increased as measured by RT-PCR (TableS1), but only expression of EPA1 EPA4/5 and EPA6 was significantly affected in the rif1Δ background, as measured by S1 nuclease protection assay (Fig. 2, TableS1 and data not shown). EPA4 and EPA5 are present in the BG2 genome as an inverted repeat and are essentially identical (De Las Penas et al., 2003); because they cannot be distinguished in our S1 assay, we treat them as a single gene (EPA4/5) for the purposes of our analysis. In the sir3Δ mutant background there was a significant transcriptional induction of EPA1, EPA6 and EPA7 as determined by S1 nuclease protection assay (Fig. 2). Although the EPA6 and EPA7 genes are highly homologous at the nucleotide level (see below), the S1 probes used are specific for each gene, as shown by controls in Fig. 2, in which the EPA6 signal, but not the EPA7 signal is lost in an EPA6 null strain, while loss of the EPA7 signal occurs only with subsequent deletion of EPA7.

These experiments moved us to focus on EPA6 and EPA7 as potential adhesins responsible for the hyperadherent phenotypes in the sir3Δ and rif1Δ mutant strains.

EPA6 and EPA7 are two novel subtelomeric EPA genes encoding adhesins

The results above suggested that EPA6 and EPA7 were potentially responsible for the adherence phenotype of the rif1Δ and sir3Δ mutants. We identified three fosmids that contain EPA6 from a previously generated fosmid library of the C. glabrata genome (Castaño et al., 2003). Sequence analysis of one of the fosmids revealed that EPA6 is predicted to encode a GPI-anchored cell wall protein (GenBank AY646925). In the N-terminal ligand-binding domain, EPA6 is a close relative of EPA1, sharing about 72% amino acid identity. By contrast, the C-terminal region is quite different and shares little homology with any of EPA1–5. We expressed full length EPA6 as well as a chimeric protein in which the Epa6 N-terminal domain is fused to the epitope-tagged C-terminal domain of Epa1 in S. cerevisiae and tested how well the strains could adhere to epithelial cells. In both cases, the constructs conferred on S. cerevisiae the ability to adhere avidly to Lec2 epithelial cell (Table 2 and data not shown). Thus, EPA6, whose transcription is strongly induced in a rif1Δ background, can mediate adherence to human epithelial cells.

We next characterized the genomic region surrounding the EPA6 locus. Sequence of the EPA6 containing fosmid showed that upstream of EPA6, there was a long intergenic region of almost 10 kb before the next ORF (GenBank AY646925). The nearest upstream gene is predicted to encode an apparent homologue of the S. cerevisiae gene DAN1(Fig. 3). Like the EPA genes, DAN1 is predicted to be a GPI-anchored cell wall protein, but it does not share any homology to the EPA genes in the N-terminal ligand-binding domain. From fosmids containing EPA6, we obtained approximately 1 kb of non-coding sequence 3′ to the EPA6 ORF and used chromosome walking to clone additional sequence downstream of EPA6 (Experimental procedures). We found that EPA6, like EPA1–5 (De Las Peñas et al., 2003) is subtelomeric, being located 2.435 kb from the telomere (Fig. 3).

We cloned and sequenced the EPA7 locus from strain BG2 as well (GenBank Accession No. AY646926). EPA7 is highly related to EPA6, sharing 94% identity over the entire length of the gene. We expressed either full length EPA7 or a chimeric protein in which the Epa7 N-terminal domain was fused to the epitope-tagged C-terminal domain of Epa1 in S. cerevisiae and tested how well the strains could adhere to epithelial cells. In both cases, the constructs conferred on S. cerevisiae the ability to adhere avidly to Lec2 epithelial cell (Table 2 and data not shown). The genomic EPA7 locus is highly homologous to the EPA6 locus. We sequenced 3.94 kb upstream of EPA7 in strain BG2, and over this region there are only 80 single nucleotide differences with the EPA6 locus (97.9% homology). In addition, there are two insertions present in the EPA7 promoter region relative to EPA6: a 20 bp insertion 2.21 kb upstream, and a 0.184 kb insertion 409 bp upstream of the start site of translation. EPA7, like EPA6, is subtelomeric, located 2.496 kb away from the telomere (Fig. 3). The sequence between EPA7 and its telomere is also highly homologous to the EPA6 locus (97.43% identity). In this region, there is also a 63 bp insertion (1296 bases downstream of the stop codon) in EPA7 relative to EPA6.

EPA6 and EPA7 mediate epithelial adherence in the rif1Δ and sir3Δ strains

We showed above that EPA6 and EPA7 can both function as adhesins when expressed in S. cerevisiae, and that they are overexpressed in a sir3Δ or rif1Δ background. We were interested in whether they were necessary for the hyperadherent phenotype of sir3Δ and rif1Δ strains, or whether there might be other adhesins also derepressed in the sir3Δ or rif1Δ strains. We made a comprehensive series of EPA gene deletions in rif1Δ or sir3Δ backgrounds and measured the ability of the multiply mutant strains to adhere in vitro to Lec2 epithelial cells. Deletion of EPA1 alone or EPA7 alone in a rif1Δ did not reduce the adherence to Lec2 cells, while deletion of EPA6 in the rif1Δ background reduced adherence by about half (Table 3). When we deleted both EPA1 and EPA7 in the rif1Δ background, adherence to Lec2 cells was not decreased, whereas deleting EPA1 and EPA6 in this background strongly reduced adherence, showing that EPA1 and EPA6 both contribute significantly to the hyperadherent phenotype (Table 3). A rif1Δ strain in which EPA1, EPA6 and EPA7 are all deleted was approximately twofold less adherent than the rif1Δepa1Δepa6Δ strain, and adhered at levels comparable to the background adherence of the wild type strain.

Our analysis of the sir3Δ mutant yielded somewhat different results. As shown in Table 3, in vitro adherence to Lec2 cells did not decrease when either EPA1, EPA6 or EPA7 was deleted in a sir3Δ background. Even deletions of any two of these three EPA genes did not reduce adherence. However, the triple deletion of EPA1, EPA6 and EPA7 decreased adherence of a sir3Δ strain 2.4-fold. We conclude that in the sir3Δ background, the EPA1, EPA6 and EPA7 genes all do contribute to adherence, but that there are probably other adhesins also induced that mediate adherence even in the absence of EPA1, EPA6 and EPA7.

Subtelomeric localization of EPA6 and EPA7 is required for silencing

As described above, we initially isolated three hyperadherent mutants: the first two were the insertions in RIF1 and SIR3, described above; the third hyperadherent strain (strain BG550) contained a large insertion 423 bp downstream of EPA7. The large insertion in the original hyperadherent mutant increased the distance between EPA7 and its telomere, placing its promoter around 20 kb away from the telomere. The phenotype of this insertion mutant suggested that the distance between EPA7 and the telomere was critical for its repression. We therefore first tested whether EPA6 and EPA7 loci are subject to subtelomeric silencing and second whether that silencing depends on proximity to the telomeric sequences.

We integrated the URA3 gene at the EPA6 and EPA7 loci and used a 5-FOA plate assay to assess if it was silenced (Fig. 4A). We found that the EPA6 and EPA7 loci, like the EPA1−3 and EPA4−5 loci (De Las Peñas et al., 2003) were subject to transcriptional silencing because strains carrying the URA3 gene integrated at the EPA6 and EPA7 loci were able to grow in the presence of 5-FOA showing that URA3 gene at these positions is transcriptionally repressed (Fig. 4B). The subtelomeric silencing at both the EPA6 and EPA7 loci is dependent on some of the proteins known to be involved in subtelomeric silencing in S. cerevisiae. As shown in Fig. 4B we generated the same URA3 insertions at the EPA6 and EPA7 loci in three mutant strains – a rif1Δ strain, a sir3Δ strain and a rap1–21 strain in which the C-terminal 28 amino acids of Rap1 have been deleted rendering it unable to interact with the Sir complex (De Las Peñas et al., 2003). In all three backgrounds, and at both loci, the silencing of the URA3 gene was eliminated, demonstrating that repression of the loci requires intact RIF1, RAP1 and SIR3. Transcriptional repression of the EPA6 and EPA7 genes themselves depends on RIF1, RAP1 and SIR3 as well, because EPA6 and EPA7 transcription was derepressed in all three mutant backgrounds (Fig. 2).

Consistent with silencing at the EPA7 locus being determined by distance from the telomere, we found that the extent of silencing, as measured by URA3 repression, changes as a function of distance from the telomere. We constructed strains with URA3 at two different positions at the EPA7 locus. Strain BG733 and BG826 contains the URA3 marker 13.9 kb and 4.89 kb from the telomere, respectively (see Fig. 4A insertions 2 and 3). As shown in Fig. 4B, the silencing of the URA3 gene, as measured by growth on 5-FOA plates, decreased as the URA3 marker was moved further from the telomere.

Deletion of RIF1 or SIR3 alters virulence in a systemic model of candidiasis

Because deletion of RIF1 and SIR3 altered expression of the EPA gene family, which we believe to be important in host colonization, we examined the role of these two genes in virulence using a systemic model of infection. We infected groups of 10 mice with five isogenic strains: BG462, a wild type C. glabrata strain, BG562, a derivative of BG462 in which the RIF1 ORF has been deleted, BG1033, a derivative of BG562 in which the RIF1 gene has been restored, BG808, a derivative of BG462 in which the SIR3 ORF has been deleted, and BG1043, a derivative of BG808 in which the wild-type SIR3 has been restored. The animals were sacrificed after 7 days and colony-forming units (cfu) from liver, kidney and spleen were determined. Interestingly, colonization of kidney by the rif1Δ or sir3Δ strains was significantly increased relative to the wild type parent strain (Fig. 5A and B). For assessing statistical significance, we calculated pooled P-values for pairwise comparisons (Mann–Whitney test). Strain BG562 (rif1Δ) showed significantly higher colonization of kidney than the wild-type strain BG462 (P = 0.023); restoration of RIF1 (strain BG1033) reversed the rif1Δ (BG562) mutant phenotype (P = 0.0065). To test if hypercolonization of kidney in the rif1Δ strain was related to derepression of EPA1 or EPA6, we compared colonization of the rif1Δ strain to strain BG1040, a derivative of the rif1Δ strain in which EPA1 and EPA6 were deleted. This strain showed some decrease in kidney colonization relative to the rif1Δ strain but this difference was not significant (P = 0.41). Deletion of SIR3 also resulted in increased colonization of kidney because strain BG808 (sir3Δ) showed significantly higher kidney colonization than the wild-type strain BG462 (P = 0.0004). No increase in kidney colonization was seen in which the SIR3 gene was restored (strain BG808) (P = 0.016). The hypercolonization of kidney in the sir3Δ strain depends, at least in part, on derepression of EPA genes because deletion of EPA1, EPA6 and EPA7 in the sir3Δ strain (strain BG1027), significantly decreased (P = 0.003) kidney colonization relative to the sir3Δ strain (strain BG808).

There was no difference in recovered cfu from liver and spleen between the wild type and sir3Δ strains. By contrast, there was a reproducible and significant reduction in colonization of liver and spleen in the rif1Δ strain. Strain BG562 (rif1Δ) showed significantly lower colonization of liver (P = 0.0004) and spleen (P = 0.0004) relative to the wild-type strain (BG462). Restoration of the RIF1 gene reversed this reduction in colonization of both liver (P = 0.04) and spleen (P = 0.09); however, the reduction in liver and spleen colonization was not reversed by deletion of EPA1 and EPA6 in the rif1Δ background suggesting that the decreased colonization of liver and spleen in vivo may not be related to inappropriate derepression of EPA6 and EPA1 in the rif1Δ strain.

Strains

All strains used in this study are listed in TableS2. Strain BG676 (sir3Δ) has been previously described (De Las Peñas et al., 2003).

Plasmids

All plasmids used in this work are summarized in TableS3.

Primers

All primers used for cloning, RT-PCR, and sequencing are described in TableS4.

Epithelial cell lines and growth and fixing conditions

Lec2 cells are CHO-derived epithelial cells (ATCC No. CRL-1736). A-498 cells are human epithelial cells from kidney (ATCC No. HTB-44), and T24 cells are human epithelial cells from bladder (ATCC No. HTB-4).

All epithelial cell lines were grown according to supplier's instructions, in media supplemented with 10% fetal bovine serum. Cells were seeded in 24-well plates and allowed to grow for 24–48 h at 37°C until the monolayer was confluent (∼10⁵cells per well). The media was aspirated and cells were fixed with 2% formaldehyde in PBS for 1–2 h. Each well was washed 4–5 times with PBS. One millilitre of PBS supplemented with Pen/Strep (Penicillin 100 U ml⁻¹, Streptomycin 100 µg ml⁻¹) was added to each well for storage at 4°C for several months.

Media and growth conditions

Yeast cells were grown on standard yeast media as described (Sherman et al., 1986), and 2% agar was added for plates. Synthetic complete (SC) contains YNB without NH₂SO₄, 1.7 g l⁻¹, NH₂SO₄ 5 g l⁻¹, and supplemented with 0.6% of casamino acids and 2% glucose. When needed, SC was supplemented with 25 mg l⁻¹ of uracil. To score for resistance to 5-fluorotic acid, 1.1 g l⁻¹ 5-FOA and 25 mg l⁻¹ of uracil was added to SC. YPD contains yeast extract 10 g l⁻¹, peptone 20 g l⁻¹, supplemented with 2% glucose. To plate tissue homogenates, YPD plates were supplemented with Pen/Strep (Penicillin 100 U ml⁻¹, Streptomycin 100 µg ml⁻¹). Bacterial media were prepared as described (Ausubel et al., 1987), and 1.5 g l⁻¹ of agar was added for plates. Luria-Bertani (LB) media contains bactopeptone 10 g l⁻¹, yeast extract 5 g l⁻¹ and  NaCl 10 g l⁻¹. When needed LB plates  were supplemented with 30 mg l⁻¹ of kanamycin, or 100 mg l⁻¹ of carbenicillin. Phosphate buffered saline (PBS) contains 8 g l⁻¹ NaCl, 0.2 g l⁻¹ KCl, 1.15 g l⁻¹ Na₂HPO₄.7H₂O, 0.2 g l⁻¹ KH₂PO₄.

Escherichia coli strain BW23473 (Table S2) was used for maintenance of conditional replicons carrying R6Kγ origin of replication. Otherwise, strain DH10 (Gibco BRL) was used routinely for electroporation of plasmids.

Media for growth of the CHO-derived epithelial cell line Lec2 (ATCC CRL-1736) is Alpha minimum essential medium (α-MEM)supplemented with 10% fetal bovine serum. Media for growth of A498 cell line was α-MEM supplemented with 2% sodium pyruvate, non-essential aminoacids and 10% fetal bovine serum. T24 cells were grown in McCoy medium supplemented with 10% fetal bovine serum.

Transformation

Yeast transformation with linear or supercoiled plasmids was done as previously described (Castaño et al., 2003).

Sequence

To retrieve DNA flanking the insertion mutation from either the library made by non-homologous insertion of YIplac211 (Cormack et al., 1999), or the library generated by homologous recombination of in vitro generated Tn7 insertions in C. glabrata genomic clones (Castaño et al., 2003), genomic DNA from the mutants of interest was prepared, digested with either HindIII or EcoRI for the YIplac21I library, or with SpeI or XbaI for the Tn7 library, ligated and transformed into E. coli strain BW23473. Plasmid DNA from the recovered mutants was prepared using Qiagen minipreps and then used as templates for sequencing with universal primers for YIplac211-derived vectors, and primers ♯661 and ♯662 (TableS4) for Tn7 vectors. These primers anneal to the Tn7 ends and face outward.

Retrieving telomeres downstream of EPA6 and EPA7

To clone the fragment downstream from EPA6 containing the telomere, we first cloned a 1.025 kb PCR fragment containing the last 448 nt of the coding sequence of EPA6 and 577 bp downstream from the EPA6 ORF into vector pNEB193 using the fosmid 6–77 (which contains EPA6) as template and primers ♯1693 and ♯1694. The resulting plasmid was mutagenized with Tn7 as described (Castaño et al., 2003) and one insertion (pSP121, TableS3) was selected to integrate by homologous recombination back into C. glabrata to generate strain BG582; insertion at the correct site was verified with primers external to the cloned and mutagenized fragment. This insertion is 98 bp upstream of the stop codon of EPA6. Genomic DNA from BG582, was digested with KpnI and blunted with T4 polymerase and religated. This plasmid was sequenced using primers ♯1685, ♯1791, ♯1799 and ♯1844; and shown to contain at least 12 repeats of the C. glabrata telomere repeats.

To clone the fragment downstream from EPA7 containing the telomere, we integrated plasmid pIC62 by homologous recombination at the NheI site of EPA7 generating a partial duplication of 1420 bp of the EPA7 locus. Insertion at the correct chromosomal site was verified by PCR as above. Genomic DNA from this strain (BG748) was prepared and digested with BstEII, blunted with T4 polymerase, religated and transformed into BW23473 electrocompetent cells. The resulting plasmid (pIC83) was sequenced with primers ♯ 1824, ♯1799 and ♯1844 and the sequence revealed at least 12 repeats of the telomere repeat of C. glabrata.

Cloning of EPA7

Because EPA7 is not represented in our C. glabrata genomic library (it is very close to the telomere and therefore under-represented), we cloned the full length EPA7 gene from strain BG826 which contains an insertion of Tn7 273 bp downstream of EPA7. Genomic DNA from this strain was digested with BseRI and BsrBI, blunted with T4 polymerase and ligated and transformed into electrocompetent BW23473 cells. The resulting plasmid (pIC90) contained full length EPA7 plus 3.944 kb of upstream sequences and 1.087 kb of downstream sequences. pIC90 was sequenced with primers ♯1617, ♯1648, ♯1697, ♯1868 and ♯1873.

Construction of deletion strains

We used the two-step gene replacement strategy to generate the deletion strains used in this work. Briefly, we cloned two fragments, one immediately upstream and one immediately downstream of the gene to be deleted in the integrative URA3 plasmid pRS406 (Sikorski and Hieter, 1989) so that they are contiguous. The plasmid was linearized with an enzyme that cuts in either of the two cloned fragments and then transformed into C. glabrata selecting for Ura⁺ transformants on SC-Ura plates. Integration at the correct locus was verified by PCR analysis of the integrant using a primer external to the cloned fragment and one primer that anneals with the plasmid. The integrants were streak purified on SC-Ura plates and then grown in YPD overnight and appropriate dilutions were plated on SC + 5-FOA plates to select for recombinants that had lost the integrated plasmid and therefore were rendered Ura^–. The resulting Ura^– segregants have either regenerated the wild-type locus, or have left the deletion constructed in the plasmid (the 5′-and the 3′ fragments adjacent to each other). Genomic DNA was prepared from several 5-FOA^R (Ura^–) segregants and PCR analysis was carried out using primers outside from the cloned fragments to confirm the genomic structure of the deletion derivatives. Strains generated by this procedure are listed in TableS2, and the plasmids used to make them are described in TableS3.

Southern blot for telomere length

Genomic DNA was isolated from wild-type strain BG14, rif1Δ (BG509), sir3::Tn7 (BG542) and EPA7::Tn7 (BG550) and digested with either ApaIl or SauIIIA, and about 5 µg of DNA were run on a 0.8% agarose gel and transferred to a Hybond-N membrane (Amersham-Pharmacia Biotech). The probe was a 32-mer oligonucleotide containing two copies of the C. glabrata telomere repeat (primer ♯1682, TableS4) end-labelled with [γ-³²P]-ATP using T4 polynucleotide kinase (New England Biolabs). Conditions for the Southern were performed as previously described (Castaño et al., 2003).

In vitro adherence assays

Yeast cells (S. cerevisiae or C. glabrata) were grown in 3 ml of SC + uracil overnight at 30°C, 0.6 µl of this overnight were inoculated in 3 ml of fresh media with [³⁵S] Express protein labelling mix (New Life Sciences Products Perkin Elmer Cat No. NEG 072), 132 µCi per culture. After 24 h growth at 30°C, cells were spun and washed three times with Hanks Balanced Salt Solution (HBSS) supplemented with 5 mM CaCl₂. Cells were resuspended in the same solution with 5 mM CaCl₂ and OD₆₀₀ was adjusted to 0.1. One millilitre of labelled yeast cell suspension was added to each well of 24-well plates containing previously fixed and washed epithelial cells in 500 µl of HBSS with CaCl₂ was added. The plates were spun down for 1 min at 1000 r.p.m. and incubated at room temperature for 10 min. Non-adherent yeast cells were washed four to five times with HBSS with CaCl₂. Adherent cells were recovered by lysing the epithelial cells with 500 µl of 0.1% triton 0.5% SDS in PBS. The cells were scraped off the plate and counted in 4 ml of scintillation liquid. Adherence is calculated as the radioactive counts that remain associated with the epithelial cells expressed as a percentage of the total counts added. Most experiments were carried out using Lec2 cells and the data shown are the result of at least two experiments performed in triplicate.

RT-PCR

RNA was extracted from the stationary phase cells (grown for 24 h) for wild-type (BG14) and the rif1Δ (BG509) strains and used for RT-PCR synthesis using the appropriate primers, shown below. cDNA synthesis and PCR were carried out as previously described (De Las Peñas et al., 2003). The following primers and temperatures that were used in this study: RT primers are EPA1, TAACAGTGTTTTCGTTTGAT; EPA2, GAATGATTTCCTTATTAAAT; EPA3, TAATTTGATCAG TAGCACCG, EPA4, GTCAAAATTCTGTAGTGAAAG; EPA4/5, GTCAAATTCTGTAGTGAAAG; EPA6 TTACAGGGTTCG GATCTGAC; EPA7, TTACAGGGTTCGAATCTGAC EPA8, TTCCTGGTTCCGTCGCACCA; EPA9, GGCCTAATGAG GCTGACCAA; EPA11, TTAGTGGAGCCGGAGAATCG; EPA12, TTATGCGCTCTGATCCATGT; EPA13, TTTGCCGA ATAGGCGAGTCA; EPA14, TGATACGTGTAGGCGCACCA; EPA15, TTATTATTGCATTTGAATGC; EPA16, TAACAATT TCTTCACTAAAT and ACT1, GGCTTTCGATTTCTCACC. The cDNA synthesis reaction was carried out at 55°C for EPA1, EPA2, EPA3, EPA4/5, EPA6, EPA7, EPA13, EPA14, EPA15, EPA16; and 62°C for EPA8, EPA9, EPA11, EPA12. The PCR primers were the same for all the reverse primers, and the forward primers were: EPA1, GGGCTCAAAAA CAGCTAAAG; EPA2, GGGATCAGATTATGCAAAAG; EPA3, GCATGTTGATAGTTCCAAAA; EPA4/5, GCTAACATTACTG TATTTCT; EPA6, GGGTTCTCAAACAGCTAAGG, EPA7, GGGTTCTCAAACAGCTAAGG; EPA8, GGGCGGGTACCA GAAAAGAC; EPA9, GGCCTAATGAGGCTGACCAA; EPA11, GGCCCAATCAAGATAAGAAT; EPA12, GGGAAGGCGA ATAGTGCGTA; EPA13, GGCCAGGCGTGAACAAAAAC; EPA14, GCATGTTGACCATTCCAAGA; EPA15, GGGCAA AAAAAGCCTCAAAA, EPA16, GGGGCTCAAAAAATGCAA AAA and ACT1, GTGGCAACGGTTTGATGC.

S1 nuclease protection assay

For the assay in Fig. 1, cells were grown for 24 h into stationary phase. RNA was extracted as previously described (De Las Peñas et al., 2003). Probes (listed in Table 4) were end labelled using [γ-³²P]-ATP with T4 polynucleotide kinase. Thirty microgram of RNA from each strain (wild-type, rif1Δ and sir3Δ) was hybridized with each end-labelled probe at 55°C overnight. The mix was digested at room temperature with 150 units of S1 nuclease (Invitrogen) for 30 min The samples were then extracted with phenol, precipitated and resuspended in 17 µl of 1X loading buffer. Five microlitre of each sample was separated by electrophoresis on a 10% acrylamide gel, and the signal detected using a phosphorimager.

Animal studies

Yeast cells from each strain were grown overnight in YPD at 30°C. Cells were collected by centrifugation and washed once with PBS. Cells were then resuspended in PBS at a concentration of ∼8 × 10⁸ cells ml⁻¹. The number of cells was determined by OD₆₀₀, by counting the cells in a hemacytometer. Total viable counts were confirmed by plating appropriate dilutions on YPD plates and assessing colony number after 24 h incubation (30°C). Eight to nine week old Balb/C mice (Taconic) were injected with ∼8 × 10⁷ yeast cells in a total volume of 0.2 ml of the cell suspension in PBS by tail vein injection. In each experiment, groups of 10 mice were infected per strain. These were sacrificed after 7 days and the kidneys, liver and spleen were harvested and homogenized. Dilutions of the homogenates were plated on YP Agar supplemented with Pen/Strep. Cfu were scored the following day. Geometric means for all the groups were calculated and wild-type values for each organ were standardized to 100%. Numbers for each mutant strain tested are expressed relative to wild-type. Each graph represents the mean of three experiments (30 mice total per strain). Pooled P-values were calculated for pair-wise comparisons from the primary data using the Mann–Whitney test.