Development of a Gene Knockout System in Candida parapsilosis Reveals a Conserved Role for BCR1 in Biofilm Formation

Candida parapsilosis is an important cause of candidiasis, yetfew molecular tools are available. We adapted a recyclable nourseothricinresistance marker gene (SAT1) originally developed for use withC. albicans in order to generate gene knockouts from C. parapsilosis.We first replaced the promoters driving expression of the FLPrecombinase and the SAT1 genes with the equivalent sequencesfrom C. parapsilosis. We then used the cassette to generatea homozygous knockout of C. parapsilosis URA3. The ura3 knockoutshave altered colony morphologies. We also knocked out both allelesof an ortholog of BCR1, a gene encoding a transcription factorknown to be required for biofilm development in C. albicans.We show that C. parapsilosis BCR1 is necessary for biofilm developmentin C. parapsilosis and for expression of the cell wall proteinencoded by RBT1. Our results suggest that there are significantsimilarities in the regulation of biofilms between the two species,despite the fact that C. parapsilosis does not generate truehyphae and that BCR1 regulates the expression of many hypha-specificadhesins in C. albicans.

INTRODUCTION

Top

INTRODUCTION

Candida species are among the most common causes of nosocomialinfection. Whereas C. albicans is generally the most commoncausative agent, several large-scale studies have shown thatother Candida species are isolated with increasing frequency.C. parapsilosis is one species of concern, and in some cases,the incidence of bloodstream infection is similar to or exceedsthat of C. albicans (3, 17, 34).

Infection with C. parapsilosis is most common in neonates, intransplant patients, and in patients receiving parenteral nutrition(1, 17, 29, 30, 40, 53). C. parapsilosis is frequently foundon the hands of healthy individuals, including hospital workers(6, 20, 21). This is associated with periodic outbreaks of C.parapsilosis infections, including those in pediatric unitsin Spain, Italy, and Finland (4, 41, 46, 47) and in communityhospitals in the United States (10, 13, 25).

The abilities of C. parapsilosis to adhere to indwelling medicaldevices and to form biofilms are strongly correlated with infection(7, 24, 48). Over the past few years, there has been a dramaticincrease in our understanding of the molecular basis of biofilmformation by C. albicans (reviewed in reference 36). Developmentof biofilms is associated with the transition to hyphal growth,and mutants that are locked in the yeast form generally produceminimal biofilms (31, 42). Regulation of adhesins and cell wallproteins is also important (reviewed in reference 52). In contrast,very little is known about the signals regulating biofilm formationin C. parapsilosis. Unlike C. albicans, C. parapsilosis doesnot form true hyphae. Biofilms appear to consist of elongatedyeast cells, and development is influenced by phenotype (26).

Genetic analysis of C. parapsilosis is hampered by the lackof suitable tools. The genome is diploid (14, 18, 27), and thereis no apparent sexual cycle (33). Protocols for high-efficiencygenetic transformation have been developed (19, 57), and cloningvectors have been constructed (39). However, to date, no methodsfor generating gene knockouts have been described. We describehere a method using the recyclable, dominant nourseothricinresistance marker (SAT1), based on protocols developed for genedisruption in C. albicans (43, 56). We used two rounds of integrationto generate a homozygous ura3 mutant. We also disrupted an orthologof the BCR1 gene, which has been shown to regulate biofilm developmentin C. albicans (35, 37). Disrupting C. parapsilosis BCR1 (CpBCR1)also inhibits the formation of biofilms, suggesting that similarpathways are required in the two species.

MATERIALS AND METHODS

Top

MATERIALS AND METHODS

Strains and growth conditions. Candida parapsilosis strains (Table 1) were grown routinelyin YPD medium (1% yeast extract, 2% peptone, 2% dextrose) at30°C. For colony selection, 2% agar was added. For selectionof transformants, nourseothricin (Werner Bioagents, Jena, Germany)was added to YPD at a final concentration of 200 µg ml^–1.To recycle the SAT1-FLP cassette, transformants were grown inYPM medium (1% yeast extract, 2% peptone, 2% maltose) overnight,and approximately 100 cells were plated onto YPD plates supplementedwith nourseothricin at 20 µg ml^–1 and incubatedat 30°C. Two sizes of colonies were readily seen after 48h of growth at 30°C. The smaller colonies were picked andrestreaked onto fresh YPD plates. URA3 mutants were examinedby growth on synthetic complete (SC)-ura plates (0.67% yeastnitrogen base, 2% dextrose, 0.075% mixture of all amino acidsexcept uridine, 2% agar). After electroporation, cells wereallowed recover by growing on YPG media (1% yeast extract, 1%Bacto peptone, 2% dextrose) with 1 M sorbitol (19). The formationof biofilms by C. parapsilosis was tested with several media,and synthetic defined (SD) medium (0.67% yeast nitrogen base)containing 50 mM dextrose was determined to yield the best results.C. albicans biofilms were grown in Spider medium (32). RNA wasisolated using a Ribopure kit (Ambion) from mid-exponentialcells grown in SD medium supplemented with 50 mM dextrose and10% fetal calf serum (FCS).

View this table:
[in this window]
[in a new window]

TABLE 1. Strains used in this study

All sequences were obtained from the C. parapsilosis genome-sequencingproject (http://www.sanger.ac.uk/sequencing/Candida/parapsilosis/).

Plasmid construction. Plasmid pSFS2 was kindly provided as a gift by Joachim Morschhauser,Universitat Wurzburg, Germany. It contains the Candida albicansSAT1 nourseothricin resistance gene (CaSAT1) under the controlof a Candida albicans ACT1 promoter and a recombinase FLP genewhose expression is driven from the CaMAL2 promoter gene (43).In this study, 443 bp of CpURA3 upstream sequences and 489 bpof CpURA3 downstream sequences were amplified by PCR from C.parapsilosis CLIB214 genomic DNA with primers But223 and But224(containing restriction sites for ApaI and XhoI) and But221and But222 (containing restriction sites for SacII and SacI).The PCR products were digested with the relevant enzymes andinserted into the flanking sites of pSFS2, generating plasmidpCD4. The entire fragment was excised from pCD4 and introducedinto C. parapsilosis CLIB214, but no nourseothricin-resistanttransformants were obtained.

The cassette from pSFS2 was altered for use with C. parapsilosisby replacing the CaMAL2 and CaACT1 promoters with equivalentsequences from C. parapsilosis. The CaSAT1 gene was amplifiedfrom plasmid pSFS2 using primers But231 and But232, containingrecognition sites for XhoI and SacII, and cloned into the vectorpRS403 (49). A 501-bp fragment of the CpACT1 upstream sequence,including the first 15 codons and the entire first intron, wasamplified from C. parapsilosis CLIB214 using primers But233and But234, containing recognition sites for XmaI and XhoI,and cloned upstream of the CaSAT1 gene. A fragment of CaFLPwas then amplified using primers But230 and But235, containingrecognition sites for EcoRI and XmaI, was cloned upstream ofthe CpACT1 promoter to produce plasmid pCD3. Plasmid pCD6 wasgenerated by exchanging the EcoRI-SacII fragment of pCD3 (fromthe center of CaFLP1 to beyond the second FRT site) with thatof pSFS2. A 548-bp fragment containing the C. parapsilosis MAL2promoter sequence was amplified from CLIB214 using primers But227and But228 containing recognition sites for BamHI and SalI andused to replace the CaMAL2 promoter in pCD6, generating theplasmid pCD8 (Fig. 1A).

View larger version (22K):
[in this window]
[in a new window]

FIG. 1. Construction of a Cpura3 ${Delta}$ strain. (A) (i) A disruption cassette was adapted from the SAT1 flipper system described by Reuss et al. (43) by replacing the MAL2 and ACT1 promoters with equivalent sequences from C. parapsilosis. (ii) Upstream and downstream sequences from CpURA3 were amplified using primer pairs But223/But224 and But221/But222 and inserted at the ApaI/XhoI and SacII/SacI sites surrounding the flipper cassette. (iii) The entire cassette and flanking regions were excised by digestion with ApaI and SacI and targeted to the CpURA3 alleles by homologous recombination. The figure shows the structure of the cassette integrated at one allele. (iv) The cassette was recycled from the first allele by induction of the FLP recombinase, inserted at the second allele and recycled once more. Restriction sites shown in bold are unique. A, ApaI; B, BamHI; E, EcoRI; H, HindIII; K, KpnI; SI, SacI; SII, SacII; Sl, SalI; X, XhoI; Xm, XmaI. (B) Insertion of the SAT1 cassette at the first CpURA3 allele was verified by PCR using primers But237 (from within the FLP gene) and But238 (from upstream of CpURA3, outside the range of the deletion). Lanes 2 to 9 show the presence of a 1.3-kb fragment in eight independent disruptants that is not present in the wild-type strain CLIB214 (lane 10). One (lane 2, strain CD9) was chosen to recycle the cassette. Lane 1 contains a molecular size marker (1 kb plus; Invitrogen). (C) Recycling of the SAT1-FLP cassette. PCR amplification using primers But237 and But238 shows that the strains in lanes 2, 3, and 4 have not lost the cassette, whereas those in lanes 5 and 6 have (top panel). This is confirmed in the bottom panel, which shows PCR amplification using the primer pair But238 and But222 surrounding CpURA3. The strains in lanes 2, 3, and 4 contain the wild-type CpURA3 allele (1.5 kb), whereas those in lanes 5 and 6 have the deleted allele (1.0 kb). One isolate (CD98) was chosen for disruption of the second allele. (D) The structure of the disruption was also confirmed by Southern hybridization using the probe indicated, derived from the downstream region of CpURA3. Genomic DNA was digested with HindIII. The probe hybridizes to one fragment of 2.2 kb (URA3) in the starting strain CLIB214 (lane 1), representing the wild-type allele. Introducing the cassette at one allele produces one larger band of 2.5 kb (ura3::SAT1-FLP) and one wild-type band (CD9, lane 2). Recycling the cassette results in a small fragment of 1.7 kb (ura3 ${Delta}$ ::FRT), corresponding to the deleted allele, and a wild-type fragment (CD98, lane 3). The cassette was introduced at the second allele, generating both large (ura3::SAT1-FLP) and small (ura3 ${Delta}$ ::FRT) fragments (CD982, lane 6). Recycling the cassette resulted in only one fragment, corresponding to the deleted CpURA3 allele (ura3 ${Delta}$ ::FRT; CDU1 and CDU6, lanes 4 and 5). (E) The strains containing SAT1-FLP integrated at the first allele of CpURA3 (CD9) or the second allele (CD982) are resistant to nourseothricin (at 200 µg ml^–1). All other strains are sensitive. The strains in which both alleles of CpURA3 were deleted (CDU1 and CDU6) failed to grow in the absence of uracil. (F) The phenotypes of the wild-type (CLIB214) and the heterozygote knockout strains (CD98) are the same when grown on YPD medium and resemble the crater phenotype described previously (26). The double-deletion strain (CDU1) has a different appearance, which reverts to the wild-type phenotype when excess uridine (50 µg ml^–1) is added.

To knock out CpURA3, the entire adapted cassette from plasmidpCD8 was digested with BamHI and SacII, and the fragment wasused to replace the equivalent region in pCD4, generating pCD10.The cassette and CpURA3-flanking sequences were excised by digestionwith ApaI and SacI and transformed into C. parapsilosis CLIB214.The same construction was used to delete the second allele.

To generate deletions of CpBCR1, a 550-bp fragment was amplifiedfrom the upstream region using primers But257 and But258 (containingrecognition sites for KpnI and ApaI), and a 502-bp fragmentwas amplified from the downstream region using primers But259and But260 (containing recognition sites for SacII and SacI).These were introduced into the relevant sites of pCD8, generatingplasmid pCD13. The entire cassette and CpBCR1 flanking sequenceswere excised by digesting with KpnI and SacI and transformedinto C. parapsilosis CLIB214. The same construct was used todelete the second allele.

Plasmid pCD20 was constructed to reintroduce CpBCR1 at its originalchromosomal location. The entire open reading frame of CpBCR1together with 82 bp of upstream and 405 bp of downstream sequenceswas amplified using primers But273_SacII and But260, containingSacII and SacI restriction sites. The PCR fragment was clonedinto plasmid pCD12, which contains a 550-bp fragment from theupstream region of CpBCR1, between the KpnI and ApaI sites ofpCD8. The plasmid was linearized by digestion with SacI beforeelectroporation.

Candida parapsilosis transformation. C. parapsilosis strains were transformed by electroporationas described previously (57), with some modifications. Overnightcell culture was diluted to an A₆₀₀ of 0.2 in 50 ml of freshYPD and incubated at 30°C until the A₆₀₀ reached approximately2.0. Cells were then pelleted at 5,000 x g for 10 min, resuspendedin 10 ml of Tris-EDTA buffer (pH 7.5) containing 10 mM dithiothreitol,and incubated at 30°C for 1 h. Water (40 ml) was added,and cells were washed twice, first with 50 ml of ice-cold waterand finally with 10 ml of ice-cold 1 M sorbitol. Finally, cellswere resuspended in 125 µl of 1 M sorbitol. Approximately1 µg of purified ApaI-SacI fragment from pCD10 or purifiedKpnI-SacI fragment from pCD13 was mixed with 50 µl ofcompetent cells and transferred into a 1-mm electroporationcuvette. The reaction was carried out at 1.25 kV, using an Eppendorf2510 electroporator. After electroporation, 950 µl ofYPG medium containing 1 M sorbitol was added immediately, andthe mixture was incubated at 30°C for 4 h. One hundred microliterswas plated onto YPD plates supplemented with nourseothricinat a concentration of 200 µg ml^–1. Transformantswere obtained following 24 h of incubation at 30°C.

Candida parapsilosis genomic DNA extraction, PCR verification, and Southern blotting. Genomic DNA was isolated as described previously (http://www.fhcrc.org/science/labs/breeden/Methods/genomic_DNAprep.html),except that cells were disrupted by using a bead beater. ForPCR, 1 µg of genomic DNA was used. Genomic DNA (15 µg)treated with RNase at 10 µg µl^–1 was usedfor Southern hybridization.

Screening of transformants carrying the SAT1-FLP cassette atCpURA3 was carried out using primers But237 and But238 (Table2). Primer But237 binds within the FLP recombinase gene, andBut238 binds upstream of the flanking region of the CpURA3 gene(Fig. 1A). Integration at CpURA3 yields a 1,280-bp PCR productthat is not present in the wild-type or the recycled strain(Fig. 1B). Primers But222 and But238 were used to verify recyclingof the SAT1-FLP cassette and deletion of the CpURA3 gene. Thewild-type CpURA3 allele generates a 1.5-kb product, and recycledstrains produce a smaller sized product of approximately 1 kb(Fig. 1B).

View this table:
[in this window]
[in a new window]

TABLE 2. Primers used in this study

Transformants carrying the SAT1-FLP cassette at CpBCR1 wereidentified using primers But237 and But261, which amplify a1.5-kb fragment from within the FLP recombinase gene to upstreamof CpBCR1 (Fig. 2). The presence of an intact CpBCR1 allelewas confirmed by using primers But261 and But262, which amplifyan 0.8-kb internal fragment. Primers But257 and But260 wereused to verify recycling of the SAT1-FLP cassette and deletionof the CpURA3 gene. The wild-type CpBCR1 allele generates a3.0-kb product, and recycled strains produce a smaller sizedproduct of approximately 1 kb.

View larger version (26K):
[in this window]
[in a new window]

FIG. 2. Construction of a Cpbcr1 ${Delta}$ strain. (A) Upstream and downstream sequences from CpBCR1 were amplified using primer pairs But257/But258 and But259/But260 and inserted at the KpnI/ApaI and SacII/SacI sites surrounding the SAT flipper cassette. The entire cassette and flanking regions were excised by digestion with KpnI and SacI and targeted to the CpBCR1 alleles by homologous recombination. The cassette was recycled from the first allele by induction of the FLP recombinase, inserted at the second allele, and recycled once more. Restriction sites shown in bold are unique. A, ApaI; B, BamHI; E, EcoRI; H, HindIII; K, KpnI; SI, SacI; SII, SacII; Sl, SalI; X, XhoI; Xm, XmaI. The CpBCR1 gene was reintroduced at one deletion allele by inserting a 550-bp fragment from the CpBCR1 promoter (plus 45 codons) upstream of the SAT1-FLP cassette and the entire open reading frame (including 82 bp of the promoter) downstream of the cassette. The cassette was removed by recombination between the FRT sites, as described before. (B) Construction of the deletion was confirmed by PCR using three different primer combinations. The order of the strains for each reaction is as follows: lane 1, CLIB214; lane 2, CDb25 (first integration); lane 3, CDb27 (first recycle); lane 3, CDb38 (second integration); lane 5, CDb71 (second recycle). Amplification with primers But261/But237 generates a 1.5-kb fragment from within the FLP gene to upstream of BCR1. Primers But261/But262 amplify a 0.8-kb fragment from the 5' region of the BCR1 gene. Primers But257/But260 amplify a 3.0-kb fragment from the wild-type BCR1 allele and a 1-kb fragment from the bcr1 ${Delta}$ allele. (C) The structure of the disruption was also confirmed by Southern hybridization using a probe derived from the downstream region of CpBCR1. Genomic DNA was digested with HindIII. The probe hybridizes to one 4.5-kb fragment (BCR1) in the starting strain, CLIB214 (lane 1), representing the wild-type allele. Introducing the cassette at one allele produces one smaller band of 2.0 kb (bcr1::SAT1-FLP) and one wild-type band (CDb25, lane 2). Recycling the cassette results in a fragment of 2.6 kb (bcr1 ${Delta}$ ::FRT), corresponding to the deleted allele, and a wild-type fragment (CDb27, lane 3). The cassette was introduced at the second allele, generating both 2.0-kb (bcr1::SAT1-FLP) and 2.6-kb (bcr1 ${Delta}$ ::FRT) fragments (CDb38, lane 4). Recycling the cassette results in only one fragment, corresponding to the deleted BCR1 allele (bcr1 ${Delta}$ ::FRT; CDb71, lane 5). (D) Reintegration of the CpBCR1 gene at one deletion allele was confirmed by PCR. Amplification with primers But261/But237 generates a 1.5-kb fragment from within the FLP gene to upstream of BCR1 in CDb72 (containing the SAT1-FLP cassette and the entire CpBCR1 gene), with no amplification from the double-deletion (CDb71) or from the reconstituted strain following recycling of the SAT1 cassette. Primers But261/But237 amplify a 0.8-kb fragment from within the CpBCR1 gene before and after recycling the cassette (CDb72 and CDb89) but not in the deletion strain (CDb71).

Transformants carrying the intact CpBCR1 gene and the SAT1-FLPcassette integrated in a double-deletion strain were identifiedusing primers But237 and But261, as described previously. Recyclingof the cassette was verified using primers But273 and But274,which amplify a 0.8-kb fragment from within CpBCR1.

Southern hybridizations were also used to verify the structureof the deletions in selected transformants, using approximately20 µg of RNase-treated genomic DNA digested with HindIII.For CpURA3, a probe was amplified from CLIB214 by using primersBut221 and But222, which binds to a region of CpURA3 downstreamfrom the integration site (Fig. 1A). For CpBCR1, a probe wasamplified using primers But259 and But260 (Fig. 2A).

Biofilm growth conditions and dry weight measurements. Biofilms were grown on silicone squares as described in thereports by Richard et al. (44) and Nobile and Mitchell (37).Silicone squares (1.5 cm by 1.5 cm) were cut from sheets (CardiovascularInstruments) and autoclaved. The squares were then placed ina 12-well plate and pretreated by the addition of 2 ml of FCSand incubated with shaking at 37°C overnight. The siliconewas then washed with 2 ml of phosphate-buffered saline (PBS)buffer to remove the residual FCS and moved to a new 12-wellplate.

Overnight cultures of C. parapsilosis grown in YPD were washedtwice with equivalent volumes of PBS buffer and then dilutedto an A₆₀₀ of 1 in SD medium with 50 mM glucose. C. albicanscultures were grown in Spider medium and treated similarly.Cultures (3 ml) were added to the 12-well plate containing thepretreated silicone. The plate was incubated on a rotor at 37°Cfor 2 h (initial adhesion phase) with gentle shaking. The mediumwas removed, and the silicone was gently washed with 2 ml ofPBS buffer. Fresh SD or Spider medium (3 ml) was added intothe 12-well plates, and they were incubated with gentle agitationon a rotary shaker at 37°C for 48 h. The resulting biofilmswere photographed.

The dry weight of the saturated biofilms was measured as follows.After 48 h, the SD and Spider media were carefully removed fromthe 12-well-plates, and the silicone squares were placed ina 50-ml centrifuge tube. PBS buffer (10 ml) was added to eachtube, and the mixture was vortexed for 30 s. The solutions containingthe biofilm cells were then filtered through predried cellulosemembrane filter paper (47-mm diameter, 0.45-µm pore size).The silicone squares were washed twice with PBS. The filterpapers were dried at 37°C for 24 h, and the dry weight wasthen calculated. The dry weight was measured for six independentreplicates of all isolates tested.

Confocal scanning laser microscopy. Biofilms were visualized using confocal scanning laser microscopy(CSLM), essentially as described by Nobile et al. (35). Biofilmson silicone squares were stained with 25 µg ml^–1concanavalin A (conA)-Alexa Fluor 594 conjugate (C-11253; Bio-science)for 45 min at 37°C. The liquid was then removed from eachwell, and the silicone squares were flipped and placed on a35-mm-diameter glass-bottomed petri dish (MatTek Corp., Ashland,MA). The biofilms were observed with a Zeiss LSM510 confocalscanning microscope equipped with Plan-Neofluar x20-magnification/0.5-numericalaperture (for measuring depth) and x40-magnification/1.3-numericalaperture (for image acquisition) oil objectives. A HeNe1 laserwas used to excite at a 543-nm wavelength. All images were capturedusing a Zeiss LSM Image Browser.

RT-PCR. Reverse transcription (RT)-PCR was performed as described previously(22). Optimum cycle numbers were determined for each primerpair. Expression of CpBCR1 generates a product of 800 bp withprimers But273 and But274; the ALS1 product is 224 bp with primersBut297 and But298, the ALS3 product is 209 bp with primers But299and But300, and the product from RBT1 is 237 bp with primersBut301 and But 302. ACT1 was amplified using primers But33 andBut34 (45).

RESULTS

Top

RESULTS

Generating a cassette-based gene knockout system for C. parapsilosis. The gene knockout system described here is based on a recyclablecassette containing a dominant selectable marker (SAT1), firstdeveloped and adapted for use with C. albicans (56, 43). Theentire cassette is flanked with direct repeats (FRT) containingthe recognition site for the FLP recombinase, allowing easyrecycling.

We first attempted to apply the C. albicans SAT1 cassettes directlyto C. parapsilosis. Fragments from upstream and downstream ofCpURA3 were amplified by PCR inserted at either end of the cassettein plasmid pSFS2. However, we failed to obtain any nourseothricin-resistanttransformants when we attempted to introduce the cassette intoC. parapsilosis.

We then replaced the promoters driving the expression of theFLP and SAT1 genes with equivalent sequences from C. parapsilosis(Fig. 1). A 548-bp fragment from upstream of the C. parapsilosisMAL2 ortholog is used to drive expression of FLP, and a 501-bpfragment from the C. parapsilosis ACT1 promoter controls expressionof CaSAT1 (Fig. 1A). The entire fragment, including the cassetteand the flanking regions from CpURA3, was isolated by digestionwith ApaI and SacI and was transformed into C. parapsilosisCLIB214 by electroporation. Nourseothricin-resistant colonieswere obtained after 48 h of growth on YPD agar plates containing200 µg ml^–1 nourseothricin.

Eight transformants were screened for the correct integrationby PCR, using primer But237, derived from within the CaFLP gene,and primer But238, from upstream of the CpURA3 gene (outsidethe range of a deletion). Amplification of a 1.3-kb fragmentindicated that the cassette had integrated at the correct chromosomallocation in all (Fig. 1B). One transformant (CD9) was inoculatedinto YPM medium without nourseothricin to allow excision ofthe SAT1 cassette. As suggested by Reuss et al. (43), we platedthe mixture on YPD agar plates containing a low level of nourseothricin(20 µg ml^–1). Sensitive colonies grow slowly onthis medium within 48 h. Several small colonies were screenedby PCR to identify isolates in which the cassette has been lost(Fig. 1C). Primers But237 and But238 were used to confirm theloss of the cassette in two isolates (Fig. 1C, lanes 5 and 6,upper panel), and But238 and But222 from upstream and downstreamof CpURA3 confirmed that the deleted URA3 allele is smallerin the same isolates (Fig. 1C, lanes 5 and 6, lower panel).We expected But238 and But222 to also amplify the second intactCpURA3 allele in these strains, but the smaller fragment waspreferentially amplified. One isolate (CD98) was chosen to repeatthe process to delete the second allele. We also confirmed thestructure of the deletions by using Southern hybridization.A downstream region of CpURA3 used as a probe detects a singlefragment from both alleles in the wild-type strain CLIB214,one wild-type allele and one increased in size when the cassetteis integrated at the second allele in strain CD9, and one wild-typeallele and one small fragment following loss of the cassettein CD98. Integration of the cassette at the second CpURA3 alleleis shown in Fig. 1D, lane 6 (isolate CD982) and the loss ofthe cassette, leaving two deleted CpURA3 alleles, in Fig. 1D,lanes 4 and 5 (isolates CDU1 and CDU6).

The isolates deleted for both alleles of CpURA3 failed to growin the absence of uracil, as expected (Fig. 1E). We also noticedthat the homozygous ura3 knockout strains differed in phenotypefrom that of the heterozygote knockout and the wild-type strain(Fig. 1F). Colonies from ura3 strains have a different appearancein the center. This occurred as soon as the cassette was introducedinto the second allele and was observed for several independentconstructs (not shown). The phenotype was rescued by the additionof uridine at 50 µg ml^–1 (Fig. 1F).

Construction of a bcr1 ${Delta}$ in C. parapsilosis. The BCR1 gene in C. albicans was identified from a screen oftranscription factor mutants and shown to be required for biofilmdevelopment (37). We identified an ortholog of CaBCR1 in thegenome of C. parapsilosis (see Fig. S1 in the supplemental material).We therefore deleted both alleles of CpBCR1 using the SAT1 flippercassette described above (Fig. 2). The wild-type CpBCR1 genewas also reintroduced into the double-deletion strain, againusing the SAT1 flipper cassette.

The effect of the Cpbcr1 ${Delta}$ on biofilm formation was tested asdescribed by Nobile and Mitchell (37). Silicone squares treatedwith FCS were placed in a sterile 12-well plate and inoculatedwith strains grown in SD medium. As shown in Fig. 3A, the wild-typeC. parapsilosis strain generates biofilms on the silicone. TheCpbcr1 ${Delta}$ /Cpbcr1 ${Delta}$ strain, however, does not adhere to the silicone,and most of the growth is confined to the surrounding liquidmedium, resembling that of the Cabcr1 ${Delta}$ strain. The heterozygoteC. parapsilosis knockout, where only one BCR1 allele has beendeleted, and the reconstituted strain (where BCR1 has been reintroducedinto the genome) both generate biofilms comparable to that ofthe wild type (Fig. 3). We compared the biofilms to those generatedby C. albicans strains in which BCR1 has been deleted and reconstituted(35). The reductions of biofilm are similar for strains carryingthe bcr1 deletions in both species (Fig. 3) (35, 37).

View larger version (85K):
[in this window]
[in a new window]

FIG. 3. Disrupting bcr1 impairs biofilm development. (A) C. parapsilosis wild-type (CLIB214), BCR1/bcr1 ${Delta}$ heterozygote (CDb27), bcr1 ${Delta}$ /bcr1 ${Delta}$ homozygote (CDb71), and reconstituted BCR1 (CDb89) strains were grown for 48 h in SD medium on silicone squares placed at the bottom of 12-well plates. Adherence to the silicone was monitored by photography and compared to that of the biofilms generated by C. albicans BCR1 reconstituted (CJN698) and bcr1/bcr1 disruption (CJN702) strains grown in Spider medium. Disrupting bcr1 greatly reduces biofilm formation in both species. (B) CLSM depth views. Biofilms were stained with conA and visualized by using a Zeiss LSM510 confocal scanning microscope, and false-color depths were constructed. Blue cells are closest to the silicone, and red cells are farthest away. The wild-type (C. parapsilosis CLIB214) biofilms are approximately 100 µm deep, and the bcr1 ${Delta}$ (C. parapsilosis CDb71) biofilms are approximately 20 µm deep. (C) The biomass of the biofilms generated was determined by measuring the dry weight. Averages of six biological replicates are shown. Ca, C. albicans; Cp, C. parapsilosis. Phenotypes are represented by strains as follows: bcr1 ${Delta}$ /bcr1 ${Delta}$ +pBCR1, C. albicans CJN698; bcr1 ${Delta}$ /bcr1 ${Delta}$ , C. albicans CJN702; wild type (WT), C. parapsilosis CLIB214; BCR1/bcr1 ${Delta}$ , C. parapsilosis CDb27; bcr1 ${Delta}$ /bcr1 ${Delta}$ , C. parapsilosis CDb71; bcr1 ${Delta}$ /bcr1 ${Delta}$ ::BCR1, C. parapsilosis CDb89. The differences between the biofilms from wild-type and bcr1 knockouts (P < 0.0001) and reconstituted and bcr1 knockouts (P = 0.003) in C. parapsilosis are statistically significant (t test, two-tailed P value). (D) Expression levels of ALS1, ALS3, and RBT1 were compared in the C. parapsilosis wild-type (1, CLIB214) and C. parapsilosis bcr1 ${Delta}$ (2, CDb71) strains using RT-PCR. Actin levels (ACT1) were measured as a control.

We used confocal microscopy to further investigate the differencesin biofilm production between the wild-type and knockout strains(Fig. 3B). Wild-type C. parapsilosis cells labeled with concanavalinA adhere to the silicone, generating a dense matrix consistingof yeast and pseudohyphal cells. There are no hyphal cells present.The Cpbcr1 deletion strain, in contrast, generates biofilmswith very few cells. Pseudohyphal cells, however, are clearlypresent, indicating that the knockout does not affect the transitionto pseudohyphae. We estimate that the wild-type biofilms havea depth of >100 µm and that the biofilms generatedby the Cpbcr1 knockout are approximately 20 µm deep.

The biofilm mass was determined by using dry weight measurements(Fig. 3C). The biofilms generated by C. parapsilosis CLIB214do contain as much mass as those generated by C. albicans CJN698,even though the medium used was selected to maximize growthof both species. Deleting BCR1 reduces the biofilm mass in both.Reintroducing CpBCR1 at its original locus does restore biofilmdevelopment but not to the same extent as that in the wild-typestrain or the heterozygote. The reconstituted strain containsan FRT site in the promoter of the CpBCR1 gene and a small duplicationof the first 45 codons (Fig. 2). The presence of the FRT siteis unavoidable with the reconstitution strategy used, chosenbecause we found it was impossible to clone the intact genein Escherichia coli. We confirmed by RT-PCR that the reconstitutedstrain expresses CpBCR1 but at a lower level than the wild-typeor the heterozygote knockout strain (not shown).

In C. albicans, BCR1 regulates the expression of at least 11cell surface proteins or cell wall modification enzymes, includingthe hyphal wall protein encoded by HWP1 and the ALS3 adhesingene (37). We could not identify an ortholog of HWP1 in C. albicans,but we did identify a close relative that is a member of thesame family (see Fig. S1 in the supplemental material). Thisis most likely an ortholog of RBT1 from C. albicans (8). Thereare four potential members of the ALS adhesin gene family inC. parapsilosis, and we tentatively identified two as ALS1 andALS3 (see Fig. S1 in the supplemental material). Deleting Cpbcr1greatly reduces the expression of RBT1, but it has no noticeableeffect on the expression of ALS1 or ALS3 (Fig. 3D).

DISCUSSION

Top

DISCUSSION

Developing tools for generating gene disruptions greatly facilitatedour understanding of the biology and pathogenesis of C. albicans.One of the first systems developed was the Ura-blaster method,consisting of a URA3 cassette flanked by direct repeats of a1.1-kb fragment from the Salmonella enterica serovar TyphimuriumhisG sequence (16). Recycling of the cassette was enabled bygrowth on 5'-fluoroorotic acid, which is toxic to Ura⁺ cells.The major disadvantage of this system relates to the use ofthe URA3 marker gene. Homozygous recycled disruptants are Ura^–,which affects phenotypes such as virulence and adhesion (2,23). Restoring the URA3 gene at a different chromosomal locationoften does not complement the defect, leading to confusion aboutthe interpretation of the role of the target gene (28, 50).This has led to a reevaluation of the virulence phenotypes ofseveral disruptions, including NOT3, BUR1, KEL1, and HWP1 (9,51). The Ura-blaster method was adapted to allow amplificationof the cassette by PCR by replacing the hisG sequences withsmaller 200-bp direct repeats (54).

One of the difficulties is the necessity of carrying out tworounds of allele disruption, as the genome is diploid. Wilsonet al. (55) first used two nonrecyclable markers (URA3 and HIS1)to disrupt both alleles of two genes in the RIM101 pathway.Simultaneous disruption was facilitated by the constructionof the UAU1 (ura3-ARG4-ura3) cassette, which contains a copyof the ARG4 gene that, when removed by recombination, leavesan intact copy of URA3 (15). Homozygous disruptions are selectedby growth on Arg^– Ura^– medium, which requires thepresence of an intact cassette at one allele and a recombinedcassette at the other (or the generation of triplication derivatives).The cassette was subsequently adapted for use in large-scalemutagenesis by incorporation into a Tn7 transposon (11). Thepotential difficulties of using URA3 as a marker were addressedby Noble and Johnson (38), who constructed a set of His^–,Leu^–, and Arg^– auxotrophic strains and associatedHIS1, LEU2, and ARG4 cassettes from species other than C. albicans.Homozygous disruptants are generated through simultaneous selectionfor two markers. A Cre-lox-based system using ARG4 and HIS1has also been reported (12).

One drawback of all these methods is that they require well-characterizedauxotrophic strains, which may also affect certain phenotypes.This can be circumvented by using dominant selectable markers,such as MPA^r and SAT1. We therefore decided to adapt the cassettesdeveloped by Reuss et al. (43), as we had no success applyingthe C. albicans constructs directly to C. parapsilosis. By drivingexpression of the FLP recombinase from the CpMAL2 promoter andthe SAT1 marker from CpACT1, we successfully generated disruptionsof URA3 and BCR1 in the C. parapsilosis type strain CLIB214.The frequency of homologous integration is high: 100% of thenourseothricin-resistant colonies tested contained the cassetteat the targeted region for the first integration. The systemhas the advantage that the disruptants are almost geneticallyidentical to the wild-type strain, apart from the excision ofthe target gene and the presence of a single FRT repeat.

One disadvantage is that the method is slow: the cassette istoo large to amplify by PCR, meaning the flanking regions mustbe inserted by cloning. Each allele must be disrupted sequentiallyfollowing recycling of the SAT1 marker. One of the goals ingenerating the CpURA3 disruption was to allow the use of auxotrophicselectable markers. However, disrupting URA3 changed the phenotypeof the strain (Fig. 1F), suggesting it may also have additionaleffects. The phenotypic change is complemented by the additionof extra uridine, suggesting it is a direct result of the disruption.As with C. albicans, it is therefore inadvisable to use URA3in a disruption cassette in C. parapsilosis, unless URA3 isreplaced at its native position. However, construction of theUra^– strain will allow the use of URA3-based replicatingplasmids. The same approach can also be applied to generatingisogenic knockouts of other markers, such as HIS1, TRP1, andARG4.

Because infection with C. parapsilosis is associated with biofilmformation (24, 48), we decided to test the effect of knockingout a candidate gene required for biofilms. We chose BCR1 becauseit was shown to be necessary for biofilm development in C. albicansand yet does not affect hyphal growth under most conditions(37). C. parapsilosis does not generate true hyphae, and itis therefore likely that some mutants affecting the hyphal transitionwill influence biofilm development in C. albicans and not inC. parapsilosis. We showed that both species require BCR1 forbiofilm development.

CLSM images of C. albicans biofilms show that the basal layerconsists of mostly yeast cells topped by a layer of pseudohyphaland hyphal cells (42, 44). Long hyphae project even furtherfrom the surface. In contrast, C. parapsilosis biofilms havebeen described as clumped blastospores (24). We have previouslyshown by light microscopy that C. parapsilosis cells form pseudohyphaewhen adhered to plastic surfaces (26). Here, we confirm usingCSLM that biofilms formed by the type strain of C. parapsilosisare a mixture of pseudohyphae and yeast cells. Unlike C. albicans(not shown), the C. parapsilosis cells exhibit some autofluorescence,affecting the quality of the image. However, it is clear thatthere are pseudohyphae present both in the basal layer and projectingaway from the surface. Deleting Cpbcr1 reduces the number ofadhered cells but does not affect the formation of pseudohyphae.It is therefore likely that BCR1 is required at an early stageof biofilm formation in both C. albicans and C. parapsilosis(5).

In C. albicans, BCR1 regulates the expression of a number ofcell wall and adhesin genes, including members of the ALS family,HWP1, and the chitinase gene (CHT2) (35, 37). As these are predominantlyexpressed in hyphal cells, it has been suggested that BCR1 regulatesthe adhesive properties of hyphal cells (37). The C. parapsilosisgenome encodes orthologs of many of the C. albicans targets,including some of the ALS family and relatives of HWP1 (fromhttp://www.sanger.ac.uk/sequencing/Candida/parapsilosis/). Theclosest match to CaHWP1 in C. parapsilosis is most likely tobe an ortholog of RBT1 (see Fig. S1 in the supplemental material).We showed that CpBCR1 is required for the expression of RBT1,but it has no noticeable effect on the expression of two ALSgenes. However, there are several members of the Als familyof GPI-linked proteins in C. albicans, and only some are regulatedby BCR1 (37). It is difficult to absolutely assign orthologyto the Als family members in C. parapsilosis at present, andit is possible that the expression of some is indeed regulatedby BCR1. We are presently actively developing whole-genome microarraysfor C. parapsilosis, and we look forward to identifying alltargets in C. parapsilosis and to comparing the activities andfunctions of the BCR1 orthologs in both species. It is possiblethat BCR1 controls adhesion through the regulation of gene expressionin yeast cells in both species. It is also likely that bothBCR1 function and biofilm development in the two species havesome shared and some different aspects.

ACKNOWLEDGMENTS

We thank Aaron Mitchell for providing the C. albicans bcr1 disruptionstrains, Joachim Morschhauser for the SAT1-FLP cassette, andthe Wellcome Trust Sanger Institute for access to the C. parapsilosisgenome sequence. We also gratefully acknowledge the assistanceof Ann Cullen for generating the confocal images.

This work was supported by the Science Foundation Ireland andthe Health Research Board.

FOOTNOTES

* Corresponding author. Mailing address: UCD School of Biomolecular and Biomedical Science, Conway Institute, University College Dublin, Belfield, Dublin 4, Ireland. Phone: 353-1-7166885. Fax: 353-1-2837211. E-mail: geraldine.butler{at}ucd.ie

Published ahead of print on 22 June 2007.

Supplemental material for this article may be found at http://ec.asm.org/.

REFERENCES

Top