EUKARYOTIC CELL, Apr. 2002, p. 249–256 Vol. 1, No. 21535-9778/02/$04.00�0 DOI: 10.1128/EC.1.2.249–256.2002Copyright © 2002, American Society for Microbiology. All Rights Reserved.

The Mold-Specific MS8 Gene Is Required for Normal HyphaFormation in the Dimorphic Pathogenic Fungus

Histoplasma capsulatumXianbin Tian and Glenmore Shearer, Jr.*

Department of Biological Sciences, Center for Molecular and Cellular Biosciences, University of Southern Mississippi,Hattiesburg, Mississippi 39406-5018

Received 2 August 2001/Accepted 22 January 2002

The dimorphic fungus Histoplasma capsulatum is the etiologic agent of one of the most common systemicmycoses of humans, histoplasmosis. In the environment, H. capsulatum grows in a differentiated mold form andshifts to an undifferentiated yeast form after mold fragments or spores are inhaled. This mold-to-yeast shift isrequired for disease. Little is known about the molecular biology of dimorphism in Histoplasma, and moststudies have been directed toward yeast-specific genes. While it is important to examine the role of genesupregulated in the yeast morphotype, genes which are silenced in the yeast (i.e., mold-specific genes) may alsoplay a critical role in dimorphism. To begin to examine this hypothesis, we report here the first misexpressionand knockout analysis of a mold-specific gene in Histoplasma. The strongly expressed MS8 gene encodes apredicted 21-kDa protein extremely rich in glycine and glutamine. Forced expression of MS8 driven by theTEF1 promoter in yeast did not alter the yeast morphology at 37°C or mold formation at 25°C. Yeast expressingMS8 did exhibit clumping in liquid medium and formed “sticky” colonies on agar plates. Allelic replacementof MS8 was accomplished by a positive-negative selection procedure. ms8 knockout mutants formed apparentlynormal yeast at 37°C but gave rise to aberrant mycelia at 25°C. The mold colonies of the knockouts were lessthan half as large as normal, had a granular surface, produced a dark-red pigment, and formed short hyphaewhich were 40% wider with a distinctive twisted “zig-zag” shape.

The dimorphic fungus Histoplasma capsulatum is the etio-logic agent of one of the most common systemic mycoses ofhumans, histoplasmosis. Integral to pathogenesis is the mold-to-yeast (M-Y) dimorphic shift. The organism grows in soil asa differentiated multicellular mold. When mold fragments orspores are inhaled, the fungus shifts to an undifferentiatedsingle-cell yeast growth form in the lungs of the host. If thisM-Y shift is blocked, the disease cannot progress (11). In thelaboratory, this reversible dimorphism can be observed bychanging the incubation temperature to favor the desiredform: 25°C for mold or 37°C for yeast.

The molecular basis of the shift from the differentiated mul-ticellular form to the undifferentiated unicellular form is un-known. Investigators have searched for genes upregulated inthe pathogenic yeast form to identify genes required for di-morphism and virulence, and several interesting genes havebeen identified in these studies, as summarized in the recentreview by Retallack and Woods (13). Keath and Abidi (4), forexample, reported that the yps-3 gene is specific to the yeastform and is expressed at higher levels in the most virulentstrains. Sebghati et al. (16) demonstrated that the calciumbinding protein gene CBP1 is required for survival within hostphagocytic cells. Neither gene is required for dimorphism,however, since strains that do not express yps-3 and cbp1knockout mutants still exhibit apparently normal dimorphism.

Although the search for yeast-specific genes as regulators of

dimorphism is quite logical, it is possible that genes critical forthe dimorphic shift are being overlooked due to the assump-tion that yeast-specific or yeast-upregulated genes are the mostimportant regulators. It may be necessary, for example, todownregulate certain mold-specific genes in order to form thein vivo yeast morphotype. One can envision a scenario in whichmold-specific (MS) genes would exert an overriding effect, i.e.,expression of a mold-specific gene(s) would shift growth to themold form and a shutdown of this gene(s) would be required toallow the shift to the undifferentiated yeast form. We havebegun to examine this hypothesis by isolating a number ofgenes which are silent in the yeast morphotype (18). Here, wereport the first misexpression and knockout analysis of a mold-specific gene, MS8, in H. capsulatum and show that ms8 knock-out mutants form aberrant mycelia.

MATERIALS AND METHODS

Strains and growth conditions. (i) Yeast strains. The H. capsulatum strainsG186AS and G184AS (7) were used in these studies. Both strains representrestriction fragment length polymorphism class 3 isolates (5) and exhibit thesame expression profile of MS8. A uracil auxotroph of G184AS, G184AS ura5-11, the kind gift of Jon Woods (University of Wisconsin, Madison), was used forknockout experiments. Routine cultures were grown in GYE medium (2% [wt/vol] glucose, 1% [wt/vol] yeast extract) at 37°C with 150-rpm gyratory shaking foryeast or at 25°C with 100-rpm gyratory shaking for mold. Cultures for transfor-mation or knockout analysis were grown in HMM, a rich synthetic medium basedon F-12 tissue culture medium supplemented with cysteine and glutamic acid andbuffered to pH 7.5 with HEPES (22). HMM plates were made by adding 1%(wt/vol) agarose (A-5093; Sigma). To reduce the chance of bacterial contamina-tion during the prolonged incubations, 50 �g of ampicillin/ml and 100 �g ofstreptomycin/ml were added to the HMM. The following additions to HMMwere made when necessary: HMM/u, 50 �g of uracil/ml; HMM/h, 200 �g ofhygromycin/ml; and HMM/f, 1 mg of 5-fluoroorotic acid/ml.

Yeast cultures for DNA or RNA extraction were grown to mid-log phase and

* Corresponding author. Mailing address: Department of BiologicalSciences, Center for Molecular & Cellular Biosciences, University ofSouthern Mississippi, Hattiesburg, MS 39406-5018. Phone: (601) 266-4722. Fax: (208) 379-0426. E-mail: glen.shearer@usm.edu.

249

on January 31, 2021 by guesthttp://ec.asm

.org/D

ownloaded from

harvested by centrifugation. Actively growing mold cultures for nucleic acidextraction were prepared as follows: 5 ml of a dense mold culture was inoculatedinto 100 ml of fresh medium, incubated at 25°C with gyratory shaking for 3 days,and then harvested by vacuum filtration.

(ii) Escherichia coli strains. E. coli SURE-2 cells {e14 (McrA) �(mcrCB-hsdSMR-mrr)171 endA1 supE44 thi-1 gyrA96 relA1 lac recB recJ sbcC umuC::Tn5 (Kanr) uvrC [F� proAB lacIqZ �(M15 Tn10) (Tetr) Amy Camr]; Stratagene},which give better results with the H. capsulatum telomeric-repeat plasmids, wereused for routine plasmid preparations. Telomere plasmids prepared in otherstrains occasionally exhibit sequence rearrangements, presumably because of thetelomeric repeats.

Preparation of DNA. Plasmids from E. coli were isolated by alkaline lysis andCelite chromatography with the Wizard kit (Promega, Madison, Wis.) accordingto the manufacturer’s directions. Purification of DNA from agarose gels or afterenzyme reactions was done by silica adsorption with the Zymoclean kit (ZymoResearch, Orange, Calif.) according to the manufacturer’s directions.

Large-scale genomic DNA preparations from Histoplasma were done as fol-lows. Approximately 35 ml of mid-log-phase yeast was harvested by centrifuga-tion at 2,000 � g for 5 min, resuspended in 35 ml of ice-cold water, and collectedby centrifugation for 5 min. The cell pellet was resuspended in 0.6 ml of DNAextraction buffer (100 mM Tris [pH 8.0], 100 mM EDTA, 250 mM NaCl).Phenol-chloroform (5:1; pH 7.9) (0.6 ml) and 0.8 ml of 0.45-mm-diameter glassbeads were added, and the tubes were vortexed on a Vortex Genie 2 (FisherScientific) at maximum speed for 1 min. The tubes were cooled on ice for 1 min,followed by vortexing for 1 min. The cycles of vortexing and cooling wererepeated for a total vortex time of 5 min. The cell debris and beads were pelletedby centrifugation at 11,000 � g for 10 min. The supernatant was removed to anew tube, and 30 �l of a 5-mg/ml RNase A solution was added. The tubes wereincubated at 37°C for 60 min to digest the RNA. The mixture was extracted withan equal volume of phenol-chloroform (5:1; pH 7.9), and the aqueous phase wasremoved to a new tube. DNA was precipitated from the aqueous phase by theaddition of 2 volumes of ethanol followed by centrifugation at 11,000 � g for 10min. The DNA pellet was dissolved in 10 mM Tris (pH 7.8)–1 mM EDTA, andthe concentration was determined by absorbance at 260 nm. The DNA typicallymigrated as a single large band slightly larger than a 23-kb lambda HindIIImarker when analyzed by agarose gel electrophoresis.

DNA miniprep for PCR testing of putative knockout colonies was as follows.A single colony was scraped from an HMM plate with a sterile toothpick andresuspended in 50 �l of H2O in a 500-�l microcentrifuge tube. DNA extractionbuffer (100 �l), 0.45-mm-diameter glass beads (50 �l), and 100 �l of phenol-chloroform (5:1; pH 8.0) were added, and the tube was vortexed for 3 min. Thetube was centrifuged at 10,000 � g for 5 min, and the aqueous phase wasrecovered. RNase A was added to the aqueous phase to a final concentration of100 �g/ml, and the tube was incubated at room temperature for 10 min. TheDNA was then ethanol precipitated as described above and dissolved in sterilewater.

Preparation of RNA. RNA was isolated by extraction with acidic phenol, whichpartitions RNA in the aqueous phase and DNA and protein in the lower phases(8). Mid-log-phase yeast cells were collected by centrifugation at 400 � g for 5min. Actively growing mold was vacuum filtered onto 0.45-�m-pore-size nitro-cellulose membranes and washed with ice-cold sterile water. The mycelium wasscraped from the filter with a plastic spatula. Yeast or mold cells were resus-pended in 0.4 ml of ice-cold RNA extraction buffer (0.1 M sodium acetate [pH5.0], 0.2 M NaCl, 0.2% [wt/vol] sodium dodecyl sulfate [SDS]) in a 1.5-ml screwcap microcentrifuge tube. An equal volume of acidic phenol-chloroform (5:1;equilibrated with RNA extraction buffer) and 300 �l of 0.45-mm-diameter glassbeads were added. The tubes were vortexed at maximum speed for 10 min atroom temperature and centrifuged for 3 min at 10,000 � g, and the aqueousphase was recovered to a new tube. RNA was precipitated from the aqueousphase by the addition of 2 volumes of ethanol followed by centrifugation at10,000 � g for 10 min. The RNA pellet was washed once with 75% ethanol anddissolved in 10 mM Tris (pH 7.5)–1 mM EDTA–0.2% (wt/vol) SDS by heatingthe solution at 65°C for 10 min. Any insoluble material was removed by centrif-ugation at 10,000 � g for 10 min, and the RNA in the supernatant was quanti-tated by absorbance at 260 nm.

Southern and Northern blotting. For Southern blotting, approximately 1 �g ofH. capsulatum genomic DNA was digested with restriction enzymes (New En-gland Biolabs) and electrophoresed on a 0.7% (wt/vol) agarose gel. The DNAwas transferred onto charged nylon membranes (Hybond N�; Amersham Corp.)by downward blotting with 0.4 N NaOH using a Turboblotter (Schleicher andSchuell, Keene, N.H.). The membrane was neutralized with several washes of 2�SSC (1� SSC is 0.15 M NaCl–0.015 M sodium citrate, pH 7.0) and allowed to airdry. The membranes were prehybridized in 0.5 M sodium phosphate (pH

7.0)–5% (wt/vol) SDS at 65°C for 1 h. Denatured radiolabeled probe was addedand allowed to hybridize at 65°C for 12 to 16 h. The membranes were washedonce at 65°C with 2� SSC–1% (wt/vol) SDS and two times with 0.1� SSC–0.1%(wt/vol) SDS. Radiolabeled probes were prepared with [�-32P]dATP with theStrip-EZ kit (Ambion). The blots were later stripped according to the manufac-turer’s directions.

Northern blots were prepared by electrophoresis of approximately 10 �g oftotal RNA per lane on a 2% denaturing formaldehyde agarose gel according tostandard methods (9). RNA was transferred onto charged nylon membranes(Hybond N�) with 20� SSC by using a Turboblotter. After the transfer, themembrane was soaked in 50 mM NaOH for 4 min and neutralized with 2� SSC.RNA was cross-linked to the membrane by exposure to 254-nm UV light for adosage of 120 mJ/cm2. Probe labeling, prehybridization, hybridization, andwashes were done as described for Southern blots.

MS8 genomic walking and sequencing. The MS8 gene and flanking sequencewere isolated by genomic-walking PCR (3). Forward and reverse primers specificfor the previously cloned MS8 cDNA (GenBank accession no. AF292398) wereused to isolate approximately 2 kb of upstream and 1.6 kb of downstreamsequence. Primers were then designed to isolate the entire genomic DNA forMS8 via PCR. DNA was sequenced by the dideoxy chain termination method. Toreduce sequence errors from PCR, a proofreading polymerase (Advantage 2;Clontech) was used and three independent PCR products were sequenced.

Plasmid construction. An H. capsulatum telomere vector was made based onthe successful plasmid constructs of Woods and Goldman (20). A DNA constructcontaining (in 5�-3� order) a NotI site, 12 H. capsulatum telomeric repeats(GGGTTA), a PmeI site, a 1.3-kb HindIII-StyI tetracycline resistance markerfrom pBR322, a PmeI site, 12 H. capsulatum telomeric repeats (TAACCC), anda NotI site was cloned into the NotI site of pGem11 Zf� (Promega) to generatepG11TelTet. A 1.7-kb EcoRI fragment containing the Pa Ura5 gene frompWU55 (21) was cloned into the EcoRI site of pG11TelTet to generate pXTU1.A 1.0-kb fragment of H. capsulatum genomic DNA (representing approximately950 nucleotides [nt] of 5� untranscribed flanking sequence and 50 nt of the 5�untranslated region of the H. capsulatum translation elongation factor 1 [TEF1]gene) containing the strong constitutive H. capsulatum TEF1 promoter wascloned into the XhoI site of pXTU1 to generate pXTU2. ApaI linkers wereligated to the MS8 cDNA (GenBank accession no. AF292398), which was theninserted into the ApaI site of pXTU2 immediately downstream of the TEF1promoter to generate pXTU3. Several clones were isolated and analyzed byrestriction cutting to select the sense orientation relative the TEF1 promoter.

An MS8 knockout vector was constructed as follows. A 4.3-kb H. capsulatumgenomic DNA PCR product containing the MS8 gene and 2 kb of 5� flankingsequence plus 1.5 kb of 3� flanking sequence was cloned into pGemTeasy (Pro-mega) to generate pMS8g. A 760-nt BclI/BstBI fragment in pMS8g representingmost of the MS8 coding sequence was replaced with a 2.9-kb BamHI/ClaIhygromycin resistance marker from pMAD93 (21) to generate pMS8g::hph.pXTU3 was cut with SmaI and ApaI to remove the TEF1 promoter and MS8cDNA. The remaining vector was blunted with Klenow and ligated to NotIadapters. A 6.4-kb NotI fragment containing the knockout construct was isolatedfrom pMS8g::hph and ligated to the resulting vector to form the knockout vectorpXTU4�. The vector pXTU3g was constructed for ms8 knockout complemen-tation experiments as follows. pXTU3 was cut with SmaI and ApaI to remove theTEF1 promoter and MS8 cDNA. The remaining vector was blunted with Klenowand ligated to NotI adaptors. The 4.3-kb MS8 genomic fragment was cut frompMS8g with NotI and ligated to the remaining pXTU3 vector to generatepXTU3g.

Histoplasma transformation. Plasmids were digested with PmeI to remove theTetr marker and expose the telomeric repeats. The resulting linear vector waspurified by agarose gel electrophoresis and recovered from the gel using a silicagel column (Zymoclean). H. capsulatum yeast was transformed by electropora-tion by a modification of the method of Woods et al. (21). Briefly, 5 ml ofmid-log-phase yeast cells from an HMM broth culture was centrifuged at roomtemperature at 300 � g for 5 min to harvest the cells. The cells were washed in5 ml of 10% (wt/vol) mannitol and collected by centrifugation. The cell pellet wasresuspended in 350 �l of 10% (wt/vol) mannitol, and 0.5 �g of linearized vector(contained in less than 10 �l of sterile water) was added and mixed. The cell-DNA mixture was transferred to a 0.2-cm electroporation cuvette (Bio-Rad).Electroporation was done in an ElectroPorator (Invitrogen) at the followingsettings: 750V; 71 �F; 150� (which yields a pulse of 3,750 V/cm for 10.6 ms). Foruracil selection, the cells were immediately plated on HMM plates. For hygro-mycin selection, the cells were mixed with 0.8 ml of HMM and incubated on aroller drum at 37°C for 12 h to allow expression of the hph marker, followed byspreading the mixture on HMM/h plates. The plates were incubated in a humid-ified chamber at 37°C. Small colonies were typically visible in about 7 days.

250 TIAN AND SHEARER EUKARYOT. CELL

on January 31, 2021 by guesthttp://ec.asm

.org/D

ownloaded from

Construction of knockout mutants. Genomic replacement of the native MS8locus with the hph-disrupted MS8 construct was accomplished by a modificationof the positive-negative selection method of Sebghati et al. (16). G184AS ura5-11yeast was transformed with linearized pXTU4�. The cells were spread on HMMand incubated for 7 to 10 days at 37°C. Ten colonies were picked and inoculatedinto 1 ml of HMM broth in 10 disposable 15-ml plastic screw cap centrifugetubes. A small hole was made in the cap of each tube with a 26-gauge needle, andthe tubes were incubated on a roller drum at 37°C for 1 week to allow time forthe double-crossover event to occur. Positive-negative selection was imposed byspreading 200 �l from each tube onto HMM/uhf plates to select for the hygro-mycin marker and against the Ura5 marker. The plates were incubated at 37°Cfor 7 to 10 days, and 10 colonies were selected for miniprep DNA extraction andanalysis by PCR to identify knockout mutants.

RESULTS

Sequence analysis of MS8. The complete nucleotide se-quence of the MS8 gene is shown in Fig. 1. The gene is inter-rupted by a single intron of 95 nt with canonical GT/AG splicejunction sequences and a putative splice box (GCTAAC),which matches the filamentous-fungus consensus sequenceRCTRAC (2, 19), starting 27 nt upstream of the 3� terminus ofthe intron. The predicted protein is 203 residues long with amass of 21,332 Da and an estimated pI of 6.76. Sequencing ofseveral cDNA clones identified two poly(A) sites: one site at nt1360, 20 nt downstream of a putative poly(A) signal

(AATAAA), and a second poly(A) site 42 nt upstream of thefirst site. Primer extension experiments (Fig 2) identified threetranscriptional start points, each starting at the A in a CAYmotif, located at �1, �9, and �12. A TATA box (TATATAA)was found starting 44 nt upstream of the first transcriptionalstart site.

Expression of MS8 during the Y-M shift. The MS8 transcriptis very abundant in the mold morphotype but is not detectableon Northern blots with RNA from the yeast growth form of H.capsulatum (18). To determine the time course of induction,yeast cells growing at 37°C were downshifted to 25°C, whichresults in a switch to the mold growth form. As shown in Fig.3, weak expression was first detected 11 h after the tempera-ture shift, corresponding to the first emergence of germ tubes.Expression then increased dramatically, particularly at day 3,corresponding to branch formation in the growing germ tubes.MS8 transcript levels remained high in stable mold cultures. Inseveral lanes, some signal was present in a slightly highermolecular weight band, which we believe represents a longertranscript from the secondary poly(A) site.

MS8 expression in yeast results in clumping and alteredcolony texture. To test the hypothesis that expression of MS8 inyeast might shift the morphology to become more mold-like at

FIG. 1. Sequence of the MS8 gene (GenBank accession no. AY049031). Nucleotide numbers are shown on the right, and amino acid numbersare shown in parentheses. The first transcriptional start site is designated �1 and underlined. Two additional start sites were found at �9 and �12.A putative TATA box (underlined) was found 44 nt upstream of the �1 site. The single intron is shown in lowercase letters and includes a putativesplice box (underlined). Two poly(A) sites were found at nt 1318 (single underline) and 1360 (double underline). A putative poly(A) signal wasfound starting at nt 1335 (double underline). Asterisk, stop codon.

VOL. 1, 2002 MOLD-SPECIFIC MS8 GENE OF H. CAPSULATUM 251

on January 31, 2021 by guesthttp://ec.asm

.org/D

ownloaded from

37°C, we forced yeast cells to express the MS8 transcript atlevels similar to that seen in mold cultures. The H. capsulatumtelomere vector pXTU3 (Fig. 4) was constructed with the MS8cDNA (GenBank accession no. AF292398), including a 30-ntpoly(A) tail, fused to the strong constitutive H. capsulatumTEF1 promoter in the sense orientation. G184AS ura5-11 cells

were transformed with the linearized plasmid, and transfor-mants were identified by uracil selection on HMM plates. Sixtransformants were selected at random, and the expressionlevel of MS8 was determined by Northern blotting, as shown inFig. 5. Transformants S1, S4, and S6 had levels similar to thatseen in mold cells, while S2, S3, and S5 showed less expressionthan in mold. All six transformants, maintained by uracil se-lection, exhibited normal yeast morphology as determined by

FIG. 2. Primer extension analysis of the MS8 gene. The four laneson the left show DNA sequencing reactions used as size markers. Therightmost lane shows a primer extension reaction, with the correspond-ing sequence of the MS8 gene.

FIG. 3. Expression analysis of MS8 during the Y-M shift. Total RNA was extracted from cells shifted from 37 to 25°C from 11 h (hr) to 11 days(d) as indicated and electrophoresed on a denaturing agarose gel (20 �g/lane). The blot was probed with MS8 cDNA (top), stripped, and probedwith a Histoplasma 18S ribosomal DNA probe (bottom) as a control. The rightmost lane (M) shows RNA extracted from fully developed mold.

FIG. 4. Plasmid for forced expression of MS8 in yeast. MS8 cDNAwas fused to the strong constitutive Histoplasma TEF1 promoter. Theplasmid was cut with PmeI (to remove the Tet marker and expose thetelomeric repeats), and cells were transformed with the linear vector.A, ApaI; E, EcoRI; P, PmeI; X, XhoI; Tel, telomeric repeats.

252 TIAN AND SHEARER EUKARYOT. CELL

on January 31, 2021 by guesthttp://ec.asm

.org/D

ownloaded from

microscopic examination. In HMM broth culture, however,yeast expressing MS8, but not yeast transformed with the vec-tor lacking MS8 sequences, exhibited pronounced clumping.Yeast colonies expressing MS8 on HMM plates had no obviousalteration in size or color but were quite sticky, i.e., whentouched with an inoculating loop, the cells tended to stick tothe loop like very thick honey. These transformants formedapparently normal mold after the growth temperature wasshifted to 25°C. No detectable change in the Y-M transforma-tion or in the microscopic morphology of the mold was noted.

Positive-negative selection of knockout mutants. To test theeffect of MS8 loss of function on H. capsulatum growth anddimorphism, knockout mutants were constructed by a modifi-cation of the positive-negative selection method of Sebghati etal. (16). H. capsulatum G184AS ura5-11 yeast (which shows thesame MS8 expression pattern as G186AS [data not shown])was transformed with a construct containing the MS8 gene witha large section of the coding sequence replaced by a hygromy-cin marker (hph), as shown in Fig. 6. The knockout constructcontained 2 kb of 5� flanking sequence and 1.5 kb of 3� flankingsequence relative to the open reading frame. Approximately0.8 kb of the MS8 coding region was replaced with the 2.9-kbhph marker. After 1 week of uracil selection followed by 1week of positive-negative selection on 5-fluoroorotic acid–hy-gromycin plates, 10 colonies were randomly selected and ex-amined for MS8 allelic replacement. Analysis by PCR andSouthern blotting showed that 3 of the 10 colonies were ms8knockout mutants. Figure 7 shows PCR analysis of two knock-out mutants. Analysis with two sets of primers (set 1, Brp plusBfp; set 2, Brp2 plus Bfp) showed that the knockout mutantsyielded a band 2.1 kb larger than that of the parental strain,

consistent with replacement of the single-copy MS8 gene withthe knockout construct (i.e., replacement of 0.8 kb of MS8sequence with the 2.9-kb hph marker). To confirm the loss ofthe native locus, DNA from the parental strain and two knock-out mutants was digested with BamHI and examined by South-ern blotting, as shown in Fig. 8. A radiolabeled probe repre-senting the 0.8 kb of MS8 sequence deleted from the knockoutconstruct hybridized to a 5.4-kb band only in the parentalstrain. In contrast, an hph probe hybridized to a 7.6-kb bandonly in the knockout mutants, as predicted for allelic replace-ment of MS8.

ms8 knockout mutants form aberrant mycelia. ms8 knock-out mutants shifted to 25°C showed no significant phenotypicchanges early in the Y-M transformation. Germ tube emer-gence and elongation occurred in both the parental strain andknockout mutants at approximately the same time and rate.After several days of growth on HMM/u plates, however, themycelia of the ms8 knockouts were substantially different fromthose of the parental strain, as shown in Fig. 9. Colonies of theknockout mutants were less than half the diameter of normalcolonies, had a brownish granular appearance, and produced adeep-red pigment particularly remarkable upon viewing theback of the agar plate. Microscopic examination showed thatthe hyphae of knockout mutants were nearly 40% wider (0.32versus 0.23 �m) than normal, as shown in Fig. 10. Whereashyphae in the parental strain generally gave rise to long,straight filaments, those in the knockout mutants gave rise tomuch shorter branches, with a dramatic “zig-zag” twisted ap-pearance.

To ensure that the aberrant mycelial morphology of the ms8knockouts was the result of loss of function of MS8, we trans-

FIG. 5. Forced expression of MS8 in yeast. Yeast was transformed with the linearized pXTU3 vector and selected on HMM plates. Total RNAwas isolated and analyzed by Northern blotting. Each lane contained 20 �g of RNA. The image on the left shows RNA from normal yeast andmold. The image on the right shows RNA from six independent transformants (S1 to S6). The blot was probed with radiolabeled MS8 cDNA.

FIG. 6. Map of MS8 knockout vector. A map of the native locus (The MS8 open reading frame is indicated by the large open box) is shownat the top, and the knockout construct is at the bottom. PCR primer sites are indicated by arrows. Approximately 0.8 kb of MS8 coding sequencewas replaced with the 2.9-kb hygromycin marker (hph). Approximately 2 kb of 5� flanking sequence and 1.5 kb of 3� flanking sequence wereincluded. This linear construct was used to replace the TEF1-MS8 sequence (XhoI to ApaI) in pXTU3 to form the knockout vector pXTU4�.

VOL. 1, 2002 MOLD-SPECIFIC MS8 GENE OF H. CAPSULATUM 253

on January 31, 2021 by guesthttp://ec.asm

.org/D

ownloaded from

formed the knockout mutants with pXTU3 (MS8 cDNA drivenby the TEF1 promoter) and pXTU3g (an MS8 genomic frag-ment with the native promoter) to complement the ms8 mu-tation. Knockout mutants were transformed with pXTU3 andpXTU3g and plated on HMM plates to maintain the plasmidsby uracil selection. Colonies grown at 25°C demonstrated nor-

mal colony and microscopic morphologies, as shown for theparental strain in the left-hand images of Fig. 9 and 10. Incontrast, ms8 knockout mutants transformed with pXTU1(which has the Ura5 marker but no MS8 sequence) still exhib-ited the aberrant morphology, as shown for the ms8 knockoutsin the right-hand images of Fig. 9 and 10. Thus, the aberrantphenotype was not incidental to the experimental protocol butdue to MS8 loss of function.

DISCUSSION

The most obvious environmental cue for the M-Y shift afterspores or mold fragments are inhaled is the upshift to 37°C.Laboratory cultures shifted from 25 to 37°C mimic this processin vitro, and nearly all studies of H. capsulatum dimorphism inthe literature have used a simple temperature shift to favormold (25°C) or yeast (37°C). Although the temperature shift isclearly an important dimorphism signal in H. capsulatum, it isnot necessary or sufficient for the morphological transition.Under certain conditions, yeast can be grown at the “nonper-missive” temperature of 25°C (10) and mold can be grown at37°C (10, 14, 15). Presumably, the temperature shift is thenatural trigger for induction of a gene regulation cascadewhich results in the dramatic differentiation process from amulticellular form to a unicellular form or vice versa. Clearly,the dimorphic shift is associated with dramatic changes in generegulation, as demonstrated by the numerous differentially reg-ulated genes found in subtracted-library and differential-dis-play experiments in our laboratory and others (6, 13, 18).

The MS8 gene was interrupted by a single intron with typicalGT-5� and AG-3� termini (Fig. 1). A putative splice box match-ing the filamentous-fungus consensus RCTRAC (GCTAAC)was found starting 27 nt upstream of the 3� end of the intron.Comparison of several cDNA clones revealed that MS8 hadtwo poly(A) sites, as do many of the H. capsulatum genes we

FIG. 7. PCR analysis of knockout mutants. Lanes A, DNA fromparental strain G184AS ura5-11; lanes B and C, DNA from putativeknockout mutants. Primer set 1, Brp and Bfp; primer set 2, Brp2 andBfp. Replacement of the native locus should yield a band 2.1 kb largerthan that in the parental strain.

FIG. 8. Southern blot analysis of knockout mutants. DNA was ex-tracted from the G184AS ura5-11 parental strain (lanes A) and twoputative knockout mutants (lanes B and C), cut with BamHI, separatedon a 0.7% gel, and blotted onto a nylon membrane. The left-handimage shows the blot hybridized with a radiolabeled probe represent-ing the 0.8 kb of MS8 sequence removed in the knockout vector. Theright-hand image shows the same blot stripped and probed with aradiolabeled hph marker.

FIG. 9. Colony morphology of G184AS ura5-11 parental strain(left) and ms8 knockout mutant (right). The colonies are shown onHMM/u agar after 6 weeks of growth at 25°C. The photographs weretaken from the same plate and with the same lighting conditions andmagnification. Bar, 5 mm.

254 TIAN AND SHEARER EUKARYOT. CELL

on January 31, 2021 by guesthttp://ec.asm

.org/D

ownloaded from

have isolated. The most distal poly(A) site was preceded by acanonical poly(A) signal sequence, AATAAA, starting 26 ntupstream. The second poly(A) site results in a transcript 42 ntshorter and is not preceded by an obvious poly(A) signal se-quence. The sequence surrounding the putative translationalstart site (boldface and underlined), TCATCATGTC, showedhigh similarity to the Kozak sequence [TCA(C/A)(A/C)ATG(G/T)C] seen in filamentous fungi (1). Primer extension exper-iments showed three closely spaced transcriptional start sites ofsimilar strengths 228, 220, and 217 nt upstream of the transla-tional start site. Each transcriptional start was the A in a CAYmotif. A consensus TATA box (TATATAA) was located start-ing 44 nt upstream of the first transcriptional start site. Thepredicted MS8 protein was 203 amino acids long with an ex-pected mass of 21.3 kDa and a pI of 6.77. Particularly strikingwas the abundance of two amino acids in the putative MS8protein. Over 40% of MS8p was composed of only two aminoacids, glycine (22%) and glutamine (19%). GenBank similaritysearches revealed only one significant match in the database: aglycine-rich protein from the plant-pathogenic fungus Colleto-trichum gloeosporioides (accession no. U94186; BLAST simi-larity, 10�21). The C. gloeosporioides protein is similar to MS8pin mass (22 kDa), G and Q content (24 and 19%, respectively),and cysteine content (0%) but is significantly more basic (pI,9.18) than MS8p. Unfortunately, the function of this protein inColletotrichum is also unknown.

Expression of MS8 is not detected on Northern blots ofRNA from the yeast morphotype but is quite abundant in moldcells (18). During the 25°C Y-M transition, expression was firstdetected at 11 h (Fig. 3), corresponding to the emergence ofgerm tubes from the yeast. As the germ tubes elongated, thelevel of the MS8 transcript increased dramatically, particularlyon day 3, which corresponded to hyphal branch formation inmost of the cells. MS8 transcript levels remained high in ac-tively growing mold cultures.

H. capsulatum has inefficient homologous recombination,and only very recently has allelic replacement become feasiblein Histoplasma. Our knockout experiments with a slight mod-

ification of the positive-negative selection methodology of Seb-ghati et al. (16) showed that nearly one-third of the colonies onthe final selection plates were ms8 knockouts. This is similar tothe frequency reported for their knockout of the calcium bind-ing protein gene (CBP1). This positive-negative selectionmethod now makes it possible to isolate H. capsulatum knock-out mutants in less than a month.

If MS8 was essential for formation of mold cells, we wouldexpect ms8 knockout mutants to be unable to form mycelia.Clearly, MS8 is not an absolute requirement for mold forma-tion, since the knockout mutants did form hyphae (albeit ofaberrant shape and size). The presence of a functional MS8gene is also not required for the shift back to the yeast mor-photype, since ms8 knockout mutants growing as mold canshift to apparently normal yeast at 37°C (not shown). Theidentity of the red pigment produced by the ms8 mutants is notknown. Red-pigmented variants of H. capsulatum have beenreported, one isolated from a canebrake (12) and one frombiopsy material from a human immunodeficiency virus-positivepatient (17). The identity of the pigment produced by thesevariants is also not known. Although it is possible that thepigments produced by the variants and the ms8 knockout mu-tants are related, we have no data at this time to allow aconclusion.

Clearly, it is not essential that MS8 be silent to maintain theyeast morphotype, since yeast forced to express high levels ofthe MS8 transcript (Fig. 5) still grew as yeast at 37°C. MS8 isrequired, however, for normal hypha formation. ms8 knockoutmutants formed aberrant short, twisted hyphae and gave rise tosmall, pigmented, slow-growing mold colonies (Fig. 9). Thefunction of MS8p within the cell is unknown. Examination ofthe putative MS8p with various proteomics structural analysisprograms (http://www.expasy.ch/tools) revealed primarily thatthe protein is expected to be quite flexible and hydrophilic,with two short regions of alpha helix in the C terminus. Thelack of diagnostic structural features and the BLAST similaritymatch to a single protein whose function is also unknown makeit impossible to postulate a particular function for this protein.

FIG. 10. Microscopic morphology of G184AS ura5-11 parental strain (left) and ms8 knockout (right) stained with lactophenol cotton blue andphotographed at the same magnification. Bar, 2 �m.

VOL. 1, 2002 MOLD-SPECIFIC MS8 GENE OF H. CAPSULATUM 255

on January 31, 2021 by guesthttp://ec.asm

.org/D

ownloaded from

It is quite possible that MS8p is a structural protein used inhyphal wall formation. This hypothesis is consistent with theMS8 expression experiments with yeast (Fig. 5). The clumpingof yeast expressing MS8 and the sticky texture of the yeast onplates indicate that the cell surface is altered in some fashion,perhaps because MS8p is localized in or on the cell wall. Infuture work, we plan to examine the cellular localization ofMS8p with GFP-MS8 fusions to help elucidate the role of thisprotein in mold formation.

ACKNOWLEDGMENT

This work was supported in part by grant AI49357 from the NationalInstitutes of Health.

REFERENCES

1. Ballance, D. J. 1986. Sequences important for gene expression in filamentousfungi. Yeast 2:229–236.

2. Gurr, S. J., S. E. Unkles, and J. R. Kinghorn. 1987. The structure andorganization of nuclear genes of filamentous fungi, p. 93–139. In J. R.Kinghorn (ed.), Gene structure in eukaryotic microbes. IRL Press, Oxford,United Kingdom.

3. Jones, D. H., and S. C. Winistorfer. 1993. Genome walking with 2- to 4-kbsteps using panhandle PCR. PCR Methods Appl. 2:197–203.

4. Keath, E. J., and F. E. Abidi. 1994. Molecular cloning and sequence analysisof yps-3, a yeast-phase-specific gene in the dimorphic fungal pathogen His-toplasma capsulatum. Microbiology 140:759–767.

5. Keath, E. J., G. S. Kobayashi, and G. Medoff. 1992. Typing of Histoplasmacapsulatum by restriction fragment length polymorphisms in a nuclear gene.J. Clin. Microbiol. 30:2104–2107.

6. Keath, E. J., A. A. Painter, G. S. Kobayashi, and G. Medoff. 1989. Variableexpression of a yeast-phase-specific gene in Histoplasma capsulatum strainsdiffering in thermotolerance and virulence. Infect. Immun. 57:1384–1390.

7. Klimpel, K. R., and W. E. Goldman. 1987. Isolation and characterization ofspontaneous avirulent variants of Histoplasma capsulatum. Infect. Immun.55:528–533.

8. Majumdar, D., Y. J. Avissar, and J. H. Wyche. 1991. Simultaneous and rapid

isolation of bacterial and eukaryotic DNA and RNA: a new approach forisolating DNA. BioTechniques 11:94–101.

9. Maniatis, T., J. Sambrook, and E. F. Fritsch. 1982. Molecular cloning: alaboratory manual. Cold Spring Harbor Laboratory, Cold Spring Harbor,N.Y.

10. Maresca, B., G. Medoff, D. Schlessinger, and G. S. Kobayashi. 1977. Reg-ulation of dimorphism in the pathogenic fungus Histoplasma capsulatum.Nature 266:447–448.

11. Medoff, G., M. Sacco, B. Maresca, D. Schlessinger, A. Painter, G. S. Koba-yashi, and L. Carratu. 1986. Irreversible block of the mycelial-to-yeast phasetransition of Histoplasma capsulatum. Science 231:476–479.

12. Morris, P. R., A. A. Terreni, and A. F. DiSalvo. 1986. Red-pigmented His-toplasma capsulatum—an unusual variant. J. Med. Vet. Mycol. 24:231–233.

13. Retallack, D. M., and J. P. Woods. 1999. Molecular epidemiology, patho-genesis, and genetics of the dimorphic fungus Histoplasma capsulatum. Mi-crobes Infect. 1:817–825.

14. Rippon, J. W. 1968. Monitored environment system to control cell growth,morphology, and metabolic rate in fungi by oxidation-reduction potentials.Appl. Microbiol. 16:114–121.

15. Sacco, M., B. Maresca, B. V. Kumar, G. S. Kobayashi, and G. Medoff. 1981.Temperature- and cyclic nucleotide-induced phase transitions of His-toplasma capsulatum. J. Bacteriol. 146:117–120.

16. Sebghati, T. S., J. T. Engle, and W. E. Goldman. 2000. Intracellular para-sitism by Histoplasma capsulatum: fungal virulence and calcium dependence.Science 290:1368–1372.

17. Staib, F., and G. Grosse. 1996. Brown-red pigment formation by the mycelialphase of a clinical isolate of Histoplasma capsulatum on Staib agar. A pre-liminary report. Zentbl. Bakteriol. 283:515–521.

18. Tian, X., and G. Shearer. 2001. Cloning and analysis of mold-specific genesin the dimorphic fungus Histoplasma capsulatum. Gene 275:107–114.

19. Unkles, S. E. 1992. Gene organization in industrial filamentous fungi, p.29–53. In J. R. Kinghorn and G. Turner (ed.), Applied molecular genetics offilamentous fungi. Blackie, Glasgow, United Kingdom.

20. Woods, J. P., and W. E. Goldman. 1993. Autonomous replication of foreignDNA in Histoplasma capsulatum: role of native telomeric sequences. J.Bacteriol. 175:636–641.

21. Woods, J. P., E. L. Heinecke, and W. E. Goldman. 1998. Electrotransforma-tion and expression of bacterial genes encoding hygromycin phosphotrans-ferase and beta-galactosidase in the pathogenic fungus Histoplasma capsu-latum. Infect. Immun. 66:1697–1707.

22. Worsham, P. L., and W. E. Goldman. 1988. Selection and characterization ofura5 mutants of Histoplasma capsulatum. Mol. Gen. Genet. 214:348–352.

256 TIAN AND SHEARER EUKARYOT. CELL

on January 31, 2021 by guesthttp://ec.asm

.org/D

ownloaded from