INFECTION AND IMMUNITY, May 1977, p. 690-700Copyright X 1977 American Society for Microbiology

Vol. 16, No. 2Printed in U.S.A.

Antigenic, Chemical, and Structural Properties of Cell WallsofHistoplasma capsulatum Yeast-Form Chemotypes 1 and 2

After Serial Enzymatic HydrolysisE. REISS,* SARA E. MILLER, W. KAPLAN, AND L. KAUFMAN

Mycology Division, Center for Disease Control, Atlanta, Georgia 30333,* and Department of Microbiologyand Immunology, Duke University Medical Center, Durham, North Carolina 27710

Received for publication 19 October 1976

Cell walls ofHistoplasma capsulatum yeast-form chemotypes 1 (chem 1) and2 (chem 2) treated sequentially with several polysaccharolytic enzymes andPronase yielded soluble, nondialyzable polysaccharides at each step, which wereanalyzed for monosaccharides, protein composition, and serological activity.Polysaccharide recovered after digestion ofchem 1 walls with f3(1-)3)-glucanasecontained glucose > mannose > glucosamine > galactose. This fraction (chem 1

,3G1) was analyzed by polyacrylamide gel electrophoresis and contained a compo-nent having an apparent molecular weight of 120,000. The chem 1 /3G1 fractionwas reactive in immunodiffusion (ID), producing an immune precipitate notidentical to the H and M factors of histoplasmin. In a side-by-side ID comparisonwith extracts of chem 2, the chem 1 j3G1 antigen contained an additionaldeterminant not found in chem 2 extracts when tested with goat antiserum to H.capsulatum. Therefore, the chem 1 antigen gave preliminary ID evidence ofantigenic group specificity. A chemical difference observed was the absence ofglucosamine from chem 2 polysaccharide. In complement fixation (CF) tests, 9 of17 sera from human histoplasmosis patients reacted with chem 1 /3G1, but somecross-reactivity with sera of patients with other systemic mycoses occurred. Theimmunoelectrophoretic patterns of chem 1 wall-derived polysaccharides showeda marked shift in mobility after Pronase digestion, implying the presence ofcovalent peptides. The ultrastructural appearance and serological activity ofintact walls and enzyme-resistant mural cores were also studied. The surface ofthe mural cores of both chemotypes was perforated and frayed. In shadow-castpreparations both fibrillar and globular areas persisted in the mural cores. TheCF end point serum dilutions showed an increase after a- and /8-glucanaseextractions ofchem 2 walls and fourfold reduction after Pronase digestion. Themural cores of both chemotypes were still reactive in CF tests and retained someability to bind fluorescent antibody. The chem 1 mural core reacted with specificfluorescein-labeled H. capsulatum antiglobulins produced by adsorption withBlastomyces dermatitidis, thus indicating at least partial retention ofH. capsu-latum-specific factors. The presence of galactose, mannose, and glucose wasdetected in the mural cores as well as enriched levels of amino sugar, despiteexposure to chitinase.

The ability of polysaccharolytic enzymes anda protease to effect 50 to 60% lysis of cell walls ofHistoplasma capsulatum yeast forms was thesubject of earlier work from this laboratory inwhich walls were hydrolyzed with a series ofenzymes consisting of a(1--3)-glucanase,,3(1-)3)-glucanase, Pronase, and chitinase (23).In addition to monosaccharide and amino acidproducts, small amounts of nondialyzable poly-saccharide were released during the procedure,which was termed serial enzymolysis. The twoserotypes studied showed differences in enzyme

susceptibility, which served to identify themwith respect to chemotype. An objective of thepresent work was to assay polysaccharide re-leased by enzymolysis from walls of both chem-otypes for serological activity, monosaccharidecomposition, and protein content. The effect ofenzymolysis on the antigenicity of the cell wallwas followed by comparing the fluorescent-antibody (FA) and complement fixation (CF)reactions of the intact as well as the partiallydigested walls with the resistant residue, re-ferred to below as the mural core. Furthermore,

690

on Decem

ber 2, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

CELL WALL ANTIGENS OF H. CAPSULATUM 691

the effect of serial enzymolysis was studied todetermine whether alterations in the mural ul-trastructure were associated with changes inFA and CF reactivities. The antigenicity ofsoluble polysaccharides and particulate glucan-ase-extracted cell walls was studied with hu-man histoplasmosis case sera to determinewhether such defined reagents might be ofvalue in the serodiagnosis of this disease.

It is useful at the outset to summarize howthe terms chemotype and serotype have beenapplied to cell wall components of H. capsula-tum. Yeast-form chemotypes were first de-scribed by Domer (4), based in part on earlierwork of Pine and Boone (22). These chemotypiccategories received substantial support fromthe recent results of Kanetsuna et al. (11). Thecell wall type designated chemotype 1 (chem 1)had an elevated chitin content, and a(1- 3)-glucan was absent, whereas in chem 2 a re-

duced level of chitin was observed, and a(1-3)-glucan occurred as a major mural component(4, 11). In the serial enzymolysis experiment(23), the walls of serotype 1,2,3 lacked a(1-+3)-glucan and had a chitin content of 30.3%, whichthus indicated the isolate was chem 1. In con-trast, walls obtained from a strain bearing sero-type 1,4 determinants were lysed to the extentof 27.3% by a(1-*3)-glucanase and had a re-duced chitin content (7.5%), fitting the criteriafor chem 2.The discovery of four serotypes among iso-

lates ofH. capsulatum was made on the basis ofFA cross-staining and adsorption studies (12,15). The significance of serotypes for the immu-nodiagnosis of histoplasmosis is that serotypicfactors 1 and 4 are shared with Blastomycesdermatitidis, whereas factors 2 and 3 are spe-cific for H. capsulatum. It would be advanta-geous to remove polysaccharide from the cellwall in a manner that would preserve the sero-type determinants for subsequent isolation andchemical characterization.

After each stage of serial enzymolysis (23),portions of the cell walls resistant to lysis werewithdrawn, and the soluble, nondialyzable ma-terial was also recovered. These fractions andthe mural core were the starting materials forthe analyses presented below.

MATERIALS AND METHODSAntigens. Soluble, nondialyzable products of se-

rial enzymatic hydrolysis obtained at each stage ofextraction of yeast-form cell walls ofH. capsulatumA811 (serotype 1,2,3; chem 1) and of H. capsulatum105 (serotype 1,4; chem 2) and the correspondingmural cores (23) were analyzed. Fractions were di-signated according to the notation shown in Table 1.The cultures were obtained from the Mycology Divi-sion, Center for Disease Control, Atlanta, Ga.

Chemical analyses. Volatile derivatives were pre-pared and analyzed in a Perkin-Elmer 990 gas chro-matograph to determine monosaccharide ratiospresent in cell fractions. For neutral sugars, vialscontaining 1-mg samples received 0.5 ml of 2 Ntrifluoracetic acid and then were sealed under N2and hydrolyzed at 100°C. Soluble fractions wereheated for 2 h, whereas insoluble samples required 4h; then acid was removed in vacuo over soda lime,and the corresponding alditols were produced byaddition of NaBH4. Conditions for the reduction andpreparation of n-butyl boronic acid ester derivativeswere those of Eisenberg (5). A column (3 m in lengthand 4 mm in diameter) of 3% OV17 (50% phenyl-substituted phenylmethyl silicone) and coated onGas-Chrom Q was programmed from 150 to 200'C;detection was by flame ionization, quantitation wasby a discontinuous integrator, and the internalstandard was fucose. Amino sugar not liberated bychitinase was determined by the Elson-Morgan re-action (2) after acid hydrolysis (100°C, 6 N HCI for 3,6 or 18 h). Methods for determination of total pro-tein, carbohydrate, N, and P were those used previ-ously (23, 24).

Serology. Eight 9-month-old female goats wereused for the production of antisera. Two were in-jected intravenously once a week with 5 x 107 live 6-day-old yeast-form cells of H. capsulatum serotype1,2,3; two others received 5 x 107 live 6-day-oldyeast-form cells ofH. capsulatum serotype 1,4 once aweek. Two were injected intravenously twice a weekwith 1 mg of yeast-form cell walls (23) ofH. capsula-tum serotype 1,2,3, and two received 1 mg of cellwalls of H. capsulatum serotype 1,4 twice a week.The live cells and cell walls used for immunizationwere suspended in 0.85% NaCl. The presence of CFantibodies and precipitins reactive in immunodiffu-sion (ID) tests with histoplasmin was monitoredweekly (8, 13, 14), and a plateau in CF titer of 1:256was reached in 8 to 9 weeks. The human sera usedcame from patients with diagnoses of fungus dis-ease; the sera were supplied by physicians and thediagnoses were supported by positive serologicalfindings in both CF and ID tests. In addition, ZellMcGee supplied serum of one patient with histoplas-mosis seen at the Vanderbilt University Hospital.For immunoelectrophoresis (IEP), 100-ug sam-

ples were placed into wells in layers (5 by 7.6 cm) of1% agarose, operated at 9 V/cm for 2 h in boratebuffer (pH 8.6; ionic strength, 0.025), and developedat 9°C overnight. In the direct FA tests, fluorescein-labeled rabbit antiglobulins to whole formaldehyde-killed yeast-form cells of H. capsulatum serotype

TABLE 1. Notation for designating solublenondialyzable fractions extracted from H.

capsulatum chem 1 and chem 2 walls during serialenzymatic hydrolysis

Code Stage of extraction and contact time (h)

.3G, f8(1- 3)-glucanase, 5fG2 ,8(1- 3)-gIucanase, 24aG, a(1- 3)-glucanase, 6aG2 a(1- 3)-glucanase, 24Pr Pronase, 10

VOL. 16, 1977

on Decem

ber 2, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

692 REISS ET AL.

1,2,3,4 were used. Fluorescein-conjugated globulinsof normal rabbits served as the control, and, inaddition, a conjugate specific for factors 2 and 3 wasprepared by adsorption of labeled antiglobulins tofactors 1, 2, and 3 with whole yeast-form cells of B.dermatitidis (12, 15). The methods used for carryingout the direct FA examinations were those of Kauf-man and Kaplan (15). A Leitz SM microscope with a

cardioid dark-field condenser was used as well as a

Leitz 250 lamp housing equipped with an OsramHBO-200 mercury vapor lamp. A BG12 primary fil-ter and a GG9 barrier filter were used. Photographyof immunofluorescence was recorded on Kodak Plus-X film developed in Kodak D76.

Polyacrylamide gel electrophoresis in the pres-

ence of sodium dodecyl sulfate (SDS-PAGE) was

performed on one antigenic fraction, chem 1 BG1,and on the f8(1- 3)-glucanase enzyme with which itwas extracted from the cell wall. The conditions forelectrophoresis in phosphate buffer and 8 M urea

were those of Obijeski et al. (21).Electron microscopy. For ultramicrotomy, lyoph-

ilized cell wall pellets were stained with 2% aqueous

potassium permanganate (20) and embedded inEpon 812 (19). Sections were cut on a Sorvall MT-2Bultramicrotome, poststained with lead citrate (25),and viewed in a Philips EM 30V electron micro-scope. Negative staining of wall components was

accomplished with saturated aqueous uranyl ace-

tate or 2% phosphotungstic acid. Other preparationswere shadowed with platinum-paladium.

RESULTS

Chemical composition of walls, muralcores, and soluble fractions released duringserial enzymolysis. Since glucose, mannose,galactose, and glucosamine were known to becomponents of the H. capsulatum cell wall (4,

INFECT. IMMUN.

6, 11, 16, 22), the proportion of each of thesemonosaccharides in the products of serial en-

zymolysis was determined in order to assess

their possible roles as antigenic determinants(Table 2). Although the enzyme-resistant mu-ral core of chem 1 comprised only 18.7% of thestarting wall weight (23), the initial and finalproportions of glucose, mannose, and galactoseremained similar (Table 2). The level of aminosugar, initially 29.8%, was elevated to 34.9% inthe core, despite extensive chitinase treatment.In the instance ofchem 2, the core as comparedwith the intact wall was poorer in glucose con-tent, but richer in amino sugar. The soluble,nondialyzable polysaccharide released at eachstage of serial enzymolysis amounted to acumulative total of 12.1% of chem 1 and 9.3%of the chem 2 wall (23). These fractions con-tained galactose, mannose and a predominantamount of glucose (Table 2). In a fraction thatwas found to be antigenic chem 1 f3G1, theratio of mannose to glucose to galactose was6.6:13.3:1. Because of the differences seen inIEP patterns of mural antigens (see Fig. 2),which are explained in detail below, it wasnoteworthy that the monosaccharide distribu-tion found in chem 1 f3G, recovered after 5 h of,3(1-+3)-glucanase digestion differed from thatfound in chem 1 f3G2 recovered after 24 h ofglucanase digestion. In interpreting the proteincontent data shown in Table 2, it is importantto note that the total protein in soluble ex-

tracts was the sum of protein indigenous to thecell wall and enzyme protein that was used toeffect serial enzymolysis. In Table 2, the con-tribution of extraneous enzyme proteins was

TABLE 2. Monosaccharide and protein composition ofH. capsulatum yeast-form cell walls, mural cores, andsoluble, nondialyzable polysaccharide products of serial enzymolysis

Cell wall or fraction Dry wt yield % of dry Wta(mg) Man Glc Gal GlcN Protein

chem 1Cell walls 400 3.3b 39.7b 1.1b 29.8c 14.5Mural core 59.5 5.8 39.6 1.7 34.9 9.3df3G1 28.8 17.8 35.8 2.7 5.4 11.0ef3G2 14.6 31.9 33.5 4.1 7.9 6.5ePr 3.7 3.8 11.5 9.5 1.7 13.4

chem 2Cell walls 500 2.9 72.5 1.3 16.8 9.4Mural core 51.5 3.5 60.3 1.3 25.4 2.1aG, 14.0 5.3 27.2 6.6 0 8.4aG2 11.8 10.5 29.8 3.3 0 8.6,1G1 9.4 25.5 18.6 5.8 Trace 0.9

a Abbreviations: Man, mannose; Glc, glucose; Gal, galactose; GlcN, glucosamine.b Gas chromatography as n-butyl boronic acid esters.c Elson-Morgan test.d Nessler test.e Folin phenol.

on Decem

ber 2, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

CELL WALL ANTIGENS OF H. CAPSULATUM 693

subtracted to give indigenous protein values.In addition to neutral sugars, soluble fractionsderived from chem 1 walls contained glucosa-mine, which was totally released after 3 hr ofhydrolysis, but not detected in fractions ofchem2 hydrolyzed for 3, 6, and 18 h. Elementalanalysis of the H. capsulatum cell walls and themural core for total P showed an increase in thecore of chem 1 (intact wall = 0.04% P; core =0.14% P); however, the result obtained withthe chem 2 core was inconclusive.ID and IEP. The reactions of wall-derived

soluble fractions of both chemotypes with se-rum from a goat injected with live yeast-formcells and with serum from a human patientwith histoplasmosis are shown in Fig. 1 andgive some evidence of serotype differences. Se-rum of a goat immunized with chem 1 yeast-form cells reacted with chem 2 extracts, and inaddition to this common factor the serumformed an immune precipitate that was relatedonly to chem 1 antigens (Fig. lB). The occur-rence of type-specific factors in extracts chem 1,3G, and 8G2 was preliminary evidence thatserotype markers up to now recognized onlywith FA could be detected by use of materialextracted from isolated cell walls. The relation-ship ofchem 1 /3G antigens and those present inhistoplasmin is shown in Fig. 1A. Lines of in-tersection between the H antigen of histoplas-min (8, 9) and chem 1 f3G components weredetected, indicating that cell wall-derived anti-gens prepared in this manner did not contain Hor M factors.The IEP patterns obtained in the reaction of

chem 1 antigens and goat antiserum are shownin Fig. 2. Each fraction was heterogeneous, andthere were differences among /3G1, f3G2, and Prantigens in the amounts and mobilities of thecomponents. The first fraction released fromthe chem 1 walls, f3G1, contained only one ma-jor anodic antigen, whereas an additionalweakly cathodic component occurred in 83G2.The pattern caused by the Pr extract was differ-ent from that due to f3G1, /3G2 due to an addi-tional strongly cathodic component. This shiftin mobility after digestion with Pronase mayhave altered the size and net charge of the cellwall antigen. A human histoplasmosis serumthat contained precipitins reactive with H. cap-sulatum H and M antigens (8, 9) was employedin IEP to compare the mobility of these factorswith that of cell wall-derived antigens. Histo-plasmin components migrated toward the an-ode in compact zones, and, on the basis of theID pattern and electrophoretic mobility, it wasconcluded that the cell wall antigens were notidentical to the H or M factors.CF. Sera from human patients with histo-

plasmosis, blastomycosis, and coccidioidomyco-sis (Table 3) were screened in CF tests withpolysaccharide extracts chem 1 83G1 and 83G2.The soluble extracts were less sensitive (9/17, 8/14) as antigens than were particulate glucan-ase-extracted cell walls (14/17), and the Pr ex-tract was not reactive. The ability of chem 1walls to react in the CF test with goat antise-rum and with serum from a human with histo-plasmosis was not diminished by serial enzym-olysis. After a(1- *3)-glucanase treatment,chem 2 walls were more reactive in the CF test,but Pronase treatment led to a fourfold de-crease in end point serum dilution (Table 4)with human serum, but not with goat serum.FA staining of intact and enzymatically

treated cell walls. Intact cell walls of the chem1 strain belonging to serotype 1,2,3 werestained with conjugated H. capsulatum anti-bodies to a 1,2,3,4 serotype. After /(1- 3)-glu-canase treatment, staining was a littlebrighter, but Pronase treatment of /3(1-*3)-glu-canase-extracted chem 1 walls decreased theimmunofluorescence. The mural core stillstained, but the intensity of fluorescence waslow. The chem 1 core also stained with specificconjugated H. capsulatum antibodies producedby adsorption with the yeast form ofB. derma-titidis, which indicated the persistence of an H.capsulatum-specific factor. chem 2 walls aftera(1-+3)-glucanase treatment stained morebrightly than untreated walls, which was unex-pected since a(1l-3)-glucanase removed 27% ofthe wall dry weight. A decline in staining oc-curred after Pronase and chitinase extractions,but the chem 2 core still was capable of reactingwith labeled H. capsulatum antibodies.SDS-PAGE of a cell wall-derived antigen.

Since significant N was associated with thepolysaccharide antigenic fractions, it was of in-terest to examine this preparation to estimateits molecular size and the number of separableprotein or glycoprotein components (Fig. 3). Apattern of four proteins was observed in thechem 1 /3G1 fraction, three of which coincided inmobility with the ones seen in a concentratedsample of /3(1-+3)-glucanase. A line nearest theorigin did not occur in the enzyme sample andwas presumed to be the antigen. A comparisonof the mobility of this line with standard-sizemarker proteins (,8-galactosidase and phospho-rylase a) indicated a molecular weight of120,000. Carbohydrate moieties were implied inthe case of all four lines that reacted positivelywith the periodic acid-Schiff reagents. Towardthe distal portion of the antigen-containing gel,a diffuse staining area was observed, whichmay also represent a cell wall component.Electron microscopy. When viewed in nega-

VOL. 16, 1977

on Decem

ber 2, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

694 REISS ET AL.

FIG. 1. ID patterns ofH. capsulatum cell wall-derived antigens versus sera ofhuman and goat origin. (A)Center well: human serum from a patient with histoplasmosis; outer wells: 1, 3, and 5, histoplasmin; 2, chem1 P8G I 4 and 6, chem 1 f3G2. (B) Center well: serum ofgoat immunized with live chem 1 yeast-form cells; 1,chem 2 aG1, 2, chem 1 /3G, 3, chem 2 aG2; 4 and 6, chem 1 1G2; 5, chem 2 83G,. Arrow refers to immuneprecipitate present in chem 1 and absent from chem 2 antigens in adjacent wells.

INFECT. IMMUN.

on Decem

ber 2, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

CELL WALL ANTIGENS OF H. CAPSULATUM 695

tively stained preparations, the enzyme-resist-ant mural cores of both chemotypes werethinned out and had pocked, cratered regions,in contrast to the smooth, denser appearance ofthe untreated walls (Fig. 4). No clear-cut differ-ences were observed between chemotypes.There were subtle differences in texture; forexample, the negatively stained chem 2 muralcore had a very prominent pocked surface.Whereas the chem 2 wall was shattered by aBraun homogenizer, chem 1 walls had a greatertendency to maintain an elliptical shape aftermechanical disruption. Shadow-casting madeglobular patches and underlying fibrils visiblein both walls and mural cores (Fig. 5). In chem1 walls, globular patches seemed more preva-lent than in chem 2 walls, and a coarser net-work of fibrils occurred in chem 1. Mural coresof both chemotypes showed the persistence ofglobular patches coexisting with fibrils. Thefibrils in the chem 1 mural core were more

clearly delineated and thicker than those ofchem 2.

DISCUSSIONThe process of unraveling the antigenic

structure ofH. capsulatum cell walls was aidedby serial enzymolysis of two strains that dif-fered in serotype and chemotype. It was possi-ble to recover relatively small amounts of solu-ble polysaccharide fragments from the enzymedigest that were antigenically reactive withgoat antiserum toH. capsulatum and sera fromhumans with histoplasmosis. The insolublewall residue after enzymolysis, i.e., the muralcore, also bore antigenic determinants detectedin CF and FA tests. Of the soluble extracts, thesoluble polysaccharide extracted from the cellwall of chem 1 at the f3(1-*3)-glucanase stage(chem 1 f3G1) was antigenically significantsince it was not identical to H or M factors ofhistoplasmin and appeared to contain a sero-

FIG. 2. IEP ofchem 1 antigens (wells). Serum ofa goat was immunized with live chem 1 yeast-form cells(troughs). H and M factors of histoplasmin are shown with respect to a rabbit antiserum (trough).

VOL. 16, 1977

on Decem

ber 2, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

696 REISS ET AL.

TABLE 3. Reactivity in the CF test of solubilizedantigens and cell wall residues of H. capsulatumafter 3(1 -1.3)-glucanase treatment versus sera ofpatients with histoplasmosis and other mycotic

infections

Antigen of chem 1Source of serum with re-spect to disease of pa- Solublea Insoluble

tient,8G, 8G2 GECWb

Histoplasmosis 9/17c 8/14 14/17Blastomycosis 2/4 2/4 4/4Coccidioidomycosis 1/5 2/4 3/5a A 5-gg dose.b GECW, f(1- 3)-glucanase-extracted cell wall

residue (1.6-,ug dose).c Number of positive reactors among total sera

tested.

INFECT. IMMUN.

1 f3G1, f3G2, and f8(1->3)-glucanase-extractedchem 1 walls. The results indicated that 57% ofthe panel reacted with 8G1, 60% with 8G2, and82% with the particulate antigen. Cross-reac-tivity was most frequent in the case of theparticulate fraction and less evident with chem1 f3G1. This screening of human sera was pre-liminary in nature. A thorough evaluation ofthese and other H. capsulatum antigens with apanel of serial serum specimens from humanswith acute, disseminated, and chronic histo-plasmosis is warranted to see whether the se-quence of precipitins directed against differentpurified histoplasmal components correlateswith the clinical status of the patient.The presence of glucose, galactose, mannose,

glucosamine, and 11.0% protein was detected inthe chem 1 8G1 extract, and the structuralrelationship of the sugars present is being

TABLE 4. End point serum dilutions in the CF test,using H. capsulatum chem 2 cell walls and

particulate fractions resistant to serial enzymolysis

Anti-H. capsulatum se-

Cell wall antigena rum titer-'Human Goatb

Pretreatment walls 32 64a(1- 3)-glucanase resistant 64 64,B(1- 3)-glucanase resistant 128 64Pronase resistant 32 128Mural core 32 64

a Antigen dose, 0.8 /tg.b Anti-chem 2.

Chem ILj3G1*

4

QIuC2riCse

type-specific component revealed by ID. In aside-by-side ID comparison with extracts of thechem 2 (1,4 serotype) isolate, the chem 1 3G,antigen contained an additional determinantnot found in chem 2 extracts when analyzedwith goat antiserum produced against the chem1 strain. There have been no previous demon-strations of serotype specificity in precipitintests with antigens ofH. capsulatum. The ini-tial description of the 1,4 serotype ofH. capsu-latum gives the geographic distribution asbeing mainly in Central America. It was alsofound in the United States, but less frequentlythan serotype 1,2,3 (12). Further studies areneeded to relate serotype, chemotype, and mat-ing type (18) to the origin of the isolate,whether from soil or from the human host.The chem 1 8G1 and 8G2 extracts were sepa-

rated into a maximum of three antigenic com-ponents by IEP, and the sugar composition var-ied with increased time of digestion with,8(1-*3)-glucanase. Several human sera thatwere reactive in CF tests with whole yeast-formH. capsulatum cells and in ID tests with histo-plasmin were screened for reactions with chem

FIG. 3. SDS-PAGE of chem 1 83G, antigenic frac-tion ofH. capsulatum cell walls (400 pg) and concen-trated sample of /3(1 -3)-glucanase (100 pAg). Aster-isk shows presumed location of antigen component.Coomassie blue G stain.

on Decem

ber 2, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

CELL WALL ANTIGENS OF H. CAPSULATUM

b

.40A

FIG. 4. Negatively stained cell walls (a and c) and mural cores (b and d) ofH. capsulatum chem 1 (a andb) and chem 2 (c and d). Marker = 1 Am.

sought. The molecular size estimate of 120,000obtained by SDS-PAGE in the instance ofchem1 f3G1 should assist in the choice of conditionsfor further separations. Every soluble antigenicfraction and the enzyme-resistant mural corescontained galactose and mannose in addition toglucose. Galactomannan has been described asan antigen ofH. capsulatum (1, 24), and in thepresent study chem 1 walls were found to bericher in it than those of chem 2. It appearedthat this heteroglycan may contribute to theantigenicity of all the fractions describedabove. Previous methods for extracting galacto-mannan with dilute alkali may not conservepeptide bridges, particularly in view of thewell-documented alkaline cleavage of O-glyco-sidic linkages to hydroxylamino acids of thepeptide moiety of glycoproteins (7). For thesereasons it is important to characterize thestructure of the fractions chem 1 j3G1 and 8G2 todetermine whether covalent bonds link glucose

oligosaccharides to galactomannan and peptidecomponents. Material extracted at the Pronasestage was found to be antigenic in IEP. Theapparent lack ofCF activity may have been dueto traces of Pronase sufficient to inactivate com-plement components, making it advisable inthe future to try immobilized Pronase. A mini-mal role of the peptide moiety ofH. capsulatumcell walls in the antigenic structure was in-ferred from the observations that Pronase-re-sistant walls were capable of binding FA andthat the soluble fraction of chem 1 released atthe Pronase stage of digestion was resolved intothree components in IEP. The only evidence ofan antigenic role for the peptide portion of thewalls studied was the fourfold reduction of CFend point serum dilution of walls of chem 2after Pronase digestion. Although the solublecell wall-derived antigens were referred to aspolysaccharide, considerable protein was asso-ciated with them (Table 2). The existence of

VOL. 16, 1977 697

... s

ok.,..Ck..g

on Decem

ber 2, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

698 REISS ET AL.

I

FIG. 5. Shadow-cast cell walls (A and C) and mural cores (B and D) ofH. capsulatum chem 1 (A and B)and chem 2 (C and D). Marker = 0.25 ,tm.

peptidopolysaccharides was given credence bythe staining reactions of the chem 1 f3G1 extractduring SDS-PAGE.The occurrence of glucosamine in extracts of

chem 1, but not ofchem 2, led to the speculationthat this sugar may have a role as a serotype-specific hapten. A previous study also impli-cated glucosamine as a significant componentof ethylenediamine extracts of H. capsulatumcell walls (6). In the present work, the com-plete liberation of glucosamine during short-term acid hydrolysis from soluble fractions sug-gested that the sugar occurred in a terminalposition. A large segment of the mural core wascomposed of amino sugar apparently resistantto chitinase. Extraction of a(1->3)-glucan to theextent of 27.3% of the chem 2 mural dry weight(23) did not result in any detectable decrease inFA and CF reactions, and thus it was not possi-ble to ascribe an antigenic role to a(1- 3)-glu-can by this indirect means.The enzyme-resistant mural cores that re-

tained some antigenicity merit further study to

define what other specific enzyme-mediatedchanges would result in additional depolymeri-zation and loss ofFA and CF activities. In bothchem 1 and chem 2 cores there was an enrich-ment in galactose, mannose, and especiallyamino sugar, despite prolonged chitinase diges-tion. Among the linkages to be considered asresistant to the serial enzymolysis procedureare those of the chitosan variety and also phos-phodiesters. In an analogous situation the man-nan core of bakers' yeast resistant to exo-a-mannanase was found to be enriched in phos-phodiesters (3). In the chem 1 mural core therewas a 3.5-fold increase in total P as comparedwith the walls before enzymolysis.chem 1 and chem 2 walls were monitored at

each stage of serial enzymolysis by CF, FA, andelectron microscopy. Although chem 1 wallslost 61.5% of the initial dry weight as monomerand polysaccharide as a result of enzyme treat-ment (3), there was no evident decrease in CFactivity in the mural core. After f8(1--3)-glu-canase extraction, the immunofluorescence of

INFECT. IMMUN.

MmEadmb7-likamlaft

on Decem

ber 2, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

CELL WALL ANTIGENS OF H. CAPSULATUM 699

chem 1 walls was brighter than that of intactwalls. It was hard to reconcile this observationwith the extraction at the 83(1-33)-glucanasestage of fragments that were antigenic in CFand ID tests. One possible explanation was thatthe digestion of ((1-*3)-glucan exposed moresurface area for binding antibody and compen-sated for the loss of antigen through solubiliza-tion. Electron microscopic observation sup-ported this view since a loosening of fibrils wasobserved in enzymatically treated walls. Theresidual immunofluorescence persisted in thechem 1 mural core stained with anti-H. capsu-latum globulins adsorbed with B. dermatitidisyeast-form cells. This immunofluorescence in-dicated that determinants responsible for themarkers Kaufman and Blumer (12) designatedas 2,3 were at least partly retained. In the caseof chem 1 walls, an increase in CF reactivitywas seen after extraction with f3(1- 3)-glucan-ase and Pronase. Since this effect was observedwith sera from both an immunized goat and ahuman histoplasmosis case, it was given somecredence. It is not known whether this in-creased CF activity was due to an unmasking ofdifferent antigenic determinants or the resultof smaller particle size that left more surfacearea for antibody binding.The conclusion drawn from electron micros-

copy of the H. capsulatum wall was that it ismade up of a combination of tightly woven finefibrils covered with a discontinuous or patchylayer of cottony, globular elements. Enzymo-lysis eroded craters in the wall matrix and re-moved some globular material, but the fibrilspersisted even after chitinase treatment. Un-like the effect in bakers' yeast (17), /8(1->3)-glucanase did not remove the globular patches.In contrast to the walls of H. farciminosum,two size classes of fibrils were not detected inthe H. capsulatum walls of either chemotype(26). In Neurospora crassa walls both amor-phous and fibrillar regions occurred, but,8(1-*3)-glucanase and Pronase treatment re-sulted in an open net of coarse fibers that wasnot seen in the H. capsulatum walls (10).

Electron microscopy was a valuable adjunctto the serological monitoring of enzymolysis.Although the cell wall carbohydrate composi-tion of the two chemotypes differed, the surfacetexture offered only minor evidence in chem 1walls of more globular patches and a coarserunderlying layer. Negative staining andshadow-casting gave different information, andboth techniques were required since globularpatches were visible only in shadow-cast prepa-rations, whereas negative staining revealedcratering in the mural cores. The persistence ofglobular patches and fibrils in the mural cores

coincided with the occurrence of galactose,mannose, and glucosamine. It is tempting tospeculate that the globular material is galacto-mannan adhering to residual chitin fibrils.

ACKNOWLEDGMENTS

We would like to thank J. F. Obijeski for conducting thegel electrophoresis studies and Maxine Clark for assistancewith CF tests.

LITERATURE CITED

1. Azuma, I., F. Kanetsuna, Y. Tanaka, Y. Yamamura,and L. M. Carbonell. 1974. Chemical and immunolog-ical properties of galactomannans obtained from His-toplasma duboisii, Histoplasma capsulatum, Paracoc-cidioides brasiliensis, and Blastomyces dermatitidis.Mycopathol. Mycol. Appl. 54:111-125.

2. Boas, N. F. 1953. Method for the determination of hex-osamines in tissues. J. Biol. Chem. 204:553-563.

3. Colonna, W. J., and J. 0. Lampen. 1974. Structure ofthe mannan from Saccharomyces FH4C, a mutantconstitutive for invertase biosynthesis. I. Signifi-cance ofphosphate to the structure and refractorinessof the molecule. Biochemistry 13:2741-2748.

4. Domer, J. 1971. Monosaccharide and chitin content ofcell walls of Histoplasma capsulatum and Blasto-myces dermatitidis. J. Bacteriol. 107:870-877.

5. Eisenberg, F., Jr. 1972. Gas chromatography of carbo-hydrates as butaneboronic acid esters, p. 168-177. InS. P. Colowick and N. 0. Kaplan (ed.), Methods inenzymol., vol. 28. Academic Press Inc., New York.

6. Garcia, J. P., and D. H. Howard. 1971. Characteriza-tion of antigens from the yeast phase ofHistoplasmacapsulatum. Infect. Immun. 4:116-125.

7. Gottschalk, A. 1972. Glycoproteins, their compositionstructure, and function, 2nd. ed., part A, p. 470.Elsevier Publishing Co., Amsterdam.

8. Harrell, W. K., H. Ashworth, L. E. Britt, J. R. George,S. B. Gray, J. H. Green, H. Gross, and J. E. Johnson.1973. Preparation ofHistoplasma capsulatum (histo-plasmin) CF and immunodiffusion antigen, p. 98-101.In Procedural manual for production of bacterial,fungal, parasitic reagents. Center for Disease Con-trol, Atlanta, Ga.

9. Heiner, D. C. 1958. Diagnosis of histoplasmosis usingprecipitin reactions in agar gel. Pediatrics 22:616-627.

10. Hunsley, D., and J. H. Barnett. 1970. The ultrastruc-tural architecture of the walls of some hyphal fungi.J. Gen. Microbiol. 62:203-208.

11. Kanetsuna, F., L. M. Carbonell, F. Gil, and I. Azuma.1974. Chemical and ultrastructural studies on the cellwalls of the yeastlike and mycelial forms of Histo-plasma capsulatum. Mycopathol. Mycol. Appl. 54:1-13.

12. Kaufman, L., and S. Blumer. 1966. Occurrence of sero-types among Histoplasma capsulatum strains. J. Bac-teriol. 91:1434-1439.

13. Kaufman, L., M. Huppert, C. Fava Netto, L. Pollak,and A. Restrepo. 1972. Manual of standardized sero-diagnostic procedures for systemic mycoses, part I,agar immunodiffusion tests, p. 1-14. Pan AmericanHealth Organization, Washington, D.C.

14. Kaufman, L., M. Huppert, C. Fava Netto, L. Pollak,and A. Restrepo. 1972. Manual of standardized sero-diagnostic procedures for systemic mycoses, part II,complement fixation tests, p. 1-39. Pan AmericanHealth Organization, Washington, D.C.

15. Kaufman, L., and W. Kaplan. 1961. Preparation of afluorescent antibody specific for the yeast phase ofHistoplasma capsulatum. J. Bacteriol. 82:729-735.

VOL. 16, 1977

on Decem

ber 2, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

700 REISS ET AL.

16. Kobayashi, G. S., and P. L. Guiliacci. 1967. Cell wallstudies of Histoplasma capsulatum. Sabouraudia5:180-188.

17. Kopecka, M., H. J. Phaff, and G. H. Fleet. 1974. Dem-onstration of a fibrillar component in the cell wall ofthe yeast Saccharomyces cerevisiae and its chemicalnature. J. Cell Biol. 62:66-76.

18. Kwon-Chung, K. J. 1974. Studies on Emmonsiella cap-

sulata (Histoplasma capsulatum). II. Distribution ofthe two mating types in 13 endemic states of the U.S.Am. J. Epidemiol. 99:44-49.

19. Luft, J. H. 1961. Improvements in epoxy resin embed-ding methods. J. Biophys. Biochem. Cytol. 9:409-414.

20. Mollenhauer, H. H. 1959. Permanganate fixation ofplant cells. J. Biophys. Biochem. Cytol. 6:431-435.

21. Obieski, J. F., A. T. Marchenko, D. H. L. Bishop, B.W. Cann, and F. A. Murphy. 1974. Comparative elec-trophoretic analysis of the virus proteins of four rhab-doviruses. J. Gen. Virol. 22:21-33.

INFECT. IMMUN.

22. Pine, L., and C. J. Boone. 1968. Cell wall compositionand serological reactivity ofHistoplasma capsulatumserotypes and related species. J. Bacteriol. 96:789-798.

23. Reiss, E. 1977. Serial enzymatic hydrolysis of cell wallsof two serotypes of yeast-form Histoplasma capsula-tum with a(1-*3)-glucanase, /8(1-3)-glucanase, Pro-nase, and chitinase. Infect. Immun. 16:181-188.

24. Reiss, E., W. 0. Mitchell, S. H. Stone, and H. F.Hasenclever. 1974. Cellular immune activity of a gal-actomannan-protein complex from mycelia of Histo-plasma capsulatum. Infect. Immun. 10:802-809.

25. Reynolds, E. S. 1963. The use of lead citrate at high pHas an electron opaque stain in electron microscopy. J.Cell Biol. 17:208-212.

26. San-Blas, G., and L. M. Carbonell. 1974. Chemical andultrastructural studies on the cell walls of the yeast-like and mycelial forms of Histoplasma farcimi-nosum. J. Bacteriol. 119:602-611.

on Decem

ber 2, 2020 by guesthttp://iai.asm

.org/D

ownloaded from