-
Vol. 29, No. 8JOURNAL OF CLINICAL MICROBIOLOGY, Aug. 1991, p.
1610-16150095-1137/91/081610-06$02.00/0Copyright C) 1991, American
Society for Microbiology
43-Kilodalton Glycoprotein from Paracoccidioides
brasiliensis:Immunochemical Reactions with Sera from Patients
with
Paracoccidioidomycosis, Histoplasmosis,or Jorge Lobo's
Disease
ROSANA PUCCIA AND L. R. TRAVASSOS*
Disciplina de Biologia Celular, Escola Paulista de Medicina,Rua
Botucatu 862, Sao Paulo, SP 04023, Brazil
Received 10 September 1990/Accepted 7 May 1991
Sera from patients with paracoccidioidomycosis (PCM),
histoplasmosis (HP), or Jorge Lobo's disease (JL)were titrated
against purified gp43 from Paracoccidioides brasiliensis by using
both enzyme-linked immuno-sorbent assay (ELISA) and
immunoprecipitation (IPP) reactions with 1251-labeled antigens. In
IPP, PCM seraand other sera could be distinguished on the basis of
serum titers, whereas in ELISA, 53% of the HP sera and29% of the JL
sera reacted similarly to the PCM sera. To investigate the possible
role of the carbohydrateepitopes in these reactions, we compared
the reactivities of sera from several patients with native
anddeglycosylated gp43. Competition experiments were carried out
with monosaccharides as inhibitors. Theresults suggest that >85%
of the reactions of the PCM sera with gp43 involved peptide
epitopes. Cross-reactions with HP and JL sera in ELISA were
predominantly attributed to periodate-sensitive
carbohydrateepitopes containing galactosyl residues. HP and JL sera
which reacted strongly with gp43 in ELISA were onlyweakly reactive
or did not react in IPP with labeled antigens in solution.
Moreover, ELISA reactions could besignificantly inhibited either by
monosaccharides or by periodate treatment. Apparently,
carbohydrateepitopes in gp43 are more accessible to the antibodies
when the molecule is bound to a plastic substrate thanwhen it is in
solution. Structural changes in the gp43 antigen arising by N
deglycosylation abolish reactivitywith PCM sera and support the
existence of conformational peptide epitopes.
Human paracoccidioidomycosis (PCM) is a deep mycosiscaused by
Paracoccidioides brasiliensis, a dimorphic funguswhich grows in the
mycelial phase at room temperature andin the yeast phase at 35 to
37°C or in infected tissues. Theimmunological test currently used
for the diagnosis of thismycosis is based on a double
immunodiffusion (ID) testrecently standardized by Camargo et al.
(2). By using ID andexoantigens from 7-day-old cultures in the
yeast form, 97%sensitivity with 100% specificity was obtained for
the diag-nosis of PCM. This is in contrast with other
immunologicalmethods, such as complement fixation (10, 16, 20)
andenzyme-linked immunosorbent assay (ELISA) (1, 4, 14),with
unfractionated reagents. With these methods, cross-reactions with
sera from patients with other deep mycoses,such as histoplasmosis
(HP), Jorge Lobo's disease (JL),candidiasis, and blastomycosis,
frequently occur.
Previously (17), we identified the antigenic molecule
re-sponsible for both the band 1 specific immunoprecipitation(IPP)
in ID (19), and the arc E precipitation band formed
inimmunoelectrophoresis (27). This antigen, called gp43, is
aconcanavalin A-binding glycoprotein of 43,000 Da whichwas isolated
from the supernatant fluid of yeast cultures byaffinity
chromatography.
Later reports (2, 3) confirmed the specificity of gp43 forthe
diagnosis of PCM by ID and Western blotting (immuno-blotting) with
a greater number of sera. Previous results fromour laboratory (18,
25) showed that sera from patients withsystemic mycoses other than
PCM also reacted with theaffinity-purified gp43 in an ELISA format,
sometimes withhigh titers. We therefore started analysis of the
gp43 mole-
* Corresponding author.
cule to search for P. brasiliensis-specific and
cross-reactingepitopes and to determine the reactivity of this
molecule indifferent immunological assays.
In the present work, we studied gp43 reactivity with serafrom
patients with systemic mycoses, focusing on the pos-sible role of
carbohydrate epitopes. The gp43 isolated byaffinity chromatography
usually copurified with smallamounts of both a
high-molecular-weight (high-MW) com-plex, which included a
galactomannan component (17), anda 72-kDa glycoconjugate (3), as
detected by sodium dodecylsulfate-polyacrylamide gel
electrophoresis (SDS-PAGE).Further purification with gel filtration
provided a highlypurified gp43 antigen. The gp43 antigen was used
either inthe native state or deglycosylated to form a 38-kDa
polypep-tide. Several sera from patients with PCM, HP, or JL
werestudied by ELISA and liquid-phase IPP reactions, using
thepurified gp43 and deglycosylated preparations of this
anti-gen.
MATERIALS AND METHODSSera. The 50 PCM sera and 32 HP sera used
were from
Brazil, Venezuela, Argentina, and the United States. Eachserum
was checked for specificity via ID with P. brasiliensisand
Histoplasma capsulatum exocellular antigens (2). PCMand HP sera
with high titers in ID tests were used as positivecontrols. All
patients exhibited various clinical forms of thedisease and were
either untreated or in different stages ofantifungal treatment. The
27 JL sera were from clinicallydiagnosed Brazilian Indians from the
Amazon region(Xingu).
Antigen preparation. Cultures of P. brasiliensis B339,from the
Mycology Section, CBI-Hospital Pablo Tobon
1610
on June 1, 2021 by guesthttp://jcm
.asm.org/
Dow
nloaded from
http://jcm.asm.org/
-
43-kDa ANTIGEN OF P. BRASILIENSIS 1611
Uribe, Medellin, Colombia, were grown in a liquid
oragar-solidified complex medium enriched with tomato juicemedium.
This medium was modified from the commercialUniversal Beer Agar
medium (GIBCO Laboratories), whichwas originally used for the
selection of Saccharomycesstrains. It consisted of 240 ml of
gauze-filtered homogenizedpulp from fresh, peeled red tomatoes
supplemented with 6.1g of yeast extract, 16.1 g of glucose, 15 g of
casein peptone,0.31 g of K2HPO4, 0.31 g of KH2PO4, 0.12 g of
MgSO47H20, and 0.006 g (each) of MnSO4. H20, NaCl, andFe2SO4 and
made up to 1 liter with distilled water. The pHwas adjusted to 6.3.
For the solid medium, 15 g of agar perliter was added. For
preparation of the liquid medium, theyeast extract was first
dialyzed in 500 ml of distilled waterovernight at 4°C. The
dialysate, free of soluble yeast man-nan, was used to dissolve the
other medium constituents andto dilute the tomato juice. The medium
was sterilized byautoclaving at 120°C for 15 min. For preparation
of theexocellular antigen, the cell mass from 3 to 4 slanted
culturesin solid tomato juice medium was inoculated in 100 ml
ofliquid tomato juice medium and incubated for 4 to 5 days at35°C.
The resulting culture in the yeast phase was trans-ferred to fresh
medium (1 liter) and further incubated for 5 to10 days. Cultures
were then killed with 0.02% (wt/vol)merthiolate and filtered
through filter paper (Klabin 80 filterpaper; Klabin Industries, Sao
Paulo, Brazil). The superna-tant fluid was concentrated 10 times in
vacuo at 45°C anddialyzed against 0.9% NaCl-0.01 M phosphate buffer
(PBS)for 2 days (two changes of 2 liters of buffer) at 4°C. It
wasthen affinity chromatographed in columns of protein A-puri-fied
rabbit anti-gp43 immunoglobulin G coupled to Affi-Gel10 (Bio-Rad),
as described previously (17). The gp43 waseluted from this column
with 50 mM citrate buffer, pH 2.8,and was immediately neutralized
with 1 M Tris-HCl, pH 9.0,and then concentrated at 4°C under N2 in
an Amicon cellwith a PM10 Diaflo membrane. Aliquots of the final
prepa-ration were kept frozen, or at 4°C, for short-term
use.Further purification of affinity-purified gp43 was achieved
bygel filtration in a Sephacryl S-200 column (1.65 by 83
cm;PBS-0.02% Na azide buffer at a flow rate of 6 ml/h) toeliminate
high-MW contaminants.
Radiolabeling. Affinity-purified gp43 (30 Rg) was iodinatedwith
125I-Na (13 mCi/mmol) to a final specific activity of 3 x106
cpm/lpg by the method using insoluble
1,3,4,6-tetra-chloro-3a,6a-diphenylglycoluril (6). Free iodine was
elimi-nated by exhaustive dialysis against PBS, and aliquots
werekept frozen. In contrast to the affinity-purified gp43,
theSephacryl-purified gp43 preparations degraded after iodina-tion
and therefore could not be used.
Deglycosylation procedures. gp43 was deglycosylated bythe
following methods: (i) with endo-p-N-acetylglucosamini-dase H (endo
H; Sigma Chemical Co.) in 50 mM sodiumacetate buffer, pH 5.5, 50 mU
of enzyme per ml (670,ug) ofcold antigen or 1.5 mU of enzyme per
100,ul (20 x 106 cpm)of labeled antigen was used; (ii) with
N-glycanase (peptide:N-glycanase F; Genzyme) at 15 U/ml, the
labeled antigen (20x 106 cpm/100 RI) was first denatured by boiling
in 0.5%SDS, and the detergent in the reaction mixture was
thendiluted to 0.17% SDS with 0.1 M sodium phosphate buffer,pH 8.6,
containing 1.25% Nonidet P-40 before addition of theenzyme.
Enzymatic digestions were carried out for 18 h at37°C, and the
products were kept frozen. Endo H- andN-glycanase-treated gp43 were
called EH38 and NG38,respectively, because they migrated in
SDS-PAGE with anapparent molecular mass of 38 kDa.
Periodate oxidation of iodinated antigens. 125I-labeled gp43
or EH38 (3 x 106 cpm/,ug) was oxidized with 10 mM
sodiummetaperiodate (Merck) in 50 mM acetate buffer, pH 4.5, for30
min at 28°C in the dark. The reaction was interrupted byadding 0.5%
glycerol (10 min, room temperature). Aldehydegroups were blocked
with 1% (final concentration) glycine in0.1 M Tris-HCl buffer, pH
7.4, with 0.2% bovine serumalbumin for 1 h at room temperature. The
reaction mixturewas dialyzed against PBS for 2 h at 4°C. It was
thenaliquoted and kept at -20°C for further use. Control reac-tions
followed the same protocol except that periodate wasnot added.
IPP. Sera were serially diluted (10-fold dilutions),
startingfrom 300-1, in PBS with 0.05% bovine serum albumin and1%
Triton X-100. Diluted sera (final volume, 200 pA) wereincubated for
1 h at 37°C with 4,000 to 5,000 cpm of125I-labeled gp43 or EH38. A
10% suspension of Staphylo-coccus aureus (Cowan 1 strain) was added
(50 ,ul) andallowed to react for 30 min at room temperature.
Boundimmunocomplexes were separated by centrifugation (Ep-pendorf
microcentrifuge, 12,000 rpm, 1 min) and washedtwice in the
PBS-BSA-Triton X-100 buffer described above.Pellets were counted
for total radioactivity (counts perminute) in a gamma counter. The
negative control corre-sponded to a pool of 10 sera from healthy
individuals, and itsreactivity was taken as the background value.
The resultswere calculated as percentages of counts per minute
precip-itated, as follows: (Pcpm - Bcpm)ITcpm x 100, where
Pcpm,Bcpm, and Tcpm are the counts per minute
corresponding,respectively, to the immunoprecipitate, background
reactiv-ity (negative sera), and total radioactivity added.
Serumtiters were recorded as the greatest serum dilutions
immu-noprecipitating more than 10% of the total radioactivity.Two
determinations were made simultaneously for eachserum, using the
same antigen preparation. The positivecontrol serum (from a
patient) at a 300-1 dilution immuno-precipitated ca. 95% of the
total radioactivity added,whereas 2 to 5% precipitation was
obtained with the negativecontrol serum. The PCM sera
immunoprecipitated at least50% (2,000 to 2,500 cpm) of the total
radioactivity at dilu-tions higher than 300-1.The antigenicities of
EH38, periodate-treated EH38, and
gp43 in the presence of carbohydrate were tested with alimited
number of sera (seven PCM sera). The workingdilution of these sera
was that sufficient to immunoprecipi-tate 50% of the 1251-gp43.
Results are means of threedeterminations, with standard deviations
not higher than 5%of the mean. When autoradiograms were to be
taken, over30,000 cpm of the total antigen was added in each
reaction,and the immunocomplexes were processed as
describedpreviously (17). Competitive assays were carried out
withthe following monosaccharides at 0.1 or 0.2 M: D(+)-galac-tose,
a-methyl-D-galactopyranoside, P-methyl-D-galactopy-ranoside,
a-methyl-D-mannopyranoside, D( - )-arabinose,and D(+)-xylose (all
from Sigma). Inhibition values for bothIPP and ELISA were the
percent differences between theresults obtained with the control
and the experimental sys-tems.ELISA. ELISA was carried out as
described previously
(1), but with the following minor changes. ELISA plateswere
coated with 4 ,ug of gp43 or EH38 per ml (50 ,ul perwell) overnight
at 4°C in 0.1 M sodium carbonate-bicarbon-ate buffer, pH 9.6.
Alternatively, a 4-h incubation at 37°Cwas used when gp43 was
tested alone. Blockade of the freesites on the plate was achieved
with PBS-0.1% Tween20-0.7% gelatin for 2 h at 37°C. Diluted sera
and a 500-1dilution of anti-human immunoglobulin G conjugated
with
VOL. 29, 1991
on June 1, 2021 by guesthttp://jcm
.asm.org/
Dow
nloaded from
http://jcm.asm.org/
-
1612 PUCCIA AND TRAVASSOS
horseradish peroxidase (Sigma) were added in
successiveincubation steps of 1 h at 37°C in the same buffer.
Betweenthe incubation steps, the wells were washed four times
withPBS-0.1% Tween 20. The substrate o-phenylenediamine(0.04%) and
0.05% H202 were added to the complexedconjugate, and the reaction
was allowed to develop for 5 minat room temperature in 0.04 M
phosphate-0.025 M citratebuffer, pH 5.0. The A492 was read in a
Titertek MultiskanMCC/340.
Sera were serially diluted (threefold dilutions) startingfrom
300-1, and the titer was that which gave an A492 at least0.1 U
higher than that for the negative control serum at thesame
dilution. The negative and positive controls were thesame as
described for IPP. All sera were also tested forbackground
reactivity without the antigen. Two or threetitrations were carried
out for each serum on different days.
In competition assays, reactions were run with monosac-charides
at 0.07 or 0.2 M (final concentration) as describedfor IPP.
Treatment of gp43 and EH38 with 10 mM periodatewas carried out in
the ELISA plates, essentially as describedpreviously (26), for 30
min at 28°C in the dark.PAGE of labeled antigens. SDS-PAGE of the
I251-antigens
was carried out by the method of Laemmli (13). Gels
withradiolabeled samples were dried and autoradiographed onKodak
XK-1 films at -70°C with an intensifying screen.
RESULTS
Titration of sera for IPP and ELISA. Homologous sera
andheterologous sera (from patients other than those with PCM)(50
PCM, 32 HP, and 27 JL sera) were titrated with gp43 byELISA; for
IPP, 40 PCM, 30 HP, and 26 JL sera were used.Quantitative
differences were seen in the distribution of theend-point dilutions
of these sera (Fig. 1). In the case of thePCM sera, this was
expected from the heterogeneity of thespecimens and the fact that
they had been collected ingeographically diverse locations. For the
same PCM serum,however, the titers obtained by ELISA and IPP were
similarin 80% of the cases.
In ELISA (Fig. 1B), 53% of HP sera and 29% of JL serahad titers
higher than 2,700'. These reactions were similarto those of the PCM
sera and contrasted with the resultsobtained in IPP reactions, in
which a distinct separation ofthe serum reactivities was clear
(Fig. 1A). A few HP and JLsera reactive with gp43 in ELISA gave IPP
reactions of 10 to18% of the total radioactivity added when they
were diluted300-1. They differed from the PCM sera, which
immunopre-cipitated at least 50% of the total radioactivity at the
samedilution. Among HP sera reactive in ELISA (titer higherthan
2,700-1), none was from Venezuela, 25% were from theUnited States,
and 75% were from Argentina. The lowreactivity of 10 sera from
Venezuela differed significantly (P< 0.01) from that of the
other HP sera. Substituting 1251_EH38 for 1251-gp43 gave the same
distribution as in Fig. 1A.
Antigenicity of carbohydrate structures in ELISA. To as-sess the
possible involvement of the carbohydrate epitopesof gp43 in the
reactions obtained by ELISA, we tested somesera with
periodate-treated gp43, EH38, and periodate-treated EH38. When the
protocol for binding antigen de-scribed in Materials and Methods
was used, 60% of the totalprotein bound to the plates, as was
determined with iodi-nated gp43 and EH38. Oxidation with periodate
was carriedout in the plates, as described previously (26). A 10 mM
finalconcentration of periodate was chosen, since at this
concen-tration most concanavalin A-binding residues in the
antigenwere destroyed (data not shown), leaving the protein
core
_ 300.
0, 30-iwu
1-3.
'0.3.
>2,1 87 -
2,187-
729-
243-0
81w
' 27-I9-
-
43-kDa ANTIGEN OF P. BRASILIENSIS 1613
** **
**AA
** ** 0* *
* 00000 ** 00000 * * @0000-0 0000. *@000
**
A *&*
@0*6^ **- AA00000 * * 00000 * *
gp43+ per I EH38 I EH38+ per I GaI I M a n I Ara / Xyl
FIG. 2. Distribution ofPCM (0), HP (*), and JL (A) sera
according to the percentages of inhibition of their reactivities
with gp43 in ELISAby the modified antigens periodate-treated gp43
(gp43 + per), EH38, and periodate-treated EH38 (EH38 + per) or by
the carbohydrateD-galactose (Gal), a-methylmannoside (Man),
D(-)arabinose (Ara), or D(+)xylose (Xyl). The percentage of
inhibition of the reaction of eachserum was calculated as the
difference between the mean A492 for the original reaction with
gp43 (taken as 100%) and the A492 of the inhibitedreactions with
modified antigens or carbohydrates. Each point gives the results
for one serum sample.
reactions, which were common when monosaccharides wereused at
high concentration (0.2 M), the monosaccharidesbeing tested as
inhibitors were used at 0.07 M. The resultsshown in Fig. 2 are
means of three independent determina-tions with standard deviations
not higher than 5% of themean value. Six of eight heterologous sera
had their reac-tions inhibited by 50 to 100% in the presence of
galactose,but the same pattern of inhibition was obtained with
either a-or ,-methylgalactopyranoside. Other sugars were
generallypoor inhibitors. In two cases, inhibition was observed
withmannose and arabinose as well as with galactose.Another
exception was one HP serum whose reactivity
was strongly inhibited by EH38 and by periodate-treatedgp43 or
EH38 but only weakly reduced by 0.2 M galactose.In IPP, the
reaction of this serum was also poorly inhibitedby
monosaccharides.
Generally, the ELISA reactions of the PCM sera werelittle
affected by addition of monosaccharides. The reactionsof two sera
with periodate-treated EH38, which were re-duced by 35 to 45% in
comparison with that of the untreatedEH38 control, were not
inhibited by any of the carbohy-drates tested.To explain the
differences in the reactivities of heterolo-
gous sera with gp43 either in solution or immobilized inplastic,
we carried out inhibitions of ELISA reactions withsoluble gp43 at
two concentrations (0.1 and 0.5 ,uM). Thereactivities of two HP
sera were not significantly affected byaddition of highly purified
gp43 at 0.1 ,uM, whereas the sameantigen solution strongly
decreased the reactions of PCMsera (50 to 75% inhibition).
Inhibition of the reactions ofPCM and HP sera with gp43 at 0.5 1xM
was 70 to 95% and 30to 40%, respectively.
Importance of carbohydrate epitopes in IPP reactions. Asalready
mentioned, the serum IPP titration curves with125I-gp43 (Fig. 1A)
were the same when 125I-EH38 wastested in parallel. In a second set
of experiments, we studiedthe reactions of 7 PCM sera at a single
dilution with EH38,periodate-treated EH38, and gp43 in the presence
ofmonosaccharides. The pattern of the labeled antigens andtheir
deglycosylated derivatives used in IPP can be seen inFig. 3.
'25I-labeled preparations of gp43 contained smallamounts of a
high-MW antigenic complex described previ-ously (17). Removal of
the periodate-sensitive high-MWcomplex from gp43 by gel filtration
did not significantlyinterfere with the reactivity of the PCM sera
with gp43 butdid clearly contribute to the responses of HP sera in
IPPreactions. Periodate-treated purified 1251-gp43 could not be
tested in these studies because after dialysis its recovery
wasmuch reduced (Fig. 3, lane 3). In contrast, the concentrationof
periodate-treated 1251-labeled EH38 (Fig. 3, lane 6) wasnot
significantly affected by dialysis.The reactivities of the PCM sera
in IPP reactions were
basically the same when gp43, EH38, or periodate-treatedEH38 was
used as the antigen. Monosaccharides at 0.2 Malso failed to inhibit
IPP of gp43 with PCM sera. In one case,a 25% inhibition was
obtained with a-methylmannopyrano-side. The IPP reaction of this
serum was weakly (14%)inhibited by EH38, whereas a 30% inhibition
was observedin ELISA with periodate-treated gp43. The IPP
reactivitiesof seven PCM sera decreased by 35% after gp43
wasdenatured with 0.5% SDS. They were further reduced tobasal
levels (those obtained with normal human sera) whenthe antigen was
deglycosylated with N-glycanase (notshown), suggesting that
conformation alterations leading toinactivation of the antigen
probably took place.Although the IPP reactions of HP and JL sera
were
quantitatively not comparable to those with PCM sera, westudied
a few HP sera which reacted weakly with gp43 todetermine their
reactivities with periodate-treated EH38 andalso with gp43 in the
presence of monosaccharides at 0.1 and0.2 M. When the results were
quantitatively significant interms of counts per minute in the
immunoprecipitates,antigens in the immunocomplexes were checked by
SDS-
1 2 3 4 5 6 7 8
FIG. 3. Autoradiogram of iodinated antigens prepared for
IPP.Lanes: 1, native gp43; 2, control for gp43 in the periodate
reaction;3, periodate-treated gp43; 4, EH38; 5, control for EH38 in
theperiodate reaction; 6, periodate-treated EH38; 7, NG38; 8,
controlfor NG38 (see Materials and Methods) in the N-glycanase
reaction.The arrow indicates a molecular mass of 43 kDa.
61 - 100 -
z0I-
31 - 60-
zz
* 30- *0*00*000000000
VOL. 29, 1991
on June 1, 2021 by guesthttp://jcm
.asm.org/
Dow
nloaded from
http://jcm.asm.org/
-
1614 PUCCIA AND TRAVASSOS
W.' AN
l-
A a B C c Dd E e Ff
FIG. 4. Autoradiogram showing the gp43 IPP by four differentHP
sera at a dilution of 3001 (lanes C through F and c through f)
inthe presence (lowercase letters) or absence (uppercase letters)
of 0.2M D-galactose. The serum in lane A (dilution, 6,000-1) is
from aPCM patient; that in lane B (dilution, 300-') is from a
healthyindividual. A total of 70,000 cpm of iodinated antigen was
used ineach reaction. The arrowhead indicates the position of
gp43.
PAGE to determine the relative contribution of the
high-MWcomponent. In these experiments, the partially
purifiedrather than the highly purified (by gel filtration) gp43
wasused.By measuring the radioactivities of the immunoprecipi-
tates (means ± standard deviations, with standard devia-tions
< 5% of the mean) from three independent determina-tions for
each serum, we verified that galactose was the bestinhibitor of
most reactions with heterologous sera, causing a50% decrease in the
total counts per minute of the precipi-tate. Inhibition studies
with either a- or P-methylgalactopy-ranoside had similar results.
With some sera (Fig. 4, serumin lane F), the radioactivity
precipitated was entirely due tothe high-MW component, and this
reaction probably in-volved epitopes containing galactose. Other
sera (Fig. 4,lanes C, D, and E) immunoprecipitated both gp43 and
thehigh-MW component. With the serum in Fig. 4, lane D,galactose
inhibited only the reaction with gp43.The reactivities of various
sera with periodate-treated
EH38 are shown in Fig. 5. The heterologous sera reactedonly with
the high-MW component present in the EH38preparation, but not with
EH38 itself. This reaction wastotally abolished by treatment with
periodate. The PCMserum tested reacted strongly with EH38. Only a
weakreaction observed with the high-MW component was inhib-ited by
periodate treatment.
Sip:.., 04
EH38-
A a 8 C c D d E e
FIG. 5. Autoradiogram of IPP of periodate-treated EH38
(low-ercase letters) and its control (uppercase letters) by HP
sera. Threerepresentative HP sera (lanes C through E and c through
e) areshown. The serum in lane A (dilution, 2,000-') is from a
patient withPCM; that in lane B (dilution, 300-') is from a healthy
individual. Atotal of 35,000 cpm of iodinated antigen was used in
each reaction.
DISCUSSION
To determine the role of carbohydrate epitopes in
theantigenicity of gp43, we tested a number of sera frompatients
with PCM, HP, or JL against native purified ordeglycosylated
antigens from P. brasiliensis yeast phase byELISA and IPP. Since
the gp43 preparations obtained fromaffinity columns could still
contain small amounts ofhigh-MW and 72-kDa antigenic components,
our first con-cern was to further purify the gp43 by Sephacryl
S-200filtration to determine whether the strong reactivity of
somesera in ELISA was due to those other exoantigens. Ourresults
show that PCM sera recognize primarily gp43 peptideepitopes that
are independent of carbohydrate epitopes.With a few sera, however,
the carbohydrate epitopes ac-counted for up to 45% of the total
reactivity in ELISA and15% of that in IPP. By competition studies
with monosac-charides, we found that a-methylmannopyranoside was
themain inhibitor of these reactions.Our results are in good
agreement with those of Taba et al.
(22), who cloned in bacteria a fungal gene(s) coding forepitopes
present in the gp43 molecule. Gene expression wasdetected by
Western blotting and IPP with rabbit hyperim-mune anti-P.
brasiliensis antiserum as well as with a pool of10 human PCM sera
with high titers by ID. Since cloning andexpression of the P.
brasiliensis gene(s) were carried out inEscherichia coli, the
positive reactivity with human serainvolved only peptide
epitopes.HP and JL sera, on the other hand, cross-reacted in
ELISA mainly with galactose-containing carbohydrateepitopes
present in N-linked carbohydrate chains of gp43.When tested by IPP,
the HP and JL sera reacted very wellwith the high-MW component,
binding in all cases to perio-date-sensitive epitopes. The nature
of the specific residuesinvolved, however, is unclear, since not
all sera reactivewith this glycoconjugate were inhibited by
galactose.The expression of carbohydrate epitopes recognizable
by
heterologous antibodies depends on whether the gp43 isbound to
the ELISA plates or is in solution. The strong HPand JL serum
reactions observed with the plastic-immobi-lized antigen are not
obtained when the soluble antigen isused in reactions such as IPP,
inhibition-of-gp43 radioim-mune assay (21), and ID (2). Moreover,
soluble gp43 signif-icantly inhibits ELISA reactions with PCM sera,
whereasthe reactions with HP sera remain strong even in thepresence
of excess soluble gp43. Obviously, gp43 assumesdifferent
conformations when bound to the plastic or insolution.
Immunological tests with the antigen in differentconformations,
such as ELISA and immunoblotting, shouldbe analyzed with caution.
As has been shown previously (5,9, 11, 15), globular protein and
peptide antigens change theirconformations when immobilized in
ELISA plates.The reactions with PCM sera decreased to the level of
the
negative control when gp43 deglycosylated with N-glyca-nase was
used as the antigen. Although this result may seemcontradictory
when compared with those involving period-ate and endo H
treatments, it is probable that in the case ofthe N-glycanase
treatment a conformational change of theantigen occurs as a result
of both the denaturation by SDSand the removal of carbohydrate
chains by cleavage of theasparagine-N-acetylglucosamine bond (23).
Such cleavage,which, as suggested elsewhere (7), causes a
conformationalchange, is not observed in the deglycosylation
reaction withendo H, which cleaves carbohydrate chains only
betweentwo internal residues of N-acetylglucosamine (24).
In conclusion, the present work shows that PCM sera
J. CLIN. MICROBIOL.
on June 1, 2021 by guesthttp://jcm
.asm.org/
Dow
nloaded from
http://jcm.asm.org/
-
43-kDa ANTIGEN OF P. BRASILIENSIS 1615
recognize predominantly conformational peptide epitopes ofthe
purified gp43 antigen, in contrast to sera from patientswith HP or
JL, which react preferentially with the galactose-containing
carbohydrate epitopes of this antigen.The identification of the
specific peptide epitopes recog-
nized by the PCM sera is a logical complement to the
presentwork. Recently, we have obtained mouse monoclonal
anti-bodies specific for peptide epitopes of gp43
(unpublishedresults). These monoclonal antibodies can be used in
com-petition experiments with the human antibodies and asreagents
to help determine the structure of the gp43 antigenand, eventually,
its sequence.
Practically, the intact gp43 molecule present in concen-trated
culture filtrates of P. brasiliensis will continue to belargely
used for the specific ID serodiagnosis of PCM. If theELISA method
is introduced, the loss of specificity of thegp43 antigen can be
minimized only by adding galactose tothe reaction mixture or by
using a deglycosylated form of theantigen. However, since the
deglycosylation procedure isexpensive, it is not a practical method
for routine testing.Perhaps future efforts should be directed to
obtaining bacte-rial recombinant clones with high expression of the
ungly-cosylated antigen or to sequencing the protein core with
theaim of producing a synthetic peptide. An attempt to clonegenes
encoding epitopes of the gp43 antigen has already beensuccessful in
our laboratory (22).
ACKNOWLEDGMENTS
We are extremely grateful for the generous gift of sera from Z.
P.Camargo, R. Tewari, and R. Baruzzi, as well as for the gift of
P.brasiliensis B339 culture from A. Restrepo-Moreno. We also
thankM. M. Rodrigues for helpful discussions throughout this
investiga-tion.
This work was supported by CNPq, FAPESP, and FINEP.
REFERENCES1. Camargo, Z. P., J. L. Guesdon, E. Drouhet, and L.
Improvisi.
1984. Enzyme-linked immunosorbent assay (ELISA) in
theparacoccidioidomycosis. Comparison with
counterimmunoelec-trophoresis and erythro-immunoassay.
Mycopathologia 88:31-37.
2. Camargo, Z. P., C. Unterkircher, S. P. Campoy, and L.
R.Travassos. 1988. Production of Paracoccidioides
brasiliensisexoantigens for immunodiffusion tests. J. Clin.
Microbiol. 26:2147-2151.
3. Camargo, Z. P., C. Unterkircher, and L. R. Travassos.
1989.Identification of antigenic polypeptides of
Paracoccidioidesbrasiliensis by immunoblotting. J. Med. Vet. Mycol.
27:407-412.
4. Cano, L. E., E. Brummer, D. A. Stevens, and A. Restrepo.
1986.An evaluation of the enzyme-linked immunosorbent assay(ELISA)
for quantitation of antibodies to Paracoccidioidesbrasiliensis. J.
Med. Vet. Mycol. 24:467-475.
5. Dierks, S. E., J. E. Butler, and H. B. Richerson. 1986.
Alteredrecognition of surface-adsorbed compared to
antigen-boundantibodies in the ELISA. Mol. Immunol. 23:403-411.
6. Fraker, P. J., and J. C. Speck, Jr. 1978. Protein and
cellmembrane iodinations with a sparingly soluble
chloroamine,1,3,4,6-tetrachloro-3a,6a-diphenylglycoluril. Biochem.
Bio-phys. Res. Commun. 80:849-857.
7. Hakimuddin, T. S., and 0. P. Bahl. 1987. A chemical method
forthe deglycosylation of proteins. Arch. Biochem. Biophys.
259:52-57.
8. Hamburger, J., S. Lustigman, T. K. A. Siongok, J. H. Ouma,and
A. A. F. Mahmoud. 1982. Characterization of a purifiedglycoprotein
from Schistosoma mansoni eggs: specificity, sta-bility, and the
involvement of carbohydrate and peptide moi-eties in its serologic
activity. J. Immunol. 128:1864-1869.
9. Jemmerson, R. 1987. Antigenicity and native structure of
glob-ular proteins: low frequency of peptide reactive
antibodies.Proc. Natl. Acad. Sci. USA 84:9180-9184.
10. Kaufman, L. 1972. Evaluation of serological tests of
paracoc-cidioidomycosis: preliminary report. Proceedings of the
FirstInternational Symposium on Paracoccidiodomycosis. Pan
Am.Health Organ. Sci. Publ. 254:221-223.
11. Kennel, S. J. 1982. Binding of monoclonal antibody to
proteinantigen in fluid phase or bound to solid supports. J.
Immunol.Methods 55:1-12.
12. Krotoski, W. A., and H. E. Weimer. 1966.
Peptide-associatedand antigenic changes accompanying periodic acid
oxidation ofhuman plasma orosomucoid. Arch. Biochem. Biophys.
115:337-344.
13. Laemmli, U. K. 1970. Cleavage of structural proteins during
theassembly of the head of bacteriophage T4. Nature
(London)227:680-685.
14. Mendes-Gianini, M. J. S., M. E. Camargo, C. S. Lacaz, andA.
W. Ferreira. 1984. Immunoenzymatic adsorption test forserodiagnosis
of paracoccidioidomycosis. J. Clin. Microbiol.20:103-108.
15. Muller, S., S. Plaue, M. Couppez, and M. H. V. Van
Regem-mortel. 1986. Comparison of different methods for
localizingantigenic regions in histone H2A. Mol. Immunol.
23:593-601.
16. Negroni, R., C. I. Flores, and A. M. Robles. 1976. Estudio
de lasreacciones serol6gicas cruzadas entre antigenos de
Paracoccid-ioides brasiliensis e Histoplasma capsulatum. Rev. Asoc.
Ar-gent. Microbiol. 8:68-73.
17. Puccia, R., S. Schenkman, P. A. J. Gorin, and L. R.
Travassos.1986. Exocellular components of Paracoccidioides
brasiliensis:identification of a specific antigen. Infect. Immun.
53:199-206.
18. Puccia, R., and L. R. Travassos. 1988. Arq. Biol. Tecnol.
31:106.19. Restrepo, A., and L. H. Moncada. 1974. Characterization
of the
precipitin bands detected in the immunodiffusion test for
para-coccidioidomycosis. Appl. Microbiol. 28:138-144.
20. Restrepo, A., M. Robledo, F. Gutierrez, M. Sanclemente,
E.Castaneda, and G. Calle. 1970. Paracoccidioidomycosis
(SouthAmerican blastomycosis). A study of 39 cases observed
inMedellin, Colombia. Am. J. Trop. Med. Hyg. 19:68-76.
21. Stambuk, B. U., R. Puccia, M. L. C. de Almeida, L.
R.Travassos, and S. Schenkman. 1988. Secretion of the 43
kDaglycoprotein antigen by Paracoccidioides brasiliensis. J.
Med.Vet. Mycol. 26:367-373.
22. Taba, M. R. M., J. F. da Silveira, L. R. Travassos, and
S.Schenkman. 1989. Expression in Escherichia coli of a genecoding
for epitopes of a diagnostic antigen of
Paracoccidioidesbrasiliensis. Exp. Mycol. 13:223-230.
23. Tarentino, A. L., C. M. G6mez, and T. H. Plummer, Jr.
1985.Deglycosylation of asparagine-linked glycans by
peptide:N-glycosidase F. Biochemistry 24:4665-4671.
24. Tarentino, A. L., and F. Maley. 1974. Purification and
propertiesof an endo-p-N-acetylglucosaminidase from Streptomyces
gri-seus. J. Biol. Chem. 249:811-817.
25. Unterkircher, C. S. 1988. Ph.D. thesis. Escola Paulista
deMedicina, Sao Paulo, Brazil.
26. Woodward, M. P., W. W. Young, Jr., and R. A. Bloodgood.1985.
Detection of monoclonal antibodies specific for carbohy-drate
epitopes using periodate oxidation. J. Immunol.
Methods78:143-153.
27. Yarzabal, L. A., D. Bout, F. Naquira, J. Fruit, and S.
Andrieu.1977. Identification and purification of the specific
antigen ofParacoccidioides brasiliensis responsible for
immunoelectro-phoretic band E. Sabouraudia 15:79-85.
VOL. 29, 1991
on June 1, 2021 by guesthttp://jcm
.asm.org/
Dow
nloaded from
http://jcm.asm.org/