The Mycobacterium tuberculosis Complex-Restricted Gene cfp32 Encodes an Expressed Protein That Is Detectable in Tuberculosis Patients and Is Positively Correlated with Pulmonary Interleukin-10

Abstract

Human tuberculosis (TB) is caused by the bacillus Mycobacterium tuberculosis, a subspecies of the M. tuberculosis complex (MTC) of mycobacteria. Postgenomic dissection of the M. tuberculosis proteome is ongoing and critical to furthering our understanding of factors mediating M. tuberculosis pathobiology. Towards this end, a 32-kDa putative glyoxalase in the culture filtrate (CF) of growing M. tuberculosis (originally annotated as Rv0577 and hereafter designated CFP32) was identified, cloned, and characterized. The cfp32 gene is MTC restricted, and the gene product is expressed ex vivo as determined by the respective Southern and Western blot testing of an assortment of mycobacteria. Moreover, the cfp32 gene sequence is conserved within the MTC, as no polymorphisms were found in the tested cfp32 PCR products upon sequence analysis. Western blotting of M. tuberculosis subcellular fractions localized CFP32 predominantly to the CF and cytosolic compartments. Data to support the in vivo expression of CFP32 were provided by the serum recognition of recombinant CFP32 in 32% of TB patients by enzyme-linked immunosorbent assay (ELISA) as well as the direct detection of CFP32 by ELISA in the induced sputum samples from 56% of pulmonary TB patients. Of greatest interest was the observation that, per sample, sputum CFP32 levels (a potential indicator of increasing bacterial burden) correlated with levels of expression in sputum of interleukin-10 (an immunosuppressive cytokine and a putative contributing factor to disease progression) but not levels of gamma interferon (a key cytokine in the protective immune response in TB), as measured by ELISA. Combined, these data suggest that CFP32 serves a necessary biological function(s) in tubercle bacilli and may contribute to the M. tuberculosis pathogenic mechanism. Overall, CFP32 is an attractive target for drug and vaccine design as well as new diagnostic strategies.

Full text

INFECTION AND IMMUNITY, Dec. 2003, p. 6871–6883 Vol. 71, No. 12
0019-9567/03/$08.00⫹0 DOI: 10.1128/IAI.71.12.6871–6883.2003
Copyright © 2003, American Society for Microbiology. All Rights Reserved.
The Mycobacterium tuberculosis Complex-Restricted Gene cfp32
Encodes an Expressed Protein That Is Detectable in Tuberculosis
Patients and Is Positively Correlated with Pulmonary Interleukin-10
Richard C. Huard,
1,2
Sadhana Chitale,
1
Mary Leung,
1
Luiz Claudio Oliveira Lazzarini,
1
Hongxia Zhu,
1
Elena Shashkina,
3
Suman Laal,
4
Marcus B. Conde,
5
Afraˆnio L. Kritski,
5
John T. Belisle,
6
Barry N. Kreiswirth,
3
Jose´ Roberto Lapa e Silva,
5
and John L. Ho
1
*
Division of International Medicine and Infectious Diseases, Department of Medicine, Joan and Sanford I. Weill Medical College,
1
and Graduate School of Medical Sciences,
2
Cornell University, and Department of Pathology, New York University School of
Medicine, and Research Center for AIDS and HIV Infection, Veterans Affairs Medical Center,
4
New York, New York;
New Jersey Medical School, National Tuberculosis Center, University of Medicine and Dentistry of New Jersey,
Newark, New Jersey
3
; Instituto de Doenc¸asdoTo´rax, Hospital Universita´rio Clementino Fraga Filho,
Universidade Federal do Rio de Janeiro, Rio de Janeiro, Brazil
5
; and Mycobacteria Research
Laboratories, Department of Microbiology, Immunology, and Pathology,
Colorado State University, Fort Collins, Colorado
6
Received 7 May 2003/Returned for modification 17 June 2003/Accepted 9 September 2003
Human tuberculosis (TB) is caused by the bacillus Mycobacterium tuberculosis, a subspecies of the M.
tuberculosis complex (MTC) of mycobacteria. Postgenomic dissection of the M.tuberculosis proteome is ongoing
and critical to furthering our understanding of factors mediating M.tuberculosis pathobiology. Towards this
end, a 32-kDa putative glyoxalase in the culture filtrate (CF) of growing M.tuberculosis (originally annotated
as Rv0577 and hereafter designated CFP32) was identified, cloned, and characterized. The cfp32 gene is MTC
restricted, and the gene product is expressed ex vivo as determined by the respective Southern and Western blot
testing of an assortment of mycobacteria. Moreover, the cfp32 gene sequence is conserved within the MTC, as
no polymorphisms were found in the tested cfp32 PCR products upon sequence analysis. Western blotting of
M.tuberculosis subcellular fractions localized CFP32 predominantly to the CF and cytosolic compartments.
Data to support the in vivo expression of CFP32 were provided by the serum recognition of recombinant CFP32
in 32% of TB patients by enzyme-linked immunosorbent assay (ELISA) as well as the direct detection of CFP32
by ELISA in the induced sputum samples from 56% of pulmonary TB patients. Of greatest interest was the
observation that, per sample, sputum CFP32 levels (a potential indicator of increasing bacterial burden)
correlated with levels of expression in sputum of interleukin-10 (an immunosuppressive cytokine and a
putative contributing factor to disease progression) but not levels of gamma interferon (a key cytokine in the
protective immune response in TB), as measured by ELISA. Combined, these data suggest that CFP32 serves
a necessary biological function(s) in tubercle bacilli and may contribute to the M.tuberculosis pathogenic
mechanism. Overall, CFP32 is an attractive target for drug and vaccine design as well as new diagnostic
strategies.
The Mycobacterium tuberculosis complex (MTC) is a group
of highly related pathogenic mycobacteria that include M.tu-
berculosis,Mycobacterium africanum (subtypes I and II), My-
cobacterium bovis (along with the attenuated M.bovis bacillus
Calmette-Gue´rin [BCG] vaccine strain), Mycobacterium bovis
subsp. caprae, and Mycobacterium microti (13). The MTC taxon
is extraordinary in that its members exhibit a restricted number
of fixed single-nucleotide polymorphisms between subspecies
but differ from one another by the presence or absence of large
chromosomal deletion loci, severity of disease, and mammalian
host spectra (13, 43, 66). Of the MTC members, M.tuberculosis
is the predominant etiologic agent of human tuberculosis (TB).
M.tuberculosis is arguably the most successful of human
pathogens in having achieved a worldwide penetrance of epi-
demic proportions. Estimates based on skin testing indicate
that approximately one-third of the human population have
been M.tuberculosis infected (21, 74). In most individuals the
infection progresses to a latent phase in which there are no
overt signs of disease. However, up to 10% of these persons are
expected to develop life-threatening disease over the course of
their lifetimes if untreated (21). In fact, TB claims up to 3
million lives each year, which is more than any other single
bacterial infectious agent (74). Coupled with the emergence of
drug-resistant stains and a deadly cooperation with the human
immunodeficiency virus (HIV) pandemic, the incidence of TB
cases worldwide continues to rise (M. Freire and G. Roscigno,
Editorial, Bull. W. H. O. 80:429, 2002). Therefore, research
efforts to characterize the unique biology of the tubercle ba-
cillus, to develop new pharmacological TB interventions, and
to formulate new TB vaccine strategies are of paramount im-
portance in order to eliminate this global killer.
M.tuberculosis is remarkable in that it appears to be exquis-
* Corresponding author. Mailing address: Cornell University, Joan
and Sanford I. Weill Medical College, Department of Medicine, Di-
vision of International Medicine and Infectious Diseases, Room
A-421, 525 East 68th St., New York, NY 10021. Phone: (212) 746-6316.
Fax: (212) 746-8675. E-mail: jlho@med.cornell.edu.
6871

itely adapted for human parasitization and host immune sys-
tem evasion. Following inhalation of aerosolized organisms, M.
tuberculosis sets up residence and propagates within the gen-
erally hostile environment of the alveolar macrophage. It
avoids sterilization by the subsequent adaptive immune re-
sponse that is mounted against it, and it finds sanctuary within
the inflammatory response-derived granulomas meant to con-
tain it. When immunity wanes, after years to decades of per-
sistence, M.tuberculosis reactivates and exploits inflammation-
mediated lung tissue destruction to enable its transmission to
new persons (26). At present, the correlates of protection from
active TB and the molecular mechanisms of infection and
pathogenesis that account for the success of M.tuberculosis
remain largely unknown but are likely to incorporate a com-
plex interplay of multiple host and pathogen factors.
A key component of protective immunity to active TB is the
timely and orchestrated production of proinflammatory cyto-
kines such as tumor necrosis factor alpha, interleukin-12 (IL-
12), IL-1␤, and gamma interferon (IFN-␥) (16). In order to
prevent overzealous proinflammatory responses and to protect
against undue immune-mediated damage, counteractive im-
munosuppressive cytokines are also secreted as part of a bal-
anced immune response and include transforming growth fac-
tor ␤and IL-10 (49). However, the premature or
disproportionate secretion of inhibitory cytokines may unde-
sirably benefit the pathogen, as elevated IL-10 levels have been
associated with poor resolution of infections by HIV, human
rhinovirus, Leishmania spp., and Mycobacterium leprae (36, 61,
67, 68). Recent studies have suggested that this may also be the
case in M.tuberculosis infection (10, 11, 25, 46, 49).
A major advance for TB research came in 1998 with the
publication of the complete genome sequence of the M.tuber-
culosis H37Rv laboratory strain (15). Of the approximately
4,000 open reading frames identified, close to 48% have not
been assigned a function, nor have most been proven to code
for expressed proteins (14). The recent advent of improved
molecular tools for mycobacteria has allowed the systematic
study of the M.tuberculosis genomic blueprint in order to
identify genes of importance and to characterize their products
(50, 60). Given the astounding success of M.tuberculosis,itis
reasonable to anticipate that M.tuberculosis genes devoted to
defense against host mycobacteriocidal immune mechanisms,
or genes that promote disturbances in effective immune func-
tion, will be found. In fact, M.tuberculosis genes implicated in
persistence, resistance to oxidative stress, and immune activa-
tion have been identified (18, 28, 30, 42, 57). Several of these
putative virulence factors are secreted or released by growing
M.tuberculosis into the culture filtrate (CF) compartment and
are thereby strategically positioned as molecular effectors to
the detriment of the host and/or for the benefit of the pathogen
(28, 52, 64). A coincident characteristic of many individual CF
proteins (as well as the CF as a whole) is their strong immu-
nostimulatory capacity. This feature may be important in the
M.tuberculosis life cycle strategy but may also contribute to
immune control of infection. Many studies have illustrated the
presence of specific antisera as well as the development of
specific Th1-like responses (lymphoproliferation and/or IFN-␥
secretion) and cytotoxic T-cell activity to CF proteins in TB
patients and/or immunized animals (8, 9, 18, 31, 62). Indeed,
the production of CF proteins is believed to account for the
heightened efficacy of live, as opposed to killed, M.tuberculosis
vaccines in animal models (3, 31). Containing in the range of
200 to 800 different proteins (many of which remain unidenti-
fied and whose functions are uncharacterized) (35, 53, 64), the
CF presents an abundance of candidates for drug intervention,
for incorporation into a TB vaccine, or to serve as TB diag-
nostic markers. Further systematic dissection and characteriza-
tion of the constituents of CF by the TB scientific community
will undoubtedly uncover useful information about the unique
biology of M.tuberculosis and will provide fundamental knowl-
edge of the immunological parameters associated with protec-
tive immunity against M.tuberculosis in humans.
In this study we detail the identification, cloning, and char-
acterization of a 32-kDa CF protein that we have designated
CFP32 (originally known as Rv0577). Comparative analyses
suggest that the cfp32 gene product may be important to the
biology of the MTC subspecies. Moreover, patient data suggest
that CFP32 is expressed in M.tuberculosis-infected individuals
and may be useful as a diagnostic, drug, and/or vaccine target.
Surprisingly, levels of CFP32 in TB patient lung sputum were
positively correlated with levels of IL-10 but not of IFN-␥in
the same sputum sample, thereby suggesting that a link be-
tween M.tuberculosis and IL-10 may play a role in the patho-
genic mechanism leading to active TB.
(This study contributed to the fulfillment of the Ph.D. re-
quirements of R.C.H.)
MATERIALS AND METHODS
PCR and Southern blotting for CFP32. Primer pairs suited to evaluate the
cfp32 (Rv0577) locus were created using the DNASTAR program (DNASTAR,
Inc., Madison, Wis.) and GenBank sequence database information (http://www
.ncbi.nlm.nih.gov). These primers amplified either the upstream region of cfp32
(577proF, 5⬘-GTG GCT TGG CGG GCA CGG TGG AG-3⬘; 577proR, 5⬘-TTT
TGG CGG CGG ACT GAT CGG TGG TCT-3⬘), the full coding region of cfp32
(Rv0577F, 5⬘-ATG CCC AAG AGA AGC GAA TAC AGG-3⬘[F]; Rv0577R,
5⬘-CTA TTG CTG CGG TGC GGG CTT CAA-3⬘[R1]), or the extended full
coding region of cfp32 (577pMS3F, 5⬘-CCC TTA ATT AAT GTC CGC CAC
CTA ACG AAA G-3⬘; 577pMS3R, 5⬘-CCC AAG CTT CTA GCA TTC TCC
GAA-3⬘[R2]). Each PCR mixture was prepared, each reaction was run using
PCR program 1 (with an initial denaturation step of 5 min at 94°C followed by
25 cycles of 1 min at 94°C, 1 min at 60°C, and 1 min at 72°C and ending with a
final elongation step for 10 min at 72°C), and results were analyzed as previously
described (33). Likewise, direct sequencing of PCR fragments was performed
and results were analyzed as recently described (33). PCR amplicons were
sequenced using their respective amplification primers, and a minimal single
overlap from two directions for each was usually achieved. Additional sequenc-
ing primers internal to cfp32 were also used (⫺605F, 5⬘-CGA ATC ATT GGC
ACG TCT ACT TTG-3⬘;⫺281R, 5⬘-ACC ACC TTG TCC ACC ACC GCA
T-3⬘). Southern blot analysis for cfp32 was done as previously described (37) and,
as the hybridization probe, used M.tuberculosis H37Rv cfp32 PCR products
generated using the Rv0577F and Rv0577R primer pair.
PAGE followed by gel staining or Western blotting for CFP32. All polyacryl-
amide gel electrophoresis (PAGE) and Western blot assays were performed as
follows. NuPage 12% Bis-Tris 10-well gels (Invitrogen, Carlsbad, Calif.) under-
went PAGE and transfer using the Xcell II apparatus (Novex, San Diego, Calif.),
per the manufacturers’instructions. In some experiments select samples were not
preboiled or mixed with reducing agent (1 ␮l of 1 M dithiothreitol) prior to gel
loading, as indicated. Full-range rainbow (Amersham, Piscataway, N.J.),
midrange (Promega, Madison, Wis.), or kaleidoscope prestained (Bio-Rad, Her-
cules, Calif.) molecular weight protein markers were used as standards. For
antibody detection of CFP32, nitrocellulose membranes were first blocked with
3% bovine serum albumin in 1⫻TBSt (Tris-buffered saline with 0.1% Tween 20)
for 1 h following transfer. Afterwards, the membranes were probed with a CFP32
antiserum for 1 h and then washed three times with TBSt. The membranes were
then probed with either anti-rabbit immunoglobulin (Ig)-horseradish peroxidase
(HRP)-linked whole antibody (Amersham; used when anti-recombinant CFP32
6872 HUARD ET AL. INFECT.IMMUN.

[anti-rCFP32], anti-PepC, or anti-Pep7 was the primary antiserum) or anti-
mouse Ig-HRP (Amersham) (used when IT-44 was the primary antibody),
washed three times with TBSt, developed using ECL Western blot detection
reagents (Amersham), and then exposed to Kodak BioMax film. Mycobacterial
lysates were generated in a mini-BeadBeater (Biospec Products Inc., Bartlesville,
Okla.), whereby growing cultures were spun down, the supernatant was removed,
the pellet was resuspended with Tris-EDTA buffer, and six 3-mm-diameter glass
beads were added to lyse the bacteria in five 30-s pulses. These lysates were
subsequently heated at 80°C for 30 min and then gamma-irradiated. A total of 37
MTC strains and 29 mycobacteria other than the MTC (MOTT) isolates were
evaluated for CFP32 by Western blotting including 8 strains of M.tuberculosis,7
strains of M.bovis, 3 strains of M.bovis BCG, 8 strains of M.microti, 6 strains of
M.africanum subtype I, 4 strains of M.africanum subtype II (Uganda), 1 strain
of M.bovis subsp. caprae, 2 strains of Mycobacterium smegmatis, 8 strains of
Mycobacterium avium subsp. avium, 2 strains of Mycobacterium avium subsp.
intracellulare, 1 isolate of M.leprae, 1 strain of Mycobacterium marinum, 1 strain
of Mycobacterium xenopi, 2 strains of Mycobacterium chelonae, 2 strains of My-
cobacterium gordonae, 4 strains of Mycobacterium abscessus, and 6 strains of
Mycobacterium fortuitum. For each strain, MTC subspecies identity was con-
firmed by a recently developed MTC PCR typing protocol (33) and MOTT
species identity was confirmed by 16S rRNA sequencing, also as described
previously (33). M.leprae lysate was kindly provided by P. Brennan as part of the
Colorado State University (CSU) NIH NIAID Leprosy Research Support Con-
tract (http://www.cvmbs.colostate.edu/mip/leprosy). Lysates of pelleted
pQE31.577-transformed IPTG (isopropyl-␤-D-thiogalactopyranoside)-induced
Escherichia coli were prepared by a method of multiple freeze-thaws with inter-
mittent water bath sonication in native condition lysis buffer (50 mM NaH
2
PO
4
[pH 8.0]; 300 mM NaCl; 1 mM phenylmethylsulfonyl fluoride; 1 ␮g of lysozyme/
ml; and 5 ␮g each of aprotinin, chymostatin, leupeptin, and pepstatin/ml) (Sigma,
St. Louis, Mo.). The protein content of mycobacterial and E.coli lysates was
quantified using the Bio-Rad protein assay and an Ultraspec 2100 Pro spectro-
photometer (Amersham Pharmacia Biotech, Cambridge, United Kingdom). All
M.tuberculosis subcellular components and CF fractions were generated at CSU
as part of the NIH NIAID TB Research Materials and Vaccine Testing Contract
(http://www.cvmbs.colostate.edu/microbiology/tb/top.htm). Stocks of the murine
IT-44 monoclonal antibody (MAb) (39) are also distributed through CSU. Silver
staining (Invitrogen) and 1% Coomassie blue staining of polyacrylamide gels
followed standard protocols. Internal sequencing of a protein band cut from a
silver-stained gel that was identified as CFP32 was done by the Rockefeller
University Protein/DNA Technology Center (23). Computer analysis of CFP32
and its homologues employed the GenBank and SwissProt (http://us.expasy.org/
sprot) websites. Basic summary information on CFP32 can be found in GenBank
(given as Rv0577) as well as the TubercuList website (http://genolist.pasteur.fr/
TubercuList/index.html) (given as TB27.3).
CFP32 cloning, expression, and purification. Acfp32 PCR fragment repre-
senting the entire open reading frame was generated with PCR program 1 from
purified M.tuberculosis H37Rv DNA by using primers that were engineered to
introduce BamHI and HindIII restriction enzyme sites into the resulting PCR
product (SMC-1, 5⬘-GAA AGG ATG AGG ATC CCC AAG AGA AGC G-3⬘,
and SMC-2, 5⬘-CGG GAT GCT CAA GCT TGC TGC GGT GC-3⬘). By stan-
dard procedures, the amplified product was restriction digested, ligated into the
pQE31 vector (Qiagen, Valencia, Calif.) to create the pQE31.577 plasmid, and
introduced into M15 E.coli, and the sequence was confirmed. The production of
N-terminal hexahistidine (His)-tagged rCFP32 followed the methodology de-
scribed in the QiaExpressionist handbook (Qiagen) but was further optimized by
growing the bacteria in Terrific Broth (Sigma) and inducing the pQE31.577
transformants with 0.5 mM IPTG (Sigma) for 4.5 h and shaking at 30°C. The
predicted amino acid sequence of rCFP32 is RGS-6⫻H-TD-(CFP32)-A. His-
tagged rCFP32 was purified by nickel affinity chromatography, using nickel-
nitrilotriacetic acid spin columns (Qiagen), under native conditions, per the
manufacturer’s protocol. A single difference was that nickel-nitrilotriacetic acid-
bound rCFP32 was washed three times using buffer containing 1 mM imidazole
prior to elution. The rCFP32 was then washed free of the imidazole and con-
centrated using Centriplus centrifugal filter devices (30-kDa cutoff) (Millipore,
Bedford, Mass.). PAGE, followed by 1% Coomassie blue staining and/or silver
staining, was done to qualify the purity of the preparation. A standard Bio-Rad
protein assay was done to quantify yield. The identity of the recombinant protein
was verified by electrospray tandem mass spectrometry of rCFP32 digested with
trypsin and interrogation of the mass spectrometry data against the M.tubercu-
losis genome by using Sequest software (22). Rabbit antisera were generated by
a commercial provider (Covance Research Products, Denver, Pa.). Candidate
rabbits for immunization with rCFP32, or CFP32-derived synthetic peptides,
were prescreened for serum reactivity to M.tuberculosis whole-cell lysate by
Western blotting, and only rabbits with low to absent reactivity were chosen. The
Pep7 immunogen was generated by a commercial provider (Sigma Genosys,
Houston, Tex.) while PepC was kindly provided by Shibo Jiang, New York Blood
Center. These synthetic peptides were covalently linked to keyhole limpet he-
mocyanin prior to injection.
Enzyme-linked immunosorbent assay (ELISA) detection of CFP32 and hu-
man antibody to CFP32. For the detection of human anti-CFP32 serum spec-
ificity, two different ELISAs were fashioned. In the first (Cornell laboratory),
the IT-44 murine MAb (1:10
5
in phosphate-buffered saline [PBS], 50 ␮l per
well) was used to coat a 96-well ELISA plate (Corning International, Corn-
ing, N.Y.) and was incubated overnight at 4°C. PBSt (PBS with 0.1% Tween
20) was used to wash the plate four times followed by 2.5 h of incubation with
200 ␮l of blocking buffer (PBS with 10% fetal calf serum) and with shaking
at room temperature. Next, rCFP32 (2.5 ng/ml in blocking buffer, 100 ␮l per
well) was incubated for 2.5 h, with gentle shaking at room temperature, and
subsequently washed four times with PBSt. Duplicate samples of each test
human serum from a Brazilian cohort (1:5 ⫻10
4
in blocking buffer, 100 ␮l per
well) were then incubated for 2 h, with gentle shaking at room temperature,
and subsequently washed four times with PBSt. Biotinylated anti-human Ig
(1:10
4
in blocking buffer, 100 ␮l per well) (Amersham) was then input and
incubated for 1 h, with gentle shaking at room temperature, and washed four
times with PBSt. Extravidin peroxidase conjugate (1:2 ⫻10
3
in PBS, 100 ␮l
per well) (Sigma) was then applied to the plate, shaken gently for 2 h at room
temperature, and subsequently washed four times with PBSt. 3,3⬘,5,5⬘-Tetra-
methylbenzidine (TMB) (Sigma) acted as the enzymatic substrate (100 ␮l per
well). Once the blue color had sufficiently developed, the reaction was
stopped using 0.5 M H
2
SO
4
(100 ␮l) and read at 450 nm with an EL 340
Biokinetics Reader (BioTek Instruments Inc., Winooski, Vt.). The absor-
bance values for each donor sample were then averaged. The detection of
human anti-CFP32 antisera from a cohort of patients in India, by the New
York University laboratory, was performed as previously described (59) using
rCFP32 (2 ␮g/ml in PBS, 50 ␮l per well) to coat a 96-well ELISA plate and
capture the specific antibodies. The international standard of ⱖ10-mm indu-
ration following the injection of 5 TU of M.tuberculosis purified protein
derivative (PPD) was used to define a positive skin test. To get a measure of
CFP32 in the lungs of TB patients, a variation on the ELISA to detect
anti-CFP32 antisera (Cornell laboratory) was used. Patients living in Brazil
who presented at the Pulmonary Service with “lung disease suggestive of TB”
and who failed to provide a spontaneous sputum sample or for whom a
sample was negative for acid-fast bacilli (AFB) underwent sputum induction
using 3% saline in an ultrasonic nebulizer. The induced sputum remaining
from diagnostic workup was treated with dithiothreitol (Sigma) and centri-
fuged, and the supernatant was stored at ⫺80°C prior to use. For the CFP32
ELISA, 50 ␮l of sputum per well, one sample per donor, was input in place
of a single set amount of rCFP32. Duplicate twofold dilutions of rCFP32 (5
⫻10
3
to 5 ⫻10
1
pg/ml) were also used to establish a standard curve for
CFP32 at this stage. As a final difference, anti-rCFP32 (1:10
4
, with gentle
shaking at room temperature overnight) was used as the second antiserum (as
opposed to the human antisera), thereby necessitating the use of anti-rabbit
Ig-HRP and TMB substrate for detection. For these experiments, TB case
patients were defined as having a positive solid medium culture or treatment
response with resolution of clinical and radiological features of TB. Sus-
pected TB cases were defined as patients with clinical and radiological fea-
tures compatible with TB for whom cultures were negative, contaminated, or
not available and who had insufficient follow-up or had prior TB without
sufficient follow-up. Non-TB cases with other lung diseases (OLD) were
defined as those patients who were AFB smear and TB culture negative, for
whom another diagnosis was established, and/or who showed clinical im-
provement after a short course of non-TB antibiotics. The detection of CFP32
by the testing laboratory was independent of knowledge of the clinical clas-
sification of each patient. For the quantification of lung cytokine levels,
ELISAs were performed upon the same lung sputum samples as those eval-
uated for CFP32. For these experiments, an anti-IL-10 antibody pair (Pierce
Endogen, Rockford, Ill.) was used in an otherwise identical protocol as given
for the evaluation of sputum CFP32 levels. IFN-␥was measured using a
commercial kit (Immunotech, Marseille, France) and converted to picograms
per milliliter using the relationship1UofIFN-␥⫽33.33 pg of IFN-␥. The
quantification of CFP32 in M.tuberculosis subcellular compartments followed
the ELISA protocol for sputum CFP32 measurement. All serum and sputum
donors signed informed consent papers, and the study was approved by the
Internal Review Boards of Cornell University, Hospital Universita´rio Clem-
entino Fraga Filho, and New York University.
VOL. 71, 2003 M.TUBERCULOSIS PROTEIN CFP32, IL-10, AND TUBERCULOSIS 6873

RESULTS AND DISCUSSION
Identification of a novel M.tuberculosis CF protein. To iden-
tify a novel extracellular M.tuberculosis antigen of prospective
pathological and/or immunological importance, CF proteins
were first separated by anion-exchange chromatography into
analyzed pools of fractions. Of these pooled fractions, PAGE
under nondenaturing conditions followed by silver staining
revealed that CF-fraction pool 9 (fx9) contained a predomi-
nant band at approximately 24 kDa (Fig. 1A). To determine
the identity of the major CF-fx9 protein, PAGE and silver
staining were repeated for CF-fx9 alone, and the same pre-
dominant band was excised (data not shown). N-terminal se-
quence analysis of the gel fragment failed; however, internal
protein sequencing identified a peptide that was identical to
amino acids 4 to 25 of the predicted product of the M.tuber-
culosis gene annotated as Rv0577 (Fig. 1B). Rv0577 was a
hypothetical gene of unknown function that was identified
upon completion of the M.tuberculosis H37Rv genome se-
quence (15). Rv0577 also corresponds to the MT0606 locus of
M.tuberculosis strain CDC1551 (GenBank accession no.
AE000516). In being a novel CF protein the Rv0577 gene
product was of sufficient interest that we decided to pursue its
characterization. In keeping with previous convention (54, 73),
and based upon the PAGE mobility of its gene product under
denaturing conditions (see below), the Rv0577 gene is hereaf-
ter designated cfp32 and the protein that it encodes is desig-
nated CFP32. While this study was ongoing, a separate group
independently identified CFP32 by microsequencing CF spots
in silver-stained two-dimensional polyacrylamide gels (given as
TB27.3), thereby confirming the synthesis and export of CFP32
to the CF (55).
Information predicted by the sequences of cfp32 and CFP32.
The gene for CFP32 is transcribed in the forward direction
(Fig. 2A), and the start of cfp32 is preceded by an AAGGA
putative Shine-Dalgarno ribosomal binding site (RBS) (Fig.
2B). The region upstream of cfp32 also contains several puta-
tive regulatory elements that are homologous to previously
described mycobacterial ⫺10 and ⫺35 RNA polymerase con-
tact sites (n⫽3 and 5, respectively), including three ⫺35 sites
identified for the CF virulence factor katG (44; data not
shown). The predicted TAG stop codon of cfp32 is followed by
15 intervening codons and a second in-frame TAG. The bio-
logical significance of such an arrangement of two stop codons
is not known but has been noted previously for the glutamine
synthetase gene of M.tuberculosis,E.coli, and Salmonella
enterica serovar Typhimurium (27). Rv0576 is the hypothetical
FIG. 1. Identification of CFP32 (Rv0577) from fractionated M.tu-
berculosis CF. (A) Silver-stained gel of CF fractions. The CF of grow-
ing M.tuberculosis H37Rv was fractionated by anion-exchange chro-
matography (using QAE Sepharose resin and an increasing NaCl
concentration), and 15 ␮l of each fraction pool (fx), as well as 1 ␮gof
unfractionated CF (whole), was subjected to nondenaturing PAGE
(without preboiling) followed by silver staining. Molecular mass stan-
dards (Bio-Rad; values in kilodaltons) are provided in the lane labeled
marker. (B) Amino acid sequence of CFP32. The predominant band in
CF-fx9 was excised and internally sequenced to obtain a peptide (bold-
face) matching the hypothetical M.tuberculosis H37Rv gene Rv0577.
Synthetic peptides, based upon the underlined amino acid sequences,
were used to derive the anti-Pep7 (amino acids 121 to 145) and anti-
PepC (amino acids 231 to 161) rabbit antisera.
FIG. 2. Characterization of the cfp32 locus. The predicted coding
region of cfp32 is 786 bp long and located at nucleotide coordinates
671166 to 671951 (relative to the M.tuberculosis H37Rv genome se-
quence, accession no. AL123456). (A) Illustration of genes in the
vicinity of cfp32. (B) Depiction of the DNA sequences upstream and
downstream of cfp32. Shown are the putative RBS, ATG start codon,
TAG stop codon, and a second in-frame TAG stop codon, as well as a
potential stem-loop structure–transcription stop signal for cfp32.
(C) PCR evaluation of the region 3⬘of cfp32 indicates the presence of
secondary DNA structure by differences in PCR amplicon intensity.
PCR products and a 100-bp ladder (in the first lane) were visualized by
agarose gel electrophoresis and ethidium bromide staining. The F
sense primer (Rv0577F) was used in combination with either the R1
(Rv0577R; product size, 786 bp) or R2 (577pMS3R; product size, 838
bp) antisense primer. Amplification was also done in the presence (⫹)
or absence (⫺) of DMSO. An additional 2.5 ␮l of water was included
in the reaction mixtures that purposely excluded DMSO. One repre-
sentative example of four experiments is shown.
6874 HUARD ET AL. INFECT.IMMUN.

open reading frame found upstream of cfp32 (Fig. 2A). This
element could cotranscribe with cfp32, as it is the only other
local gene also predicted to be transcribed in the forward
direction. However, the presence of an RBS for cfp32 suggests
that one level of its regulation is monocistronic. Downstream
of cfp32 is the inversely transcribed PE-PGRS gene Rv0578c as
well as a putative stem-loop structure(s) that may act as the
transcriptional stop signal for both cfp32 and Rv0578c (Fig.
2B). Evidence for the presence of secondary DNA structure in
the intergenic region of cfp32 and Rv0578c was provided in
PCR amplification experiments (Fig. 2C). Herein, a single for-
ward primer (F) was used in combination with either of two
reverse primers that are complementary to sequences flanking
each side of the putative stem-loop. For further comparison,
the amplification was done in the presence or absence of di-
methyl sulfoxide (DMSO), a PCR recipe additive that helps to
open secondary structure and improve PCR efficiency. As a
result, the PCR product band intensity was significantly re-
duced using the reverse primer that was 3⬘to the putative
stem-loop (R2; Fig. 2C) compared to that with the reverse
primer that was 5⬘to the stem-loop (R1). In addition, the
amplification efficiency was further reduced in the absence of
DMSO in the case of the F-R2 but not F-R1 primer pair,
thereby indicating that amplification using R2 is inhibited by
secondary DNA structure.
In silico modeling revealed several intriguing insights into
the nature of the cfp32 gene product. Analysis for functional
regions indicated that CFP32 is a bimodular protein with two
homologous domains (N-terminal residues ⬃2 to 129 and C-
terminal residues ⬃130 to 261; 26% identity) each with struc-
tural similarity to members of the glyoxalase-dioxygenase su-
perfamily of enzymes (GenBank; Fig. 3A). CFP32 may
FIG. 3. Alignment of the bimodular CFP32 and its homologues with a divergent glyoxalase reveals conserved amino acids that may be related
to the catalytic mechanism. The CFP32 polypeptide is predicted to contain 261 amino acids, to have a molecular mass of 27.3 kDa, to have an
isoelectric point (pI) of 4.24, and to be a compact globular protein (SwissProt). (A) Cartoon to illustrate the two predicted glyoxylase domains of
each CFP32 module. (B) M.tuberculosis CFP32 was aligned with its homologues (encoded by sequences with GenBank accession numbers in
parentheses; 15 to 58% overall range of homology). The alignment results for CFP32 with two bimodular [R.equi (CorD1, CAC44898) and
Streptomyces peucetius (DnrV, AAD04716)] and two unimodular [Mesorhizobium loti (BAB53970) and Caulobacter crescentus (AAK25386)]
representative homologues are illustrated. Not shown are the additional CFP32 homologues from M.tuberculosis (Rv0911, CAB08509), Strepto-
myces spp. (SgaA, BAA14012; BAA08202; CAA15810; CAB42934; CAB45588; CAB55527; CAB92885; CAB95980; CAC08431), Corynebacterium
glutamicum (CAC26380), M.loti (BAB48973), Vibrio cholerae (AAF96246; AAF96546), C.crescentus (AAK23809), Bacillus halodurans
(BAB04023), Pseudomonas aeruginosa (AAG05061), Myxococcus xanthus (AAL56603), Agrobacterium tumefaciens (AAK86662; AAK87322;
AAL41869), Brucella melitensis (AAL54026), and Sinorhizobium meliloti (CAC47416). A contrasting glyoxalase from Arabidopsis thaliana
(BAB17665) that had low sequence identity (13%) to CFP32 is also provided for comparison. Amino acids that were highly conserved among the
set of CFP32-like proteins are shaded. Glutamic acids that substitute for conserved aspartic acids are also shaded. Asterisks above residues indicate
those that are present almost without exception. Homologous residues in the A.thaliana glyoxylase are also specified by shading. Putative aspartic
acid nucleophiles are shaded black. DnrV, dnrV gene product; CorD1, corD1 gene product; hypoth., hypothetical protein.
VOL. 71, 2003 M.TUBERCULOSIS PROTEIN CFP32, IL-10, AND TUBERCULOSIS 6875

therefore be a bifunctional enzyme and catalyze more than one
reaction. GenBank BLAST searches found that CFP32 shows
significant pairwise homology (15 to 58% identity) to many
other unimodular and bimodular polypeptides from a variety
of microorganisms (described for Fig. 3B). As with CFP32,
many of these polypeptides are of unknown function and an-
notated as “probable hydrolases”or “hypothetical proteins.”
Notably, CFP32 homologues were most plentiful in Streptomy-
ces spp., while CorD1 of Rhodococcus equi (a close phyloge-
netic relative of M.tuberculosis and an intracellular pathogen
that causes a TB-like pulmonary disease in foals and immuno-
compromised patients [47]) had the highest percent identity
(58%). Of additional note were the Streptomyces spp. dnrV and
sgaA gene product homologues of CFP32. The dnrV-encoded
protein plays a role in the synthesis of the polyketide antibiotic
doxorubicin (40), and the sgaA gene encodes a regulatory fac-
tor of growth and osmotic stress responses, as well as strepto-
mycin production and resistance (4). By analogy, CFP32 may
therefore have similar physiological activities. Remarkably, the
alignment of CFP32 homologues revealed several highly con-
served amino acids, several of which were also present in a
surprising number of ostensibly paralogous glycosyl hydrolases
(one example is given in Fig. 3B). Of these residues, the ty-
rosines and aspartic acids may be important to the catalytic
mechanism with the most significant being CFP32 (module 1)
Asp
118
and CFP32 (module 2) Asp
252
. These aspartic acids
were each in the context of a DPXG motif analogous to that
for the determined enzymatic nucleophile (the residue that
forms the enzyme-substrate intermediate during cleavage) of
the well-characterized class II (family 38) ␣-mannosidases
(32). Compare human Golgi ␣-mannosidase II (PRSGWQID
PFGHSA), jack bean ␣-mannosidase (PRAGWAIDPFGHSP),
CFP32 module 1 (GRMSFITDPTGAAV), and CFP32 module
2(GRFAVLSDPQGAIF) whereby the conserved amino acids
are in boldface and the nucleophiles (known and putative) are
underlined. Perhaps residues Asp
118
and Asp
252
serve similar
mechanistic roles for CFP32 and its homologues.
Development of antisera to CFP32 and confirmation of
CFP32 as the IT-44-reactive antigen. The cfp32 gene was
cloned from M.tuberculosis H37Rv and expressed in E.coli,
and rCFP32 was purified by nickel column affinity chromatog-
raphy. The IPTG-induced pQE31.577 transformant expressed
rCFP32 at a band size of ⬃33 kDa (Fig. 4A). This band was
absent from the uninduced transformant and was absent from
both the induced and uninduced E.coli transformed with the
naked pQE31 plasmid (data not shown). Soluble rCFP32 of
high purity, as determined by PAGE and silver staining (Fig.
4B), was readily obtained, supporting the predicted soluble
nature of CFP32. Mass spectrometry of trypsin-digested
rCFP32 derived four peptide sequences (18 to 36 amino acids
long), each of which perfectly matched separate stretches of
amino acids in the expected sequence of CFP32 (data not
shown). Rabbits were then immunized either with the purified
rCFP32 or with CFP32-based synthetic peptides. Peptide 7
(Pep7) is identical to an internal length of amino acids while
peptide C (PepC) parallels the C terminus of CFP32 (Fig. 1B).
In Western blotting, each of the three rabbit-raised antisera
(anti-rCFP32, anti-PepC, and anti-Pep7) recognized rCFP32
at a band size of 33 kDa (Fig. 4C). Importantly, the preimmu-
nization sera of these rabbits did not show any reactivity in
parallel Western blots (data not shown). The trio of anti-
CFP32 antisera also recognized a band at ⬃32 kDa from the
whole-cell lysate of M.tuberculosis H37Rv that is presumably
CFP32 (Fig. 4C). Subsequent Western blotting of the whole
CF fraction, as well as CF-fx9 (from whence CFP32 was first
identified), also showed a 32-kDa band (Fig. 4D). The enrich-
ment of CFP32 in CF-fx9 is given by its relative band strength
in Western blotting (at 10
3
-fold-less input CF-fx9 sample com-
pared to whole CF). The difference between expected (27.3
kDa) and observed (32 kDa) molecular masses of CFP32 in
PAGE under denaturing conditions has been noted previously
for other CF factors (73) and may be due to anomalous mi-
gration and/or unidentified posttranslational modifications
such as myristoylation, glycosylation, or phosphorylation.
FIG. 4. Cloning of CFP32 and the derivation of anti-CFP32 anti-
sera. For each of the following, all samples were boiled prior to being
loaded in the gel, and molecular mass protein markers (Amersham;
values in kilodaltons) are shown in the first lane. (A) Coomassie
blue-stained polyacrylamide gel of lysate from IPTG-induced
pQE31.577-transformed M15 E.coli expressing rCFP32. The rCFP32
band appears at ⬃33 kDa. A similar band was absent from parallel
IPTG-induced M15 E.coli lysates that were either wild type or trans-
formed with the pQE31 vector (data not shown). (B) Silver-stained gel
following PAGE of His-tagged purified rCFP32. Additional protein
molecular mass markers (Promega; values in kilodaltons) are shown in
the third lane. (C) Each of the three rabbit-derived antisera recognized
both CFP32 and rCFP32. Separate parallel sets of purified rCFP32 (10
ng) and M.tuberculosis (M.tb) lysate (1 ␮g) were probed with either
anti-rCFP32 (1:10
3
), anti-PepC (1:10
3
), or anti-Pep7 (1:250) antiserum
in a Western blot. (D) Antiserum raised against purified rCFP32 rec-
ognized, and was specific for, M.tuberculosis CFP32. Samples of CF (1
␮g) and CF-fx9 (1 ng) were probed with the rabbit-derived anti-
rCFP32 antiserum (1:10
3
) in a Western blot. The M.tuberculosis
CFP32 band appears at ⬃32 kDa. (E) CFP32 is the IT-44 MAb-
reactive antigen. IT-44 is a mouse-derived IgG2a MAb raised upon
mouse challenge with M.tuberculosis CF (39). Western blotting, sim-
ilar to the preceding, was done using IT-44 (1:2.5 ⫻10
4
) to probe for
CFP32.
6876 HUARD ET AL. INFECT.IMMUN.

Overall, the combined data confirm the correct cloning and
exogenous expression of CFP32 from fractionated CF and
argue for the specificity of the developed antisera.
IT-44 (also known as HBT7) is an IgG2a murine MAb that
was derived from the immunization of inbred mouse strains
with the CF of M.tuberculosis H37Rv (39). A GenBank sub-
mission of unpublished data from T. Oettinger (accession no.
AJ007737) identified the IT-44-reactive antigen as being the
gene product of cfp32 (given as the cfp30B gene). IT-44 was
also shown previously to react with three spots in two-dimen-
sional PAGE of M.tuberculosis CF (64). However, the proteins
were clustered at ⬃32 kDa and migrated within a narrow pI
range of 4.75 to 4.93, thereby suggesting that the antibody was
reacting with multiple isoforms of the same antigen. As a result
of this information, IT-44 was obtained and was evaluated for
CFP32 reactivity by Western blotting. Bands for rCFP32 and
CFP32 were seen in the same position as in the Western blots
probed with rabbit anti-rCFP32 antisera (Fig. 4E) while IT-44
Western blot reactivity could be blocked by blot preincubation
with anti-rCFP32 (data not shown), thereby verifying CFP32 as
the IT-44-reactive antigen. This finding has been indepen-
dently confirmed in CF mapping studies (53, 55). It should also
be noted that Western blot assays probing for CFP32, similar
to previous silver stain gel results (Fig. 1A), suggested that
CFP32 and rCFP32 exist in two states: the respective linearized
32- or 33-kDa form that was seen when samples were prepared
under denaturing conditions (by being heated to 100°C for 5
min in the presence of dithiothreitol) and a predominant ⬃24-
kDa form that was visible in parallel nondenatured samples
(data not shown). It was therefore thought that native CFP32
maintains a compacted hydrophodynamic volume that is un-
folded upon boiling, the likes of which were also noted previ-
ously for CFP25 (73). However, the CFP32 sequence contains
but a single cysteine residue, and so forces other than intramo-
lecular disulfide bonds must maintain the globular three-di-
mensional structure of monomeric CFP32.
Distribution of CFP32 among M.tuberculosis subcellular
compartments. To localize CFP32, M.tuberculosis H37Rv sub-
cellular compartments were evaluated for the presence of
CFP32 by Western blotting with the developed antisera (Fig.
5). On a per-microgram basis, the greatest quantity of CFP32
was found in the CF followed by a very strong CFP32 band in
the cytosolic and whole-cell lysate fractions. Small amounts
were also detected in the cell wall, soluble cell wall proteins,
and membrane fractions but not in the purified mannosylated
lipoarabinomannan. At least one additional lot of each com-
ponent was tested by Western blotting and gave a similar result
(data not shown). ELISA measurement of CFP32 levels in the
illustrated components supported the Western blot data, indi-
cating relative amounts of CFP32 in each by band intensity
(Fig. 5). These data suggest a directed movement of CFP32
from the cytosol to the CF despite the lack of a clear gener-
alized signal peptide for bacterial secretory proteins in the
CFP32 N terminus (51; data not shown). It is further notewor-
thy that the original sequencing of the CF-fx9 CFP32 band did
not indicate the occurrence of N-terminal cleavage associated
with export signal peptides. Even so, there are several other
known CF protein genes that do not code for the classical
signal peptides, including superoxide dismutase (28), glu-
tamine synthetase (27), and CFP29 (54) as well as ESAT-6 and
CFP10 (65). Whether CFP32 or other such proteins are ac-
tively exported, excreted, or released during cell division or
autolysis is unresolved but may involve an uncharacterized
mycobacterial secretory mechanism (27, 28, 70). Moreover, by
analogy to other CF proteins (27, 28), CFP32 may serve intra-
cellular, in addition to extracellular, functions, thus explaining
the necessity of its partial cytoplasmic retention.
CFP32 is MTC restricted. For CFP32 to be characterized as
a unique biofactor of tuberculous mycobacteria, it should be
present in all members of the MTC and absent from MOTT. In
a previous study, 8 MTC subspecies (representing 72 strains)
and 12 MOTT species (comprising 46 strains) were evaluated
for the presence of cfp32 by PCR (given as Rv0577 and using
primers Rv0577F and Rv0577R [33]). A cfp32 PCR fragment
was observed only in the MTC groupings, indicating that cfp32
is an MTC-restricted gene. To determine the degree of poly-
morphism in cfp32 among the MTC subspecies, sequence anal-
ysis of PCR products representing the full and/or extended full
cfp32 coding sequence from M.tuberculosis (n⫽8), M.africa-
num (subtypes I and II, n⫽4 and 4, respectively), M.bovis (n
⫽4), M.bovis BCG (n⫽3), M.bovis subsp. caprae (n⫽1), and
M.microti (n⫽4) was done. Likewise, the 321 bp of the
putative cfp32 upstream promoter region was also PCR am-
plified and sequenced from M.tuberculosis (n⫽5), M.africa-
num (subtypes I and II, n⫽2 and 4, respectively), M.bovis (n
⫽1), M.bovis BCG (n⫽2), M.bovis subsp. caprae (n⫽1), and
M.microti (n⫽3). In all cases, the entire cfp32 locus (1,160 bp,
representing nucleotides 670843 to 672002 of the M.tubercu-
losis H37Rv genome sequence, accession no. AL123456) was
completely nonpolymorphic relative to the M.tuberculosis
H37Rv genome sequence (data not shown). This observation is
in keeping with previous studies that have noted a remarkable
FIG. 5. CFP32 localizes predominantly to the CF and cytosol frac-
tions of M.tuberculosis. Western blotting was done to probe the lysate,
CF, mannosylated lipoarabinomanan (manLam) glycolipid, cell wall,
soluble cell wall proteins (SCWP), membrane, and cytosol components
of M.tuberculosis (at 1 ␮g each) for the presence of CFP32 by using the
anti-rCFP32 antiserum (1:10
3
). Molecular mass protein markers (Am-
ersham; values in kilodaltons) are shown in the first lane. The amount
of CFP32, as measured by ELISA (average for duplicate samples in
three experiments), in each sample is given below each respective lane
(in picograms per microgram of sample ⫾standard error [SE]). M.
tuberculosis PPD was negative for CFP32 by Western blotting (data not
shown) and by ELISA was measured as having 51 ⫾28 pg of CFP32
per ␮g of sample. CF-fx9 was also tested by ELISA and had 610 ⫾53
ng of CFP32 per ␮g of sample.
VOL. 71, 2003 M.TUBERCULOSIS PROTEIN CFP32, IL-10, AND TUBERCULOSIS 6877

paucity of single-nucleotide polymorphisms in the structural
genes of the MTC subspecies (66).
Southern blotting was employed next to verify that cfp32 is
restricted to the MTC organisms. Of the evaluated species,
only M.tuberculosis strains H37Rv and W, as well as the ad-
ditional MTC subspecies M.africanum subtype I, M.bovis, and
M.bovis BCG, were positive for a single copy of cfp32, while all
13 MOTT species and strains evaluated were negative (Fig.
6A). M.smegmatis was also repeatedly evaluated and found to
be negative for cfp32 by Southern blotting (data not shown).
Moreover, a cfp32 homologue could not be found in the M.
smegmatis or the M.leprae genome sequences (http://www.tigr
.org and http://www.sanger.ac.uk). Further cfp32 Southern
blotting probed a comprehensive range of M.tuberculosis clin-
ical isolates (n⫽70) previously coded by IS6110-restriction
fragment length polymorphism pattern classification (Fig. 6B
subpanels i to iv) (37). Included in this evaluation were 36
strains prototypic for their particular IS6110 fingerprint (Fig.
6B subpanels i and ii). Remarkably, every M.tuberculosis strain
was positive for a single band that ran at approximately the
same location for all but two strains, for which it ran slightly
lower than the others (Fig. 6B subpanels iii and iv). This
difference most likely relates to the emergence of a new a PvuII
cutting site outside cfp32 since sequencing of the strain
TN13475 cfp32 did not uncover any polymorphisms. As such, it
is impressive that cfp32 was completely conserved within the
MTC given that subspecies- and strain-defining large chromo-
somal deletions are increasingly found in the MTC genomes
(13, 33, 43). These deletions are emerging as potentially sig-
nificant determinants of MTC pathobiological diversity but do
not appear to include cfp32. Therefore, the complete conser-
vation of cfp32 and its sequence for the tested isolates and its
absence from MOTT species suggest that this gene may play an
important role that is unique to M.tuberculosis and the other
MTC groupings.
To probe for the mycobacterial expression of CFP32, West-
ern blotting was done against a panel of MTC subspecies and
strains (n⫽37), as well as a range of MOTT species and
strains (n⫽29), which are listed in Materials and Methods.
Upon completion, a CFP32 band was detected only from M.
tuberculosis and the other MTC subspecies and not from any of
the 10 MOTT species that were evaluated (Fig. 7; select strains
are illustrated). The combined data support the idea that cfp32
is an expressed MTC-restricted gene and are in keeping with
an integral role for CFP32 in the lifestyle of tubercle bacilli.
Detection of specific antisera to CFP32 in clinical samples
by ELISA. CFP32 is immunogenic for mice as given by the
development of the murine IT-44 MAb following immuniza-
tion with M. tuberculosis CF (39). As an indirect indicator that
CFP32 is expressed in TB patients, the human serologic re-
sponse to CFP32 was evaluated. Sera from a cohort of patients
with active TB from Brazil (n⫽35) and their healthy house-
hold contacts (n⫽11; four PPD skin test positive, seven PPD
skin test negative) were tested for antibodies that recognize
rCFP32. Altogether, 34% (12 of 35) of TB case patients had
detectable antibodies to CFP32 while none of the healthy con-
trols were positive (P⬍0.05, Fisher’s exact test) (Fig. 8A);
PPD status did not segregate the healthy household contacts.
Notably, only half of these patients had a documented chest
X-ray in their medical record, and of these, only 25% of the
CFP32 antiserum-positive patients had cavitary TB, thereby
indicating that cavitary TB status did not increase the likeli-
hood of having a positive anti-CFP32 serologic response. The
anti-CFP32 positivity rate was similar in those with and those
without a documented chest X-ray. To extend these observa-
tions to a population from India, a modified ELISA (without
the primary coating MAb) was used to test sera from AFB
smear-positive cavitary TB patients (n⫽30) and PPD skin
FIG. 6. Southern blot analysis for cfp32. (A) The cfp32 gene is
MTC restricted by Zoo blotting. DNA from an assortment of MTC
subspecies (n⫽5; namely, M.tuberculosis [M.tb] strains H37Rv and
W, M.africanum subtype I, M.bovis, and M.bovis BCG) and myco-
bacteria other than MTC (MOTT; n⫽13) was evaluated using cfp32
PCR fragments as the probe in Southern blotting. (B) Each clinical
isolate of M.tuberculosis tested possesses the cfp32 gene. Subpanels i
and ii illustrate the cfp32 Southern blot results for 36 M.tuberculosis
isolates with unique IS6110-restriction fragment length polymorphism
patterns prototypical of their lineages. An additional 35 M.tuberculosis
clinical isolates were also evaluated (72 M.tuberculosis strains tested in
total), examples of which are shown in subpanels iii and iv.
6878 HUARD ET AL. INFECT.IMMUN.

test-positive healthy controls (n⫽29). In this set, 30% (9 of
30) of the cavitary TB patients and 3% (1 of 29) of the PPD
skin test-positive controls were reactive to rCFP32 (P⬍0.013,
Fisher’s exact test) (Fig. 8B). Hence, there was comparatively
limited variation between the two TB patient populations.
Combined, 32% of TB case patients (n⫽65) and only 2.5% of
healthy controls (n⫽40) exhibited a significant serological
response to E.coli-expressed rCFP32 (P⬍0.003, Fisher’s
exact test). These data are in the range of those of other
studies that examined humoral immunity to recombinant M.
tuberculosis proteins, whereby, depending upon the antigen
evaluated and its method of production, the sera of 12 to 58%
of TB patients were found to contain specific antibodies (38,
41). Overall, the available data indicate that the spectrum of M.
tuberculosis antigens recognized humorally varies dramatically
between patients (41, 58). One contributing factor is the ex-
pansion of the repertoire of humoral specificities to M.tuber-
culosis antigens as TB progresses (2, 58). For example, serum
antibodies to the IT-44-reactive antigen (i.e., CFP32) have
been identified in cavitary TB patients but not in noncavitary
TB patients or PPD-positive individuals by Western blotting
(58). (The apparent inconsistency with regard to our detection
of anti-CFP32 specificities in noncavitary TB patients may well
be a factor of the different means of evaluation.) Moreover, the
CFP32 spot was one of only 26 serum-reactive CF spots that
were seen in the study by Samanich et al. (58). As such, CFP32
appears to be a promising humoral antigen for inclusion in a
multiantigen serodiagnostic test and may provide a useful in-
dicator of worsening disease. It is also important that signifi-
cantly fewer patients recognize certain M.tuberculosis proteins
when they are expressed in E.coli compared to the counterpart
native proteins (59). Since E.coli-expressed rCFP32 was used
in these evaluations as the antibody capture antigen, this may
have been a factor in reducing the sensitivity of our assays for
CFP32 humoral specificity.
Detection of CFP32 in lung sputum samples of TB patients
by ELISA and positive correlation of CFP32 levels with IL-10.
If immunogenic, a candidate in vivo-expressed M.tuberculosis
antigen should be present at the site of disease. To detect lung
CFP32, the induced sputa from patients in Brazil, who pre-
sented for diagnostic workups for “lung disease or suspected
TB,”were tested for the presence of CFP32 by ELISA. This
cohort included patients with TB (n⫽41), suspected TB (n⫽
16), or OLD (n⫽18), as defined in Materials and Methods.
Sequential samples were taken at days 0, 15, 30, 60, and 180
following first presentation with one to five visits per patient. A
significant number of TB patients had detectable amounts of
CFP32 in their sputum. In contrast to the TB cases, none of the
patients with OLD had detectable CFP32 in their sputum,
resulting in a specificity of 100%. In total, 56% (23 of 41) of TB
case patients were positive for CFP32 in at least one sample (P
⬍0.0001, Fisher’s exact test), of which 32% (13 of 41) were
positive at study entry for CFP32 (P⫽0.0057, Fisher’s exact
test). Overall, 59% (10 of 17) of cavitary TB patients were
CFP32 sputum positive in at least one sample and 30% (42 of
140) of all samples from TB patients were positive for CFP32
(P⫽0.0007, Fisher’s exact test) (Fig. 9A). Among TB cases,
AFB smear was positive in 54% (22 of 41) while culture was
positive in 80% (33 of 41). Sputum CFP32 was detected in at
least one sample in 61% (20 of 33) of culture-positive TB cases
and 50% (3 of 6) culture-negative (n⫽5) and culture-unavail-
able (n⫽1) TB cases. When cross-correlated with AFB smear,
64% (14 of 22) of AFB smear-positive and 47% (9 of 19) of
AFB smear-negative case patients had detectable sputum
CFP32. There were only five TB case patients who were AFB
and culture negative and whose diagnosis was based on treat-
ment response with resolution of clinical and radiological fea-
tures of TB. Of these case patients, 40% (two of five) were
positive for CFP32. In the suspected TB category, for whom
follow-up data were not available to establish a diagnosis,
CFP32 was positive in 56% (9 of 16) of cases. For these ex-
periments the standard curve of the CFP32 ELISA was con-
sistently linear and sensitive to the level of ⬃5 to 10 pg/ml
(data not shown). The data strongly support the idea that
CFP32 is present in the diseased lung. Although the function
of CFP32 here remains unknown, we speculate that it may
contribute to the pathobiology of M.tuberculosis. However, as
proposed for other M.tuberculosis CF proteins (5, 12, 52), the
actions of CFP32 could include both direct enzymatic activity
upon host cells or structures and/or bacterial components and
FIG. 7. Western blot analysis for CFP32. (A) CFP32 is MTC re-
stricted. Mycobacterial lysates (7 ␮g of each sample) and purified
rCFP32 (10 ng) were probed by Western blotting with the anti-rCFP32
antiserum (1:10
3
). A total of 37 MTC isolates (M.tuberculosis H37Rv
and one strain each of M.africanum subtype I, M.bovis,M.bovis BCG,
and M.microti are shown) and 29 MOTT isolates (one isolate each of
M.avium subsp. avium,M.smegmatis, and M.leprae is illustrated) were
tested. A breakdown by MOTT species and MTC subspecies is given in
Materials and Methods. (B) Both laboratory and clinical isolates of M.
tuberculosis express CFP32. The lysates of M.tuberculosis strains (3 ␮g
of each sample) were probed by Western blotting with anti-rCFP32
antiserum (1:10
3
). For both panels, parallel silver-stained gels (with
10-fold-more sample per isolate) confirmed that approximately the
same amount of protein was loaded for each Mycobacterium isolate
illustrated (data not shown). Molecular mass protein markers (Amer-
sham; values in kilodaltons) are shown in the first lane of each blot.
VOL. 71, 2003 M.TUBERCULOSIS PROTEIN CFP32, IL-10, AND TUBERCULOSIS 6879

antigenicity-based local tissue damage via immunohyperstimu-
lation. By virtue of its expression in vivo and given that CFP32
could be detected in the suspected TB subset of patients as well
as a proportion of AFB smear-negative and/or culture-negative
patients, these data further present CFP32 as a strong candi-
date antigen for inclusion in a next-generation diagnostic strat-
egy as a marker of increasing bacterial burden or as an indi-
cation of the effectiveness of TB pharmacologic therapy.
Previously, the coexpression of mRNA for IFN-␥with IL-10,
IL-2, and inducible nitric oxide synthase, by lung cells from
patients with active pulmonary TB, was described [48; M. D.
Bonecini-Almeida, J. R. Lapa e Silva, S. Nicholson, J. Geng, N.
Boechat, C. Linhares, L. Rego, and A. L. Kritski, abstract from
the American Thoracic Society Annual Meeting 1997, Am. J.
Respir. Crit. Care Med. 155(Suppl.):A441, 1997]. To further
dissect the in vivo human immunologic parameters associated
with TB, ELISA was done to quantify the IL-10 and IFN-␥in
the same induced sputum samples from TB patients (n⫽34)
previously assayed for CFP32. A significant correlation be-
tween CFP32 and IL-10 in the sputum of patients was found by
linear regression analysis (n⫽112 samples; r
2
⫽0.60, P⬍
0.0001) (Fig. 9B subpanel i). No convincing association was
identified between CFP32 and IFN-␥(n⫽125 samples) (Fig.
9B subpanel ii) nor between IL-10 and IFN-␥(n⫽110 sam-
ples; data not shown). IFN-␥is regarded as a pivotal cytokine
in the protective immune response against M.tuberculosis in-
fection, acting as the major mediator of macrophage activation
and as a crucial component in the development of specific
counter-M.tuberculosis adaptive immunity (16). IL-10, on the
other hand, is a pleiotropic immunosuppressive cytokine that
opposes many IFN-␥-mediated effects including macrophage-
mediated mycobacteriocidal activity (10, 19, 24). Indeed, evi-
dence is accumulating linking IL-10 to the failure in immunity
that results in the progression to active TB. For example,
infection studies with either IL-10 gene-knockout mice or
IL-10 transgenic mice have shown that IL-10 is an inhibitor of
anti-tubercle bacillus responses (34, 45, 46, 71). In humans,
healthy persons reactive to PPD produce high concentrations
of IFN-␥from M.tuberculosis antigen-stimulated peripheral
blood mononuclear cells (PBMCs) while TB patients with se-
vere disease, and without reactivity to PPD (i.e., anergized),
exhibit impaired IFN-␥production in association with in-
creased IL-10 (11, 17). Moreover, increased levels of IL-10, in
the presence of IFN-␥, have been detected in the serum of TB
patients, as well as from ex vivo M.tuberculosis antigen-stim-
ulated PBMCs of TB patients (20, 49, 63, 72). In this report, we
found a novel association of bacterial physiological activity (as
reflected by CFP32 antigen levels) and IL-10 production in the
lungs of patients with the failure of counter-M.tuberculosis
immunity (which was also notably coincident with continued
IFN-␥production). These data contrast with those of Barnes et
al. (7), who found elevated IFN-␥in association with IL-10 in
the pleural fluid of patients with tuberculous pleuritis (a form
of TB that resolves without chemotherapy). Together these
data suggest that the immunosuppressive actions of IL-10 may
come to predominate and eliminate the protective immune
system-activating properties of IFN-␥in the lungs of persons
with advanced TB. The mechanism underlying the elevated
IL-10 levels is likely multifactorial and involves contributions
from both the host and the pathogen. In this regard, naïve
human PBMCs, monocytes, and dendritic cells are known to
produce IL-10 when stimulated with M.tuberculosis or with its
cell wall constituents (6, 10, 29, 69). Hence, as an immune
evasion strategy, M.tuberculosis may deliberately induce the
FIG. 8. Antiserum specificity to CFP32 is detectable in a significant proportion of human TB patients. (A) Cohort of patients living in Brazil.
Sera from 35 active TB case patients, along with the sera of 11 healthy household contacts (seven PPD skin test negative and four PPD skin test
positive), were used in an ELISA to identify human humoral specificity for CFP32. The serologic reactivity of the healthy controls was used to set
the cutoff value above which samples were deemed positive (mean [M] A
450
⫹1.5 standard deviation). The results of statistical analysis of the data
are shown (P⬍0.05, Fisher’s exact test). (B) Cohort of patients living in India. Sera from 30 active TB case patients, as well as from 29
PPD-positive controls, were used in a variant ELISA to confirm the existence of human humoral response to CFP32. The serologic reactivity of
the healthy PPD-positive controls was used to set the cutoff value (mean A
490
⫹1.5 standard deviation). The results of statistical analysis of the
data are shown (P⬍0.013, Fisher’s exact test).
6880 HUARD ET AL. INFECT.IMMUN.

production of IL-10 and thereby depress cellular responses to
IFN-␥and promote M.tuberculosis intramacrophage survival
(10). In relation to this idea, increased local IL-10 is also
thought to promote the development of the IL-10-producing
CD4
⫹
T regulatory 1 (Tr1) cells (1). In fact, the majority of
bronchoalveolar lavage-derived CD4
⫹
T-cell clones from TB
patients are reminiscent of Tr1 cells (25; although in that study
they also produced IFN-␥), and Tr1-like cells have been im-
plicated in the PPD anergy of TB patients (11, 17). Therefore,
Tr1 cells may act in their turn to further stifle local anti-M.
tuberculosis innate and T-cell-mediated adaptive immune re-
sponses and potentiate a positive feedback loop of IL-10 se-
cretion that supports M.tuberculosis persistence and/or reac-
tivation. Since IL-10 is also known to promote B-cell
production of antibody (56), the increased lung IL-10 levels
may be responsible for the expansion of anti-M.tuberculosis
serum specificities and enhanced antibody titers as TB
progresses (2, 56). Therefore, lung IL-10 level bears further
investigation as an immunological correlate for the develop-
ment of pulmonary TB.
In summary, this work adds cfp32 to the growing list of M.
tuberculosis genes proven to code for an expressed protein.
This CFP32 protein appears to be a bimodular glyoxalase lo-
calized to both the cytosolic and CF compartments of M.tu-
berculosis. Although the biological role(s) of CFP32 remains to
be elucidated, the accumulated data suggest that CFP32 is an
important biofactor since it is MTC restricted, the nonpoly-
morphic cfp32 gene is present in all M.tuberculosis strains that
have been evaluated, and it is expressed in the lungs of a
significant proportion of TB patients in addition to being a
FIG. 9. CFP32 is detectable in the lungs of a significant number of TB patients and is positively correlated with IL-10 but not IFN-␥levels.
(A) Excess induced lung sputum not used for diagnostic purposes was tested for the presence of CFP32 by ELISA. Donors were either TB case
patients (TB; defined as having a positive solid medium culture or treatment response with resolution of clinical and radiological features of TB)
(n⫽41; 140 samples), suspected TB case patients (Suspected TB; defined as patients with clinical and/or radiological features compatible with
TB for whom cultures were negative, contaminated, or not available and who had insufficient follow-up or had prior TB without sufficient
follow-up) (n⫽16; 17 samples), or non-TB case patients with other lung diseases (OLD; defined as those patients who were AFB smear and TB
culture negative, for whom another diagnosis was established, and/or who showed clinical improvement after a short course of non-TB antibiotics)
(n⫽18; 25 samples). The mean (M) value for the non-TB group was used to set the cutoff value (M ⫹2.5 standard deviations) above which
samples were deemed positive. Statistical analyses found a significant difference between TB and non-TB cases (P⫽0.0007, Fisher’s exact test).
(B) CFP32 levels were then correlated with IL-10 and IFN-␥in matched sputum samples by ELISA; n, number of samples. Linear regression
analyses revealed that 60% of the variance in the measured amounts of sputum CFP32 corresponds to variation in IL-10 levels (r
2
⫽0.60, P⬍
0.0001).
VOL. 71, 2003 M.TUBERCULOSIS PROTEIN CFP32, IL-10, AND TUBERCULOSIS 6881

recognized humoral antigen. That CFP32 levels in the lungs of
active TB patients correlated with measured IL-10, but not
IFN-␥, levels supports the hypothesis that M.tuberculosis-stim-
ulated local IL-10 secretion precipitates the immunodysregu-
lation that contributes to the success of M.tuberculosis as a
human pathogen. Determining the role of CFP32, if any, in this
particular pathogenic strategy of M.tuberculosis is a priority
interest of our laboratory.
ACKNOWLEDGMENTS
We thank Howard Doo, Kelley Sookraj, Krishna Menon, Patricia
Lago, Vera Batista, and Brianna Holland for technical assistance and
Albert Ko for aiding the statistical analyses. We are also grateful to
Lee W. Riley, Sabine Ehrt, and Warren D. Johnson for suggestions,
support, and encouragement.
Funding support was provided by NIH grants R0-1 AI39606 and
R0-1 HL61960 (J.L.H.), NIH NIAID contract N01 AI-75320 (J.T.B.),
NIH Fogarty International Center Training grant (FICTG) D43
TW00018, a grant from the Coordenac¸a˜o de Aperfeic¸oamento de
Pessoal de Nivel Superior (CAPES; Ministry of Education-Brazil), and
a grant from the Laura Cook Hull Trust Fund (LCHTF) (Warren D.
Johnson, Principal Investigator). R.C.H. was supported by the
LCHTF, H.Z. was a FICTG trainee, and L.C.O.L. was an FICTG and
CAPES trainee. S.L. was supported by a VA merit award. M.B.C.,
A.L.K., and J.R.L.S. were funded in Brazil by the following grants:
Brazilian TB Research Network 62.0055/01-4-PADCT III/MILLEN
IUM (CNPq/Brazilian Research Council and World Bank; M.B.C.,
A.L.K., J.R.L.S.), “Excellence Research Nuclei for TB Control”
66.1028/1998-4 (PRONEX/Brazilian Research Council; A.L.K.,
J.R.L.S.), “Scientists of Our State”2000 and 2003 (Rio de Janeiro
Research Council/FAPERJ; A.L.K., J.R.L.S.), “Small Grants Pro-
gram”(Rio de Janeiro Research Council/FAPERJ; M.B.C.), and
“Small Grants Program”(Fundac¸a˜o Universita´ria Jose´Bonifa´cio/
FUJB; J.R.L.S.).
REFERENCES
1. Akbari, O., R. H. DeKruyff, and D. T. Umetsu. 2001. Pulmonary dendritic
cells producing IL-10 mediate tolerance induced by respiratory exposure to
antigen. Nat. Immunol. 2:725–731.
2. Amara, R. R., and V. Satchidanandam. 1996. Analysis of a genomic DNA
expression library of Mycobacterium tuberculosis using tuberculosis patient
sera: evidence for modulation of host immune response. Infect. Immun.
64:3765–3771.
3. Andersen, P. 1997. Host responses and antigens involved in protective im-
munity to Mycobacterium tuberculosis. Scand. J. Immunol. 45:115–131.
4. Ando, N., K. Ueda, and S. Horinouchi. 1997. A Streptomyces griseus gene
(sgaA) suppresses the growth disturbance caused by high osmolality and a
high concentration of A-factor during early growth. Microbiology 143:2715–
2723.
5. Armitige, L. Y., C. Jagannath, A. R. Wanger, and S. J. Norris. 2000. Dis-
ruption of the genes encoding antigen 85A and antigen 85B of Mycobacte-
rium tuberculosis H37Rv: effect on growth in culture and in macrophages.
Infect. Immun. 68:767–778.
6. Barnes, P. F., D. Chatterjee, J. S. Abrams, S. Lu, E. Wang, M. Yamamura,
P. J. Brennan, and R. L. Modlin. 1992. Cytokine production induced by
Mycobacterium tuberculosis lipoarabinomannan. Relationship to chemical
structure. J. Immunol. 149:541–547.
7. Barnes, P. F., S. Lu, J. S. Abrams, E. Wang, M. Yamamura, and R. L.
Modlin. 1993. Cytokine production at the site of disease in human tubercu-
losis. Infect. Immun. 61:3482–3489.
8. Bhaskar, S., S. P. Khanna, and R. Mukherjee. 2000. Isolation, purification
and immunological characterization of novel low molecular weight protein
antigen CFP 6 from culture filtrate of M.tuberculosis. Vaccine 18:2856–2866.
9. Boesen, H., B. N. Jensen, T. Wilcke, and P. Andersen. 1995. Human T-cell
responses to secreted antigen fractions of Mycobacterium tuberculosis. Infect.
Immun. 63:1491–1497.
10. Bonecini-Almeida, M. G., S. Chitale, I. Boutsikakis, J. Geng, H. Doo, S. He,
and J. L. Ho. 1998. Induction of in vitro human macrophage anti-Mycobac-
terium tuberculosis activity: requirement for IFN-gamma and primed lym-
phocytes. J. Immunol. 160:4490–4499.
11. Boussiotis, V. A., E. Y. Tsai, E. J. Yunis, S. Thim, J. C. Delgado, C. C.
Dascher, A. Berezovskaya, D. Rousset, J. M. Reynes, and A. E. Goldfeld.
2000. IL-10-producing T cells suppress immune responses in anergic tuber-
culosis patients. J. Clin. Investig. 105:1317–1325.
12. Brodin, P., K. Eiglmeier, M. Marmiesse, A. Billault, T. Garnier, S. Niemann,
S. T. Cole, and R. Brosch. 2002. Bacterial artificial chromosome-based com-
parative genomic analysis identifies Mycobacterium microti as a natural
ESAT-6 deletion mutant. Infect. Immun. 70:5568–5578.
13. Brosch, R., S. V. Gordon, M. Marmiesse, P. Brodin, C. Buchrieser, K.
Eiglmeier, T. Garnier, C. Gutierrez, G. Hewinson, K. Kremer, L. M. Par-
sons, A. S. Pym, S. Samper, D. van Soolingen, and S. T. Cole. 2002. A new
evolutionary scenario for the Mycobacterium tuberculosis complex. Proc.
Natl. Acad. Sci. USA 99:3684–3689.
14. Camus, J. C., M. J. Pryor, C. Medigue, and S. T. Cole. 2002. Re-annotation
of the genome sequence of Mycobacterium tuberculosis H37Rv. Microbiology
148:2967–2973.
15. Cole, S. T., R. Brosch, J. Parkhill, T. Garnier, C. Churcher, D. Harris, S. V.
Gordon, K. Eiglmeier, S. Gas, C. E. Barry III, F. Tekaia, K. Badcock, D.
Basham, D. Brown, T. Chillingworth, R. Connor, R. Davies, K. Devlin, T.
Feltwell, S. Gentiles, N. Hamlin, S. Holroyd, T. Hornsby, K. Jegels, A.
Krogh, J. McLean, S. Moule, L. Murphy, K. Oliver, J. Osborne, M. A. Quail,
M.-A. Rajandream, J. Rogers, S. Rutter, K. Seeger, J. Skelton, R. Squares,
S. Squares, J. E. Sulston, K. Taylor, S. Whitehead, and B. G. Barrell. 1998.
Deciphering the biology of Mycobacterium tuberculosis from the complete
genome sequence. Nature 393:537–544.
16. Collins, H. L., and S. H. Kaufmann. 2001. The many faces of host responses
to tuberculosis. Immunology 103:1–9.
17. Delgado, J. C., E. Y. Tsai, S. Thim, A. Baena, V. A. Boussiotis, J. M. Reynes,
S. Sath, P. Grosjean, E. J. Yunis, and A. E. Goldfeld. 2002. Antigen-specific
and persistent tuberculin anergy in a cohort of pulmonary tuberculosis pa-
tients from rural Cambodia. Proc. Natl. Acad. Sci. USA 99:7576–7581.
18. Demissie, A., P. Ravn, J. Olobo, T. M. Doherty, T. Eguale, M. Geletu, W.
Hailu, P. Andersen, and S. Britton. 1999. T-cell recognition of Mycobacte-
rium tuberculosis culture filtrate fractions in tuberculosis patients and their
household contacts. Infect. Immun. 67:5967–5971.
19. Dickensheets, H. L., S. L. Freeman, M. F. Smith, and R. P. Donnelly. 1997.
Interleukin-10 upregulates tumor necrosis factor receptor type-II (p75) gene
expression in endotoxin-stimulated human monocytes. Blood 90:4162–4171.
20. Dlugovitzky, D., A. Torres-Morales, L. Rateni, M. A. Farroni, C. Largacha,
O. Molteni, and O. Bottasso. 1997. Circulating profile of Th1 and Th2
cytokines in tuberculosis patients with different degrees of pulmonary in-
volvement. FEMS Immunol. Med. Microbiol. 18:203–207.
21. Dye, C., S. Scheele, P. Dolin, V. Pathania, and M. C. Raviglione. 1999.
Consensus statement. Global burden of tuberculosis: estimated incidence,
prevalence, and mortality by country. WHO Global Surveillance and Mon-
itoring Project. JAMA 282:677–686.
22. Eng, J. K., A. L. McCormack, and J. R. Yates. 1994. A approach to correlate
tandem mass-spectral data of peptides with amino-acid sequences in a pro-
tein database. J. Am. Soc. Mass. Spectrom. 5:976–989.
23. Fernandez, J., L. Andrews, and S. M. Mische. 1994. An improved procedure
for enzymatic digestion of polyvinylidene difluoride-bound proteins for in-
ternal sequence analysis. Anal. Biochem. 218:112–117.
24. Flesch, I. E., J. H. Hess, I. P. Oswald, and S. H. Kaufmann. 1994. Growth
inhibition of Mycobacterium bovis by IFN-gamma stimulated macrophages:
regulation by endogenous tumor necrosis factor-alpha and by IL-10. Int.
Immunol. 6:693–700.
25. Gerosa, F., C. Nisii, S. Righetti, R. Micciolo, M. Marchesini, A. Cazzadori,
and G. Trinchieri. 1999. CD4
⫹
T cell clones producing both interferon-
gamma and interleukin-10 predominate in bronchoalveolar lavages of active
pulmonary tuberculosis patients. Clin. Immunol. 92:224–234.
26. Glickman, M. S., and W. R. Jacobs, Jr. 2001. Microbial pathogenesis of
Mycobacterium tuberculosis: dawn of a discipline. Cell 104:477–485.
27. Harth, G., and M. A. Horwitz. 1997. Expression and efficient export of
enzymatically active Mycobacterium tuberculosis glutamine synthetase in My-
cobacterium smegmatis and evidence that the information for export is con-
tained within the protein. J. Biol. Chem. 272:22728–22735.
28. Harth, G., and M. A. Horwitz. 1999. Export of recombinant Mycobacterium
tuberculosis superoxide dismutase is dependent upon both information in the
protein and mycobacterial export machinery. A model for studying export of
leaderless proteins by pathogenic mycobacteria. J. Biol. Chem. 274:4281–
4292.
29. Henderson, R. A., S. C. Watkins, and J. L. Flynn. 1997. Activation of human
dendritic cells following infection with Mycobacterium tuberculosis. J. Immu-
nol. 159:635–643.
30. Heym, B., Y. Zhang, S. Poulet, D. Young, and S. T. Cole. 1993. Character-
ization of the katG gene encoding a catalase-peroxidase required for the
isoniazid susceptibility of Mycobacterium tuberculosis. J. Bacteriol. 175:4255–
4259.
31. Horwitz, M. A., B. W. Lee, B. J. Dillon, and G. Harth. 1995. Protective
immunity against tuberculosis induced by vaccination with major extracellu-
lar proteins of Mycobacterium tuberculosis. Proc. Natl. Acad. Sci. USA 92:
1530–1534.
32. Howard, S., S. He, and S. G. Withers. 1998. Identification of the active site
nucleophile in jack bean ␣-mannosidase using 5-fluoro-␤-L-gulosyl fluoride.
J. Biol. Chem. 273:2067–2072.
33. Huard, R. C., L. C. de Oliveira Lazzarini, W. R. Butler, D. van Soolingen,
and J. L. Ho. 2003. A PCR-based method to differentiate the subspecies of
6882 HUARD ET AL. INFECT.IMMUN.

the Mycobacterium tuberculosis complex on the basis of genomic deletions.
J. Clin. Microbiol. 41:1637–1650.
34. Jacobs, M., N. Brown, N. Allie, R. Gulert, and B. Ryffel. 2000. Increased
resistance to mycobacterial infection in the absence of interleukin-10. Im-
munology 100:494–501.
35. Jungblut, P. R., U. E. Schaible, H. J. Mollenkopf, U. Zimny-Arndt, B.
Raupach, J. Mattow, P. Halada, S. Lamer, K. Hagens, and S. H. Kaufmann.
1999. Comparative proteome analysis of Mycobacterium tuberculosis and
Mycobacterium bovis BCG strains: towards functional genomics of microbial
pathogens. Mol. Microbiol. 33:1103–1117.
36. Kane, M. M., and D. M. Mosser. 2001. The role of IL-10 in promoting
disease progression in leishmaniasis. J. Immunol. 166:1141–1147.
37. Kurepina, N. E., S. Sreevatsan, B. B. Plikaytis, P. J. Bifani, N. D. Connell,
R. J. Donnelly, D. van Soolingen, J. M. Musser, and B. N. Kreiswirth. 1998.
Characterization of the phylogenetic distribution and chromosomal insertion
sites of five IS6110 elements in Mycobacterium tuberculosis: non-random
integration in the dnaA-dnaN region. Tuber. Lung Dis. 79:31–42.
38. Lim, R. L., L. K. Tan, W. F. Lau, M. C. Ming, R. Dunn, H. P. Too, and L.
Chan. 2000. Cloning and expression of immunoreactive antigens from My-
cobacterium tuberculosis. Clin. Diagn. Lab. Immunol. 7:600–606.
39. Ljungqvist, L., A. Worsaae, and I. Heron. 1988. Antibody responses against
Mycobacterium tuberculosis in 11 strains of inbred mice: novel monoclonal
antibody specificities generated by fusions, using spleens from BALB.B10
and CBA/J mice. Infect. Immun. 56:1994–1998.
40. Lomovskaya, N., S. L. Otten, Y. Doi-Katayama, L. Fonstein, X. C. Liu, T.
Takatsu, A. Inventi-Solari, S. Filippini, F. Torti, A. L. Colombo, and C. R.
Hutchinson. 1999. Doxorubicin overproduction in Streptomyces peucetius:
cloning and characterization of the dnrU ketoreductase and dnrV genes and
the doxA cytochrome P-450 hydroxylase gene. J. Bacteriol. 181:305–318.
41. Lyashchenko, K., R. Colangeli, M. Houde, H. Al Jahdali, D. Menzies, and
M. L. Gennaro. 1998. Heterogeneous antibody responses in tuberculosis.
Infect. Immun. 66:3936–3940.
42. McKinney, J. D., K. Honer zu Bentrup, E. J. Munoz-Elias, A. Miczak, B.
Chen, W. T. Chan, D. Swenson, J. C. Sacchettini, W. R. Jacobs, Jr., and D. G.
Russell. 2000. Persistence of Mycobacterium tuberculosis in macrophages and
mice requires the glyoxylate shunt enzyme isocitrate lyase. Nature 406:735–
738.
43. Mostowy, S., D. Cousins, J. Brinkman, A. Aranaz, and M. A. Behr. 2002.
Genomic deletions suggest a phylogeny for the Mycobacterium tuberculosis
complex. J. Infect. Dis. 186:74–80.
44. Mulder, M. A., H. Zappe, and L. M. Steyn. 1997. Mycobacterial promoters.
Tuber. Lung Dis. 78:211–223.
45. Murray, P. J., L. Wang, C. Onufryk, R. I. Tepper, and R. A. Young. 1997. T
cell-derived IL-10 antagonizes macrophage function in mycobacterial infec-
tion. J. Immunol. 158:315–321.
46. Murray, P. J., and R. A. Young. 1999. Increased antimycobacterial immunity
in interleukin-10-deficient mice. Infect. Immun. 67:3087–3095.
47. Navas, J., B. Gonzalez-Zorn, N. Ladron, P. Garrido, and J. A. Vazquez-
Boland. 2001. Identification and mutagenesis by allelic exchange of choE,
encoding a cholesterol oxidase from the intracellular pathogen Rhodococcus
equi. J. Bacteriol. 183:4796–4805.
48. Nicholson, S., M. D. Bonecini-Almeida, J. R. Lapa e Silva, C. Nathan, Q. W.
Xie, R. Mumford, J. R. Weidner, J. Calaycay, J. Geng, N. Boechat, C.
Linhares, W. Rom, and J. L. Ho. 1996. Inducible nitric oxide synthase in
pulmonary alveolar macrophages from patients with tuberculosis. J. Exp.
Med. 183:2293–2302.
49. Othieno, C., C. S. Hirsch, B. D. Hamilton, K. Wilkinson, J. J. Ellner, and Z.
Toossi. 1999. Interaction of Mycobacterium tuberculosis-induced transform-
ing growth factor beta and interleukin-10. Infect. Immun. 67:5730–5735.
50. Parish, T., B. G. Gordhan, R. A. McAdam, K. Duncan, V. Mizrahi, and N. G.
Stoker. 1999. Production of mutants in amino acid biosynthesis genes of
Mycobacterium tuberculosis by homologous recombination. Microbiology
145:3497–3503.
51. Pugsley, A. P. 1993. The complete general secretory pathway in gram-neg-
ative bacteria. Microbiol. Rev. 57:50–108.
52. Raynaud, C., G. Etienne, P. Peyron, M. A. Laneelle, and M. Daffe. 1998.
Extracellular enzyme activities potentially involved in the pathogenicity of
Mycobacterium tuberculosis. Microbiology 144:577–587.
53. Rosenkrands, I., A. King, K. Weldingh, M. Moniatte, E. Moertz, and P.
Andersen. 2000. Towards the proteome of Mycobacterium tuberculosis. Elec-
trophoresis 21:3740–3756.
54. Rosenkrands, I., P. B. Rasmussen, M. Carnio, S. Jacobsen, M. Theisen, and
P. Andersen. 1998. Identification and characterization of a 29-kilodalton
protein from Mycobacterium tuberculosis culture filtrate recognized by mouse
memory effector cells. Infect. Immun. 66:2728–2735.
55. Rosenkrands, I., K. Weldingh, S. Jacobsen, C. V. Hansen, W. Florio, I.
Gianetri, and P. Andersen. 2000. Mapping and identification of Mycobacte-
rium tuberculosis proteins by two-dimensional gel electrophoresis, microse-
quencing and immunodetection. Electrophoresis 21:935–948.
56. Rousset, F., E. Garcia, T. Defrance, C. Peronne, N. Vezzio, D. H. Hsu, R.
Kastelein, K. W. Moore, and J. Banchereau. 1992. Interleukin 10 is a potent
growth and differentiation factor for activated human B lymphocytes. Proc.
Natl. Acad. Sci. USA 89:1890–1893.
57. Ruan, J., G. St. John, S. Ehrt, L. Riley, and C. Nathan. 1999. noxR3, a novel
gene from Mycobacterium tuberculosis, protects Salmonella typhimurium
from nitrosative and oxidative stress. Infect. Immun. 67:3276–3283.
58. Samanich, K. M., J. T. Belisle, M. G. Sonnenberg, M. A. Keen, S. Zolla-
Pazner, and S. Laal. 1998. Delineation of human antibody responses to
culture filtrate antigens of Mycobacterium tuberculosis. J. Infect. Dis. 178:
1534–1538.
59. Samanich, K. M., M. A. Keen, V. D. Vissa, J. D. Harder, J. S. Spencer, J. T.
Belisle, S. Zolla-Pazner, and S. Laal. 2000. Serodiagnostic potential of cul-
ture filtrate antigens of Mycobacterium tuberculosis. Clin. Diagn. Lab. Im-
munol. 7:662–668.
60. Sambandamurthy, V. K., X. Wang, B. Chen, R. G. Russell, S. Derrick, F. M.
Collins, S. L. Morris, and W. R. Jacobs. 2002. A pantothenate auxotroph of
Mycobacterium tuberculosis is highly attenuated and protects mice against
tuberculosis. Nat. Med. 8:1171–1174.
61. Sieling, P. A., J. S. Abrams, M. Yamamura, P. Salgame, B. R. Bloom, T. H.
Rea, and R. L. Modlin. 1993. Immunosuppressive roles for IL-10 and IL-4 in
human infection. In vitro modulation of T cell responses in leprosy. J. Im-
munol. 150:5501–5510.
62. Smith, S. M., M. R. Klein, A. S. Malin, J. Sillah, K. Huygen, P. Andersen,
K. P. McAdam, and H. M. Dockrell. 2000. Human CD8
⫹
T cells specific for
Mycobacterium tuberculosis secreted antigens in tuberculosis patients and
healthy BCG-vaccinated controls in The Gambia. Infect. Immun. 68:7144–
7148.
63. Song, C. H., H. J. Kim, J. K. Park, J. H. Lim, U. O. Kim, J. S. Kim, T. H.
Paik, K. J. Kim, J. W. Suhr, and E. K. Jo. 2000. Depressed interleukin-12
(IL-12), but not IL-18, production in response to a 30- or 32-kilodalton
mycobacterial antigen in patients with active pulmonary tuberculosis. Infect.
Immun. 68:4477–4484.
64. Sonnenberg, M. G., and J. T. Belisle. 1997. Definition of Mycobacterium
tuberculosis culture filtrate proteins by two-dimensional polyacrylamide gel
electrophoresis, N-terminal amino acid sequencing, and electrospray mass
spectrometry. Infect. Immun. 65:4515–4524.
65. Sorensen, A. L., S. Nagai, G. Houen, P. Andersen, and A. B. Andersen. 1995.
Purification and characterization of a low-molecular-mass T-cell antigen
secreted by Mycobacterium tuberculosis. Infect. Immun. 63:1710–1717.
66. Sreevatsan, S., X. Pan, K. E. Stockbauer, N. D. Connell, B. N. Kreiswirth,
T. S. Whittam, and J. M. Musser. 1997. Restricted structural gene polymor-
phism in the Mycobacterium tuberculosis complex indicates evolutionarily
recent global dissemination. Proc. Natl. Acad. Sci. USA 94:9869–9874.
67. Stockl, J., H. Vetr, O. Majdic, G. Zlabinger, E. Kuechler, and W. Knapp.
1999. Human major group rhinoviruses downmodulate the accessory func-
tion of monocytes by inducing IL-10. J. Clin. Investig. 104:957–965.
68. Stylianou, E., P. Aukrust, D. Kvale, F. Muller, and S. S. Froland. 1999. IL-10
in HIV infection: increasing serum IL-10 levels with disease progression—
down-regulatory effect of potent anti-retroviral therapy. Clin. Exp. Immunol.
116:115–120.
69. Thoma-Uszynski, S., S. M. Kiertscher, M. T. Ochoa, D. A. Bouis, M. V.
Norgard, K. Miyake, P. J. Godowski, M. D. Roth, and R. L. Modlin. 2000.
Activation of toll-like receptor 2 on human dendritic cells triggers induction
of IL-12, but not IL-10. J. Immunol. 165:3804–3810.
70. Tullius, M. V., G. Harth, and M. A. Horwitz. 2001. High extracellular levels
of Mycobacterium tuberculosis glutamine synthetase and superoxide dis-
mutase in actively growing cultures are due to high expression and extracel-
lular stability rather than to a protein-specific export mechanism. Infect.
Immun. 69:6348–6363.
71. Turner, J., M. Gonzalez-Juarrero, D. L. Ellis, R. J. Basaraba, A. Kipnis,
I. M. Orme, and A. M. Cooper. 2002. In vivo IL-10 production reactivates
chronic pulmonary tuberculosis in C57BL/6 mice. J. Immunol. 169:6343–
6351.
72. Verbon, A., N. Juffermans, S. J. Van Deventer, P. Speelman, H. Van Deu-
tekom, and T. Van Der Poll. 1999. Serum concentrations of cytokines in
patients with active tuberculosis (TB) and after treatment. Clin. Exp. Immu-
nol. 115:110–113.
73. Weldingh, K., I. Rosenkrands, S. Jacobsen, P. B. Rasmussen, M. J. Elhay,
and P. Andersen. 1998. Two-dimensional electrophoresis for analysis of
Mycobacterium tuberculosis culture filtrate and purification and characteriza-
tion of six novel proteins. Infect. Immun. 66:3492–3500.
74. World Health Organization. 1996. Report of the tuberculosis epidemic.
World Health Organization, Geneva, Switzerland.
Editor: S. H. E. Kaufmann
VOL. 71, 2003 M.TUBERCULOSIS PROTEIN CFP32, IL-10, AND TUBERCULOSIS 6883