JOURNAL OF CLINICAL MICROBIOLOGY, Sept. 2002, p. 3398–3405
0095-1137/02/$04.00?0 DOI: 10.1128/JCM.40.9.3398–3405.2002
Copyright © 2002, American Society for Microbiology. All Rights Reserved.
Vol. 40, No. 9
Mycobacterium africanum Subtype II Is Associated with Two Distinct
Genotypes and Is a Major Cause of Human Tuberculosis in
S. Niemann,1* S. Ru ¨sch-Gerdes,1M. L. Joloba,2C. C. Whalen,3D. Guwatudde,2,3J. J. Ellner,4
K. Eisenach,5N. Fumokong,5J. L. Johnson,3T. Aisu,2R. D. Mugerwa,2A. Okwera,2and
S. K. Schwander4
National Reference Center for Mycobacteria, Research Center Borstel, Borstel, Germany1; Departments of Medicine and Medical
Microbiology, Makerere University, Kampala, Uganda2; Departments of Epidemiology and Biostatistics and of Medicine, Case
Western Reserve University School of Medicine, Cleveland, Ohio 441063; Department of Medicine and Ruy V. Lourenco
Center for the Study of Emerging and Reemerging Pathogens, University of Medicine and Dentistry–New Jersey
Medical School, Newark, New Jersey 071034; and Departments of Pathology and Immunology, University
of Arkansas for Medical Sciences, Little Rock, Arkansas 722055
Received 31 January 2002/Returned for modification 18 March 2002/Accepted 14 June 2002
The population structure of 234 Mycobacterium tuberculosis complex strains obtained during 1995 and 1997
from tuberculosis patients living in Kampala, Uganda (East Africa), was analyzed by routine laboratory
procedures, spoligotyping, and IS6110 restriction fragment length polymorphism (RFLP) typing. According to
biochemical test results, 157 isolates (67%) were classified as M. africanum subtype II (resistant to thiophen-
2-carboxylic acid hydrazide), 76 isolates (32%) were classified as M. tuberculosis, and 1 isolate was classified as
classical M. bovis. Spoligotyping did not lead to clear differentiation of M. tuberculosis and M. africanum, but
all M. africanum subtype II isolates lacked spacers 33 to 36, differentiating them from M. africanum subtype I.
Moreover, spoligotyping was not sufficient for differentiation of isolates on the strain level, since 193 (82%)
were grouped into clusters. In contrast, in the IS6110-based dendrogram, M. africanum strains were clustered
into two closely related strain families (Uganda I and II) and clearly separated from the M. tuberculosis isolates.
A further characteristic of both M. africanum subtype II families was the absence of spoligotype spacer 40. All
strains of family I also lacked spacer 43. The clustering rate obtained by the combination of spoligotyping and
RFLP IS6110 analysis was similar for M. africanum and M. tuberculosis, as 46% and 49% of the respective
isolates were grouped into clusters. The results presented demonstrate that M. africanum subtype II isolates
from Kampala, Uganda, belong to two closely related genotypes, which may represent unique phylogenetic
branches within the M. tuberculosis complex. We conclude that M. africanum subtype II is the main cause of
human tuberculosis in Kampala, Uganda.
Mycobacterium africanum is a member of the Mycobacterium
tuberculosis complex, which also comprises the closely related
species M. tuberculosis, M. bovis, M. microti, and M. canetti (21,
24). Since its first description in 1968 (3), M. africanum has
been found in several regions of Africa, where it represents up
to 60% of clinical strains obtained from patients with pulmo-
nary tuberculosis (7, 18, 19, 23).
Recent surveys show highly variable prevalences of M. afri-
canum in different African regions. For example, M. africanum
was found in approximately 5% of patients with tuberculosis in
the Ivory Coast and in at least 60% of patients in Guinea-
Bissau (2, 10). Most of the studies presented so far have ana-
lyzed small numbers of strains from different regions, and
systematic studies of prevalence and geographic distribution of
M. africanum are still infrequent.
In contrast to M. tuberculosis and M. bovis, M. africanum
strains show a higher variability of phenotypic attributes, com-
prising characteristics common to both M. tuberculosis and M.
bovis. This phenotypic heterogeneity of M. africanum compli-
cates its unequivocal identification and may lead to misclassi-
fication of clinical strains. According to their biochemical char-
acteristics, two major subgroups of M. africanum have been
described, corresponding to their geographic origin in western
(subtype I) or eastern (subtype II) Africa. Numerical analyses
of biochemical characteristics revealed that M. africanum sub-
type I is more closely related to M. bovis, whereas subtype II
more closely resembles M. tuberculosis (5).
In our recent work, we determined diagnostic criteria, in-
cluding phenotypic and biochemical characteristics as well as
results of the molecular spoligotyping technique, that permit
the accurate differentiation of M. africanum subtypes I and II
(15). Spoligotyping is a rapid molecular test based on the
detection of various nonrepetitive spacer sequences located
between small repetitive units (direct repeats) in the direct
repeat locus of M. tuberculosis complex strains. However, spo-
ligotyping does not allow differentiation of M. africanum sub-
types from M. tuberculosis without additional routine labora-
tory procedures. This drawback led us to evaluate the
usefulness of gyrB DNA sequence polymorphisms as a further
molecular marker for differentiation of the species of the M.
tuberculosis complex (16).
* Corresponding author. Mailing address: National Reference Cen-
ter for Mycobacteria, Forschungszentrum Borstel, Parkallee 18,
D-23845 Borstel, Germany. Phone: (49) 4537 188658. Fax: (49) 4537
188311. E-mail: email@example.com.
We established a rapid PCR-restriction fragment length
polymorphism (RFLP) assay that allows the differentiation of
M. bovis subsp. bovis, M. bovis subsp. caprae, and M. microti as
well as the clear identification of M. africanum subtype I
strains. M. africanum subtype II and M. tuberculosis, however,
displayed identical gyrB DNA sequences and were indistin-
guishable in this analysis. Thus, differentiation of M. africanum
subtype II from M. tuberculosis continues to be based on phe-
notypic characteristics such as growth on bromocresol purple
medium (16). This finding, in accordance with previous reports
(5, 7), reiterates the close relationship between M. africanum
subtype II and M. tuberculosis and questions the taxonomic
status and phylogenetic position of M. africanum subtype II
within the M. tuberculosis complex.
The present study investigated the population structure of
M. tuberculosis complex strains obtained from patients with
tuberculosis who were recruited at the Mulago Hospital in
Kampala, Uganda, because the presence of M. africanum sub-
type II in limited study populations in Uganda has been re-
ported (7, 15). The aim of this study was to assess the genetic
relationship of M. tuberculosis and M. africanum subtype II in
order to verify a genetic basis for this repeatedly described M.
africanum subtype. Furthermore, the intention was to analyze
the prevalence of M. africanum in a large study group of well-
defined patients with tuberculosis. Based on the results ob-
tained, we hypothesize that the majority of M. tuberculosis
complex strains in Uganda belong to M. africanum subtype II
and that this subtype contains two distinct genotypes (Uganda
I and Uganda II) that may represent two closely related phy-
logenetic branches within the M. tuberculosis complex.
MATERIALS AND METHODS
Strains analyzed. A total of 234 M. tuberculosis complex strains isolated from
sputum samples that had been collected between 1995 and 1997 were analyzed.
Sputum samples were obtained from 234 adult patients with newly diagnosed
initial episodes of sputum smear-positive pulmonary tuberculosis from the Na-
tional Tuberculosis/Leprosy Program (NTLP) clinic (the largest tuberculosis
clinic in Uganda) at Mulago Hospital. After decontamination (N-acetyl-L-cys-
teine–sodium hydroxide), the sediment was inoculated on Lo ¨wenstein-Jensen
slants at 37°C as described elsewhere (12). Nearly all patients were ambulatory,
and a few were hospitalized. The households of these patients were subsequently
studied in the context of a household contact study. All strains were confirmed
as M. tuberculosis complex by spoligotyping (11).
Biochemical tests and susceptibility testing. Biochemical analysis for differ-
entiation within the M. tuberculosis complex included colony morphology, nitrate
reduction on modified Dubos broth, niacin accumulation test (INH test strips;
Difco, Detroit, Mich.), growth in the presence of thiophen-2-carboxylic acid
hydrazide (TCH, 1 ?g/ml), catalase activity at room temperature, and growth
characteristics on Lebek and on bromocresol purple medium, as described pre-
viously (12, 15).
IS6110 RFLP and spoligotyping analysis. Extraction of DNA from mycobac-
terial strains and DNA fingerprinting with IS6110 as a probe was performed
according to a standardized protocol described elsewhere (14, 20). Spoligotypes
and IS6110 fingerprint patterns of mycobacterial strains were analyzed with the
Gelcompar software (Windows NT, version 4.2; Applied Maths, Kortrijk, Bel-
gium) as described previously (14, 20). Clusters were defined as groups of
patients with bacterial strains showing identical spoligotype and/or IS6110 RFLP
patterns. Spoligotyping of strains was performed as described by Kamerbeek et
Quality control. The laboratory participated twice a year in external profi-
ciency testing (national and international). For all test panels, negative and
positive internal controls were included.
PvuII-digested total DNA of reference strain Mt.14323 (obtained from the
National Institute of Public Health and Environmental Protection, Bilthoven,
The Netherlands) was used in each Southern blot experiment as an external size
standard and for quality control and quality assurance of IS6110 RFLP experi-
ments. M. tuberculosis H37 (ATCC 27294) and M. bovis BCG (ATCC 27289)
were used as control strains in each spoligotype experiment performed. The
accuracy of each experiment and the normalization procedure performed were
controlled by comparing the IS6110 fingerprint patterns or spoligotype patterns
of reference strains present on each autoradiogram with those stored in the
In this study, M. tuberculosis complex strains from 234 pa-
tients with pulmonary tuberculosis were investigated by bio-
chemical tests, spoligotyping, and IS6110 RFLP analysis.
Species differentiation. According to their phenotypic char-
acteristics and biochemical test results, the 234 M. tuberculosis
complex strains were classified as 157 M. africanum strains, 76
M. tuberculosis strains, and 1 M. bovis strain. M. africanum was
identified on the basis of its colony morphology on Lo ¨wen-
stein-Jensen slants (dysgonic growth), microaerophilic growth
on Lebek medium (Lebek is a semisolid medium which can be
used to test the oxygen preference of mycobacterial strains),
low catalase reaction, and lack of induction of a color change
of bromocresol purple medium (pH-dependent change of
color from blue to yellow, e.g., in the case of M. tuberculosis
strains) (15). The biochemical characteristics of the strains
analyzed and of the type strains M. tuberculosis H37 (ATCC
27294), M. bovis (ATCC 19210), and M. africanum (ATCC
25420) are summarized in Table 1. All strains classified as M.
africanum showed resistance to TCH and were therefore dif-
ferentiated as M. africanum subtype II, which more closely
resembles M. tuberculosis (15).
Spoligotyping analysis. All strains were analyzed by spoli-
gotyping, and the patterns obtained were digitized and ana-
lyzed for similarity with the Dice coefficient (position toler-
ance, 1.0%).A dendrogram
unweighted pairgroup method
(UPGMA) for the M. tuberculosis and M. africanum strains.
The only M. bovis strain found showed the typical absence of
spacers 39 to 43 and the presence of spacers 3 to 16. This strain
could thus be identified as M. bovis subsp. bovis (pyrazinamide
In accordance with previous results (15), all M. africanum
subtype II strains showed no hybridization to the M. bovis-
derived spacers 33 to 36. Although M. africanum strains were
found mainly in two large groups of strains with similar spoli-
gotype patterns, differentiation from M. tuberculosis could not
be achieved in the dendrogram based on spoligotyping results
(Fig. 1). Several M. tuberculosis and M. africanum strains
showed only minor differences (one or two spacers) in the
spoligotype patterns (see groups on top of dendrogram in Fig.
1 and 2), which resulted in adjacent positions in the dendro-
Considering the clustering rate among the strains analyzed,
unique spoligotype patterns were found in only 40 (17%) of the
233 M. tuberculosis and M. africanum strains. The remaining
193 (83%) strains had a spoligotype pattern identical to that of
one or more of the M. tuberculosis or the M. africanum strains.
Among the M. tuberculosis strains, 57 (74%) were grouped into
17 clusters with identical spoligotype patterns. Each of the
clusters contained between two and eight strains. Of the 157 M.
africanum strains, 136 (87%) were grouped into 18 clusters,
VOL. 40, 2002M. AFRICANUM GENOTYPE FAMILIES3399
with 2 to 37 strains per cluster. This indicated high genetic
homogeneity among the M. africanum strains, an observation
that was further supported by the inclusion of more than 50%
of all M. africanum strains within the three largest clusters,
which contained 17, 29, and 37 strains (Fig. 1).
IS6110 RFLP analysis. In order to further analyze the ge-
netic relationship of the strains, DNA fingerprinting with
IS6110 as a probe was performed. The IS6110 RFLP patterns
of the M. tuberculosis and M. africanum subtype II strains were
analyzed for similarity with the Dice coefficient (position tol-
erance, 1.3%), and a dendrogram was calculated, which is
shown in Fig. 3.
In contrast to the dendrogram that was based on the spoli-
gotyping results (Fig. 1), RFLP analysis grouped the M. africa-
num subtype II strains into two closely related genotype fam-
ilies (Uganda I [n ? 55] and Uganda II [n ? 102]). RFLP
patterns among strains of these genotypes showed a similarity
of at least 75% and were distinctly separated from the M.
tuberculosis strains (Fig. 3). Even though M. africanum subtype
II and M. tuberculosis strains showed very similar spoligotype
patterns, they could be clearly distinguished by IS6110 RFLP
typing (Fig. 2b). Overall, the M. tuberculosis IS6110 RFLP
patterns were more variable than those of M. africanum strains,
as was depicted by large differences in the IS6110 copy num-
bers, ranging from only 1 to 17 per strain (Fig. 3). In contrast,
the IS6110 RFLP patterns among the M. africanum strains
were more homogeneous, with copy numbers ranging between
approximately 14 and 20 IS6110 bands per strain.
Separate evaluation of the spoligotype patterns of strains of
the M. africanum subtype II genotypes Uganda I and II showed
that the absence of spacer 40 is an obvious marker of both
genotypes (Fig. 4). In addition, all strains of genotype Uganda
I lack spacer 43. In contrast to these findings in M. africanum
subtype II, most of the M. tuberculosis strains (44 of 57) showed
hybridization to spacers 40 and 43 (data not shown). All of the
M. tuberculosis strains, which lack one or both of spacers 40
and 43, were clearly distinguishable from the M. africanum
strains by their IS6110 RFLP patterns. This further confirmed
our species differentiation based on phenotypic and biochem-
ical characteristics. Thus, lack of spacers 40 and 43 is not an
exclusive marker of M. africanum subtype II but might repre-
sent a useful additional criterion for M. africanum subtype
identification in combination with biochemical test results.
When the results from spoligotyping and IS6110 RFLP anal-
ysis were combined, rates of strains in clusters with identical
spoligotype and IS6110 RFLP patterns were reduced to 47%
(110 of 234). Among the 76 M. tuberculosis strains in this study,
37 strains (49%) showed identical IS6110 and spoligotype pat-
terns and were grouped into 13 clusters containing two to
seven strains each. Among the 157 M. africanum strains, 72
(46%) were grouped in 28 such clusters with two to seven
strains per cluster that consisted mainly (75%) of pairs of
strains. Although the fingerprint polymorphism detected by
spoligotyping was lower than that of IS6110 RFLP typing, an
overall correlation between the two techniques was observed.
All strains with identical IS6110 RFLP patterns also displayed
identical or very similar spoligotype patterns (data not shown),
confirming the genetic relationship of the strains determined
by IS6110 RFLP typing. The accurate classification of the M.
africanum subtype II genotypes Uganda I and II by IS6110
RFLP typing was further supported by the shared characteris-
tic spoligotype features of the strains.
This study systematically analyzed the population structure
of M. tuberculosis complex strains isolated between 1995 and
1997 from tuberculosis patients living in Kampala, Uganda.
Sixty-seven percent of the strains were M. africanum subtype
II, suggesting that the main cause of human tuberculosis in
Kampala is M. africanum subtype II. We further demonstrated
that M. africanum subtype II strains from Kampala, Uganda,
belong to two closely related genotypes (Uganda I and II) that
share specific spoligotyping characteristics and are clustered
into two IS6110 RFLP strain families.
Geographic variants of M. africanum had initially been de-
scribed in studies by David et al. (5) and were more recently
noted by Haas et al. (7). The results of these studies indicated
TABLE 1. Biochemical characteristics of type strains M. tuberculosis H37 (ATCC 27294), M. bovis (ATCC 19210), and M. africanum
(ATCC 25420) and the strains analyzeda
Organism and group
(no. of strains)
Test result (% of isolates)
Growth in presence of:
Change of color of
M. tuberculosis H37
M. bovis (ATCC 19210)
M. tuberculosis (76)
M. bovis (1)
M. africanum subtype II
Uganda I (55)
M. africanum subtype II
Uganda II (102)
? (7), ? (87), ? (6)
? (97), ? (3)c
3.0 ? 1.2
0.5 ? 0.2
? (5), ? (91), ? (4)
? (100) Dysgonic (100)
? (100) Mircroaerophilic
0.5 ? 0.2
aAbbreviations and symbols: ?, positive test result; ?, negative test result; ?, weakly positive; PZA, pyrazinamide. One M. tuberculosis isolate, three M. africanum
subtype II Uganda I, and two Uganda II isolates were resistant to isoniazid and cross-resistant to TCH.
bCentimeters of foam production at room temperature.
cStrains were resistant to isoniazid, streptomycin, and pyrazinamide.
3400 NIEMANN ET AL.J. CLIN. MICROBIOL.
FIG. 1. Spoligotype patterns of the 233 M. tuberculosis (darker shading) and M. africanum subtype II (lighter shading) strains. Banding patterns
are ordered by similarity in a dendrogram. The position of each spoligotyping hybridization spot is normalized so that banding patterns of all strains
are mutually comparable. The scale depicts similarity of patterns calculated with the Dice coefficient and the UPGMA method.
that M. africanum subtype I predominantly originated from
West African countries, whereas M. africanum subtype II was
found predominantly in East Africa.
Systematic studies analyzing larger numbers of strains from
one study region are still rare, and their interpretation is com-
plicated by the lack of clear characteristics for the differenti-
ation of M. africanum and its two subtypes. In our own studies
(15, 16), we analyzed a collection of M. africanum strains from
western and eastern African countries and found criteria which
allowed the accurate differentiation of the two M. africanum
subtypes in accordance with the geographic origin of the
strains. The main criteria for the differentiation of the two M.
africanum subtypes are susceptibility to TCH, hybridization to
at least two of the M. bovis-derived spacers 33 to 36, and a
specific gyrB DNA sequence for subtype I and resistance to
TCH and lack of hybridization to spacers 33 to 36 for subtype
Recent studies in West African countries have shown M.
africanum prevalence rates with high regional variability, rang-
ing from approximately 5% in the Ivory Coast (2) to 61%
(biovars 2, 3, and 4) in Guinea-Bissau (10). Because of their
susceptibility to TCH, the M. africanum strains in these two
studies were confirmed as M. africanum subtype I. Considering
the IS6110 RFLP patterns obtained, an obvious characteristic
of the M. africanum subtype I strains in both studies was the
presence of an intermediate or small number of IS6110 bands,
which has also been observed by Haas et al. (7) for M. africa-
num subtype I. In accordance with our previous results (15),
the spoligotyping analysis performed by Ka ¨llenius et al. (10)
confirmed that M. africanum subtype I strains are character-
ized by a specific spoligotype pattern which is intermediate
between those of M. bovis and M. tuberculosis (hybridization to
spacers 33 to 36 as well as to spacers 39 to 43).
This typical genotype, the combination of an intermediate
spoligotype pattern together with a small number of IS6110
bands, was further observed in a very recent study by Viana-
Niero and coworkers (23), who analyzed a collection of M.
africanum strains from several West African countries. All
these studies verify the presence of M. africanum subtype I in
West Africa, which is characterized by certain phenotypic
properties as well as a characteristic spoligotype and IS6110
RFLP patterns. Our previous results indicate that this subtype
may be identified by a specific gyrB DNA sequence, but this
finding still remains to be analyzed for a larger number of M.
africanum subtype I strains.
In accordance with our preliminary observations during a
study of 49 M. tuberculosis complex strains from Kampala (18),
the present study confirms that M. africanum, and particularly
its subtype II, represents a major cause of human tuberculosis
in this African region. This finding is in contrast to the results
obtained in a study performed from 1992 to 1993 in the region
of Buluba, Uganda, in which only 16% of the strains analyzed
were differentiated as M. africanum (19). These contrasting
results may be simply explained by a variable prevalence of M.
africanum subtype II in different regions of Uganda or differ-
ences in the sampling procedures applied. A further possible
reason is an increase in the prevalence of M. africanum subtype
II in recent years, which might have resulted from other con-
tributing factors, such as the increased rate of human immu-
nodeficiency virus type I (HIV-1) in the Ugandan population.
In contrast to M. africanum subtype I, subtype II strains were
resistant to TCH and showed no hybridization to spoligotype
spacers 33 to 36. The most striking finding of this investigation
is that the M. africanum subtype II strains from Kampala,
Uganda, clustered in two closely related genotypes, which
could be clearly separated from the M. tuberculosis strains
FIG. 2. Spoligotype (a) and IS6110 RFLP (b) patterns of four pairs of M. tuberculosis and M. africanum subtype II strains. M. tuberculosis and
M. africanum subtype II strains had very similar spoligotype patterns but were clearly separated by IS6110 RFLP typing.
3402NIEMANN ET AL.J. CLIN. MICROBIOL.
FIG. 3. IS6110 DNA fingerprint patterns of the 233 M. tuberculosis (darker shading) and M. africanum subtype II (lighter shading) strains.
Banding patterns are ordered by similarity in a dendrogram. M. africanum subtype II strains were clustered in two closely related strain families
(genotypes Uganda I and II) and were clearly separated from the M. tuberculosis strains.
analyzed by their RFLP pattern. Within both subtype II geno-
types, the strains showed very homogenous IS6110 RFLP pat-
terns, but with a large number of IS6110 copies per strain
(approximately 16 to approximately 20), clearly differentiating
these strains from M. africanum subtype I. A further charac-
teristic of genotypes Uganda I and II is the absence of spoli-
gotype spacer 40 and also the absence of spacer 43 in strains of
Uganda I. These results indicate that the strains of these two
genotypes are closely related and may have diverged from an
M. tuberculosis-like ancestor.
In contrast to the homogenous IS6110 RFLP patterns ob-
served for the M. africanum subtype II strains, M. tuberculosis
strains from Kampala showed a high variability of IS6110
banding patterns as well of IS6110 copy numbers. One can
speculate that the differences in homogeneity patterns between
the M. africanum subtype II strains and the M. tuberculosis
strains result from closely related indigenous mycobacterial
populations in the region of Kampala and a high degree of
influx from abroad resulting in highly diverse IS6110 RFLP
patterns, respectively. In accordance, Daniel (4) presented an
interesting study on the early history of tuberculosis in central
Africa, which demonstrates that tuberculosis was present in
central East Africa at the time of the earliest European entries
in the region of Kampala.
The clustering rate obtained by the combination of spoligo-
typing and IS6110 RFLP analysis was similar for M. tubercu-
losis and M. africanum subtype II (46% and 49%, respectively)
and indicates a high rate of recent human-to-human transmis-
sion for strains of both species. Similar clustering rates have
recently been measured by IS6110 typing in other African
countries such as Botswana (42% ), Namibia (47% ),
and South Africa (45% ). Only slightly lower or compara-
ble clustering rates have been reported from other areas of the
world with a low incidence of tuberculosis, such as New York
(37% ) and The Netherlands (47% ). This somewhat
surprising observation may be due to short study periods or
limited numbers of patients with pulmonary tuberculosis that
were analyzed in these studies.
Considering the discriminatory power of both typing meth-
ods, the results in this study clearly indicate that spoligotyping
alone is not well suited for differentiation of M. tuberculosis
complex strains on the strain level in this high-incidence com-
munity. Also, spoligotyping did not facilitate an accurate anal-
ysis of the genetic relationship of the strains, as M. tuberculosis
and M. africanum strains with similar spoligotype patterns were
clearly separated by their IS6110 RFLP patterns and biochem-
ical characteristics. In contrast to IS6110 RFLP patterns, for
which modifications appear to occur by changes of single bands
as a function of time (6), large alterations of spoligotype pat-
terns seem to be possible in relatively short time periods.
FIG. 4. Representative spoligotype patterns of M. africanum subtype II strains of genotypes Uganda I and II (C and D) compared to spoligotype
patterns of type strains M. tuberculosis H37 (ATCC 27294), M. bovis (ATCC 19210), M. bovis BCG (ATCC 27289), M. africanum (ATCC 25420),
and a collection of M. africanum subtype I (A) and M. africanum subtype II (B) isolates from our previous work (15). In contrast to M. bovis, all
M. africanum strains showed hybridization to several of the spacers 39 to 43 which were derived from the direct repeat (DR) region of M.
tuberculosis H37. In the case of M. africanum subtype II, no hybridization was observed to the M. bovis BCG-derived spacers 33 to 36, whereas M.
africanum subtype I isolates as well as the M. africanum type strain (ATCC 25420) showed hybridization to at least two of these spacers. All M.
africanum subtype II strains showed a characteristic lack of hybridization to spacer 40. Strains of genotype Uganda I lack spacer 43 in addition
(arrows). In contrast, M. africanum subtype I strains lack spacer 39.
3404 NIEMANN ET AL.J. CLIN. MICROBIOL.
Alterations of spoligotype patterns thus do not necessarily
represent the overall rate of change of the genome. Hence,
spoligotyping appears not to be a useful method for determi-
nation of the genomic relatedness of M. tuberculosis complex
strains for phylogenetic purposes.
In conclusion, the results presented here and in earlier stud-
ies clearly confirm the existence of M. africanum subtype I
(West Africa) and subtype II (East Africa, Uganda), which
have been previously proposed by numerical analysis of the
phenotypic characteristics. M. africanum subtype I and M. af-
ricanum subtype II represent two unique phylogenetic
branches within the M. tuberculosis complex that originates in
West and East Africa, respectively. Both M. africanum sub-
types have been verified to represent a high portion of M.
tuberculosis complex strains in certain regions of Africa, as we
confirmed that more than 60% of the tuberculosis cases in
Kampala are due to M. africanum subtype II and not to M.
A high prevalence of M. africanum strains in human tuber-
culosis in Africa might have important implications for tuber-
culosis control, considering the enormous burden of tubercu-
losis and HIV-1/AIDS in Africa. Based on the clustering rates
observed in our study, no difference in transmission patterns
between M. africanum subtype II and M. tuberculosis could be
verified. A preliminary result obtained by analyzing 13 patients
indicated that presentations and responses to short-course che-
motherapy are comparable for M. africanum and M. tubercu-
A more detailed analysis of the clinical presentation, therapy
outcome, and epidemiological characteristics of more than 300
cases of M. africanum- and M. tuberculosis-induced tuberculo-
sis that includes the strains presented in this study is in prep-
aration. Further studies in larger study populations will be
needed for more detailed analyses of the regional prevalence
and transmission of M. africanum-induced tuberculosis, espe-
cially in the context of factors such as coinfection with HIV-1.
We thank I. Radzio, B. Schlu ¨ter, P. Vock, and A. Zyzik, Borstel,
Germany, for excellent technical assistance.
1. Alland, D., G. E. Kalkut, A. R. Moss, R. A. McAdam, J. A. Hahn, W.
Bosworth, E. Druckner, and B. R. Bloom. 1994. Transmission of tuberculosis
in New York City. An analysis by DNA fingerprinting and conventional
epidemiological methods. N. Engl. J. Med. 330:1710–1716.
2. Bonard, D., P. Msellati, L. Rigouts, P. Combe, D. Coulibaly, I. M. Coulibaly,
and F. Portaels. 2000. What is the meaning of repeated isolation of Myco-
bacterium africanum? Int. J. Tuberc. Lung Dis. 4:1176–1180.
3. Castets, M., H. Boisvert, F. Grumbach, M. Brunel, and N. Rist. 1968.
Tuberculosis bacilli of the African type: preliminary note. Rev. Tuberc.
4. Daniel, T. M. 1998. The early history of tuberculosis in central East Africa:
insights from the clinical records of the first twenty years of Mengo Hospital
and review of relevant literature. Int. J. Tuberc. Lung Dis. 2:784–790.
5. David, H. L., M. T. Jahan, A. Jumin, J. Grandry, and E. H. Lehman. 1978.
Numerical taxonomy analysis of Mycobacterium africanum. Int. J. Syst. Bac-
6. De Boer, A. S., M. W. Borgdorff, P. E. de Haas, N. J. Nagelkerke, J. D. A. van
Embden, and D. van Soolingen. 1999. Analysis of rate of change of IS6110
RFLP patterns of Mycobacterium tuberculosis based on serial patient strains.
J. Infect. Dis. 180:1238–1244.
7. Haas, W. H., G. Bretzel, B. Amthor, K. Schilke, G. Krommes, S. Ru ¨sch-
Gerdes, V. Sticht-Groh, and H. J. Bremer. 1997. Comparison of DNA fin-
gerprint patterns of strains of Mycobacterium africanum from east and west
Africa. J. Clin. Microbiol. 35:663–666.
8. Haas, W. H., G. Engelmann, B. Amthor, S. Shyamba, F. Mugala, M. Felten,
M. Rabbow, M. Leichsenring, O. J. Oosthuizen, and H. J. Bremer. 1999.
Transmission dynamics of tuberculosis in a high-incidence country: prospec-
tive analysis by PCR DNA fingerprinting. J. Clin. Microbiol. 37:3975–3979.
9. Joloba, M. L., J. L. Johnson, A. Namale, A. Morrissey, A. E. Assegghai, S.
Rusch-Gerdes, R. D. Mugerwa, J. J. Ellner, and K. D. Eisenach. 2001.
Quantitative bacillary response to treatment in Mycobacterium tuberculosis
infected and M. africanum infected adults with pulmonary tuberculosis. Int.
J. Tuberc. Lung Dis. 5:579–582.
10. Ka ¨llenius, G., T. Koivula, S. Ghebremichael, S. E. Hoffner, R. Norberg, E.
Svensson, F. Dias, B. L. Marklund, and S. B. Svenson. 1999. Evolution and
clonal traits of Mycobacterium tuberculosis complex in Guinea-Bissau. J. Clin.
11. Kamerbeek, J., L. Schouls, A. Kolk, M. van Agterveld, D. van Soolingen, S.
Kuijper, A. Bunschoten, H. Molhuizen, R. Shaw, M. Goyal, and J. D. A. van
Embden. 1997. Simultaneous detection and strain differentiation of Myco-
bacterium tuberculosis for diagnosis and epidemiology. J. Clin. Microbiol.
12. Kent, P. T., and G. P. Kubica. 1985. Public health mycobacteriology: a guide
for the level III laboratory. U.S. Department of Health and Human Services,
Centers for Disease Control, Atlanta, Ga.
13. Lockman, S., J. D. Sheppard, C. R. Braden, M. J. Mwasekaga, C. L. Wood-
ley, T. A. Kenyon, N. J. Binkin, M. Steinman, F. Montsho, M. Kesupile-Reed,
C. Hirschfeldt, M. Notha, T. Moeti, and J. W. Tappero. 2001. Molecular and
conventional epidemiology of Mycobacterium tuberculosis in Botswana: a
population-based prospective study of 301 pulmonary tuberculosis patients.
J. Clin. Microbiol. 39:1042–1047.
14. Niemann, S., E. Richter, and S. Ru ¨sch-Gerdes. 1999. Stability of Mycobac-
terium tuberculosis IS6110 restriction fragment length polymorphism patterns
and spoligotypes determined by analyzing serial strains from patients with
drug-resistant tuberculosis. J. Clin. Microbiol. 37:409–412.
15. Niemann, S., E. Richter, and S. Ru ¨sch-Gerdes. 2000. Differentiation among
members of the Mycobacterium tuberculosis complex by molecular and bio-
chemical features: evidence for two pyrazinamide-susceptible subtypes of M.
bovis. J. Clin. Microbiol. 38:152–157.
16. Niemann, S., D. Harmsen, S. Ru ¨sch-Gerdes, and E. Richter. 2000. Differ-
entiation of clinical Mycobacterium tuberculosis complex strains by gyrB DNA
sequence polymorphism analysis. J. Clin. Microbiol. 38:3231–3234.
17. Niemann, S., E. Richter, and S. Ru ¨sch-Gerdes. 2002. Biochemical and ge-
netic evidence for the transfer of Mycobacterium tuberculosis subsp. caprae
Aranaz et al. 1999 to the species Mycobacterium bovis Karlson and Lessel
1970 (Approved Lists 1980) as Mycobacterium bovis subsp. caprae comb. nov.
Int. J. Syst. Evol. Microbiol. 52:433–436.
18. Schwander, S., S. Ru ¨sch-Gerdes, A. Mateega, T. Lutalo, S. Tugume, C.
Kityo, R. Rubaramira, P. Mugyenyi, A. Okwera, R. Mugerwa T. Aisu, R.
Moser, K. Ochen, B. M’Bonye, and M. Dietrich. 1995. A pilot study of
antituberculosis combinations comparing rifabutin with rifampicin in the
treatment of HIV-1 associated tuberculosis. A single-blind randomized eval-
uation in Ugandan patients with HIV-1 infection and pulmonary tuberculo-
sis. Tuberc. Lung Dis. 76:210–218.
19. Sticht-Groh, V., G. Bretzel, S. Ru ¨sch-Gerdes, S. Bwire, and H. J. S. Kawuma.
1994. M. africanum strains isolated in East Africa, Uganda. Tuber. Lung Dis.
20. Van Embden, J. D. A., M. D. Cave, J. T. Crawford, J. W. Dale, K. D.
Eisenach, B. Gicquel, P. Hermans, C. Martin, R. McAdam, T. M. Shinnick,
and P. M. Small. 1993. Strain identification of Mycobacterium tuberculosis by
DNA fingerprinting: recommendations for a standardized methodology.
J. Clin. Microbiol. 31:406–409.
21. Van Soolingen, D., T. Hoogenboezem, P. E. de Haas, P. W. Hermans, M. A.
Koedam, K. S. Teppema, P. J. Brennan, G. S. Besra, F. Portaels, J. Top,
L. M. Schouls, and J. D. A. van Embden. 1997. A novel pathogenic taxon of
the Mycobacterium tuberculosis complex, Canetti: characterization of an ex-
ceptional isolate from Africa. Int. J. Syst. Bacteriol. 47:1236–1245.
22. Van Soolingen, D., M. W. Borgdorff, P. E. de Haas, M. M. Sebek, J. Veen, M.
Dessens, K. Kremer, and J. D. A. van Embden. 1999. Molecular epidemiol-
ogy of tuberculosis in the Netherlands: a nationwide study from 1993 through
1997. J. Infect. Dis. 3:726–736.
23. Viana-Niero, C., C. Gutierrez, C. Sola, I. Filliol, F. Boulahbal, V. Vincent,
and N. Rastogi. 2001. Genetic diversity of Mycobacterium africanum clinical
strains based on IS6110-restriction fragment length polymorphism analysis,
spoligotyping, and variable number of tandem DNA repeats. J. Clin. Micro-
24. Wayne, L. G., and G. P. Kubica. 1986. The mycobacteria, p. 1435–1457. In
P. H. A. Sneath and J. G. Holt (ed.), Bergey’s manual of systematic bacte-
riology, vol. 2. The Williams Co., Baltimore, Md.
25. Wilkinson, D., M. Pillay, J. Crump, C. Lombard, G. R. Davies, and A. W.
Sturm. 1997. Molecular epidemiology and transmission dynamics of Myco-
bacterium tuberculosis in rural Africa. Trop. Med. Int. Health 2:747–753.
VOL. 40, 2002M. AFRICANUM GENOTYPE FAMILIES 3405