JOURNAL OF BACTERIOLOGY, Oct. 2009, p. 60676074 Vol. 191, No.

19
0021-9193/09/$08.000 doi:10.1128/JB.00762-09
Copyright 2009, American Society for Microbiology. All Rights Reserved.

Comparative Sequence Analysis of Mycobacterium leprae and the New

Leprosy-Causing Mycobacterium lepromatosis
Xiang Y. Han,1 Kurt C. Sizer,1 Erika J. Thompson,2 Juma Kabanja,3 Jun Li,3 Peter Hu,3
Laura Gomez-Valero,4 and Francisco J. Silva5,6*
Department of Laboratory Medicine,1 DNA Analysis Core Facility,2 and School of Health Sciences,3 The University of
Texas M. D. Anderson Cancer Center, Houston, Texas; Institut Pasteur, Unite de Biologie des Bacteries Intracellulaires,
Paris, France4; Institut Cavanilles de Biodiversitat i Biologia Evolutiva, Universitat de Valencia, Valencia,
Spain5; and CIBER en Epidemiologa y Salud Publica (CIBERESP), Madrid, Spain6
Received 12 June 2009/Accepted 20 July 2009

Mycobacterium lepromatosis is a newly discovered leprosy-causing organism. Preliminary phylogenetic anal-

ysis of its 16S rRNA gene and a few other gene segments revealed significant divergence from Mycobacterium
leprae, a well-known cause of leprosy, that justifies the status of M. lepromatosis as a new species. In this study
we analyzed the sequences of 20 genes and pseudogenes (22,814 nucleotides). Overall, the level of matching of
these sequences with M. leprae sequences was 90.9%, which substantiated the species-level difference; the levels
of matching for the 16S rRNA genes and 14 protein-encoding genes were 98.0% and 93.1%, respectively, but the
level of matching for five pseudogenes was only 79.1%. Five conserved protein-encoding genes were selected to
construct phylogenetic trees and to calculate the numbers of synonymous substitutions (dS values) and
nonsynonymous substitutions (dN values) in the two species. Robust phylogenetic trees constructed using
concatenated alignment of these genes placed M. lepromatosis and M. leprae in a tight cluster with long terminal
branches, implying that the divergence occurred long ago. The dS and dN values were also much higher than
those for other closest pairs of mycobacteria. The dS values were 14 to 28% of the dS values for M. leprae and
Mycobacterium tuberculosis, a more divergent pair of species. These results thus indicate that M. lepromatosis
and M. leprae diverged 10 million years ago. The M. lepromatosis pseudogenes analyzed that were also
pseudogenes in M. leprae showed nearly neutral evolution, and their relative ages were similar to those of M.
leprae pseudogenes, suggesting that they were pseudogenes before divergence. Taken together, the results
described above indicate that M. lepromatosis and M. leprae diverged from a common ancestor after the massive
gene inactivation event described previously for M. leprae.

Leprosy, one of the oldest human diseases, remains a sig-
nificant public health problem in many developing countries
(8). Mycobacterium leprae was the only known cause of leprosy
until recently, when a new mycobacterium, Mycobacterium lep-
romatosis, was found to be the cause of diffuse lepromatous
leprosy (DLL), a unique form of leprosy endemic in Mexico
and the Caribbean (17). The discovery of this new species may
provide an explanation for the clinical and geographic variabil-
ity of leprosy.
The initial phylogenetic analysis of M. lepromatosis was car-
ried out using the sequences of the 16S rRNA gene and seg-
ments of groEL, rpoB, and other genes (total, 4.99 kb) (17).
This study revealed significant sequence differences between
M. lepromatosis and all known Mycobacterium species and
placed M. lepromatosis closest to M. leprae. However, the se-
quence variation justified assigning a new species for the new
organism instead of classifying it as a variant of M. leprae. All
M. leprae strains collected worldwide have been found to be
clonal and to differ by only single-nucleotide polymorphism or
variable numbers of tandem repeats (24). Also, the genomes of
two M. leprae strains, strain TN from India (GenBank acces-
* Corresponding author. Mailing address: Institut Cavanilles de Biodi-
versitat i Biologia Evolutiva, Universitat de Valencia, Apartat 22085,
46071 Valencia, Spain. Phone: 34 963543650. Fax: 34 963543670. E-mail:
francisco.silva@uv.es.
Published ahead of print on 24 July 2009.
sion numbers AL583917 to AL583926) (4) and strain Br4923
from Brazil (GenBank accession number FM211192) (N.
Honore et al., unpublished data) share 99.98% identity.
Like M. leprae, M. lepromatosis has not been cultivated on
artificial media. In addition, our previous study also showed
other similarities between these organisms, such as degenera-
tion of mmaA3 into a pseudogene, the presence of unique
AT-rich inserted sequences in the 16S rRNA gene, identical
six-base tandem repeats in rpoT, similar GC contents, and
great evolutionary distance from other mycobacteria (17).
The M. leprae genome (3.3 Mb) is much smaller than the
Mycobacterium tuberculosis genome (4.4 Mb) (3, 4). More in-
triguingly, the M. leprae genome has undergone reductive evo-
lution; 40% of the genes are inactivated (4), and 50% of
the genes of the last common ancestor of M. leprae and M.
tuberculosis have been lost (13). On the other hand, the M.
leprae genome has been far more stable than the M. tubercu-
losis genome, and the worldwide clonality of the M. leprae
strains paralleled the global spread of M. leprae strains that
occurred via human activity and migration during the last
100,000 years (24). Recently, by comparing the genomes of
M. leprae and M. tuberculosis and by analyzing the ages of the
M. leprae pseudogenes, Gomez-Valero et al. (13) estimated
that a massive gene inactivation event took place in the M.
leprae genome in the last 20 million years.
The discovery of M. lepromatosis and its differences from M.
leprae make it relevant for further study for diagnosis, treat-

6067
6068 HAN ET AL.
ment, and prevention of DLL. Likewise, the many similarities
between these two organisms prompted questions about their
evolutionary histories and about how M. lepromatosis became
endemic mainly in Mexico, while M. leprae occurs worldwide.
In this study, we extended and refined our previous phyloge-
netic study by determining and analyzing the sequences of 20
genes and pseudogenes of M. lepromatosis. Our findings solid-
ified the phylogeny of this new organism and provided new
insights into the history of pseudogenes.
MATERIALS AND METHODS
DNA isolation and sequencing. DNA was extracted from the autopsy liver of
a patient who died of DLL (17). Briefly, approximately 200 mg of freshly frozen
infected liver tissue was homogenized and processed with N-acetyl-L-cysteine and
NaOH to enrich the mycobacterium, decontaminate the preparation, and disrupt
the host cells, the usual procedure for mycobacterial culture. After essentially
pure mycobacterial cells were obtained, a commercial DNA extraction solution
(PrepMan Ultra; Applied Biosystems, CA) was added, and the mixture was
boiled to disrupt the bacterial cells. The supernatant contained the released
DNA and was used for PCRs.
Many sets of PCR primers were designed to target various genes and pseu-
dogenes of M. lepromatosis. Without knowledge of the specific sequences of this
organism, the primers were designed using the conserved sequences of the
corresponding genes of M. leprae and M. tuberculosis after alignment. For a long
gene, multiple sets of forward and reverse primers were designed, with each set
amplifying 500 to 600 bp, with overlaps. The PCR amplicons were sequenced
directly using an ABI sequencer. All sequences were matched with GenBank
data and inspected visually to ensure accuracy.
The full-length rpoB gene of Mycobacterium haemophilum ATCC 29548T was
also sequenced for phylogenetic analysis.
Retrieval of other sequences. Gene and pseudogene sequences of M. lepro-
matosis were used to extract putative orthologous genes from completely se-
quenced mycobacterial genomes. The extraction was performed using the Mi-
crobial Genome Database for Comparative Analysis (30). The genomes used
included those of Mycobacterium abscessus ATCC 19977T, Mycobacterium avium
strain 104, M. avium subsp. paratuberculosis, Mycobacterium bovis BCG Pasteur
1173P2, Mycobacterium gilvum, M. leprae, Mycobacterium marinum M, Mycobac-
terium smegmatis, Mycobacterium sp. strain JLS, Mycobacterium sp. strain KMS,
Mycobacterium sp. strain MCS, M. tuberculosis CDC1551, Mycobacterium ulcer-
ans, and Mycobacterium vanbaalenii. Because there were six genomes of M.
tuberculosis complex organisms in the database and all of them have essentially
identical sequences, only two organisms, M. bovis BCG Pasteur 1173P2 and M.
tuberculosis CDC1551, were used.
Sequence alignment. The MAFFT program (18) was used to align nucleotide
sequences of genes and pseudogenes. The ClustalW program implemented in
MEGA 3.1 (19) was used to align the amino acid sequences of protein-encoding
genes for the corresponding nucleotide alignments. Each alignment was visually
inspected to detect short misalignments or poorly aligned regions. Doubtful
regions, usually at both ends, were trimmed.
To calculate the numbers of synonymous (dS) and nonsynonymous (dN) nu-
cleotide substitutions per site in five M. lepromatosis pseudogenes (citA, gnd2,
icd1, lld1, and mmaA3), alignment with corresponding M. leprae pseudogenes
was performed using the following parameters: (i) the alignment length corre-
sponded to the available length of each M. lepromatosis pseudogene; (ii) to
maintain continuous reading frames, like those in protein-encoding genes, the
nucleotides at triplet codon positions covering any indel in either organism were
not counted; and (iii) nucleotide positions at which there were stop codons in
either organism were removed. The final lengths of the aligned regions, with
15% reduction of the original alignments, were 918, 852, 552, 483, and 540
nucleotides for citA, gnd2, icd1, lld1, and mmaA3, respectively.
Phylogenetics and relative rate tests. To determine the relationship of M.
lepromatosis to other mycobacteria, phylogenetic trees were constructed using a
concatenated alignment of five conserved genes. These genes, rpoB, ligA, groEL,
gnd1, and bfrA, were chosen from 14 protein-encoding genes based on the
following criteria: functional importance or conservation, the presence of or-
thologous genes in the 17 completely sequenced mycobacterial genomes, full
length or nearly full length, and alignment accuracy. In particular, rpoB and
groEL have frequently been used in phylogenetic studies (6). The alignments
were also visually inspected to trim more variably aligned ends (on average, 9%
of the alignments). The final lengths of the rpoB, ligA, groEL, gnd1, and bfrA
J. BACTERIOL.
genes were 3,429, 2,013, 1,575, 1,125, and 405 nucleotides, respectively, and there
was total concatenation of 8,547 nucleotide sites.
A maximum likelihood phylogenetic analysis was performed with the concat-
enated alignment using the program PHYML (16) with 100 bootstrap replicates
and the HKY model. An additional neighbor-joining phylogeny was also ob-
tained with the program MEGA 3.1 (19) with the Tamura-Nei model and 1,000
bootstrap replicates. To include M. haemophilum (without genome data) in the
analysis, additional phylogenies with the same conditions were constructed using
the full-length rpoB gene of this organism. Tajimas relative rate tests (28) were
implemented in MEGA (19). Both M. leprae and M. lepromatosis branches were
compared with M. haemophilum using M. tuberculosis CDC1551 as the outgroup.
The analyses included the first and second codon positions (usually associated
with nonsynonymous substitutions) and the third codon positions (usually result-
ing in synonymous substitutions due to codon wobbling).
Estimation of dN and dS values. The dS and dN values, along with standard
errors, were estimated for gene-gene, gene-pseudogene, or pseudogene-pseudo-
gene pairs using yn00 from PAML (32). When a pseudogene was involved, dS
and dN values were used only for comparison, and no protein-encoding function
of the pseudogene was indicated.
To estimate the relative age of an M. lepromatosis pseudogene, the formula for
the p parameter developed for M. leprae (13) was used:
p dNilps fdNit/dSilps fdNit
where dNilps is the dN value for the ancestor (i) of M. tuberculosis and M. leprae
and the present pseudogene (lps) of M. leprae (or M. lepromatosis), dNit is the dN
value for the same ancestor (i) and the gene of M. tuberculosis, and f(dNit) is a
function that makes a small correction for the expected number of substitutions
in the lineage as if the pseudogene had not formed and is estimated from the
complete set of pairs of orthologous genes for M. tuberculosis and M. leprae. This
formula is based on the fact that the dN or dS value for a gene-pseudogene pair
(for example, an M. tuberculosis gene and an M. leprae pseudogene) is the sum for
the period of evolution as a gene (for the complete lineage of M. tuberculosis and
part of the lineage of M. leprae) and the period of evolution as a pseudogene
(which depends on the moment of inactivation in the M. leprae lineage). It also
takes into consideration the fact that the rates of evolution for nonsynonymous
changes are gene specific and that to determine these rates in M. leprae (or M.
lepromatosis), the dNilps and dNit for each gene had to be compared. Because of
the lack of a data set for M. lepromatosis, the function estimated for M. leprae was
used. For the substitution values, M. avium subsp. paratuberculosis was used as an
outgroup.
Nucleotide sequence accession numbers. The M. lepromatosis nucleotide se-
quences have been deposited in the GenBank database under the accession
numbers shown in Table 1. The GenBank accession number for the nucleotide
sequence of the full-length rpoB gene of M. haemophilum ATCC 29548T is
GQ245966.
RESULTS
M. lepromatosis sequences and matches with M. leprae. The
20 genes and pseudogenes (22,814 nucleotides) from M. lep-
romatosis and their matches with M. leprae genes and pseudo-
genes are shown in Table 1. These genes and pseudogenes
were chosen from the eight initial sequences (17) with exten-
sion of the 16S rRNA, rpoB, groEL, and rpoT genes and 16
other metabolic genes or pseudogenes in the M. leprae and M.
tuberculosis genomes as previously described (4). Of the 16 new
genes or pseudogenes, 12 were successfully amplified and se-
quenced, but 4 pseudogenes (ligB, ummA, bfrB, and pfkB) were
not amplified and sequenced despite many attempts and rede-
sign of the primers, implying that there is too much variation or
these pseudogenes are not present in M. lepromatosis. The
GC content of the nucleotide sequences was 58.6%, similar
to the 58.8% for the corresponding M. leprae sequences or the
M. leprae genome (57.8%) (4).
The average levels of nucleotide identity between M. lepro-
matosis and M. leprae were 93.1% and 79.1% for the coding
genes and pseudogenes, respectively, and the overall level of
VOL. 191, 2009 SEQUENCE ANALYSIS OF M. LEPROMATOSIS 6069

TABLE 1. Sequences for 22,814 bp of M. lepromatosis and matches with M. leprae

GenBank Nucleotide matches Amino acid matches

Length
Gene or pseudogene (description) accession No. No. of No. of gaps No. No. of
(bp) % %
no. aligned matches (%) aligned matches

All 14 protein-encoding genes 17,318 17,250 16,061 93.1 47 (0.27) 5,527 5,260 95.2
rpoT (RNA polymerase sigma EU203595 881 881 775 88.0 47 (5.3) 252 199 79.0
factor)
rpsO (ribosomal protein O) EU203592 415 398 359 90.2 0a 89 82 92.1
ligA (DNA ligase) EU839554 2,138 2,130 1,970 92.5 0 685 636 92.8
ribF (riboflavin kinase) EU203592 352 343 319 93.0 0 114 108 94.7
bfrA (bacterioferritin) EU839559 421 419 387 92.4 0 136 129 94.9
icd2 (isocitrate dehydrogenase) EU839560 2,255 2,240 2,095 93.5 0 736 699 95.0
lld2 (L-lactate dehydrogenase 2) EU839561 1,292 1,290 1,194 92.6 0 400 380 95.0
mmaA4 (mycolic acid synthase 4) EU203591 280 280 265 94.6 0 84 80 95.2
gnd1 (6-phosphogluconate EU839562 1,545 1,543 1,441 93.4 0 481 461 95.8
dehydrogenase)
pfkA (6-phosphofructokinase) EU839563 845 844 789 93.5 0 281 272 96.8
groEL (heat shock protein 65) EU203593 1,615 1,614 1,488 92.2 0 536 520 97.0
umaA2 (mycolic acid synthase 2) EU839564 417 413 390 94.4 0 137 133 97.1
gltA2 (citrate synthase 1) EU839565 1,266 1,265 1,197 94.6 0 418 406 97.1
rpoB (RNA polymerase beta EU203594 3,596 3,590 3,392 94.5 0 1,178 1,155 98.0
subunit)
All five pseudogenes 3,958 4,077 3,224 79.1 207 (5.1)
gnd2 (6-phosphogluconate EU839556 1,147 1,220 911 74.7 82 (6.7)
dehydrogenase 2)
lld1 (L-lactate dehydrogenase 1) EU839555 564 581 448 77.1 40 (6.9)
icd1 (isocitrate dehydrogenase 1) EU839557 615b 190 149 78.4 13 (6.8)
433 355 82.0 12 (2.8)
citA (citrate synthase A) EU839558 978 975 802 82.3 21 (2.1)
mmaA3 (mycolic acid synthase 3) EU203591 654 678 559 82.4 39 (5.7)
16S rRNA gene EU203590 1,538 1,542 1,511 98.0 10 (0.65)
Total 22,814 22,869 20,796 90.9 264 (1.2)
a
There is a 3-bp gap before the start codon.
b
There is a 128-bp deletion between the better matched regions.

nucleotide identity was 90.9% (Table 1). This level is well
below the 95% genome-level cutoff used to distinguish species
(15). In addition, the level of nucleotide identity for the rpoB
gene was 94.5%, which is also less than the 98% cutoff pro-
posed recently (1). Therefore, M. lepromatosis meets the def-
inition of a species.
The levels of identity for the protein-encoding genes were
88.0 to 94.6% at the nucleotide level and 79.0 to 98.0% at the
amino acid level (overall level of identity, 95.2%). Thirteen of
the 14 genes studied (16,369 bp) were aligned without gaps; the
exception was rpoT, which had 47 gap sites (5.3% of 881 bp).
The rpoT gene of M. lepromatosis contained a unique 74-bp tan-
dem repeat, (CGAGCCACCAATACAGCATCT)3CGAGCCA
CCAA, which corresponded to amino acids (RATNTAS)3RAT.
A 42-bp portion of this sequence, (AATACAGCATCTCGA
GCCACC)2, which corresponded to (NTASRAT)2, was not
present in the M. leprae rpoT gene. Slightly downstream of this
sequence was a 3-bp insertion in M. lepromatosis. Upstream, a
2-bp insert in M. lepromatosis shifted the reading frame and
likely led to a different start codon. Thus, rpoT is the most
variable of the 14 genes.
The levels of identity for the pseudogenes were much lower,
ranging 74.7 to 82.4%. Due to indels, the sequence of each of
the five pseudogenes had many gaps in the alignment. Alto-
gether, there was 207 bp of gaps (5.1% of 4,077 bp), accounting
for 24.3% of the 853 mismatches. In addition, the M. leproma-
tosis icd1 gene had a 128-bp deletion in the middle. Therefore,
in these pseudogenes, the lack of constraints of reading frames
allowed much faster accumulation of indels.
Finally, the level of identity for the 16S rRNA gene was
98.0%, which was the highest level of identity, but there were
10 gap sites (0.65% of 1,542 bp). In all mycobacteria, 16S
rRNA genes are highly conserved, and there are relatively few
indels or indels are not present in closely related mycobacteria,
such as M. marinum and M. ulcerans, M. avium and Mycobac-
terium intracellulare, and other pairs of species (data not
shown). Thus, indel mutations are also a feature of 16S rRNA
gene evolution in the two leprosy organisms. Other 16S rRNA
gene features were analyzed previously (17).
Phylogenetic position of M. lepromatosis. As shown in Fig.
1A, phylogenetic trees constructed by using the maximum like-
lihood and neighbor-joining methods produced the same to-
pology with bootstrap values of 100 at all nodes except one. M.
leprae and M. lepromatosis were the most closely related spe-
cies, yet in both trees the terminal branches leading to these
species were much longer than those leading to other tight
groups, such as the M. tuberculosis complex and the M. avium
group. These long branches could not be explained by faster
evolution or high substitution rates. On average, a gene evolv-
ing in the lineage leading to M. leprae had a slightly higher dS
value than a gene in the M. tuberculosis lineage (from the last
common ancestor) (0.82 and 0.72, respectively) (10), while the
average dN values were 0.065 and 0.045, respectively (13).
Because the dS values are similar and their contribution to
6070 HAN ET AL. J. BACTERIOL.

FIG. 1. Maximum likelihood phylogenies of selected mycobacteria based on (A) concatenated alignment of the rpoB, ligA, groEL, gnd1, and
bfrA genes, (B) alignment of the rpoB gene, and (C) alignment of the 16S rRNA gene. Bootstrap values are indicated at the nodes. At some nodes
there are two values, a value obtained by the neighbor-joining method (left value) and a value obtained by the maximum likelihood method (right
value). The bars indicate the numbers of nucleotide substitutions per site. The trees were rooted to separate the slow-growing mycobacterial clade
(A and B) and the M. avium branch (C).

branch lengths is much more important than that of dN values, difference was the decrease in the GC content, mainly at the
we believe that a much earlier divergence time is the best third codon positions (Table 2).
explanation. A phylogeny was also obtained using the 16S rRNA gene
Because M. haemophilum was recently reported to be a close sequences, which resulted in a different topology (Fig. 1C); the
relative of M. leprae (22) and M. lepromatosis (17), phyloge-
netic trees based on rpoB (3,429 bp) were also reconstructed
(Fig. 1B). The same topology was obtained despite some lower TABLE 2. GC contents of rpoB gene
bootstrap values. It was observed that M. haemophilum was GC content (%)
much closer to the leprosy organisms than to other mycobac- Species Codon positions Codon
teria. Yet the branch lengths indicated that there was faster 1 and 2a position 3a
Total
evolution of the leprosy mycobacteria than of M. haemophilum.
Using Tajimas relative rate tests for this gene, the branches M. lepromatosis 51.4 77.9 60.2
M. leprae 51.9 79.1 61.0
leading to M. leprae and M. lepromatosis were found to be M. haemophilum 52.7 88.0 64.4
significantly longer than the branch leading to M. haemophilum M. tuberculosis 53.0 86.9 64.3
for the third codon position in both leprosy organisms a
The first- and/or second-codon substitutions are frequently nonsynonymous,
(P0.001 for both) and for the first and second codon posi- whereas the third-codon substitutions are usually synonymous due to codon
tions in M. lepromatosis (P 0.042). The main reason for this wobbling.
VOL. 191, 2009
FIG. 2. Numbers of nucleotide substitutions per site for five pairs of Mycobacterium species. (Left panel) dS values. (Right panel) dN values.
The genes analyzed were rpoB, ligA, groEL, gnd1, and bfrA (bars from left to right). Abbreviations: map, M. avium subsp. paratuberculosis; mav,
M. avium 104; mbo, M. bovis AF2122/97; mle, M. leprae; mlp, M. lepromatosis; mmi, M. marinum M; mtc, M. tuberculosis CDC1551; mul, M.
ulcerans. The dN/dS ratios for the rpoB, ligA, groEL, gnd1, and bfrA genes for the M. lepromatosis-M. leprae pair were 0.022, 0.155, 0.031, 0.075,
and 0.089, respectively; the average dN/dS ratio was 0.074.
branch between the leprosy organisms was much longer, and
the M. marinum-M. ulcerans clade was closer to M. tuberculosis
than to the clade formed by M. haemophilum and the leprosy
organisms. This tree was less realistic than the tree obtained
using the concatenated protein-encoding genes (Fig. 1A). Phy-
logenomic studies have shown that in a gene phylogeny, when
one lineage evolves much faster than the other lineages, an
incorrect topology may be obtained, in which the fast-evolving
branch has a tendency to move toward the base of the tree.
Thus, the true topology may be obtained only with a small
subset of genes that evolve slowly (2). The 16S rRNA genes of
the leprosy organisms evolved much faster than those of other
mycobacteria, as shown by more indels, as noted above and
previously (17).
Evolutionary distance between M. lepromatosis and M. lep-
rae. To estimate the evolutionary distance between M. lepro-
matosis and M. leprae, dS and dN values were calculated using
the five conserved genes described above. These values were
compared with the dS and dN values for other pairs of most
closely related species or subspecies, including M. avium sub-
species, M. marinum and M. ulcerans, and M. tuberculosis and
M. bovis. The more divergent pair M. tuberculosis and M. leprae
was also included as a gauge because these most medically
important mycobacteria diverged 66 million years ago (13).
The results are shown in Fig. 2.
Since they are nearly unaffected by natural selection, the dS
values are similar for many genes and thus more suitable for
estimating relative divergence times. The dS values for M.
tuberculosis and M. bovis were zero, consistent with the clade-
level divergence of these species and a genome level of identity
of 99.95% (7). Similarly, subspecies-level differences for the M.
marinum-M. ulcerans and M. avium subspecies pairs were re-
flected by low dS values. Conversely, for M. lepromatosis and
M. leprae the dS values were far higher and for the five genes
ranged from 14 to 28% (average, 20%) of the values for the
much more divergent organisms M. tuberculosis and M. leprae.
Similar results were obtained for the dN values, although in
this case natural selection affected each gene or lineage differ-
ently, which resulted in some variability. For example, the
highest dN value for ligA meant faster evolution of this gene in
SEQUENCE ANALYSIS OF M. LEPROMATOSIS 6071
the M. lepromatosis and M. leprae lineages. Together, these
results suggest that the divergence of the two leprosy-causing
organisms was not recent and that the length of their diver-
gence was around 20% of the age of divergence of M. tuber-
culosis and M. leprae or 10 million years.
Age of M. lepromatosis pseudogenes. For all five pseudogenes
(citA, gnd2, icd1, lld1, and mmaA3) in both species there were
corresponding functional genes in M. tuberculosis and other
mycobacteria. So when were the genes inactivated, in the last
common ancestor or after divergence? This question was ad-
dressed from a few angles.
First, the dN/dS ratio for both species was estimated for each
pseudogene. For an ancestral pseudogene, continued evolu-
tion after divergence in both lineages would be neutral, result-
ing in similar dN and dS values with a ratio of 1 or close to 1.
Otherwise, selective evolution as a gene in both lineages after
divergence with late inactivation would result in a much
smaller dN/dS ratio. The longer the selective evolution, the
smaller the dN/dS ratio would be. For example, Gomez-Valero
et al. (13) showed that in the M. leprae lineage after divergence
from M. tuberculosis, the average dN/dS ratio for genes was
0.079.
As shown in Table 3, the dN/dS ratios for M. lepromatosis
and M. leprae were close to 1 (range, 0.680 to 0.789) for the
citA, gnd2, lld1, and mmaA3 pseudogenes, indicating that there
was likely neutral evolution and thus the sequences were pseu-
dogenes prior to divergence. It is noteworthy that removal of
stop codons, which was required for calculation, resulted in
underestimation of the dN values, leading to smaller ratios.
Mutations from codons to stop codons, which are common in
pseudogenes, are essentially dN events. Yet similar dN and dS
values were also indicated by the overlapping confidence in-
tervals (mean 1.96 standard errors) for these values. The
only exception was icd1, for which the dN value was signifi-
cantly smaller than the dS value (a ratio much less than 1). The
pseudogene dN/dS ratios contrasted with the dN/dS ratios for
the five coding genes analyzed, for which the average dN/dS
ratio was 0.074 (Fig. 2). The dN/dS ratios were also estimated
for M. lepromatosis and M. tuberculosis, and as expected, they
6072 HAN ET AL. J. BACTERIOL.

TABLE 3. Synonymous and nonsynonymous substitutions in pseudogenesa

M. lepromatosis- M. lepromatosis- M. lepromatosis-M. leprae
Pseudogene M. leprae M. tuberculosis
dN/dS ratio dN/dS ratio dN dS P value

citA 0.6804 0.1669 0.1639 0.0175 0.2409 0.0372 NS

gnd2 0.7148 0.1947 0.1950 0.0200 0.2728 0.0428 NS
mmaA3 0.7690 0.1023 0.1415 0.0198 0.1840 0.0490 NS
lld1 0.7890 0.1458 0.1858 0.0264 0.2355 0.0496 NS
icd1 0.3234 0.0699 0.1072 0.0169 0.3315 0.0694 0.05
a
The confidence intervals for dN and dS (mean 1.96 standard errors) overlap (NS, not statistically significant) or do not overlap (P 0.05).

were small (Table 3), reflecting the evolution of the sequences
entirely as genes in the M. tuberculosis lineage.
Second, the relative ages of M. lepromatosis pseudogenes
were estimated by using the formula used for the M. leprae
pseudogenes (13). This formula required the presence of cor-
responding orthologous genes in M. tuberculosis and M. avium.
Thus, it could not be used for mmaA3 because this pseudogene
is not present in M. avium. The relative ages of the remaining
four pseudogenes (Table 4) were found to be similar to the
relative ages in M. leprae and also to the average age (0.13) of
the 611 M. leprae pseudogenes analyzed previously (13). Thus,
the similar ages of pseudogenes in the two species may indicate
that they were pseudogenes in the ancestor.
Finally, indels that were present in both species and thus
likely of ancestral origin were examined using sequence align-
ments. Several common indels were detected, except in citA,
and they included a 2-nucleotide insertion in gnd2, a 5-nucle-
otide deletion in icd1, a 10-nucleotide deletion in lld1, and a
5-nucleotide insertion and a 8- or 9-nucleotide deletion in
mmaA3. These indels were not in triplet nucleotides; as a
result, they could alter the codon reading frames and possibly
create stop codons. Thus, these findings in essence show the
ancestral presence of these pseudogenes, rather than conver-
gent evolution. Many unique indels in each species represented
continued evolution of these pseudogenes after divergence.
Together, these analyses provided ample evidence showing
that (i) all or most of the pseudogenes analyzed were inacti-
vated prior to the divergence of M. lepromatosis and M. leprae
and (ii) the massive gene inactivation described for M. leprae
(4, 13) before the discovery of M. lepromatosis (17) in fact took
place in the last common ancestor of these two species.
DISCUSSION
The discovery of the new leprosy-causing organism M. lep-
romatosis and its close relationship to M. leprae provided an
TABLE 4. Relative pseudogene agea
Pseudogene M. lepromatosis
icd1 0.105
citA 0.106
lld1 0.111
gnd2 0.175
Avg 0.124
a
The relative age ranges from 1 (the age equals the time of divergence of M.
lepromatosis-M. leprae from M. tuberculosis) to 0 (pseudogene inactivation at the
present time).
opportunity to discern the evolutionary features of these or-
ganisms. Using in-depth analysis of 20 genes and pseudogenes
from M. lepromatosis, this study fulfilled some of our goals.
First, the species-level divergence between these two organ-
isms was confirmed to be on a larger genetic scale despite the
current lack of culture and phenotypic features. Second, diver-
gence 10 million years ago is proposed based on the topology
of robust phylogenetic trees and the levels of nucleotide sub-
stitutions in the protein-encoding genes. This time is orders of
magnitude (88 times) greater than the relatively well-estab-
lished divergence time for M. tuberculosis and M. bovis
(113,000 years ago), which share 99.95% genome identity (7)
and had no resolvable difference in our analyses (Fig. 1 and 2).
And third, the age of the pseudogenes of M. lepromatosis,
which is nearly identical to the age of M. leprae pseudogenes,
their nearly neutral evolution, and the presence of codon-
altering common indels suggest these pseudogenes were pseu-
dogenes in the last common ancestor of the two species.
Bacterial taxonomy has changed along with technological
advances, from morphological classification a century ago to
phenotype-based chemotaxonomy to a polyphasic approach
involving assessment of genetic variation, such as genomic
DNA hybridization (70% cutoff to differentiate species) and
divergence of the 16S rRNA gene (in general, 97% cutoff) in
recent decades (21). Numerous bacterial genome sequences
are now readily accessible, making genetic comparisons accu-
rate and comprehensive for the purpose of species identifica-
tion (15). Yet the phenotype of an organism, which includes
complex traits involving many genetic elements, is still essen-
tial. For instance, historically as well as practically, several
species with essentially identical genome sequences are re-
tained, such as M. tuberculosis for humans and M. bovis for
cows. In the case of M. lepromatosis, status as a novel species
can be justified because of its remarkable genetic variation, as
well as uniqueness of the disease that it causes (DLL). How-
ever, to satisfy the bacteriological code for a full description of
this new species, phenotypic characterization after cultivation
and/or animal passage is needed, and efforts to do this are
being made.
Practically, the sequences enable us to separate the two
M. leprae
leprosy organisms to better define the disease. For instance, we
0.079 have devised a reliable PCR assay based on the species-specific
0.089 16S rRNA gene sequences to detect the presence of each
0.080
0.168
species in a skin biopsy (X. Y. Han, K. C. Sizer, F. Aung, H. H.
0.104 Tan, and B. Werner, unpublished data; X. Y. Han, K. C. Sizer,
S. Velarde-Felix, and F. Vargas-Ocampo, unpublished data).
Using this assay in a follow-up study of M. lepromatosis infec-
tions in Mexico (X. Y. Han, K. C. Sizer, S. Velarde-Felix, and
VOL. 191, 2009
F. Vargas-Ocampo, unpublished data), we found that far more
leprosy cases were caused by M. lepromatosis than by M. leprae,
suggesting that M. lepromatosis is the dominant leprosy-causing
species in Mexico. We also noted that, in addition to leproma-
tous leprosy, M. lepromatosis also specifically caused DLL in
30% of the infections. DLL is endemic in Mexico but rare
elsewhere. These findings, as well as the lack of DLL reports or
description in Europe, more specifically in Spain, suggest that
DLL is native to Mexico and was not brought over by the
Spaniards over the past 500 years. Native Mongoloid Mexicans
initially migrated from Asia 12,000 years ago through the
Alaska Bering land bridge and settled in the Pacific coastal
areas of Mexico. Some of them migrated further south and
settled in the Caribbean, other central American countries,
and South America. Our finding of M. lepromatosis in Brazil
and Singapore Chinese (X. Y. Han, K. C. Sizer, F. Aung, H. H.
Tan, and B. Werner, unpublished data) and the presence of
DLL in the Caribbean (8) are also consistent with the possi-
bility that there was Mongoloid spread of this organism. There-
fore, the studies of M. lepromatosis have provided new insights
into the puzzling clinical and geographic features of leprosy.
While the overall features of M. lepromatosis and M. leprae
are very similar in terms of culture difficulty, GC content, and
the functional status of genes and pseudogenes, M. lepromato-
sis likely has some unique phenotypic traits that have yet to be
discovered that may be responsible for some distinct clinical
features of M. lepromatosis infections, such as DLL. DLL is
chronic, and the organism is strictly intracellular (17, 20, 26).
Vargas-Ocampo (31) recently studied the histopathology of
199 patients with DLL in Mexico and concluded that DLL is
essentially a vascular disease caused by mycobacterial invasion,
with early vaculitis to late-stage vascular occlusion, which is
responsible for the clinical manifestations of the disease that
include skin purpura, ulceration, and necrosis.
Humans are the only known natural host of M. leprae and
have probably been infected for 100,000 years according to
genomic evidence based on many M. leprae strains obtained
worldwide (24). Consistent with the strictly parasitic lifestyle of
this organism, the M. leprae genome is the smallest genome of
all of the mycobacterial genomes sequenced to date. This ge-
nome has also undergone reductive evolution with massive
inactivation of genes (4) estimated to have occurred during the
last 20 million years (13). The present study further establishes
that this genome decay took place in the last common ancestor
of M. leprae and M. lepromatosis.
What led to the genome decay in the ancestor 20 million
years ago? We speculate that there is some relationship to
adaptation from a free-living lifestyle, like that of almost all
other mycobacteria (100 species) except M. tuberculosis and
M. bovis, to an increasingly parasitic lifestyle in a host (an early
ancestor of hominids?). During the last decade, sequencing
and comparative analysis of many bacterial genomes have
shown that changes in lifestyle from free living to host depen-
dence or restriction to a specific host may lead to relaxation of
natural selection and make many genes nonessential, initiating
the process of genome downsizing. In the early stages, many
genes (up to thousands of genes), usually less selectively con-
strained genes, are inactivated (4, 5, 13, 29). This inactivation
occurs through missense or nonsense nucleotide substitutions,
indels, or mobilization of insertion elements (9). There is then
SEQUENCE ANALYSIS OF M. LEPROMATOSIS 6073
a tendency to lose the nonfunctional DNA (12, 23, 27). Anal-
ysis of some bacterial endosymbionts of insects, such as the
aphid endosymbiont Buchnera aphidicola and the ant endo-
symbiont Blochmannia floridanus, has permitted workers to
establish a stepwise scenario for genome reduction involving
very slow and gradual degradation with scattered punctual
large deletion events (11, 14). This model has been confirmed
by complete sequencing of several B. aphidicola genomes (25).
In this regard, future genome sequencing of M. lepromatosis
and comparative genomic analyses with M. leprae may shed
light on the genome decay in these highly specialized patho-
gens.
Why did the ancestor evolve into two species 10 million
years ago if it had become parasitic in a primate host? The
answer to this question is even more elusive. There are a few
possible explanations, including subtle niche differences in the
host, eventually leading to separate evolution paths; coevolu-
tion with different host populations of the same species that
might be affected variably by the availability of nutrients, sus-
ceptibility, geographic restriction, and other factors; and a shift
from one host species to another. The niche possibility can be
tested in future clinical studies by examining tissue infected by
each organism. Both organisms are known to be intracellular in
macrophages; however, remarkable endothelial invasion by M.
lepromatosis in DLL has been documented (17, 26, 31). In our
clinical studies, we also found many cases of dual M. leproma-
tosis-M. leprae infection (X. Y. Han, K. C. Sizer, F. Aung, H. H.
Tan, and B. Werner, unpublished data; X. Y. Han, K. C. Sizer,
S. Velarde-Felix, and F. Vargas-Ocampo, unpublished data).
Further studies of the coexistence of these two organisms may
provide insight as well.
ACKNOWLEDGMENTS
This work was supported in part by a University Cancer Foundation
grant from the University of Texas M. D. Anderson Cancer Center
(MDACC) (to X.Y.H.), by the School of Health Sciences of MDACC,
and by National Institutes of Health grant CA16672 for the MDACC
DNA Core Facility.
We thank the DNA Core Facility staff for DNA sequencing and
John Spencer for helpful discussions. X.Y.H. initiated and orches-
trated this study; X.Y.H., K.C.S., E.J.T., J.K., J.L., and P.H. deter-
mined the sequences; F.J.S., L.G.V., and X.Y.H. performed data anal-
ysis; and F.J.S. and X.Y.H. wrote the paper.
REFERENCES
1. Adekambi, T., T. M. Shinnick, D. Raoult, and M. Drancourt. 2008. Complete
rpoB gene sequencing as a suitable supplement to DNA-DNA hybridization
for bacterial species and genus delineation. Int. J. Syst. Evol. Microbiol.
58:18071814.
2. Canback, B., I. Tamas, and S. G. Andersson. 2004. A phylogenomic study of
endosymbiotic bacteria. Mol. Biol. Evol. 21:11101122.
3. Cole, S. T., R. Brosch, J. Parkhill, T. Garnier, C. Churcher, D. Harris, S. V.
Gordon, K. Eiglmeier, S. Gas, C. E. Barry, F. Tekaia, K. Badcock, D.
Basham, D. Brown, T. Chillingworth, R. Connor, R. Davies, K. Devlin, T.
Feltwell, S. Gentles, N. Hamlin, S. Holroyd, T. Hornby, K. Jagels, 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 393:537544.
4. Cole, S. T., K. Eiglmeier, J. Parkhill, K. D. James, N. R. Thomson, P. R.
Wheeler, N. Honore, T. Garnier, C. Churcher, D. Harris, K. Mungall, D.
Basham, D. Brown, T. Chillingworth, R. Connor, R. M. Davies, K. Devlin, S.
Duthoy, T. Feltwell, A. Fraser, N. Hamlin, S. Holroyd, T. Hornsby, K. Jagels,
C. Lacroix, J. Maclean, S. Moule, L. Murphy, K. Oliver, M. A. Quail, M. A.
Rajandream, K. M. Rutherford, S. Rutter, K. Seeger, S. Simon, M. Sim-
monds, J. Skelton, R. Squares, S. Squares, K. Stevens, K. Taylor, S. White-
6074 HAN ET AL.
head, J. R. Woodward, and B. G. Barrell. 2001. Massive gene decay in the
leprosy bacillus. Nature 409:10071011.
5. Delmotte, F., C. Rispe, J. Schaber, F. J. Silva, and A. Moya. 2006. Tempo
and mode of early gene loss in endosymbiotic bacteria from insects. BMC
Evol. Biol. 6:56.
6. Devulder, G., M. P. de Montclos, and J. P. Flandrois. 2005. A multigene
approach to phylogenetic analysis using the genus Mycobacterium as a model.
Int. J. Syst. Evol. Microbiol. 55:293302.
7. Garnier, T., K. Eiglmeier, J. C. Camus, N. Medina, H. Mansoor, M. Pryor,
S. Duthoy, S. Grondin, C. Lacroix, C. Monsempe, S. Simon, B. Harris, R.
Atkin, J. Doggett, R. Mayes, L. Keating, P. R. Wheeler, J. Parkhill, B. G.
Barrell, S. T. Cole, S. V. Gordon, and R. G. Hewinson. 2003. The complete
genome sequence of Mycobacterium bovis. Proc. Natl. Acad. Sci. USA 100:
78777882.
8. Gelber, R. H. 2005. Leprosy (Hansens disease), p. 966972. In D. L. Kasper,
E. Braunwald, A. S. Fauci, S. L. Hauser, D. L. Longo, and J. L. Jameson
(ed.), Harrisons principles of internal medicine. McGraw-Hill, Inc., New
York, NY.
9. Gil, R., E. Belda, M. J. Gosalbes, L. Delaye, A. Vallier, C. Vincent-Monegat,
A. Heddi, F. J. Silva, A. Moya, and A. Latorre. 2008. Massive presence of
insertion sequences in the genome of SOPE, the primary endosymbiont of
the rice weevil Sitophilus oryzae. Int. Microbiol. 11:4148.
10. Gomez-Valero, L. 2006. Evolucion reductiva del tamano del genoma en
bacterias intracelulares. Ph.D. thesis. Universitat de Valencia, Valencia,
Spain.
11. Gomez-Valero, L., A. Latorre, R. Gil, J. Gadau, H. Feldhaar, and F. J. Silva.
2008. Patterns and rates of nucleotide substitution, insertion and deletion in
the endosymbiont of ants Blochmannia floridanus. Mol. Ecol. 17:43824392.
12. Gomez-Valero, L., A. Latorre, and F. J. Silva. 2004. The evolutionary fate of
nonfunctional DNA in the bacterial endosymbiont Buchnera aphidicola.
Mol. Biol. Evol. 21:21722181.
13. Gomez-Valero, L., E. P. C. Rocha, A. Latorre, and F. J. Silva. 2007. Recon-
structing the ancestor of Mycobacterium leprae: the dynamics of gene loss and
genome reduction. Genome Res. 17:11781185.
14. Gomez-Valero, L., F. J. Silva, J. C. Simon, and A. Latorre. 2007. Genome
reduction of the aphid endosymbiont Buchnera aphidicola in a recent evo-
lutionary time scale. Gene 389:8795.
15. Goris, J., K. T. Konstantinidis, J. A. Klappenbach, T. Coenye, P. Vandamme,
and J. M. Tiedje. 2007. DNA-DNA hybridization values and their relation-
ship to whole-genome sequence similarities. Int. J. Syst. Evol. Microbiol.
57:8191.
16. Guindon, S., and O. Gascuel. 2003. A simple, fast, and accurate algorithm to
estimate large phylogenies by maximum likelihood. Syst. Biol. 52:696704.
17. Han, X. Y., Y. H. Seo, K. C. Sizer, T. Schoberle, G. S. May, J. S. Spencer, W.
Li, and R. G. Nair. 2008. A new Mycobacterium species causing diffuse
lepromatous leprosy. Am. J. Clin. Pathol. 130:856864.
J. BACTERIOL.
18. Katoh, K., K. Misawa, K. Kuma, and T. Miyata. 2002. MAFFT: a novel
method for rapid multiple sequence alignment based on fast Fourier trans-
form. Nucleic Acids Res. 30:30593066.
19. Kumar, S., K. Tamura, and M. Nei. 2004. MEGA3: integrated software for
molecular evolutionary genetics analysis and sequence alignment. Briefs
Bioinform. 5:150163.
20. Latapi, F., and A. Chevez-Zamora. 1948. The spotted leprosy of Lucio: an
introduction to its clinical and histological study. Int. J. Lepr. 16:421437.
21. Ludwig, W., and H.-P. Klenk. 2001. Overview: a phylogenetic backbone and
taxonomic framework for prokaryotic systematics, p. 4965. In D. R. Boone,
R. W. Castenholz, and G. M. Garrity (ed.), Bergeys manual of systematic
bacteriology, 2nd ed., vol. I. The archaea and the deeply branching and
phototrophic bacteria. Springer-Verlag, New York, NY.
22. Mignard, S., and J. P. Flandrois. 2008. A seven-gene, multilocus, genus-wide
approach to the phylogeny of mycobacteria using supertrees. Int. J. Syst.
Evol. Microbiol. 58:14321441.
23. Mira, A., H. Ochman, and N. A. Moran. 2001. Deletional bias and the
evolution of bacterial genomes. Trends Genet. 17:589596.
24. Monot, M., N. Honore, T. Garnier, R. Araoz, J. Y. Coppee, C. Lacroix, S.
Sow, J. S. Spencer, R. W. Truman, D. L. Williams, R. Gelber, M. Virmond,
B. Flageul, S. N. Cho, B. Ji, A. Paniz-Mondolfi, J. Convit, S. Young, P. E.
Fine, V. Rasolofo, P. J. Brennan, and S. T. Cole. 2005. On the origin of
leprosy. Science 308:10401042.
25. Moran, N. A., H. J. McLaughlin, and R. Sorek. 2009. The dynamics and time
scale of ongoing genomic erosion in symbiotic bacteria. Science 323:379382.
26. Rea, T. H., and R. S. Jerskey. 2005. Clinical and histologic variations among
thirty patients with Lucios phenomenon and pure and primitive diffuse
lepromatosis (Latapis lepromatosis). Int. J. Lepr. Other Mycobact. Dis.
73:169188.
27. Silva, F. J., A. Latorre, and A. Moya. 2001. Genome size reduction through
multiple events of gene disintegration in Buchnera APS. Trends Genet.
17:615618.
28. Tajima, F. 1993. Simple methods for testing the molecular evolutionary clock
hypothesis. Genetics 135:599607.
29. Toh, H., B. L. Weiss, S. A. H. Perkin, A. Yamashita, K. Oshima, M. Hattori,
and S. Aksoy. 2006. Massive genome erosion and functional adaptations
provide insights into the symbiotic lifestyle of Sodalis glossinidius in the tsetse
host. Genome Res. 16:149156.
30. Uchiyama, I. 2007. MBGD: a platform for microbial comparative genomics
based on the automated construction of orthologous groups. Nucleic Acids
Res. 35:D343D346.
31. Vargas-Ocampo, F. 2007. Diffuse leprosy of Lucio and Latapi: a histologic
study. Lepr. Rev. 78:248260.
32. Yang, Z. H., and R. Nielsen. 2000. Estimating synonymous and nonsynony-
mous substitution rates under realistic evolutionary models. Mol. Biol. Evol.
17:3243.