110 A. Smyth and S.V. Gordon
numbered 1 to 4 in M. tuberculosis, with M. bovis
missing the locus encoding Mce3 (Zumarraga
et al., 1999). While the function of Mce3 is cryp-
tic, inactivation of the locus was shown to atten-
uate M. tuberculosis (Senaratne et al., 2008),
suggesting M. bovis may have developed com-
pensatory mechanisms to adjust for the loss of
this locus. The one clear function linked to the
Mce proteins is the role of the genes of the mce4
locus in the uptake of cholesterol and sterols;
these have been suggested to be key carbon
sources during persistent infection.
The pore-forming protein OmpATb has
been demonstrated in both M. tuberculosis and
M. bovis to be involved in virulence as its deletion
results in significantly reduced multiplication
of mutants in macrophages (Raynaud et al.,
2002b). The transcription of ompAtb was also
increased in acidic conditions, suggesting it may
play a role in surviving phagosomal maturation
as the compartment acidifies. On the subject of
channel proteins, an intriguing example of a
dual function protein located at the cell surface
is provided by the ‘channel protein with necrosis-
inducing toxin’, CpnT, which plays roles in both
nutrient uptake and induction of host cell death
by M. tuberculosis (Danilchanka et al., 2014).
The CpnT protein consists of an N- terminal
channel domain that is involved in the uptake of
nutrients across the outer membrane, and a
secreted toxic C-terminal domain that causes
necrotic cell death in eukaryotic cells. The mech-
anism by which the C-terminal domain induces
necrosis is unknown, but CpnT represents the
only secreted ‘toxin’ described to date in the
MTBC.
The ‘exported repetitive protein’, aka Erp or
P36, is cell wall associated and secreted (Berthet
et al., 1998a; de Mendonca-Lima et al., 2003).
When it was knocked out in M. bovis it resulted
in a mutant strain with impaired replication in
macrophages and decreased lung pathology in
mice that was fully restored with complementa-
tion (Bigi et al., 2005). Virulence of this protein
has been linked to its central domain that con-
tains multiple repeats, so it is tempting to specu-
late that variation in the repeats may impact on
its role in virulence. Variation in the multiple
lipoproteins located at the bacterial cell surface
between M. bovis and M. tuberculosis is also of
note, with the genes for lipoproteins LppQ, LpqT,
LpqG and LprM, all deleted or frameshifted in
M. bovis, while the gene for lipoprotein LppA is
duplicated. This variation could well be involved
in altering the way the bacteria interacts with
the host and its environment, facilitating differ-
ent tissue tropisms or manipulation of different
immune responses.
The MTBC contains many polyketide syn-
thases (PKS), a diverse family of multifunctional
enzymes that are involved in the synthesis of
secondary metabolites in a wide array of bacte-
ria (O’Hagan, 1993; Sirakova et al., 2003). In
mycobacteria, the PKSs are often closely associ-
ated with genes involved in fatty acid metabo-
lism. Many PKS genes in MTBC bacilli have now
been linked to pathways synthesizing lipids and
glycolipid conjugates that are essential for viru-
lence, as well as key components of the complex
cell envelope (Kolattukudy et al., 1997). It was
shown that PKS enzymes are involved in the
synthesis of lipids in the dimycocerosyl phthioc-
erol family (DIMs) (Trivedi et al., 2005; Quadri,
2014). The role these enzymes play in virulence
is not fully understood, but there are some
interesting avenues to pursue in differentiating
M. bovis from M. tuberculosis. For example,
M. bovis cannot synthesize PKS6 due to a frame-
shift mutation in the encoding gene (Garnier
et al., 2003), but when the pks6 is knocked out in
M. tuberculosis this results in an attenuated
strain in the murine model; hence, M. bovis may
have developed compensatory systems to accom-
modate for the loss of PKS6, or its loss may be
beneficial to the bacillus.
One of the most abundant components of
the outer lipid layer are the aforementioned
DIMs, primarily phthiocerol dimycocerosate
(PDIM) (Azad et al., 1997). These lipids on the
outer surface are not covalently bound to the
inner membrane, unlike the mycolic acids. Many
groups have shown that knocking out the syn-
thesis of PDIM results in an attenuated pheno-
type, with reduced growth or reduced bacterial
burdens in infected cells and model hosts (Cox
et al., 1999; Camacho et al., 2001). The role of
PDIM was originally thought to be mainly struc-
tural, playing a part in the fluidity of the cell
membrane. It was considered an indirect viru-
lence factor as it was thought that its removal
altered the ability of other proteins to affect the
host, but that it did not itself have a direct influ-
ence on the ability of the bacteria to cause dis-
ease. However, later work has shown that this is