MICROBIOLOGY
PE/PPE proteins mediate nutrient transport across
the outer membrane ofMycobacterium tuberculosis
Qinglan Wang^1 , Helena I. M. Boshoff^1 , Justin R. Harrison^2 , Peter C. Ray2,3, Simon R. Green^2 ,
Paul G. Wyatt^2 , Clifton E. Barry III1,4*
Mycobacterium tuberculosishas an unusual outer membrane that lacks canonical porin proteins for the
transport of small solutes to the periplasm. We discovered that 3,3-bis-di(methylsulfonyl)propionamide
(3bMP1) inhibits the growth ofM. tuberculosis, and resistance to this compound is conferred by mutation
within a member of the proline-proline-glutamate (PPE) family, PPE51. Deletion of PPE51 rendered
M. tuberculosiscells unable to replicate on propionamide, glucose, or glycerol. Growth was restored
upon loss of the mycobacterial cell wall component phthiocerol dimycocerosate. Mutants in other
proline-glutamate (PE)/PPE clusters, responsive to magnesium and phosphate, also showed a
phthiocerol dimycocerosate–dependent growth compromise upon limitation of the corresponding
substrate. Phthiocerol dimycocerosate determined the low permeability of the mycobacterial outer
membrane, and the PE/PPE proteins apparently act as solute-specific channels.
M
ycobacterium tuberculosisthrives in
the hostile environment of the human
lung despite an onslaught of host re-
sponses that prevent the growth of near-
ly all other bacteria ( 1 ).M. tuberculosis
survives because it is covered by a highly im-
permeable envelope and is able to manipulate
the host defense mechanisms ( 2 ). Outside of the
cytoplasmic membrane is the mycobacterial
envelope, which consists of a complex hetero-
polymer composed of peptidoglycan covalent-
ly attached to arabinogalactan terminated by
mycolic acids, which are specific to myco-
bacteria ( 3 ). Cryo–electron microscopy shows
that the outer lipid layer of the mycobacterial
outer membrane is ~8 to 9 nm in width ( 4 , 5 ).
No canonical porin protein has been identi-
fied in slow-growing mycobacteria, such as
M. tuberculosis, although a major porin (MspA)
has been identified in fast-growing mycobac-
terial species, such asM. smegmatis( 6 ). How-
ever, porin-like activity has been observed in
membrane preparations from slow-growing
mycobacteria ( 7 , 8 ), but no protein determinants
of this activity have been identified. Express-
ion of the porin from fast-growing mycobac-
teria inM. tuberculosissubstantially reduces
virulence and enhances antibiotic suscepti-
bility ( 9 , 10 ). A canonical porin is incompatible
with the need for a permeability barrier that is
extreme enough to allowthe intracellular sur-
vival ofM. tuberculosis.
As part of our drug discovery efforts, we
identified a small-molecule whole-cell inhibi-
tor from a library of agrichemicals. This simple
molecule, 3,3-bis-di(methylsulfonyl)propionamide
(3bMP1), displayed a modest minimum inhibitory
concentration (MIC) of ~3mg/ml independ-
ent of growth media and was highly bacteri-
cidal, reducing starting colony forming units
(CFU) by 3 logs after 4 days and achieving
near-sterilization after 8 days of incubation
(Fig. 1, A and B). To identify the mechanism ofaction of this compound, we selected for resist-
ant mutants by plating wild-typeM. tuberculosis
cells on media containing 10-fold the MIC of
3bMP1, and we observed mutants growing
after 5 weeks of incubation. The resistance of
these mutants was confirmed in liquid cul-
tures, showing an MIC of 25mg/ml in 7H9/OAD
medium containing 0.05% Tween 80 [a deter-
gent typically added to tuberculosis (TB) cul-
tures to disperse the bacteria] and a much
higher MIC in the same medium without Tween
80(Fig.1,CandD).Whole-genomesequencing
of 10 clones from this experiment revealed that
all had scattered mutations in a single pro-
tein, proline-proline-glutamate (PPE) family
protein 51 (PPE51) (fig. S1A). Complementation
of these mutants with a wild-type copy of the
ppe51gene fully restored sensitivity to 3bMP1
(Fig. 1, C and D), which is consistent with what
were apparently loss-of-function mutations in
ppe 51 (fig. S1A).
The PE/PPE family of proteins, named for
the homologous proline-glutamate (PE) or
proline-proline-glutamate (PPE) repeated
regions in their N terminus, comprises 169 or-
thologs. These proteins represent nearly 10%
of theM. tuberculosisgenome, and many are
substrates for the type VII secretion systems of
M. tuberculosis( 11 ). Because they are secretedRESEARCH
Wanget al.,Science 367 , 1147–1151 (2020) 6 March 2020 1of5
(^1) Tuberculosis Research Section, Laboratory of Clinical
Immunology and Microbiology, National Institute of Allergy
and Infectious Diseases, National Institutes of Health,
Bethesda, MD 20892, USA.^2 Drug Discovery Unit, College of
Life Sciences, James Black Centre, University of Dundee,
Dundee DD1 5EH, UK.^3 Exscientia Ltd., Oxford OX1 3LD, UK.
(^4) Institute of Infectious Disease and Molecular Medicine,
University of Cape Town, Cape Town 7925, South Africa.
*Corresponding author. Email: [email protected]
S
S
O O
O
O
NH 2
O
initialDMSO
Rif 5 μg/ml
100
102
104
106
108
1010
3 log
8 days
initialDMSO
3bMP1 1X3bMP1 5X3bMP1 10XRif 5 μg/ml
100
102
104
106
108
CFU/ml
3 log
4 days
3bMP1
A
B
*
3bMP1 1X3bMP1 5X3bMP1 10X
10 -1
0
50
100
(^150) wt
R1
R2
R1+ppe51
R2+ppe51
100 101 102 103
10 -1 100 101 102 103
0
50
100
150
Bacteria Growth (%)
Drug concentration (μg/ml)
Drug concentration (μg/ml)
Bacteria Growth (%)
D
C
Fig. 1. Loss-of-function mutations inppe51result in the resistance to compound 3bMP1.(A) The chemical
structure of 3bMP1. (B) Logarithmically growingM. tuberculosiscells were exposed for 4 and 8 days to 3bMP1
at 1×, 5×, and 10× MIC values. Rifampicin (Rif) and dimethyl sulfoxide (DMSO) were used as positive and
negative controls, respectively. Data are generated from two independent experiments and shown as mean ± SD
(P<0.05,**P< 0.01, versus DMSO control, unpairedttest). (CandD) Susceptibility ofM. tuberculosisstrains to
3bMP1 in 7H9/OAD media with (C) or without (D) 0.05% Tween 80. Bacterial growth was quantified using an
alamarBlue-based assay. R1 contains a G18D (Gly^18 →Asp) mutation in the PPE51. R2 contains insertion sequence
IS6110 at codon 247 in theppe51gene. Data are representative of two independent experiments, both done as
technical duplicates, and error bars represent SD [in (C):P= 0.0007, wild type (wt) versus R1;P=0.0107,wtversus
R2; no significant difference among wt, R1+ppe51,andR2+ppe51;in(D):P< 0.0001, wt versus R1;P< 0.0001,
wt versus R2; no significant difference among wt, R1+ppe51, and R2+ppe51; data were analyzed using two-way
analysis of variance (ANOVA) of matched values].