Mycobacterium tuberculosis EspK Has Active but Distinct Roles in the Secretion of EsxA and EspB

Abstract

The Mycobacterium tuberculosis type-7 protein secretion system ESX-1 is a major driver of its virulence. While the functions of most ESX-1 components are characterized, many others remain poorly defined. In this study, we examined the role of EspK, an ESX-1-associated protein that is thought to be dispensable for ESX-1 activity in members of the Mycobacterium tuberculosis complex. We show that EspK is needed for the timely and optimal secretion of EsxA and absolutely essential for EspB secretion in M. tuberculosis Erdman. We demonstrate that only the EsxA secretion defect can be alleviated in EspK-deficient M. tuberculosis by culturing it in media containing detergents like Tween 80 or tyloxapol. Subcellular fractionation experiments reveal EspK is exported by M. tuberculosis in an ESX-1-independent manner and localized to its cell wall. We also show a conserved W-X-G motif in EspK is important for its interaction with EspB and enabling its secretion. The same motif, however, is not important for EspK localization in the cell wall. Finally, we show EspB in EspK-deficient M. tuberculosis tends to adopt higher-order oligomeric conformations, more so than EspB in wild-type M. tuberculosis. These results suggest EspK interacts with EspB and prevents it from assembling prematurely into macromolecular complexes that are presumably too large to pass through the membrane-spanning ESX-1 translocon assembly. Collectively, our findings indicate M. tuberculosis EspK has a far more active role in ESX-1-mediated secretion than was previously appreciated and underscores the complex nature of this secretion apparatus. IMPORTANCE Mycobacterium tuberculosis uses its ESX-1 system to secrete EsxA and EspB into a host to cause disease. We show that EspK, a protein whose role in the ESX-1 machinery was thought to be nonessential, is needed by M. tuberculosis for optimal EsxA and EspB secretion. Culturing EspK-deficient M. tuberculosis with detergents alleviates EsxA but not EspB secretion defects. We also show that EspK, which is exported by M. tuberculosis in an ESX-1-independent manner to the cell wall, interacts with and prevents EspB from assembling into large structures inside the M. tuberculosis cell that are nonsecretable. Collectively, our observations demonstrate EspK is an active component of the ESX-1 secretion machinery of the tubercle bacillus.

Keywords: ESX-1; EspK; Mycobacterium tuberculosis; protein secretion; virulence.

FIG 1

Characterization of the M. tuberculosis espK::Tn mutant. (A) Schematic of transposon insertion into the espK gene in the esx-1 locus of M. tuberculosis Erdman. (B) Growth comparison of wild-type M. tuberculosis Erdman (WT) and isogenic espK::Tn mutant strain in 7H9 (left) and Sauton’s (right) medium (average of 2 independent experiments; error bars represent the standard deviations). (C) Immunoblots of CF (10 μg/well) and CL total protein (5 μg/well) of wild-type M. tuberculosis Erdman (WT), espK::Tn, and 5′ Tn::pe35 strains cultured 3 days in Sauton’s medium without detergent to examine indicated proteins (representative of 3 independent experiments). (D) Cytotoxicity assay in THP-1 cells infected with wild-type M. tuberculosis Erdman (WT), espA::Tn, and espK::Tn strains at an MOI of 5 (data represent means and standard errors of means from 4 independent experiments with replicate wells; the y axis indicates cytotoxicity values relative to uninfected control; significance in difference calculated using Student's t test). (E) Bacterial burden after 3 and 6 weeks postinfection in lungs (top) and spleens (bottom) of mouse aerosol challenged with wild-type M. tuberculosis Erdman (WT) and espK::Tn strain (each data point indicates the bacterial burden per mouse, with median and standard deviations from 4 mice per group per time point shown; significance in difference calculated using Student's t test; NS, not significant).

FIG 2

Cellular localization of EspK in M. tuberculosis Erdman strains. (A to C) Immunoblots of increasing amounts of total CF protein of wild-type M. tuberculosis Erdman (WT) versus espK::Tn mutant strain (A), 10 μg/well each of CF, CW, CM, and cytosol proteins from wild-type M. tuberculosis Erdman (WT) versus espK::Tn strain (B), and wild-type M. tuberculosis Erdman (WT) versus 5′ Tn::pe35 strain (C) (each set of blots is representative of at least 2 independent experiments; L, protein ladder).

FIG 3

Secretion and expression of ESX-1-associated proteins in liquid culture media over time. Immunoblots of CF (10 μg/well) and CL proteins (5 μg/well) of wild-type M. tuberculosis Erdman (WT) and espK::Tn mutant cultured 3, 5, and 7 days in Sauton’s medium without detergent (A) and modified 7H9 medium without ADC and detergent supplementation (B) (each set of blots is representative of 3 independent experiments).

FIG 4

Secretion and expression of ESX-1 associated proteins in WT M. tuberculosis and espK::Tn cultured in liquid media with and without detergents. (A and B) Immunoblots of CF (10 μg/well) and CL (5 μg/well) total protein of wild-type M. tuberculosis Erdman (WT) and espK::Tn mutant cultured for 3 days in Sauton’s medium with and without Tween 80 (A) and Sauton’s medium with and without tyloxapol (B) to examine indicated proteins (each set of blots are representative of 3 independent experiments).

FIG 5

Secretion and expression of ESX-1 associated proteins by WT M. tuberculosis and complemented espK::Tn. (A and B) Immunoblots of CL proteins (5 μg/well) from WT M. tuberculosis and espK::Tn strain complemented with pMD31 (empty vector), pMDespK^WT, pMDespK^W62R, and pMDespK^G64R cultured 3 days in Sauton’s medium without detergent to examine EspK levels (A) and CF (10 μg/well) and CL proteins (5 μg/well) from WT M. tuberculosis and espK::Tn strain complemented with pMD31, pMDespK^WT, pMDespK^W62R, and pMDespK^G64R cultured 3, 5, and 7 days in Sauton’s medium without detergent (B) (each set of blots is representative of 3 independent experiments).

FIG 6

Importance of the conserved W-X-G motif to EspK export and cell wall localization. Immunoblots of 10 μg/well each of CF, CW, CM, and cytosol proteins from espK::Tn strain complemented with pMDespK^WT, pMDespK^W62R, and pMDespK^G64R (blot is representative of at least 2 independent experiments; L, protein ladder).

FIG 7

Pulldown analysis of EspB and EspK coexpressed in E. coli. (A and B) Immunoblots of EspK pulled down by His tag Dynabeads from the lysates of E. coli/His tag EspB^WT + S tag EspK^WT and E. coli/S tag EspK^WT (A) and EspB pulled down by S-protein agarose beads from the lysates of E. coli/His tag EspB^WT + S tag EspK^WT and E. coli/His tag EspB^WT (B). (C and D) Immunoblots of EspB and EspK pulled down by His tag Dynabeads (C) and S-protein agarose beads (D) from the lysates of E. coli/His tag EspB^WT + S tag EspK^WT, E. coli/His tag EspB^WT + S tag EspK^W62R, and E. coli/His tag EspB^WT + S tag EspK^G64R (each set of blots is representative of at least 2 independent experiments; I, input 5 μg of lysates; E, all proteins eluted from beads in 50 μL of SDS-PAGE sample buffer).

FIG 8

EspB native conformation analysis. (A and B) Native PAGE and EspB immunoblots of CL proteins (5 μg/well) from wild-type M. tuberculosis Erdman (WT) and espK::Tn strain cultured for 3 days in Sauton’s medium without Tween 80 (A) and CF protein (10 μg/well) of espK::Tn complemented with pMDespK^WT and CL proteins (5 μg/well) of espK::Tn complemented with pMDespK^WT, pMDespK^W62R, pMDespK^G64R, and pMD31 cultured for 4 days in Sauton’s medium without detergent (B) (each set of blots are representative of at least 2 independent experiments; asterisks indicate higher-order oligomers of EspB).