Instability of toxin A subunit of AB(5) toxins in the bacterial periplasm caused by deficiency of their cognate B subunits

Abstract

Shiga toxin (STx) belongs to the AB(5) toxin family and is transiently localized in the periplasm before secretion into the extracellular milieu. While producing outer membrane vesicles (OMVs) containing only A subunit of the toxin (STxA), we created specific STx1B- and STx2B-deficient mutants of E. coli O157:H7. Surprisingly, STxA subunit was absent in the OMVs and periplasm of the STxB-deficient mutants. In parallel, the A subunit of heat-labile toxin (LT) of enterotoxigenic E. coli (ETEC) was absent in the periplasm of the LT-B-deficient mutant, suggesting that instability of toxin A subunit in the absence of the B subunit is a common phenomenon in the AB(5) bacterial toxins. Moreover, STx2A was barely detectable in the periplasm of E. coli JM109 when stx2A was overexpressed alone, while it was stably present when stxB was co-expressed. Compared with STx2 holotoxin, purified STx2A was degraded rapidly by periplasmic proteases when assessed for in vitro proteolytic susceptibility, suggesting that the B subunit contributes to stability of the toxin A subunit in the periplasm. We propose a novel role for toxin B subunits of AB(5) toxins in protection of the A subunit from proteolysis during holotoxin assembly in the periplasm.

Fig. 1

TEM and IB analyses for characterization of the ΔstxB mutants. (A) TEM visualization of OMVs from Sakai-DM/Δstx2B mutant. The image showed typical sizes of OMVs ranges 20–100 nm in diameter. (B, C, and D) Immunoblots for detection of STx2A subunit in the PPE and OMV samples. STx2A was identified in the periplasm and OMVs of the parental O157 strains (B) by using anti-STx2A mAb, however, it was absent in the PPE samples (lanes 2 and 4, C) and the OMVs (lane 2, D) of the Δstx2B mutants. (E) STx1B subunit was probed with anti-STx1B mAb in the periplasm of the Δstx2B mutant (lane 1) but was absent in the Δstx1B mutant (lane 2).

Fig. 2

IB analysis of periplasmic proteins extracted from STEC and ETEC mutants. (A) The STx2A subunit was detected in the PPE samples of STEC O157 Sakai derivatives. STx2A band was present only in the wild type strain (lane 1) but absent in the both Δstx2B and Δstx2B/ΔdegP::Cm mutants (lanes 2 and 3, respectively). Inactivation of degP in the Δstx2B mutant background did not rescue the STx2A band in the periplasm (lane 3), suggesting that DegP might be not associated with the absence of STx2A. BAP (bacterial alkaline phosphatase as a periplasmic protein marker) was used as an internal control for the PPE samples. (B) PPE samples of ETEC H10407 and its LT-B-deficient mutant (ETEC/dLT-B) were analyzed by using anti-LT-A mAb (Abcam). As shown, the LT-A subunit (~28 kDa band) was absent in the Δlt-B mutant (lane 2), compared to wild type strain (lane 1).

Fig. 3

IB analysis of PPE samples of JM109 transformants. (A) PPE samples prepared from the JM109 cells harboring the respective pBAD-stx2A and pBAD-stx2AB clone were immunoblotted with anti-STx2A mAb. (B) For the complementation of JM109/pBAD-stx2A, plasmid pAY-stx1B or pAY-stx2B was introduced. STx2A band was strongly detected in the complementation with either pAY-stx2B or pAY-stx1B, whereas STx2A subunit overexpressed in the JM109/pBAD-stx2A strain (lane 1) was hardly detectable by IB with anti-STx2A mAb. (C) IB was performed to see whether A1 fragment of STx2A can exist alone in the periplasm of JM109. PPE of JM109/pBAD-stx2A1 contained the A1 fragment, albeit at reduced amount (lane 4) compared to the A1 fragment complemented with STx2B (lane 5).

Fig. 4

In vitro STx2A susceptibility to proteolytic degradation by periplasmic proteases. (A and B) The purified STx2 holotoxin (Toxin Technology Inc., USA) and the free STx2A subunit fractionated by HPLC as described previously [19] were subjected to proteolytic susceptibility assay with the PPE from JM109. The inset (A) showed STx2-A and -B subunit separated by SDS-PAGE. (C) Approximately half of the free STx2A (2.2-fold reduction in the band intensity) was degraded in 5 min co-incubation with the PPE (10 μg) from JM109 at 37 °C, while the holotoxin (5 μg) remained intact after the co-incubation with the source of periplasmic proteases.

Fig. 5

Ribbon structure of STx2 toxin. Structures of STx2 holotoxin (A) and STx2A (B) were obtained from the Protein Data Bank (accession code: 1R4P) and displayed after manipulation by using the PyMOL Molecular Graphics System, Version 1.3, Schrödinger, LLC. The 32 kDa STx2A subunit, subdivided as A1 (yellow) and A2 (magenta) fragments in the panel A, is known to be non-covalently associated with five STx2B (7.7-kDa) subunits (green). In the absence of the B-pentamer (panel B), the A2-domain (magenta) of STx2A would be then exposed to periplasmic proteases. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)