Host cell cytotoxicity and cytoskeleton disruption by CerADPr, an ADP-ribosyltransferase of Bacillus cereus G9241

Abstract

Bacillus cereus G9241 was isolated from a welder suffering from an anthrax-like inhalation illness. B. cereus G9241 encodes two megaplasmids, pBCXO1 and pBC210, which are analogous to the toxin- and capsule-encoding virulence plasmids of Bacillus anthracis. Protein modeling predicted that the pBC210 LF homologue contained an ADP-ribosyltransferase (ADPr) domain. This putative bacterial ADP-ribosyltransferase domain was denoted CerADPr. Iterative modeling showed that CerADPr possessed several conserved ADP-ribosyltransferase features, including an α-3 helix, an ADP-ribosyltransferase turn-turn loop, and a "Gln-XXX-Glu" motif. CerADPr ADP-ribosylated an ~120 kDa protein in HeLa cell lysates and intact cells. EGFP-CerADPr rounded HeLa cells, elicited cytoskeletal changes, and yielded a cytotoxic phenotype, indicating that CerADPr disrupts cytoskeletal signaling. CerADPr(E431D) did not possess ADP-ribosyltransferase or NAD glycohydrolase activities and did not elicit a phenotype in HeLa cells, implicating Glu431 as a catalytic residue. These experiments identify CerADPr as a cytotoxic ADP-ribosyltransferase that disrupts the host cytoskeleton.

Figure 1. CerADPr selectively ADP-ribosylates a high molecular weight host protein in HeLa cell membranes

(A) A HeLa cell lysate (Total) was centrifuged to generate cytosol (Cytoplasm) and membranes. Membranes were extracted with 0.1% Triton X-100 and centrifuged to remove the insoluble cell matrix (soluble membranes were termed Tx-100). Fractions were normalized for protein and incubated alone (−ADPr) or with 50ng CerADPr or 20ng ExoS for 1 hr at 25°C alone (−biotin-NAD) or with 4 µM biotin-NAD. The reaction was subjected to SDS-PAGE and proteins transferred to a PVDF membrane, which was probed with streptavidin-conjugated HRP. (B) Tx-100 HeLa cell fraction was incubated alone (0) or with the indicated amount of CerADPr (ng) or ExoS (ng) for 1hr at 25°C with 4 µM biotin-NAD or with 10-fold excess unlabeled NAD (100 +U). The reaction mixture was treated as above. The amount of biotin labeling from the 50ng ExoS lane was set to 100% and other signals were normalized, error bars display SEM. An unpaired t-test was performed on selected columns *p<0.05 **p<0.01

Figure 2. NAD glycohydrolase activity of CerADPr

(A) Recombinant CerADPr was incubated with 5 mM NAD+ for 24, 36, or 48 hr. An aliquot of each reaction was analyzed by PEI-cellulose thin layer chromatography. (B) CerADPr, CerADPr(E431D), or ExoS was incubated with 5 mM NAD+ for 24 hr. An aliquot of each reaction was then subjected to PEI-cellulose TLC. Plates were imaged and the amount of ADP-ribose produced was quantified by densitometry and normalized to an ADP-ribose standard. Quantification is from 3 independent experiments, error bars display SEM. An unpaired t-test was performed on selected columns *p<0.05 **p<0.01

Figure 3. CerADPr domain contains sequence and predicted structural conservation of other bacterial ADP-ribosyltransferases

(A) A predicted structure-based alignment of the Cereus toxin ADPr domain (CerADPr) was generated with the crystal structure of Iota toxin ADP-ribosyltransferase domain Ia (PDB: 1GIR). Three regions of alignment that are common components of ADP-ribosyltransferase family members, including an Arg, the Ser-Thr-Ser motif, and the Gln/Glu-XXX-Glu sequence (“RSE” motif) are shown. The latter defines an active site glutamic acid that is conserved among ADP-ribosyltransferases. Secondary structure labeling is based off the Iota toxin Ia ADP-ribosyltransferase domain, starting with α-helix1 at residues 214–232. (B) CerADPr(E431D) ADP-ribosyltransferase activity was measured relative to CerADPr by incubation with Tx-100 HeLa cell lysate and 4 µM biotin-NAD for 1 hr at 25°C. The reaction mixture was processed as described in Figure 1 and the ADP-ribosylation of the high-molecular weight substrate is shown in the inset image.

Figure 4. Expression of EGFP-CerADPr elicits early actin rearrangement in HeLa cells

HeLa cells were transfected with pEGFP-CerADPr (A) or pEGFP-CerADPr(E431D) (B) for 4 or 5 hr when cells were fixed, permeabilized, and incubated with Alexa Fluor®647-phalloidin to label the actin cytoskeleton. Cells were subjected to epifluorescence microscopy. Representative images are shown for each condition. Inset shows close-up of actin spikes seen at cell periphery of EGPF-CerADPr transfected cells.

Figure 5. CerADPr does not covalently modify actin or tubulin

(A) HeLa cells were transfected with either pEFGP, pEGFP-CerADPr(E431D), or pEGFP-CerADPr. At 5 and 20 hr post-transfection (HPT), cells were harvested and subjected to SDS- PAGE. The gel was transferred to PVDF and probed independently with α-GFP, α-actin and α-tubulin primary antibodies. Membranes were processed as described in Figure 1. (B) Cell lysates were prepared from HeLa cells transfected with either pEFGP, pEGFP-CerADPr(E431D), or pEGFP-CerADPr and incubated with the TX-100-soluble membrane fraction of HeLa cells and biotin-NAD. After 2 hr, the reaction was subjected to SDS-PAGE and proteins were transferred to PVDF membranes and probed for the ADP-ribosylation of the high-molecular weight substrate of CerADPr and actin reactivity measured to normalize the cell lysates.

Figure 6. CerADPr ADP-ribosylates the 120KDa target within HeLa cells

(A) HeLa cells were transfected with pEGFP-CerADPr, pEGFP-CerADPr(E431D), or pExoS for 2 hr. Cells were treated with 200 ng/ml of tetanolysin and then incubated with biotin-NAD for 2.5 hr. Cell lysates were resolved by SDS-PAGE and biotin-ADP-ribose incorporation was visualized by immunoblotting with streptavidin-HRP. Actin was used as a loading control. (B) HeLa cells were treated with 200 ng/ml of tetanolysin, then incubated with 3 µg/ml recombinant CerADPr, CerADPr(E431D), or ExoS for 2 hr. Cell lysates prepared and then incubated with biotin-NAD and recombinant CerADPr for an hour. Lysates were resolved by SDS-PAGE, transferred to PVDF membranes, where biotin-ADP-ribose incorporation was visualized with streptavidin-HRP. Actin was used as a loading control.

Figure 7. CerADPr is cytotoxic for HeLa cells

HeLa cells were grown to ~70 % confluence in 6-well dishes and transfected with 500 ng of either pExoS, pEGFP-CerADPr(E431D), or pEGFP-CerADPr. At 5 and 20 hr post-transfection (HPT), media was removed, cells were washed and stained for 5 min with trypan blue. Trypan blue was removed, cells were washed, and color images were obtained as described in Methods. Cytotoxicity was determined by counting the number of trypan blue positive HeLa cells / total number of HeLa cells per field. Five random fields were assayed from three biological replicates, and the results of this determination shown. 1-way ANOVA was performed on all columns ** p<0.01.