Cytosolic Delivery of Multidomain Cargos by the N Terminus of Pasteurella multocida Toxin

The zoonotic pathogen Pasteurella multocida produces a 146-kDa modular toxin (PMT) that enters host cells and manipulates intracellular signaling through action on its Gα protein targets. The N terminus of PMT (PMT-N) mediates cellular uptake through receptor-mediated endocytosis, followed by the delivery of the C-terminal catalytic domain from acidic endosomes into the cytosol.

ABSTRACT

The zoonotic pathogen Pasteurella multocida produces a 146-kDa modular toxin (PMT) that enters host cells and manipulates intracellular signaling through action on its Gα protein targets. The N terminus of PMT (PMT-N) mediates cellular uptake through receptor-mediated endocytosis, followed by the delivery of the C-terminal catalytic domain from acidic endosomes into the cytosol. The putative native cargo of PMT consists of a 710-residue polypeptide with three distinct modular subdomains (C1-C2-C3), where C1 contains a membrane localization domain (MLD), C2 has an as-yet-undefined function, and C3 catalyzes the deamidation of a specific active-site glutamine residue in Gα protein targets. However, whether the three cargo subdomains are delivered intact or undergo further proteolytic processing during or after translocation from the late endosome is unclear. Here, we demonstrate that PMT-N mediates the delivery of its native C-terminal cargo as a single polypeptide, corresponding to C1-C2-C3, including the MLD, with no evidence of cleavage between subdomains. We show that PMT-N also delivers nonnative green fluorescent protein (GFP) cargo into the cytosol, further supporting that the receptor-binding and translocation functions reside within PMT-N. Our findings further show that PMT-N can deliver C1-C2 alone but that the presence of C1-C2 is important for the cytosolic delivery of the catalytic C3 subdomain by PMT-N. In addition, we further refine the minimum C3 domain required for intracellular activity as comprising residues 1105 to 1278. These findings reinforce that PMT-N serves as the cytosolic delivery vehicle for C-terminal cargo and demonstrate that its native cargo is delivered intact as C1-C2-C3.

INTRODUCTION

Toxinogenic capsular type D and some capsular type A strains of the zoonotic pathogen Pasteurella multocida produce a 146-kDa modular protein toxin (PMT), which is the major virulence factor responsible for the symptoms of dermonecrosis, atrophic rhinitis, and pasteurellosis in infected animals and humans (reviewed in reference 1). PMT acts intracellularly by gaining entry into host cells via binding to the membrane phospholipids phosphatidylcholine and sphingomyelin and an as-yet-unknown trypsin-sensitive protein partner (2). Once bound to its receptor, PMT is endocytosed, and its toxic cargo is delivered from the late endosome into the cytosol in a pH-dependent manner (3, 4). In the cytosol, the catalytic cargo of PMT modulates downstream signaling pathways (reviewed in references 5 and 6) through the selective deamidation of a specific active-site Gln residue in target Gα subunits that results in their activation (7–11). The initial activation of Gq protein signaling is subsequently followed by downregulation through membrane redistribution and depletion of the Gα_q protein (12).

Discrete functional domains mediate the binding and entry of PMT into host cells, followed by the cytosolic delivery of the components responsible for intracellular activity (reviewed in reference 4). The structure of the C terminus of PMT (residues 569 to 1285) (13), which harbors the intracellular toxic activity, revealed three distinct subdomains (Fig. 1) that serve as functional modules: C1 (residues 575 to 719), with membrane-targeting function, where amino acid residues 589 to 668 comprise the putative membrane localization domain (MLD) (14, 15); C2 (residues 720 to 1104), with an as-yet-unknown function; and C3 (residues 1105 to 1285), with the minimal catalytic domain having Gln deamidase activity (16, 17).

FIG 1.

Crystal structure of the C-terminal catalytic domain of PMT. Shown is the crystal structure of the PMT C terminus (amino acid residues 575 to 1285), rendered in PyMOL from the structure reported under PDB accession number 2EBH. The C1 subdomain is depicted in red (residues 575 to 668) and blue (residues 669 to 719), where red highlights the region containing the membrane localization domain (MLD). The C2 subdomain (amino acid residues 721 to 1104) is depicted in green. The C3 catalytic subdomain is depicted in magenta (residues 1105 to 1262), black (residues 1263 to 1277), and orange (residues 1278 to 1285).

Although there is no crystal structure available for the N terminus of PMT (PMT-N), several biochemical and mutagenesis studies have demonstrated that PMT-N harbors the functional module(s) responsible for binding to cell receptors, triggering cellular uptake and trafficking to endosomes, membrane translocation, and cytosolic delivery of cargo (2, 3, 18–20). Previous studies have defined the minimal cargo delivery domain of PMT-N as consisting of amino acid residues 1 to 505 (19). Mutations of residues in a predicted helix-loop-helix motif (residues 402 to 457) blocked pH-dependent entry into the cytoplasm, suggesting the involvement of this region in the translocation of the activity cargo (18). PMT-N also mediates the cellular uptake of nonnative cargo, including green fluorescence protein (GFP) (3), and the delivery of catalytic domains of other toxins (e.g., diphtheria toxin [DTa]) (19). Some insights into the molecular events involved in PMT intracellular trafficking have been gained (3). However, we do not have a complete understanding of the role of each of the discrete C1-C2-C3 cargo subdomains of PMT in efficient cytosolic delivery, which subdomains or fragments thereof are actually delivered, or whether the delivered subdomains undergo further processing after cellular uptake or once in the cytosol.

Here, we present experimental evidence that PMT delivers its cargo as an intact single unit comprised of subdomains C1-C2-C3, including the MLD, from endosomes into the cytosol in a pH-dependent manner, without further processing. Furthermore, we show that truncated PMT, lacking C3 or a portion thereof, also delivers the respective truncated cargo fragments, and we demonstrate that C1-C2 subdomains are important for the cytosolic delivery of C3 by PMT-N. We also refine the minimal C3 module required for intracellular signaling activity. Moreover, we show that PMT-N delivers exogenous C-terminal GFP cargo into the cytosol in a pH-dependent manner. This study offers insight into the functional organization of single-chain A-B-like protein toxins with regard to the nature and contribution of the cargo in the cytosolic delivery process. This information may be useful for the protein engineering of bacterial toxin-based platforms, such as the previously described bacterial toxin-inspired drug delivery (BTIDD) platform (21), for cytosolic delivery of novel therapeutics.

RESULTS

Cytosolic cargo delivery of C-terminal PMT fragments by PMT-N.

We first asked whether the entire, intact C terminus of PMT (comprised of C1-C2-C3) or portions thereof are delivered into the cytosol of host cells. For sensitive immunoblot detection, we used hemagglutinin (HA)-tagged protein fusions at the C terminus or N terminus of PMT to examine toxin localization within cells by subcellular fractionation and immunoblot analysis. Labeling of the N terminus of PMT with an HA tag was found to interfere with cell binding (data not shown) and so was not further explored. While designing C-terminally HA-tagged protein fusions, we considered that lysine and arginine residues present near or within loops between residues 1260 and 1285 (K1263, K1278, R1280, and R1283) might be subject to trypsin-like proteases, such that the HA tags could be lost upon proteolytic cleavage, thus rendering the toxin undetectable via immunoblotting. We also considered whether this positively charged region was critical for the stability or cytosolic delivery of the C terminus of PMT. Finally, we were concerned that PMT-induced cell rounding and detachment might diminish the cellular uptake and accumulation of HA-tagged proteins. To address these issues, we generated and purified a series of recombinant toxin fusions with HA tags directly at the C terminus (Fig. 2A and B): after residue 1285 for full-length PMT (PMT-HA) and a catalytically inactive mutant (PMT^C1165S-HA) and, for truncations, after residue 1277 (PMT-1277-HA) and residue 1262 (PMT-1262-HA).

FIG 2.

Cytosolic cargo delivery of C-terminal catalytic domain fragments by PMT-N. (A) Schematic diagram of recombinant His₆-tagged PMT-HA toxin and variants thereof constructed for this study. For PMT-C1-C2-HA-C3 and PMT-C1-C2-HA-HA-C3, HA tags we inserted at position 1105 or both positions 1105 and 1118. (B) Representative Coomassie blue-stained protein gels (left) and corresponding immunoblots (right) of the recombinant PMT-HA toxin variants used in this study, as indicated. (C) Representative immunoblot of subcellular fractions demonstrating the benchtop centrifugation technique for subcellular fractionation, as outlined in Materials and Methods. Anti-β-tubulin antibody was used as a cytosolic protein marker, and anti-Rab5 antibody was used as an endosomal vesicle marker for the same immunoblot membrane. (D) Representative immunoblot profile obtained from subcellular fractionation of Swiss 3T3 cells treated with 500 nM PMT-1262-HA for 18 h. As subcellular fractionation controls, samples of the same corresponding treatment fractions were used to blot for the cytosolic marker β-tubulin and the endosomal vesicle marker Rab5. (E) Representative immunoblot of cytosolic fractions showing the subcellular localization of HA-tagged protein fragments in Swiss 3T3 cells treated with the indicated PMT-HA variants at 500 nM for 18 h. The top and middle panels show 5-min and 15-min exposures, respectively, of the radiography film to the immunoblot nitrocellulose membrane. The bottom panel is the same immunoblot membrane blotted for the cytosol marker β-tubulin as a loading control. (F) Representative immunoblot of intracellular fractions showing HA-tagged protein fragments present in HEK-293T cells treated with the indicated PMT-HA variants at 250 nM for 4 h. The top and bottom panels are 5-min and 2-min exposures, respectively, of radiography film to the immunoblot nitrocellulose membrane. Controls are mock-treated cells with no toxin in DMEM. Single asterisks indicate the size of the expected protein band for PMT-C1-HA. Double asterisks indicate anti-HA antibody cross-reactive material. Arrows indicate noncleaved full-length toxin. Arrowheads indicate the expected cleaved HA-tagged protein fragments. Numbers directly to the left of gels and immunoblots indicate molecular masses (kilodaltons) based on standard protein markers (M). For each gel or blot of fractions from cell treatments, lanes were loaded with equal volumes of the samples.

To determine which portion of the C terminus of PMT is delivered into the cytosol, toxin-treated cells were first subjected to subcellular fractionation to separate contents into two fractions: one comprised of nuclei, large organelles, and cellular debris and the other (denoted the intracellular fraction) comprised of small membrane vesicles and soluble cytosolic components. The intracellular fraction was further separated into membrane vesicles (vesicle fraction) and soluble cytosolic components (cytosolic fraction) (a representative separation technique example is shown in Fig. 2C).

Swiss 3T3 cells were first treated with PMT-1262-HA, PMT-1277-HA, PMT-HA, and PMT^C1165S-HA. The cytosolic fractions were resolved via SDS-PAGE and then visualized by using immunoblot analysis. As expected, while the corresponding full-length HA-tagged toxin could be detected in the respective vesicle fraction, none was observed in the cytosol (a representative example profile for PMT-1262-HA is shown in Fig. 2D). The cytosolic fractions for each of the fusion proteins revealed protein bands with sizes close to the expected size of the C-terminal fragment of ∼80 kDa for intact cargo of the respective HA-tagged toxin (Fig. 2E), although the corresponding band for PMT-1277-HA was weaker than the others. These results indicate that the entire C-terminal cargo is delivered intact into the cytosol.

We also asked whether the entire, intact C1-C2-C3 module, as depicted in the crystal structure (Fig. 1), was required for the delivery of the C-terminal cargo. To address this, we made a truncation of PMT where the C3 subdomain was deleted (PMT-C1-C2-HA). A truncation where C2-C3 was deleted (PMT-C1-HA) was not stably expressed (Fig. 2A and B). The cytosolic fraction of cells treated with PMT-C1-C2-HA showed a protein band near the expected C-terminal fragment size of ∼60 kDa for intact cargo of PMT-C1-C2-HA (Fig. 2E), indicating that C1-C2 alone could be delivered intact.

Delivered PMT cargo does not undergo further proteolytic processing between C-terminal subdomains.

Next, we asked if C-terminal fragments of PMT cargo undergo any further proteolytic cleavage between subdomains. To address this question, we generated full-length PMT with one or two HA tags inserted between subdomains C2 and C3 near potential trypsin-like protease cleavage sites (PMT-C1-C2-HA-C3 and PMT-C1-C2-HA-HA-C3) (Fig. 2A and B). The intracellular fractions of HEK-293T cells treated with PMT-C1-C2-HA-C3, PMT-C1-C2-HA-HA-C3, PMT-1262-HA, or PMT-C1-C2-HA were resolved via SDS-PAGE and then visualized by using immunoblot analysis. No additional fragments with a size smaller than the size of the expected cargo were detectable for any of the HA-tagged toxins in the intracellular fractions (Fig. 2F).

Delivered PMT C-terminal cargo includes the MLD.

Although we confirmed that the C-terminal cargos of PMT-N were delivered to the cytosol intact, and it appeared that the delivered cargo was not cleaved between subdomains C1 and C2 or between subdomains C2 and C3, we noticed that the observed protein bands were near, but not exactly, where we expected them to appear on the immunoblots. Fragment bands consistently migrated at lower molecular masses than expected, suggesting a smaller size. For instance, the PMT-C1-C2-HA, PMT-1262-HA, PMT-1277-HA, and PMT-HA fragments were expected to appear at 61, 79, 80, and 81 kDa, respectively, but consistently migrated at lower molecular masses in each case (Fig. 2E). This size discrepancy led us to hypothesize that part of the N terminus of the expected fragment was not being delivered into the cytosol or was cleaved off once in the cytosol. We speculated that the loss of the MLD from the C1 subdomain could account for the observed size discrepancy. To address whether the MLD was being lost during or after delivery, we generated and purified recombinant HA-tagged C1-C2 and C1-C2-C3(Δ1263–1285) truncated proteins (Fig. 3A and B), where the MLD was either removed or left in place (denoted ΔMLD_C1-C2-HA, MLD_C1-C2-HA, ΔMLD_1262-HA, and MLD_1262-HA). We chose to examine the delivered cargos from treatment with PMT-C1-C2-HA and PMT-1262-HA because the smaller cargos would provide better size resolution on immunoblots. Swiss 3T3 cells treated with PMT-C1-C2-HA and PMT-1262-HA were subjected to subcellular fractionation, and the resulting cytosolic fractions were resolved and analyzed alongside the purified recombinant proteins ΔMLD_C1-C2-HA, MLD_C1-C2-HA, ΔMLD_1262-HA, and MLD_1262-HA (Fig. 3C). Delivered cytosolic PMT-1262-HA and PMT-C1-C2-HA cargo fragments migrated on the gel at approximately the same distance as the corresponding purified MLD_C1-C2-HA and MLD_1262-HA proteins, where the difference between observed and expected migrations is attributed to the presence of the 1.5-kDa N-terminal His₆ tag and linker on the recombinant MLD_C1-C2-HA and MLD_1262-HA proteins. These results indicate that cargo delivered by PMT-N into the cytosol is intact and includes the MLD.

FIG 3.

The membrane localization domain is included in the delivered cytosolic cargo. (A) Schematic diagram of the His₆-tagged MLD-HA and ΔMLD-HA protein variants used in this study. (B) Representative Coomassie blue-stained protein gels (left) and immunoblots (right) for the indicated MLD proteins used in this study. Asterisks indicate impurities. (C) Representative immunoblot of cytosolic fractions showing cleaved toxin fragments from cells treated with the indicated PMT-HA variants. Cytosolic fractions were resolved alongside the indicated purified MLD and ΔMLD proteins. Double asterisks indicate anti-HA antibody cross-reactive material. Numbers directly to the left of gels and immunoblots indicate molecular masses (kilodaltons) based on standard protein markers (M). For each gel or blot of fractions from cell treatments, lanes were loaded with equal volumes of the samples.

Cytosolic delivery of C-terminal cargos by PMT-N is time and dose dependent.

To further confirm that C-terminal cargo is delivered into the cytosol by PMT-N by a dose- and time-dependent process, we treated Swiss 3T3 cells with PMT-C1-C2-HA or PMT-1262-HA at various doses or for various times before subjecting the cells to subcellular fractionation, separation of proteins by SDS-PAGE, and immunoblot analysis. After a 4-h treatment at the highest dose tested (5 μM), we were able to detect a prominent protein band in the cytosolic fraction corresponding to the expected HA-tagged protein fragment of each respective protein, while the full-length protein was confined to the vesicle fraction (Fig. 4A). Cargo delivery was both dose dependent (Fig. 4B) and time dependent (Fig. 4C). Taken together, these results show that PMT-N delivers C-terminal cargos into the cytosol of host cells through a time- and dose-dependent mechanism.

FIG 4.

Time, dose, and endosome acidification dependence of cytosolic delivery of cargo by PMT-HA variants. (A) Representative immunoblot of the subcellular fractionation profile for Swiss 3T3 cells treated with 5 μM PMT-C1-C2-HA (top three panels) or PMT-1262-HA (bottom three panels) for 4 h. Samples of the same corresponding treatment were used to blot for the cytosol marker β-tubulin and the endosomal vesicle marker Rab5. (B) Representative immunoblot of cytosolic fractions for the dose response of Swiss 3T3 cells treated with PMT-C1-C2-HA or PMT-1262-HA at the indicated dosages for 4 h. Samples of the same corresponding treatment were used to blot for the cytosol marker β-tubulin. (C) Representative immunoblot of cytosolic fractions for the time course of Swiss 3T3 cells treated with 5 μM PMT-C1-C2-HA or PMT-1262-HA at the indicated times. Samples of the same corresponding treatment were used to blot for the cytosol marker β-tubulin. (D) Representative immunoblot of cytosolic fractions for the response of Swiss 3T3 cells pretreated with the indicated dosages of ammonium chloride for 60 min, followed by treatment with 1 μM PMT-C1-C2-HA for 6 h. The cytosolic marker β-tubulin was used as an internal control. Plotted data are presented as the means ± SD. Filled bars indicate relative ratios of C1-C2-HA/β-tubulin (n = 6, except for 40 mM data points, where n = 3). *, P < 0.05; **, P < 0.01 (for C1-C2-HA levels versus the control, as determined by unpaired two-tailed Student's t test). Numbers directly to the left of gels and immunoblots indicate molecular masses (kilodaltons) based on standard protein markers (M). For each gel or blot of fractions from cell treatments, lanes were loaded with equal volumes of the samples.

Cytosolic delivery of C-terminal cargos is inhibited by ammonium chloride.

To confirm that the delivered C-terminal PMT HA-tagged cargo fragments were being delivered to the cytosol from acidified endosomes, we tested whether ammonium chloride, a known intracellular acidification inhibitor of endosomes, would inhibit the cytosolic delivery of cargo. Swiss 3T3 cells were pretreated for 30 min with increasing doses of ammonium chloride, followed by treatment with PMT-C1-C2-HA. After subcellular fractionation, cytosolic fractions were resolved by SDS-PAGE and subjected to immunoblot analysis. As expected, treatment with ammonium chloride decreased the amount of C1-C2-HA fragments detected in the cytosolic fractions (Fig. 4D), indicating that the inhibition of endosomal acidification prevents the cytosolic delivery of cargo by PMT-N and that translocation occurs from late endosomes.

C-terminal truncations of PMT further refine the minimal C3 catalytic subdomain required for intracellular activity.

Since the HA tags were not lost from full-length PMT (PMT-HA), and the truncations within the C3 domain (PMT-1262-HA and PMT-1277-HA) did not prevent the ability of PMT-N to deliver the respective C-terminal cargos, we were interested in further refining the core region within the C3 domain required for intracellular activity, as evidenced by the use of a previously described serum response element (SRE)-luciferase reporter assay (16, 21). In this assay, PMT-mediated stimulation of downstream mitogenic signaling is determined by measuring luciferase activity from the transcriptional activation of the SRE fused to a luciferase reporter gene. We generated and purified a series of recombinant His₆-tagged full-length PMT and PMT truncation proteins without HA tags, denoted PMT, PMT-1262, PMT-1278, PMT-1280, and PMT-1283 (Fig. 5A and B).

FIG 5.

Refining the minimal C3 subdomain responsible for intracellular activity. (A) Schematic diagram of recombinant PMT with the His₆ tag removed (no His), His₆-tagged PMT, and the indicated His₆-tagged PMT truncation proteins used in this study. (B) Representative Coomassie blue-stained gel of PMT (no His), His₆-tagged PMT, and the indicated His₆-tagged PMT truncation proteins used in this study. Numbers directly to the left of the gels indicate molecular masses (kilodaltons) based on standard protein markers. (C) Effect of the His₆ tag on PMT intracellular activity. Shown are dose-response curves for the intracellular activity of HEK-293T cells transfected with SRE-luciferase gene reporter plasmids and then treated with the indicated proteins at the indicated doses and subjected to an SRE-luciferase assay, as described in Materials and Methods. Relative PMT activation values are in comparison to untreated cells, normalized to the maximum activation of PMT (no His), as indicated and as determined by the 4PL equation for each experiment. (D) Dose-response curves for PMT intracellular activity of C-terminal truncations of PMT. Shown are the dose-response curves for the intracellular activity of the indicated proteins, performed similarly as described above for panel C. Results shown in panels C and D are from 3 independent experiments, where experiments for each data point were performed in triplicate.

To address whether the His₆ tag might interfere with the enzyme activity of the toxin, the His₆ tag was removed via thrombin cleavage from a portion of the purified PMT used in this study, denoted recombinant PMT (rPMT). While the presence of the N-terminal His₆ tag had a detectable effect on the SRE-luciferase activity of PMT versus rPMT (Fig. 5C), we concluded that this minor difference would not interfere significantly with the determination of relative activities among the recombinant proteins. PMT-1280 and PMT-1283 showed wild-type-like SRE-luciferase activity, while PMT-1278 showed approximately 1,000-fold less activity but was able to reach maximal signaling activity in the range of concentrations tested (Fig. 5D). PMT-1262 showed no detectable SRE-luciferase activity in the range tested (Fig. 5D). Previously, we defined the minimal C3 subdomain necessary for the intracellular activity of PMT as comprising residues 1105 to 1285 (16). We now further refine the minimal core C3 subdomain necessary for intracellular signaling activity as comprising residues 1105 to 1278.

PMT-N-GFP is internalized and localizes within early endosomes, and the GFP cargo is delivered into the cytosol via a pH-dependent process.

We previously showed that PMT-N fused with GFP at the C terminus (PMT-N-GFP) (Fig. 6A) is taken up through receptor-mediated endocytosis and localizes in endosomes (3). We first confirmed this by fluorescence microscopy using Swiss 3T3 cells transfected with the fluorescent early-endosome-specific marker DsRed fused to the early endosome antigen 1 (EEA1) protein. After 3 h of PMT-N-GFP treatment, the majority of PMT-N-GFP appeared to localize to early endosomes (Fig. 6B). We then asked if the PMT-N-GFP fusion protein could deliver the GFP cargo into the cytosol. First, we examined the dose- and time-dependent uptake of PMT-N-GFP into Swiss 3T3 cells and the cellular localization of GFP using subcellular fractionation, followed by SDS-PAGE and immunoblot analysis with anti-GFP antibody. Within 1 h of exposure to PMT-N-GFP, full-length PMT-N-GFP was detectable, and the relative amount plateaued and remained constant after 4 h. In contrast, a stable GFP-containing protein fragment of approximately 25 kDa appeared in the intracellular fraction within an hour and accumulated over time (Fig. 6C). The cellular uptake of PMT-N-GFP was also dose dependent, where increasing the amount of PMT-N-GFP increased the presence of the GFP-containing fragment in the intracellular fractions (Fig. 6D).

FIG 6.

Cytosolic delivery of GFP cargo by PMT-N. (A) Schematic diagram and representative Coomassie blue-stained gel of His₆-tagged PMT-N-GFP used in this study. (B) PMT-N-GFP colocalizes with the early endosome (EE). Shown are fluorescent micrographs of Swiss 3T3 cells expressing DsRed2-EEA1 (red), followed by treatment with PMT-N-GFP (green) for 3 h. Colocalization (yellow) was observed in merged images. (C) Time course of cellular uptake of PMT-N-GFP. Shown is a representative immunoblot of intracellular fractions from Swiss 3T3 cells treated for the indicated times with 250 nM PMT-N-GFP. (D) Dose response of cellular uptake of PMT-N-GFP. Shown is a representative immunoblot of intracellular fractions of Swiss 3T3 cells treated for 4 h with the indicated dosages of PMT-N-GFP. (E) Inhibition of endosomal trafficking of PMT-N-GFP by cytochalasin D (cytoD). Shown is a representative immunoblot of intracellular fractions of Swiss 3T3 cells pretreated for 1 h with the indicated dosages of cytochalasin D (millimolar) and then treated for 4 h with 250 nM PMT-N-GFP. (F) Ammonium chloride inhibition of cytosolic delivery of GFP by PMT-N. Shown is a representative immunoblot of intracellular fractions from Swiss 3T3 cells pretreated for 30 min with the indicated dosages of ammonium chloride (millimolar), followed by treatment for 4 h with 250 nM PMT-N-GFP. Numbers directly to the left of gels and immunoblots indicate molecular masses (kilodaltons) based on standard protein markers. For each gel or blot of fractions from cell treatments, lanes were loaded with equal volumes of the samples.

We next considered whether the accumulation of GFP was dependent on cytosolic delivery from acidified endosomes. Inhibition of endosomal trafficking would prevent the formation of late endosomes, where we expected that cargos would be subsequently delivered into the cytosol. Treatment of cells with cytochalasin D, a known actin polymerization inhibitor that prevents endosomal trafficking, prior to PMT-N-GFP treatment blocked the accumulation of the GFP-containing fragment (Fig. 6E). Since this indicated that the trafficking of PMT-N-GFP to late endosomes was inhibited, we expected the GFP-containing fragment to be delivered into the cytosol of host cells from acidified late endosomes. We confirmed this by treating cells with ammonium chloride to prevent acidification before treatment with PMT-N-GFP for 4 h. Intracellular fractions were further fractionated by differential high-speed ultracentrifugation into vesicle and cytosolic fractions. GFP cargo fragments localized in cytosolic fractions (Fig. 6F), and treatment with increasing amounts of ammonium chloride decreased the amount of GFP cargo found in the cytosolic fractions. As expected, full-length PMT-N-GFP was not found in the cytosolic fractions but was confined to the vesicle fractions. This indicated that inhibition of endosomal acidification prevents the cytosolic delivery of GFP-containing cargo by PMT-N and that translocation occurs from late endosomes.

Role of C1-C2 in cytosolic delivery of C3 by PMT-N.

Since PMT-N delivers intact native cargo and truncations thereof, such as C1-C2, and even an exogenous GFP cargo, we asked if PMT-N could deliver the C3 subdomain alone. To address this question, we generated an HA-tagged deletion mutant that has C1-C2 removed, denoted PMT-C3-HA (Fig. 7A and B). Swiss 3T3 cells were treated with 5 μM PMT-C3-HA or PMT^C1165S-HA for 18 h and subsequently subjected to subcellular fractionation, followed by SDS-PAGE and immunoblot analysis. The expected fragment corresponding to C3-HA cargo alone did not appear in the cytosolic fraction of cells treated with PMT-C3-HA. However, as expected, a fragment corresponding to the intact C1-C2-C3^C1165S-HA cargo, but no smaller C3-HA-containing fragments, was observed in the cytosolic fractions of cells treated with PMT^C1165S-HA (Fig. 7C).

FIG 7.

Role of C1-C2 in cytosolic delivery of C3 by PMT-N. (A) Schematic diagram of HA-tagged and His₆-tagged C-terminal deletion mutants of PMT used in this study. (B) Representative Coomassie blue-stained gel (left) and immunoblot (right) of His₆-tagged PMT-C3-HA used in this study. (C) Representative immunoblot of cytosolic fractions from Swiss 3T3 cells treated for 18 h with 5 μM PMT^C1165S-HA or 5 μM PMT-C3-HA, showing cleaved HA-tagged fragments. The same immunoblot membrane was blotted for the cytosolic marker β-tubulin. (D) Representative Coomassie blue-stained gel of the His₆-tagged C-terminal deletion mutants of PMT used in this study. (E) Intracellular signaling activity of C-terminal deletion mutants of PMT. Shown are dose-response curves for the intracellular activity of HEK-293T cells transfected with SRE-luciferase gene reporter plasmids and then treated with the indicated proteins at the indicated doses and subjected to an SRE-luciferase assay, as described in Materials and Methods. Relative PMT activation values are in comparison to untreated cells, normalized to the maximum activation of PMT, as indicated and as determined by the 4PL equation for each experiment. Results shown are from 3 independent experiments, where experiments for each data point were performed in triplicate.

We speculated that C3 may have been delivered but at levels that were not detectable by immunoblot analysis. To test this possibility, we explored the more sensitive SRE-luciferase assay for detecting whether C3 was delivered into the cytosol by assaying downstream intracellular activity. For this, we generated recombinant PMT-C3 without an HA tag (Fig. 7A and D) and tested for SRE-luciferase reporter activity. Intracellular signaling activity was detected only for full-length recombinant PMT but not for PMT-C3 (Fig. 7E). We also considered the possibility that recombinant proteins without HA tags, where only the C1 or C2 subdomain was deleted (Fig. 7A and D), might restore SRE-luciferase reporter activity; however, none of the resulting internal deletion mutants displayed detectable intracellular activity above the levels of the controls (Fig. 7E). In total, these results demonstrate that native cargo is delivered by PMT-N as an intact unit comprised of C1-C2-C3 and support a model whereby C1-C2 is important for the delivery of C3 by PMT-N.

DISCUSSION

PMT belongs to a growing superfamily of single-chain bacterial A-B toxins that utilize discrete modular domains to carry out specific functions of the intoxication process. A number of advancements have been made in elucidating the modular nature and functional organization of the PMT-related dermonecrotic toxin (DNT) family, which includes the cytotoxic necrotizing factors (CNFs) from Escherichia coli and Yersinia pseudotuberculosis and the DNT from Bordetella sp. (reviewed in references 5, 6, and 22). The available crystal structures of the C-terminal domains of PMT (13) and CNF1 (23) have helped define the subdomains that comprise the cargo modules responsible for intracellular toxic activity. In both instances, the far-C-terminal subdomain, corresponding to C3 in PMT, harbors the Gα protein Gln deamidase activity (13, 16, 23–25). However, no corresponding structure of the N terminus of this family of toxins is available, and so the mechanism by which the N-terminal module mediates the delivery of the cognate activity cargo into the cytosol of host cells is not fully understood.

To gain further insight into the cargo delivery process, we first sought to demonstrate the cellular uptake and localization of the cargo subdomains of PMT. We clearly demonstrated that the three C-terminal subdomains are delivered by PMT-N into the cytosol of host cells as an intact C1-C2-C3 unit, and no individual subdomains or further proteolytically cleaved products were detected in the cytosol. In agreement with previous studies demonstrating the dependence of PMT intracellular activity on the acidification of late endosomes (3), the cytosolic delivery of C1-C2-HA cargo from endocytic vesicles was prevented by treatment with ammonium chloride. Additionally, we showed that the observed cytosolic fragments possess the MLD at the N terminus of the delivered cargo. Moreover, we were able to conclude that the core C3 subdomain required for intracellular activity is comprised of residues 1105 to 1278, which suggests that the very-C-terminal residues in PMT may not be critical to toxin activity.

Since it appeared that truncations that affected or removed the C3 subdomain did not appear to hamper their respective delivery, we next asked the question of whether the nature of the cargo might play a role in the delivery process. To determine whether PMT-N alone could deliver nonnative cargo, we tracked the cellular uptake and localization of PMT-N-GFP by microscopy to endosomes and demonstrated dose- and time-dependent cytosolic delivery of GFP from late acidified endosomes in the absence or presence of cytochalasin D or ammonium chloride. These findings align with data from a previous report of cellular uptake into endosomes and subsequent cell death upon treatment with PMT or PMT-N fused with the catalytic domain of diphtheria toxin (DTa), PMT^C1165S-DTa or PMT-N-DTa, respectively (19). These results, in total, suggest that PMT-N delivers both native and nonnative cargos into the cytosol as an intact unit. Consequently, we asked whether C3 could be delivered by PMT-N without C1, C2, or C1-C2. However, we saw no evidence of C3 delivery by PMT-N for any of the C1, C2, or C1-C2 deletions. We thus conclude that in the case of the native toxin cargo, the presence of C1-C2 is important for the delivery of the catalytic subdomain C3.

Understanding the molecular organization and functional modularity of A-B toxins enables their manipulation for potential therapeutic applications as delivery vehicles for various therapeutic cargos. Work from our laboratory has begun to investigate the considerations involved in the assembly of functional modules and the limitations of swapping domains and cargos of bacterial toxins. Recent work from our laboratory indicates that the efficiency of the delivery vehicle for the transfer of cargo into the cytosol depends in part upon the intrinsic nature of the cargo itself, in such a way that perhaps the cargo plays an important role in concert with the delivery vehicle to ensure effective translocation (21). As evidenced by the importance of C1-C2 for the delivery of C3, our results allow us to draw a similar deduction, that the cargo partly contributes to its own delivery. This phenomenon may be widely applicable to other potential exogenous cargos of PMT and delivery vehicles of different A-B toxins, which ultimately should be a drug design consideration made by researchers when manipulating bacterial toxins for therapeutics purposes. The observation that PMT-C1-HA was not stably expressed but PMT-C1-C2-HA was expressed and effectively delivered the C1-C2-HA cargo and other C-terminal peptides (C3 and truncations thereof) suggests that the use of PMT-C1-C2 might be an alternative delivery vehicle for some cargos.

Based on the findings in this study and previous work from our laboratory and from other groups, we propose here a model for PMT intoxication and the cytosolic delivery of cargo (Fig. 8). Upon cell binding of the cell surface receptors, full-length PMT is taken up by endocytosis and is trafficked from early to late endosomes. Upon a decrease of the endosomal pH, instead of being trafficked to the lysosomal degradative pathway, the acidification of PMT-N triggers conformational changes and membrane interactions that facilitate the translocation of cargo. During this process, C1-C2 first translocates across the endocytic membrane and subsequently pulls C3 through into the cytosol. The translocated intact C1-C2-C3 is cleaved near the MLD of C1, and the MLD helps to localize intact C1-C2-C3 at the plasma membrane, bringing the catalytic subdomain C3 in proximity to its target Gα subunit.

FIG 8.

Proposed model for PMT intoxication and cytosolic cargo delivery by PMT-N.

MATERIALS AND METHODS

Plasmid constructs.

The genes for the PMT N-terminal fragment (residues 1 to 568) (denoted PMT-N), the GFP fusion protein of PMT-N (denoted PMT-N-GFP), and the C-terminal GFP fusion protein of full-length PMT (denoted PMT-GFP) were cloned into the pET21b expression vector (Novagen), and the corresponding recombinant proteins were expressed in E. coli BL21 cells (Novagen) under the control of an isopropyl-β-d-thiogalactopyranoside (IPTG)-inducible promoter. HA tags were introduced, and truncations were generated by several rounds of primer extension and then incorporated into pTHC-toxA (Invitrogen) or pGEP-1 expression vectors by fragment exchange. The genes for full-length PMT, the PMT truncation protein, C-terminally HA-tagged PMT, and the C-terminally HA-tagged PMT truncation protein were cloned into the pTrcHisC (Invitrogen) or pGEP-1 expression vector. Organelle-specific cellular markers, expressed as DsRed fusion proteins (nucleus, endoplasmic reticulum [ER], and mitochondria), were obtained from Clontech. The signal sequence for the early endosome marker early endosome antigen 1 (EEA1) was generated by reverse transcription-PCR (RT-PCR) of the corresponding signal sequences using total RNA obtained from human HEK-293T (ATCC CRL-11268) cells, using forward primer 5′-CTCGAGGAGAAATTGCTGTCTTAGAAGCAAC-3′ and reverse primer 5′-CTACTCTAGAGTTGTGATAACCCGTTATCCTTGC-3′, followed by fragment exchange into the pDsRed2-nucleus vector to yield pDsRed2-EEA1.

Expression, purification, and quantification of PMT, PMT-N-GFP, HA-tagged PMT variants, and HA-tagged MLD and ΔMLD variants.

All recombinant proteins were expressed in E. coli TOP10 or BL21 cells (Invitrogen) under the control of an IPTG-inducible promoter, essentially as described previously (3, 21). Briefly, the resulting recombinant proteins were purified from cellular extracts by Ni²⁺ chelation affinity chromatography using a Ni²⁺-nitrilotriacetic acid (NTA)-agarose column (Qiagen), followed by anion-exchange chromatography using a HiTrapQ column (Amersham). The His₆ tag was removed for recombinant PMT (rPMT) and PMT-N-GFP by using a thrombin cleavage capture kit (Novagen), and the protein was further purified by anion-exchange chromatography using a HiTrapQ column. The toxin-containing fractions were desalted with a PD-10 column (Amersham) using 1× phosphate-buffered saline (PBS) (10 mM phosphate, 137 mM NaCl, 2.7 mM KCl [pH 7.4]) containing 10% glycerol. Protein concentrations of the resulting purified proteins were determined by quantitative digital image analysis of SDS-PAGE gels stained with Pierce GelCode blue using NIH ImageJ software and using bovine serum albumin (BSA) as the protein standard. The toxin samples were flash-frozen and stored at −80°C until use. The biological activity of purified full-length proteins, PMT (no His) and PMT-HA, was confirmed as described previously (21, 26), with a 50% effective concentration (EC₅₀) of 0.6 nM.

Cell culture.

Swiss 3T3 or HEK-293T cells (ATCC CRL-11268) were cultured in Dulbecco's modified Eagle's medium (DMEM; Gibco-Invitrogen, Grand Island, NY, USA) supplemented with 0.37% sodium bicarbonate, 100 U/ml penicillin-streptomycin (ThermoFisher Scientific), and 10% fetal bovine serum (FBS; Atlanta Biologicals, Lawrenceville, GA, USA) or 10% bovine growth serum (BGS; HyClone). The cells were maintained in DMEM with 5% or 10% FBS or BGS, which was stepped down to 2% FBS or BGS at the time of transfection before experiments were performed. Swiss 3T3 cells (ATCC CCL-92) were cultured at 37°C with 5% CO₂ in DMEM (Gibco) supplemented with 10% heat-inactivated BGS (pH 7.4) and containing 100 U/ml penicillin G and 100 μg/ml streptomycin. For transient-transfection experiments, HEK-293T cells were plated onto 6-well plates at 70 to 80% confluence 24 h prior to transfection. Transfection was performed by using TransIT-LT1 reagent (Mirus), according to the manufacturer's protocol. In short, 250 μl of serum-free Opti-MEM was mixed with 7.5 μl TransIT-LT1 reagent by pipetting gently, 2.5 μg plasmid DNA was added to the mixture, and the mixture was incubated at room temperature for 15 to 30 min. During incubation, the medium from each well was replaced with fresh DMEM supplemented with 10% BGS, the TransIT-LT1 reagent–DNA complex was added to the cells, and incubation continued at 37°C with 5% CO₂. Medium was replaced after 24 h. Two days later, the transfection efficiency was determined by visualization using fluorescence microscopy (Olympus IX-70 inverted fluorescence microscope equipped with an Olympus DP70 digital camera or Zeiss Axiovert 200M inverted fluorescence microscope).

Endothelial HEK-293T and epithelial Swiss 3T3 cells were used to demonstrate that the delivery of PMT-N cytosolic cargo occurs irrespective of the cell type. Swiss 3T3 cells were primarily used for immunoblot analysis as they express less cross-reactive material to the anti-HA antibody. Swiss 3T3 cells were also used in the fluorescence microscopy experiments as they have a larger cytoplasmic area for ease of visualization. HEK-293T cells were used in the SRE-luciferase reporter assays as they are more easily transfected with plasmid DNA.

Fluorescence microscopy.

Swiss 3T3 cells, cultured as described above, were plated onto coverslips placed in a 12-well plate at a cell density of 30 to 50% and incubated in DMEM supplemented with 2% BGS overnight at 37°C with 5% CO₂. Cells were treated with PMT-N-GFP at a final concentration of 685 nM for the indicated times. The cells were washed three times with 1× PBS and fixed with 3.7% formaldehyde in DMEM supplemented with 10% BGS for 15 min at room temperature, followed by washing three times with 1× PBS before mounting onto slides and visualization using fluorescence microscopy (as described above). All the experiments shown were repeated independently at least three times.

Toxin treatment of cells.

Swiss 3T3 or HEK-293T cells, cultured as described above, were plated onto 6-well plates at a density of 0.3 × 10⁶ cells/well or onto 100-mm culture dishes at a density of 4 × 10⁶ cells per plate and incubated at 37°C with 5% CO₂ for 1 to 2 days prior to toxin treatment. For experiments involving PMT-N-GFP, cells were cooled on ice (4°C); medium was replaced with fresh cold DMEM supplemented with 10% BGS containing the protein, as indicated; and the mixture was incubated on ice for 30 min. The toxin-treated cells were then incubated at 37°C for the indicated times.

For experiments involving treatment with cytochalasin D (200 μM stock solution dissolved in dimethyl sulfoxide [DMSO]) (catalog number C8273; Sigma) or NH₄Cl (1 M stock solution dissolved in H₂O; J. T. Baker), the cells were first exposed to the reagents at the indicated concentrations for the indicated times prior to toxin treatment. In short, medium was replaced with fresh warm (37°C) DMEM supplemented with 10% BGS, followed by the addition of the reagent. The cells were incubated at 37°C for 1 h (cytochalasin D) or for 30 min or 60 min (NH₄Cl). The plate was then removed from the incubator and placed on ice. The toxin was added to the medium, and the cells were incubated on ice for 30 min before being returned to the incubator (37°C) for 4 h.

For experiments involving C-terminally HA-tagged PMT variants (full length or truncations), 1 to 2 days after plating, medium was replaced with fresh DMEM supplemented with 10% BGS containing toxin at the indicated concentrations and treatment times, and the mixture was incubated at 37°C. After toxin treatment, cells were washed 3 times with warm 1× PBS (2 min each), replaced with fresh DMEM supplemented with 10% BGS without toxin, and incubated at 37°C for 60 min. This step involving three washes plus toxin-free medium replacement was subsequently repeated two more times. Cells were washed three more times with warm 1× PBS (2 min each) and then washed twice with cold 1× PBS before the cells were harvested.

For all types of treatments, the cells were collected and analyzed by using a subcellular fractionation method (see below). All the experiments shown were repeated independently at least three times.

Data analysis.

Protein concentration levels were determined by quantitative digital image analysis of immunoblots or Coomassie-stained SDS-PAGE gels, as indicated, using NIH ImageJ software. All data are presented as the means ± standard deviations (SD) from at least three independent experiments or as indicated otherwise. P values were determined by using unpaired two-tailed Student's t test. Statistical significance was set at the 95% confidence level.

Subcellular fractionation.

For subcellular fractionation using high-speed ultracentrifugation, Swiss 3T3 or HEK-293T cells were cultured, plated, and treated as described above. The cells were collected in cold 1× PBS and then centrifuged for 5 min at 4°C at 3,000 × g to pellet the cells. Cell pellets were washed with cold 1× PBS and then recentrifuged for 5 min at 3,000 × g at 4°C. The cell pellets were resuspended in 0.5 ml homogenization buffer (250 mM sucrose and 0.5 mM EDTA containing 1× protease inhibitor cocktail [PI]) and homogenized at 4°C by 30 passages through a 26 3/8-gauge needle fitted using a 1-ml syringe. The homogenate was fractionated by centrifugation for 10 min at 800 × g at 4°C to separate the nuclear pellet, comprised of nuclei and cell debris, and the postnuclear supernatant (intracellular fraction), comprised of organelles and cytosolic components. For further subcellular fractionation, the intracellular fraction was centrifuged at 100,000 × g for 1 h at 4°C using a TH-660 rotor to enrich for the microsomal pellet, comprised of vesicles, which were solubilized in homogenization buffer (vesicle fraction). The cytosolic proteins remained in the supernatant (cytosolic fraction).

For subcellular fractionation using benchtop centrifuges, Swiss 3T3 or HEK-293T cells were cultured, plated, and treated as described above. The washed cell pellets were resuspended in 250 μl of homogenization buffer (250 mM sucrose and 0.5 mM EDTA containing 1× PI) and homogenized at 4°C via 10 plunges in a 2-ml Dounce homogenizer. The homogenate was centrifuged for 5 min at 750 × g at 4°C to separate the nuclear pellet, comprised of nuclei and cell debris, and the postnuclear supernatant (intracellular fraction), comprised of organelles and cytosolic components. The intracellular fraction was further centrifuged at 30,000 × g for 1 h at 4°C by using a Hettich Mikro 22R centrifuge to enrich for the microsomal pellet containing vesicles. The vesicle pellet was solubilized in homogenization buffer (vesicle fraction). The resulting supernatant was further centrifuged at 30,000 × g for 1 h at 4°C by using a Hettich Mikro 22R centrifuge to remove any residual debris. The supernatant contained cytosolic proteins (cytosolic fraction). All the experiments shown were repeated independently at least three times.

SDS-PAGE and immunoblot analysis.

The protein samples described above were separated by 10% or 12% SDS-PAGE, transferred onto polyvinyl difluoride or nitrocellulose membranes (Pall Life Sciences), and analyzed by immunoblotting using standard procedures, with the following primary antibodies against the indicated proteins: mouse monoclonal anti-GFP (catalog number sc9996; Santa Cruz Biotechnology), rabbit polyclonal anti-β-tubulin (catalog number ab6046; Abcam), mouse monoclonal anti-Rab5 (D11) (catalog number sc46692; Santa Cruz Biotechnology), or mouse monoclonal anti-HA (catalog number 26183; ThermoFisher Scientific). Specific protein bands were visualized by using the ECL Plus system (GE Life Sciences), using the appropriate horseradish peroxidase (HRP)-conjugated secondary antibodies, goat anti-mouse or goat anti-rabbit. Blots were scanned by using an HP Scanjet G4050 scanner, and images were quantified by using NIH ImageJ.

SRE-luciferase reporter assay and data analysis.

SRE-luciferase reporter assays were performed and analyzed as described previously (21). Briefly, HEK-293T cells plated onto 24-well plates at a density of 7.5 × 10⁴ cells/well were transfected with reporter plasmids. Seven hours later, cells were treated with the indicated toxins at the indicated doses and incubated for 18 h prior to assays. To calculate fold activation, firefly reporter relative luminescence units (RLU) were divided by the control Renilla RLU and then normalized to the corresponding value for untreated samples. Relative PMT activation was determined compared to the maximum activation observed for PMT (no His) or for PMT, as indicated. Data points shown are the means plus or minus SD. To calculate the dose-response curves, the relative PMT activation for each data point was compiled and analyzed with the Solver function in Microsoft Excel to generate a best-fit, four-parameter logistic (4PL) equation, y = F(x) = (A − D)/[1 + (x/C)^B] + D, where A is the minimum asymptote, B is the slope, C is the point of inflection or the EC₅₀, and D is the maximum asymptote. The best-fit curve was optimized for the least sum of the squared difference between observed and expected 4PL values.