Topological analysis of the type 3 secretion system translocon pore protein IpaC following its native delivery to the plasma membrane during infection

Many Gram-negative bacterial pathogens require a type 3 secretion system (T3SS) to deliver effector proteins into eukaryotic cells. Contact of the tip complex of the T3SS with a target eukaryotic cell initiates the secretion of the two bacterial proteins that assemble into the translocon pore in the plasma membrane. The translocon pore functions to regulate effector protein secretion and is the conduit for effector protein translocation across the plasma membrane. To generate insights into how the translocon pore regulates effector protein secretion, we defined the topology of the Shigella translocon pore protein IpaC in the plasma membrane following its native delivery by the T3SS. Using single-cysteine substitution mutagenesis and site-directed labeling with a membrane-impermeant chemical probe, we mapped residues accessible from the extracellular surface of the cell. Our data support a model in which the N-terminus of IpaC is extracellular and the C-terminus of IpaC is intracellular. These findings resolve previously conflicting data on IpaC topology that were based on non-native delivery of IpaC to membranes. Salmonella enterica serovar Typhimurium also requires the T3SS for effector protein delivery into eukaryotic cells. Although the sequence of IpaC is closely related to the Salmonella translocon pore protein SipC, the two proteins have unique functional attributes during infection. We showed a similar overall topology for SipC and IpaC and identified subtle topological differences between their transmembrane α-helixes and C-terminal regions. Together, our data suggest that topological differences among distinct translocon pore proteins may dictate organism-specific functional differences of the T3SSs during infection. Importance The type 3 secretion system (T3SS) is a nanomachine required for virulence of many bacterial pathogens that infect humans. The system delivers bacterial virulence proteins into the cytosol of human cells, where the virulence proteins promote bacterial infection. The T3SS forms a translocon pore in the membrane of target cells. This pore is the portal through which bacterial virulence proteins are delivered by the T3SS into the eukaryotic cytosol. The pore also regulates the secretion of these virulence proteins. Our work defines the topology of translocon pore proteins in their native context during infection, resolves previously conflicting reports about the topology of the Shigella translocon pore protein IpaC, and provides new insights into how interactions of the pore with the T3SS likely produce signals that activate secretion of virulence proteins.


Importance 23
The type 3 secretion system (T3SS) is a nanomachine required for virulence of many 24 bacterial pathogens that infect humans. The system delivers bacterial virulence proteins 25 into the cytosol of human cells, where the virulence proteins promote bacterial infection. 26 The T3SS forms a translocon pore in the membrane of target cells. This pore is the 27 portal through which bacterial virulence proteins are delivered by the T3SS into the 28 eukaryotic cytosol. The pore also regulates the secretion of these virulence proteins. 29 Our work defines the topology of translocon pore proteins in their native context during 30 infection, resolves previously conflicting reports about the topology of the Shigella 31 translocon pore protein IpaC, and provides new insights into how interactions of the 32 pore with the T3SS likely produce signals that activate secretion of virulence proteins. 33

Introduction 34
The type 3 secretion system is a specialized nanomachine required for the virulence of 35 more than 30 bacterial pathogens (1). The T3SS translocates bacterial virulence 36 proteins, known as effector proteins, from the bacterial cytosol into the cytosol of 37 eukaryotic cells. The T3SS is composed of a base that spans the two bacterial 38 membranes (2), a needle that is anchored in the base and extends away from the 39 bacterial surface (2,3), and a tip complex that prevents non-specific secretion (4,5). 40 Upon contact of the tip complex with a eukaryotic cell, the T3SS secretes two bacterial 41 proteins that embed in the plasma membrane (6), where they assemble into a hetero-42 oligomeric pore, known as the translocon pore (3). The translocon pore is essential for 43 T3SS activity; it functions as a conduit through which bacterial virulence proteins 44 ("effectors") traverse the plasma membrane to gain access to the eukaryotic cytosol (3), 45 and it participates in defining the timing of the secretion of these effectors by the T3SS 46 (7). 47 are predicted to interact with the needle, but it is uncertain how this interaction occurs, 57 as the topology of IpaC in the plasma membrane is controversial. The two previous 58 studies that investigated IpaC topology used purified recombinant IpaC. These studies 59 showed IpaC inserting into the membrane with its N-terminus on the extracellular 60 surface of the plasma membrane, but came to opposing conclusions about the location 61 of the C-terminus; interactions of IpaC with artificial liposomes concluded that IpaC 62 contained a single transmembrane-helix with the C-terminus present in the liposome 63 lumen (9), whereas investigation of purified IpaC incorporating into macrophage 64 membranes concluded that IpaC contained two transmembrane -helixes with the C-65 terminus accessible on the extracellular surface of the macrophage (10). 66 Here we defined the topology of the Shigella translocon protein IpaC during bacterial 67 infection following its native delivery into the plasma membrane by the T3SS. Using 68 single cysteine accessibility mutagenesis, we defined a topological map of IpaC 69 showing that the amino terminal region is extracellular, that the carboxy terminal region 70 is in the cytosol, and that a single transmembrane-helix is present. Furthermore, to 71 test whether this topology is conserved among IpaC homologs in other pathogens that 72 require T3SS for virulence, the accessibility of analogous residues for the Salmonella 73 pore protein SipC was tested. We found that the overall topologies of IpaC and SipC are 74 similar. However, we observed subtle differences between the two proteins in the 75 accessibility of the transmembrane-helixes and the C-terminal regions that may 76 contribute to organism-specific functional differences of these T3SSs during infection.

78
transcription-based reporter of T3SS activity (TSAR, (7,16)); following the secretion of 123 the translocon pore proteins and the bacterial effector protein OspD through the T3SS, 124 the TSAR reporter produces GFP in the bacterium. The promoter for GFP is regulated 125 by MxiE, which is activated upon secretion of OspD (16). S. flexneri producing cysteine 126 substituted IpaC derivatives activated the T3SS to secrete effector proteins at 127 efficiencies similar to WT IpaC (Figs. 2A, F, and G, and S1A). Both the absolute number 128 per cell of bacteria with active secretion ( Fig. 2A and F)  flexneri producing WT IpaC (Fig. S1B). In sum, substitution of selected IpaC residues 142 with cysteine did not significantly alter either IpaC activity or T3SS function during S.

Accessibility of IpaC residues in membrane-embedded S. flexneri translocon 145
pores. To measure the accessibility of IpaC residues in translocon pores embedded in 146 the plasma membrane of host cells, site-directed labeling by methoxypolyethelene 147 glycol maleimide (PEG5000-maleimide) was performed with the library of IpaC cysteine 148 substitution derivatives characterized above. Because PEG5000-maleimide is 149 membrane impermeant (12) and is too big to pass entirely through the translocon pore 150 (3,7), this approach specifically assesses the accessibility of the cysteine residue to the 151 extracellular surface of the eukaryotic cell. Cysteine substitutions within the N-terminal 152 domain (residues 1-99) of IpaC labeled efficiently with PEG5000-maleimide, as 153 demonstrated by a distinct shift to a slower migrating position on SDS-PAGE (Fig. 3). 154 Efficient labeling was also observed for one of the four residues tested within the 155 putative transmembrane-helix (residues 100-120) (Fig. 3); labeling at IpaC A106C 156 was efficient, whereas labeling of the three other residues in the putative 157 transmembrane-helix that were tested (S101C, S108C, and A119C) was essentially 158 undetectable. The labeling of IpaC cysteine substitutions C-terminal to the putative 159 transmembrane-helix was weak, except among substitutions of the 15 residues 160 closest to the C-terminus (residues 349-363) (Fig. 3). All cysteine substitutions of IpaC, 161 including the substitutions in the putative transmembrane-helix and the substitutions 162 C-terminal to it, were readily labeled by PEG5000-maleimide when the labeling was 163 performed on IpaC in solution following secretion from bacteria in vitro ( Fig. S2A-B), 164 indicating that the lack of labeling observed during infection was a function of 165 inaccessibility due to the membrane-embedded state rather than to fundamental 166 inaccessibility of the residue in the protein per se. When labeling was performed following permeabilization of the plasma membrane, accessibility of IpaC residues near 168 the C-terminus was similar to the accessibility of the residues in the N-terminal domain 169 ( Fig. S2C-D), which demonstrates that under non-permeabilizing conditions, the plasma 170 membrane inhibited access of PEG5000-maleimide to the residues near the C-171 terminus. To test whether docking of the needle on the translocon pore inhibited 172 labeling of residues near the C-terminus, we compared the accessibility of IpaC A358C, 173 which is close to the C-terminus (R362), with that of S17C, a residue in the N-terminal To explore whether the observed functional differences between IpaC and SipC were 203 associated with differences in protein topology in the plasma membrane, we compared 204 the accessibility of SipC residues with that of IpaC residues. Like IpaC, native SipC 205 lacks cysteines. The putative secondary structure of SipC is very similar to that of IpaC, 206 with a putative N-terminal extracellular domain (residues 1-119), a putative 207 transmembrane -helix (residues 120-140), and C-terminal to the transmembrane-208 helix, a putative coiled-coil domain (residues 293-320) ( Fig 4A). The second putative 209 transmembrane span identified in IpaC was not clearly delineated by in silico analyses 210 of SipC. We substituted cysteines at residues of SipC that were analogous to those 211 chosen for substitution in IpaC ( Fig. 4B-C). During S. Typhimurium infection of HeLa 212 cells, the accessibility of these residues from the extracellular surface of the plasma membrane was tested using the same method used for IpaC, site-directed labeling with 214

PEG5000-maleimide. As observed for IpaC, substitutions in the N-terminal domain of 215
SipC labeled efficiently ( Fig. 4D-E). In contrast to our findings for IpaC, labeling was not 216 observed for any other SipC substitution tested ( Fig. 4D-E); this included SipC A126C, a 217 cysteine substitution at a residue within the SipC putative transmembrane-helix that is 218 analogous to IpaC A106C, which labeled efficiently with PEG5000-maleimide, and 219 cysteine substitutions at residues within the C-terminal 15 amino acids of SipC, which in 220 IpaC labeled with intermediate efficiency. domain is extracellular, a single transmembrane-helix is present, and the C-terminus 230 is intracellular (Fig. 5). Despite the similarities in overall topology, the differences 231 observed in the accessibility of substitutions close to the C-termini of the two proteins 232 suggest that subtle topological or structural differences exist in the transmembrane-233 helix and C-terminal regions. The differences we observe between the two proteins in 234 the context of plasma membrane-embedded translocon pores may be relevant to

Discussion 237
The T3SS is essential for the pathogenesis of many important human bacterial 238 pathogens, and the translocon pore is essential for T3SS function. To generate insights 239 into how the translocon pore participates in the signaling that activates secretion of 240 effector proteins, we defined the topology of IpaC in the membrane of human-derived 241 cells. To our knowledge, this represents the first detailed topological analysis of a type 3 242 translocon pore protein following its native delivery to the plasma membrane. We show 243 that the N-terminal region of IpaC is extracellular and the C-terminal region of IpaC is in 244 the host cytosol. Moreover, we observed a similar overall topology for SipC, the IpaC 245 homolog in Salmonella. The observations that the N-terminal region of IpaC is located 246 on the extracellular side of the plasma membrane is consistent with previous 247 observations investigating the interactions of recombinant IpaC with the membrane 248 (9,10). 249 The labeling of A106C in the transmembrane span indicates that the outer portion of the 250 pore channel is at least 4.4 nm wide, the approximate size of PEG5000-maleimide. 251 Moreover, the lack of labeling for other residues in the transmembrane-helix of IpaC 252 suggest that towards the cytosolic side of the plasma membrane, either the lumen of the 253 pore becomes narrower, such that the PEG5000-maleimide is precluded from 254 accessing the IpaC residues, and/or that other IpaC residues do not line the pore 255 interior. If the latter were true, it could indicate that IpaB residues line the bulk of the 256 channel.
secondary structure (residues 121-308) and a predicted coiled-coil domain (residues 259 309-344). The minimal accessibility to PEG5000-maleimide of IpaC cysteine 260 substitutions within these sequences strongly suggests that these sequences, including 261 the coiled-coil domain, lie within the cytosol and that residues 100-120 constitute a of IpaC re-enter the pore interior (Fig 5A). 289 The overall topology of SipC and IpaC was similar, with the N-terminus of SipC 290 extracellular and the C-terminus intracellular (Fig. 5B). Both IpaC and SipC support 291 The similarity between the overall topology of IpaC and that of SipC suggests that the 303 mechanism(s) required to deliver the two proteins into the membrane, multimerize into 304 the pore, interact with the lipid membrane, support effector translocation, and generate 305 signals to activate secretion may be similar. However, in contrast to the accessibility of 306 some cysteine substitutions in the C-terminal region of IpaC, substitutions near the C-307 terminus of SipC were completely inaccessible. This indicated that multiple topologies of 308 SipC do not occur at an efficiency that was detectable in this assay and, by extension, investigations into translocon pore structure will inform the molecular mechanisms that

Materials and Methods 325
Bacterial culture. Strains used in this study are described in Table 1 were mixed by pipetting and then centrifuged again at 2,000 g for 10 minutes at 25C. suspension was centrifuged at 21,000 g for 30 minutes at 4C. The supernatant, which contains the cytosol fraction, was decanted into a fresh tube. The pellet was 415 resuspended in 50 mM Tris, pH 7.4, containing protease inhibitors and 0.5% Triton X-416 100, incubated on ice for 30 minutes, and centrifuged at 21,000 g for 15 minutes at 4C. 417 The supernatant from this spin contained the membrane fraction, and the pellet 418 consisted of the detergent insoluble fraction, which included intact bacteria. The 419 efficiency of PEG5000-maleimide labeling was monitored by assessing the gel shift of showing the putative secondary structure of IpaC (7,9). Putative transmembrane span, 560 residues 100-120 (solid zig-zag), predicted by in silico analyses and experimental data 561 Data are the mean ± SEM; dots represent independent experimental replicates. **, 598 p<0.01; ***, p<0.001 (Student's t-test). Stable docking of bacteria and secretion via the 599 T3SS for each strain producing an IpaC cysteine substitution derivative was not 600 statistically different from that of the strain producing WT IpaC (by ANOVA).