The N Terminus of Type III Secretion Needle Protein YscF from Yersinia pestis Functions To Modulate Innate Immune Responses

Abstract

The type III secretion system is employed by many pathogens, including the genera Yersinia, Shigella, Pseudomonas, and Salmonella, to deliver effector proteins into eukaryotic cells. The injectisome needle is formed by the polymerization of a single protein, e.g., YscF (Yersinia pestis), PscF (Pseudomonas aeruginosa), PrgI (Salmonella enterica SPI-1), SsaG (Salmonella enterica SPI-2), or MxiH (Shigella flexneri). In this study, we demonstrated that the N termini of some needle proteins, particularly the N terminus of YscF from Yersinia pestis, influences host immune responses. The N termini of several needle proteins were truncated and tested for the ability to induce inflammatory responses in a human monocytic cell line (THP-1 cells). Truncated needle proteins induced proinflammatory cytokines to different magnitudes than the corresponding wild-type proteins, except SsaG. Notably, N-terminally truncated YscF induced significantly higher activation of NF-κB and/or AP-1 and higher induction of proinflammatory cytokines, suggesting that a function of the N terminus of YscF is interference with host sensing of YscF, consistent with Y. pestis pathogenesis. To directly test the ability of the N terminus of YscF to suppress cytokine induction, a YscF-SsaG chimera with 15 N-terminal amino acids from YscF added to SsaG was constructed. The chimeric YscF-SsaG induced lower levels of cytokines than wild-type SsaG. However, the addition of 15 random amino acids to SsaG had no effect on NF-κB/AP-1 activation. These results suggest that the N terminus of YscF can function to decrease cytokine induction, perhaps contributing to a favorable immune environment leading to survival of Y. pestis within the eukaryotic host.

INTRODUCTION

The type III secretion (T3S) system is a protein secretion nanomachine used by many Gram-negative bacteria to deliver virulence factors into eukaryotic host cells (1–3). T3S systems are found in both animal and plant pathogens (1–3). Among these bacteria are the human pathogens Salmonella spp., Pseudomonas spp., Shigella spp., and Yesinia spp. The structural and functional properties of the T3S system are well conserved among these bacteria. However, the properties of the effectors and the resulting effect on the host organism are species specific, ranging from induction of apoptosis to suppression of host defense mechanisms (1, 3). The T3S system in all cases contributes to infection by allowing the bacteria to communicate with the eukaryotic cell and is essential to the survival and pathogenesis of many Gram-negative bacteria (1–3).

A key component of the T3S system is the needle-like structure that spans both bacterial membranes and through which effectors are delivered directly into the cytosol of host cells (1, 3, 4). The needle is formed by the polymerization of single proteins: YscF in Yersinia spp., PrgI and SsaG (Salmonella pathogenicity island 1 (SPI-1) and Salmonella pathogenicity island 2 (SPI-2), respectively) in Salmonella enterica serovar Typhimurium, PscF in Pseudomonas spp., and MxiH in Shigella flexneri (3–8). X-ray crystallography and nuclear magnetic resonance (NMR) have been utilized to analyze structures of several needle proteins: MxiH from Shigella (9), BsaL from Burkholderia pseudomallei (10), and PrgI from Salmonella Typhimurium (11). The crystal structure of MxiH was used to generate a model of T3S needle structure (10, 12, 13). In this model, the N terminus of MxiH was predicted to line to the lumen of the T3S needle (13). Contrary to the previous model, recent reports by Loquet et al. and Demers et al. have revealed that the variable N termini of needle proteins in Salmonella and Shigella, respectively, are in fact on the outside surfaces of the needles, exposing them to host elements, while the conserved carboxy ends face the lumen, reflecting a bacterial strategy to evade host response (14, 15).

The needle proteins share high sequence identity, and their predicted structures have two α-helical segments separated by a loop at a P-(S/D)-(D/N)-P motif (4, 16). The C termini of needle proteins are highly conserved between different bacterial species. As expected, the conserved residues of the C termini mediate important intra- and intermolecular interactions of the complex. Deletion of 5 residues from the C termini of PrgI, MxiH, and BsaL prevents needle polymerization (16, 17). The N termini of needle proteins in all these bacteria are suggested to be highly mobile and disordered (9), with no defined structure found for this portion of the proteins. YscF has only been crystallized in complex with its chaperones YscE and YscG (7, 16). Sun et al. reported the N terminus in this crystal structure to be largely unorganized and not representative of YscF in its needle conformation (7). The N termini of needle proteins are variable not only in amino acid composition but also in the number of amino acids.

Torruellas et al. (18) has reported that YscF is a multifunctional structure that is involved in virulence protein secretion, translocation of virulence proteins, and cell contact- and calcium-dependent regulation of T3S. However, mutations in the N terminus of the needle protein have no significant impact on these essential processes. A study by Allaoui et al. (19) reports that truncation of 12 N-terminal amino acids from YscF reduced secretion of YopB and YopD but not the other Yops (19). Additionally, other studies have also reported that the hypervariable N termini do not mediate important needle subunit interactions (16, 17).

Needle proteins were recently identified as Toll-like receptor 2 (TLR2) and TLR4 ligands that activate expression and secretion of cytokines via an MyD88-dependent pathway (20). Purified needle proteins from different bacteria induced cytokine expression to different magnitudes (20). YscF comes from a T3S system from pathogens that have an anti-inflammatory infection objective (1, 21) and induces lower levels of proinflammatory cytokines. MxiH, from a bacterium with a proinflammatory infection objective (3, 22, 23), induces high levels of proinflammatory cytokines, as does SsaG. SsaG is expressed when Salmonella is intracellular (24–26), and therefore it is not exposed to surface-expressed TLRs. Interestingly, SsaG has fewer amino acids than other needle proteins, notably lacking residues that correspond to the N termini of other needle proteins. The reported variation in the magnitude of cell activation by various needle proteins (20) coupled with the dispensability of the N terminus for needle assembly led to the development of a hypothesis that the N termini of needle proteins could be involved in modulating TLR interactions.

In this study, the role of the N termini of needle proteins in modulating host responses was investigated. Further, the region of the N terminus that contributes to the host inflammatory activity was examined in YscF. A role for the N terminus of YscF in modulation of host immune responses was demonstrated by constructing truncated, deleted, and chimeric recombinant and purified needles.

MATERIALS AND METHODS

Bacterial strains, culture conditions and plasmids.

Bacterial strains and plasmids used in this study are listed in Table 1. All strains were stored at −80°C in 25% (vol/vol) glycerol. Escherichia coli strains were grown at 37°C in LB broth (BD Difco, Sparks, MD) or on tryptose blood agar (TBA; BD Difco) plates with antibiotics added as needed. Kanamycin and carbenicillin were used at 50 μg/ml.

TABLE 1.

Bacterial strains and plasmids used in this study

Strain or plasmid	Characteristics	Source or reference
Strains
E. coli
BL21(DE3) Star	F− ompT hsdSB (rB− mB−) gal dcm (DE3)	Invitrogen
Top10	F− mcrA Δ(mrr-hsdRMS-mcrBC) Φ80dlacZΔM15 ΔlacX74 nupG recA1 araD139 Δ(ara-leu)7697 galE15 galK16 rpsL (Strr) endA1 Δ−	Invitrogen
Novablue	recA1 endA1 hsdR17 (rK− mK−) supE44 thi-1 gyrA96 relA1 lac (F′ proA+B+) lacIqZΔM15::Tn10	Novagen
Yersinia pestis
KIM8-3002	pCD1, pPCP1−, pMT1	18
KIM8-3002.P61 (ΔyscF)	pCD1 (ΔyscF), pPCP1−, pMT1	18
Salmonella enterica
14028	Salmonella enterica serovar Typhimurium	ATCC
BLS101	14028 ΔssaG	Brian Ahmer
Plasmids
pET200	T7 expression vector, N-terminal peptide containing the X-press epitope and the 6× His tag; Kmr	Invitrogen
pBAD18-Kan	araBADp cloning vector; Kmr	38
pDO1	pET200 with YscF Δ(S2-S5) ΔN5	This study
pDO2	pET200 with YscF Δ(S2-G10) ΔN10	This study
pDO3	pET200 with YscF Δ(S2-D15) ΔN15	This study
pDO4	pET200 with YscF Δ(S2-A20) ΔN20	This study
pDO5	pDO3 L67A	This study
pDO6	pDO3 D17A	This study
pDO7	pDO3 V19A	This study
pDO8	pDO3 L16A/D17A	This study
pDO9	pDO3 L16A/V19A	This study
pDO10	pDO3 D17A/V19A	This study
pDO11	pDO3 L16A/D17A/V19A	This study
pDO12	pET200 with YscF-SsaG	This study
pDO13	pET200 with RN-SsaG	This study
pDO14	pBAD18-Kan with YscFΔ(S2-D15)ΔN15	This study
pDO15	pBAD18-Kan with YscFΔ(S2-A20)ΔN20	This study
pDO16	pBAD18-Kan with YscF-SsaG	This study
pDO17	pBAD18-Kan with RN-SsaG	This study
pDO18	pBAD18-Kan with SsaG	This study
pET15B MxiH	pET15b-MxiH	W. Picking
pET200 PrgI	pET200-PrgI Kmr	20
pET200 SsaG	pET200-SsaG Kmr	20
pJM119	pET24b-YscF	53

Plasmids used in this study to overexpress needle proteins were constructed in pET200 TOPO using Champion TOPO expression kits (Invitrogen, Carlsbad, CA). Primers for gene amplification were synthesized by Eurofins Genomics (Huntsville, AL). PCR primers were designed to clone fragments of yscF missing regions for the N terminus of YscF into an expression vector, pET200 (Invitrogen); the primers used are listed in Table 2. SsaG was also N-terminally truncated to correspond to the 66-bp truncation of YscF using the forward primer 5′-CAC CCT CTC CCA CAT GGC GCA C-3′. The plasmids encoding MxiH, truncated MxiH, and truncated PrgI plasmids were a kind gift from Wendy Picking, Oklahoma State University. PCR fragments were cloned into pET200 TOPO (Invitrogen). Template DNA for amplification was generated using the DNeasy kit (Qiagen, Valencia, CA); the manufacturer's instructions were followed. PCR was performed using PFU Ultra II HS DNA polymerase (Agilent Technologies, Santa Clara, CA).

TABLE 2.

Oligonucleotides used in this study

Primer	Sequence
YscF mutants, pET200
S2-S5	CACCGGATTTACGAAAGGAACCGATATCGCAG
S2-G10	CACCACCGATATCGCAGACTTAGATGC
S2-D15	CACCTTAGATGCGGTGGCTCAAAC
S2-A20	CACCCAAACGCTCAAGAAGCCAG
YscF R	TTATGGGAACTTCTGTAGGATGCCTTGC
SsaG chimera, pET200
YscF F	CACCATGAGTAACTTCTCTGGATTTACGAAAGGAACC
YscF R	TAATTGTGCAATATCTAAGTCTGCGATATCGGTTCCTTTCG
SsaG F	GGAACCGATATCGCAGATATTGCACAATTAGTGGATATGCTC
ssaG R	TCAGATTTTAGCAATGATTCCACTAGCATATCCTTG
RN-SsaG F	TCACCATGAGGTACCTTCTCTGGATTTACG
RN-SsaG R	GTGATTAACACGTTAAGATTCAGACGC
YscF mutants, pBAD
Δ S2-D15	GGAATTCAGGAGGAAACGATGTTAGATGCGGTGGCAAACG
Δ S2-A20	GGAATTCAGGAGGAAACGATGCAAACGCTCAAGAAGCCAG
YscF R	CGCGGATCCTTATGGGAACTTCTGTAGGATGCCTTG
SsaG chimera, pBAD
SsaG F	CGGAATTCATGATTGCACAATTAGTGGATATGCTCTCC
SsaG R	CGCGAGCTCTCAGATTTTAGCAATGATTCCACTAAGC
YscF-SsaG F	CGGAATTCATGAGTAACTTCTCTGGATTTACGAAAGGAAC
YscF-SsaG R	CGCGAGCTCTCAGATTTTAGCAATGATTCCACTAAGC
RN-SsaG F	CGGAATTCATGAGGTACCTTCTCTGGATTTACCGAAAGG
RN-SsaG R	CGCGAGCTCTCAGATTTTAGCAATGATTCCACTAAGC

In-frame deletions in yscF that removed DNA sequences encoding residues S2 to S5, S2 to G10, S2 to D15, and S2 to A20 of YscF were constructed by PCR. Amplified DNA fragments were cloned into pET200, generating plasmids pDO1 to pDO4, respectively. Site-directed mutagenesis of yscF in plasmid pDO3 was performed using the QuikChange site-directed mutagenesis kit (Agilent Technologies) according the manufacturer's instructions. Complementary oligonucleotides were designed to change one, two, or three yscF codons, generating plasmids pDO5 to pDO11. Plasmid pDO12 encoding the YscF-SsaG chimera was constructed by amplifying the sequence encoding the 15 N-terminal residues of YscF from pCD1 and the entire SsaG sequence from SPI-2 of Salmonella enterica serovar Typhimurium by PCR, using primers listed in Table 2. The 2 PCR products were hybridized to each other by overlap extension PCR (27), purified, and cloned into pET200 for protein expression. Plasmid pDO13 containing the RN-SsaG chimera was constructed by making a +1 frameshift insertion in the region coding for N-terminal YscF of the YscF-SsaG chimera to change the amino acid sequence, and following the yscF sequence, the reading frame was corrected with a −1 deletion. Plasmids pDO16 to pDO18 were constructed by cloning EcoRI- and SacI-cleaved SsaG, YscF-SsaG, or RN-SsaG into pBAD-18 Kan (28). All of the gene constructs were verified by sequencing by Eurofins Genomics.

All plasmid constructs were transformed into commercially obtained E. coli TOP10 (Invitrogen) by chemical transformation. Plasmids for protein expression were purified from E. coli TOP10 with a Qiaprep Miniprep kit (Qiagen). Purified plasmid DNA was then transformed into the expression host, BL21(DE3) Star (Invitrogen).

Purification of His-tagged proteins.

Protein purification was performed as reported previously (20). Briefly, E. coli BL21(DE3) Star (Invitrogen) carrying plasmids for a given protein were grown overnight in noninducing medium (MDG medium from Table 1 in reference 29) supplemented with antibiotic. Bacteria were then inoculated into autoinducing medium (ZYM-5052 medium from Table 1 in reference 29) with antibiotic and grown to an optical density at 620 nm (OD₆₂₀) of 0.6 to 0.8. Cells were harvested by centrifugation at 4,000 × g for 10 min at 4°C and resuspended on ice in wash buffer (50 mM NaH₂PO₄, 300 mM NaCl, 10% [wt/vol] glycerol). The resulting cellular suspension was then French pressed at 20,000 lb/in² twice to lyse cells. The lysate was clarified by centrifugation at 12,000 × g for 20 min. The clarified supernatant was collected and diluted with 1,000 ml of wash buffer before application to a preequilibrated TALON metal affinity resin (Clontech, Mountain View, CA) column. The lysates were applied to the columns twice before washing with new wash buffer. Bound protein was eluted in buffer containing 50 mM sodium phosphate, 200 mM NaCl, 150 mM imidazole, and 20% (wt/vol) glycerol. Purified protein was concentrated with Amicon Ultra centrifugal filters (Millipore, Billerica, MA) and dialyzed against phosphate-buffered saline (PBS) plus 10% (wt/vol) glycerol in Slide-A-Lyzer dialysis cassettes (Thermo Scientific, Rockford, IL). Protein concentrations were determined with the Bradford protein assay kit (Thermo Scientific), and purified proteins were stored at −20°C for future use. Purified proteins were visualized by Coomassie blue staining of 15% SDS-PAGE gels and 10% native gels (GelCode blue stain; Thermo Scientific), followed by immunoblotting with anti-YscF.

Cell culture.

The human monocyte cell line THP-1 (ATCC TIB-202) and THP1-XBlue cells (InvivoGen, San Diego, CA) were maintained in RPMI 1640 (Corning Cellgro, Manassas, VA) containing 10% (vol/vol) heat-inactivated fetal calf serum (Invitrogen), 25 mM HEPES (Fisher Scientific, Pittsburgh, PA), 2 mM l-glutamine (Corning Cellgro), 1 mM sodium pyruvate (Corning Cellgro), and 50 μg/ml of Pen-Strep (Corning Cellgro) at 37°C with 5% CO₂. The THP1-XBlue cells contain the secreted embryonic alkaline phosphatase (SEAP) reporter gene under the control of NF-κB and AP-1. Stimulation of THP1 and THP1-XBlue cells was performed as reported previously (20). Briefly, THP1-XBlue cells were seeded at 3 × 10⁶ cells/ml into 96-well plates and THP-1 cells were seeded at 8 × 105 cells/ml into 24-well plates. Cells were suspended in infection medium as described by the manufacturer. Proteins were added at a final concentration of 1 μg/ml. Cells were stimulated at 37°C with 5% CO₂ for 5 h or 24 h.

SEAP reporter assays.

Quantification of SEAP from the supernatant was performed using Quanti-Blue reagent (InvivoGen) according to the manufacturer's protocol. A microplate reader (Synergy HT; BioTek, Winooski, VT) was used to quantify SEAP activity by measuring absorbance at 630 nm; data were collected using KC4 v3.3 software (BioTek).

Cytokine analysis.

THP-1 cells were stimulated with PBS, 1 μg/ml of heat-killed Listeria monocytogenes (HKLM; InvivoGen), 1 μg/ml of lipopolysaccharide (LPS-EK Ultrapure; InvivoGen), 1 μg/ml of flagellin (FLA-ST; InvivoGen), or 1 μg/ml of needle proteins for 5 to 24 h, and cellular supernatants were collected and stored at −20°C before analysis with Quantikine enzyme-linked immunosorbent assay (ELISA) kits from R&D systems (Minneapolis, MN). Human tumor necrosis factor alpha (TNF-α), interleukin 6 (IL-6), and IL-8 kits were used as instructed by the manufacturer.

Preparation of purified needle samples.

Bacterial cultures were grown at 37°C for 8 h in 60 ml of heart infusion broth (HIB; BD Difco) subcultured 1:100 into fresh HIB and incubated overnight at 37°C. Samples were prepared as reported previously (20). Briefly, overnight cultures were harvested by centrifugation (10 min at 10,000 × g) and washed twice in 20 mM Tris-HCl (pH 7.5). The cell suspension was transferred to a 40-ml Dounce glass-glass tissue grinder (Wheaton, Millville, NJ) that shaves needles from the surface of cells through the exertion of sheer force for 60 cycles. Unbroken cells and debris were removed (10 min at 10,000 × g), and the supernatant was passed through a 0.45-μm cellulose acetate membrane filter (Whatman; GE Healthcare Life Sciences, Pittsburgh, PA) to remove all bacteria. The supernatant was then centrifuged at 60,000 × g for 30 min using a Beckman Coulter JA 25.15 rotor. The sediment was suspended in 50 μl of 20 mM Tris-HCl (pH 7.5) plus 5% sucrose and subsequently loaded onto a 6-ml step gradient of 70%, 20%, and 10% sucrose (2 ml each). One-milliliter fractions were collected and analyzed by Coomassie blue staining of 15% SDS-PAGE gels and 10% native gels (GelCode blue stain; Thermo Scientific) and immunoblotting for YscF.

Transcomplementation.

Plasmids were isolated using the Qiaprep Miniprep kit (Qiagen). DNA fragments encoding the N-terminal deletion YscF constructs or the SsaG chimeras were amplified by PCR and purified using the QiaQuick PCR purification kit (Qiagen). The primers used to amplify the DNA fragments are listed in Table 2. Gene amplification was performed with PFU Ultra II HS DNA polymerase (Agilent Technologies, Santa Clara, CA) in a Mastercycler gradient thermal cycler (Eppendorf, Hauppauge, NY). The amplified sequences were digested with EcoRI and ligated into EcoRI- and SmaI-cleaved pBAD18-kan (28). Mutations were confirmed by sequencing and the resulting plasmids transformed into ΔyscF Y. pestis or ΔSsaG Salmonella entrica serovar Typhimurium. Transformants were selected on TBA plates with kanamycin, streak purified, and subsequently used for growth curves, cytotoxicity assay, or gentamicin protection assay.

Growth curves.

Growth curves were performed as described by Straley and Bowmer (30). Briefly, bacteria were grown in TMH medium (25) at 26°C overnight, subcultured into fresh TMH medium at an A₆₂₀ of 0.1 (plus 2.5 mM CaCl₂ and 0.1% l-arabinose [BD Difco, Sparks, MD]), and grown until the A₆₂₀ reached 0.2. The temperature was shifted to 37°C, and bacteria were incubated for 6 h while A₆₂₀ readings were taken at 1-h intervals. One milliliter of sample was aliquoted into 1.5-ml tubes and centrifuged at 20,000 × g for 5 min at 4°C, and pellets were separated from the supernatant, which contains secreted proteins. The pellet and supernatants were precipitated with 10% (wt/vol) trichloroacetic acid (TCA), followed by SDS-PAGE and silver staining.

HeLa cell infections.

Yop translocation was monitored visually by cytotoxicity (cells rounding up) as described previously (18, 31). Briefly, HeLa cells were grown in Dulbecco's modified Eagle minimum essential medium (DMEM) supplemented with 10% fetal calf serum (Invitrogen) and 50 μg/ml of penicillin-streptomycin (Corning Cellgro) at 37°C with 5% CO₂. HeLa cells were seeded into 24-well tissue culture plates, and after the HeLa cells reached near confluence, the growth medium was removed and the cells washed twice with L15 medium and placed in fresh L15 medium containing 0.1% (wt/vol) l-arabinose (BD Difco, Sparks, MD). Bacteria were added to the cell monolayer at a multiplicity of infection (MOI) of 10:1. The plates were centrifuged at 800 × g for 5 min to allow cell contact. The plates were incubated at 37°C for 2 to 6 h to check for cytotoxicity and photographed at 4 h. Micrographs were captured on a Nikon D70 digital camera.

Gentamicin protection survival assay.

For macrophage infections, THP-1 cells were seeded at 5 × 10⁵ cells per well in 24-well tissue culture dishes and were differentiated into macrophages by addition of 50 mM phorbol 12-myristate 13-acetate for 48 h. Macrophage infections were performed as described by Forest et al. (32), with minor modifications. Briefly, bacteria were grown overnight without shaking in LB broth. The overnight culture was subcultured to an A₆₀₀ of 0.1 and grown to an A₆₀₀ of 0.6. Bacteria were added to cell monolayer at an MOI of 10:1 and centrifuged at 800 × g for 5 min to synchronize bacterial uptake. After incubation at 37°C for 2 h, extracellular bacteria were removed by washing cells with PBS and incubated again with medium containing 100 μg/ml of gentamicin to kill extracellular bacteria for 2 h; then the cells were washed and the medium was replaced with fresh medium containing 10 μg/ml of gentamicin and 0.1% (wt/vol) l-arabinose (33) for another 20 h. Counts of gentamicin-resistant intracellular bacteria were determined by lysis of cells for 30 min at 37°C with 1 ml of 0.1% (vol/vol) Triton X-100 in H₂O, and bacteria were enumerated by serial dilution in PBS and plating on LB agar containing kanamycin.

Data analysis and statistics.

Data were assembled into graphs using GraphPad Prism, version 5.0d (GraphPad Software). Statistical analysis was completed using one-way analysis of variance with Bonferroni or Dunnett's multiple-comparison posttest as indicated below.

RESULTS

Alignment of needle proteins.

Type III secretion systems found in animal pathogens can be divided into three main families: Ysc-type injectisomes (e.g., Yersinia and Pseudomonas), Shigella and Salmonella SPI-1-type injectisomes, and E. coli and Salmonella SPI-2-type injectisomes (2). Our assessment of needle proteins includes at least one needle protein homolog from each T3S system family and includes needle proteins from Yersinia pestis (YscF), Salmonella enterica serovar Typhimurium SPI-1 (PrgI) and SPI-2 (SsaG), and Shigella flexneri (MxiH). Needle proteins of the T3S system are highly conserved, except for the N termini (Fig. 1). Representative needle protein sequences were aligned, and this information was used to guide the construction and expression of truncated proteins for a selected set of needle proteins. Truncated needle proteins were made by deletion of residues from the N termini; the truncations correspond to the 22nd amino acid of Y. pestis YscF. This resulted in truncation of amino acids A2 to K15 from PrgI, S2 to T18 from MxiH, and D2 to M9 from SsaG.

FIG 1.

Multiple-sequence alignment of needle proteins demonstrates that the N termini of T3S needle proteins are not conserved. Needle protein sequences from several species of bacteria were aligned with Megalign from the DNAStar Lasergene package (v. 10.1), using the Jotun Hein algorithm with the PAM250 matrix. Identical residues are shown in shaded boxes. Aligned needle proteins are from Y. pestis (YpYscF), S. enterica (PrgI and SsaG), S. flexneri (MXIH_SHIFL), P. aeruginosa (PscF), Aeromonas hydrophila (AhYscF), Vibrio parahaemolyticus (VparaYscF), Burkholderia pseudomallei (BsaL), and E. coli (EscF).

N-terminally truncated needle proteins induce NF-κB/AP-1 activation.

Recently, T3S needle proteins were demonstrated to function as pathogen-associated molecular pattern (PAMPs) for TLR2 and TLR4 (20). Published work suggests that the N termini of needle proteins are not necessary for needle assembly or function (16, 17). These observations led to a hypothesis that the nonconserved N terminus may function in the recognition of needle proteins by the host. To examine this possibility, activation of NF-κB/AP-1 (using a reporter system) was used to evaluate the effect of removing the N termini of needle proteins on immune stimulation. Recombinant whole needle proteins and truncated forms of needle proteins were used to treat THP1-XBlue cells, and NF-κB/AP-1 activation was assessed by measuring secreted embryonic alkaline phosphatase (SEAP) production. SEAP expression in THP1-XBlue cells is under the control of the transcriptional activators NF-κB and AP-1, which activate cytokine and chemokine expression critical for innate immune responses. Therefore, an increase in SEAP expression equated to an increase in NF-κB and/or AP-1 activity. All proteins were applied at 1 μg/ml to cells, and as expected, all of the full-length proteins induced NF-κB/AP-1 activation. Truncated forms of YscF and PrgI significantly activated NF-κB/AP-1 more than their full-length counterparts (Fig. 2A). However, the truncated form of MxiH had lower levels of NF-κB/AP-1 activation than did the full-length MxiH (Fig. 2A). These results showed that the truncated forms of YscF and PrgI increase NF-κB/AP-1 activation to higher levels in comparison to the full-length proteins. SsaG, which is expressed only within eukaryotic cells, activated cells only slightly more than the truncated form. MxiH activated NF-κB/AP-1 more than the truncated form of MxiH. The results suggested that the N termini (notably residues corresponding to the first 15 amino acids of YscF) of needle proteins act to influence host recognition of the proteins. The responses from the N-terminal truncations of YscF and PrgI suggested that the presence of the nonconserved N termini may inhibit NF-κB/AP-1 activation in some cases, whereas the presence of the N terminus of MxiH may increase NF-κB/AP-1 activation by MxiH in THP1-XBlue cells.

FIG 2.

Full-length and truncated needle proteins activate NF-κB/AP-1 in THP1-XBlue cells in an MyD88-dependent manner. THP-1 (A) and THP-1 defMyd88 (B) cells were seeded in wells and treated with PBS, 1 μg/ml of HKLM (Heat Killed Listeria monocytogenes), 1 μg/ml of LPS, 1 μg/ml of flagellin, 1 μg/ml of l-Ala-γ-d-Glu-meso-diaminopimelic acid (tri-DAP), or 1 μg/ml of needle protein dissolved in PBS. SEAP levels were measured as representation of NF-κB/AP-1 activation. Error bars represent SEMs. n = 3. Data are representative of at least three experiments. *, P is between 0.05 and 0.01; ***, P is between 0.001 and 0.0001; ****, P < 0.0001.

When needle proteins (both truncated and full length) were incubated with THP1-XBlue cells deficient in MyD88, an adaptor protein required for most TLR-mediated responses (34, 35), the activation of NF-κB/AP-1 was abolished (Fig. 2B). The lack of NF-κB/AP-1 activation in cells lacking MyD88 indicated that the NF-κB/AP-1 activation by the truncated needle proteins was most likely occurring through TLR2 and TLR4 recognition of the needle proteins, as previously demonstrated (20) for the full-length proteins.

Definition of amino acids within the N terminus involved in TLR activation.

Having demonstrated an effect by the N termini of the needle proteins in modulating NF-κB/AP-1 activation, the specific amino acids in the N terminus of YscF involved in NF-κB/AP-1 activation were sought. A panel of yscF truncations was constructed that resulted in deletions of amino acids S2 to S5, S2 to G10, S2 to D15, or S2 to A20 from the N terminus of YscF. Truncated YscF proteins were expressed and subsequently purified: their purity was assessed by Coomassie blue staining following SDS-PAGE (see Fig. 4). The ability of truncated YscF and full-length YscF to induce cytokines from THP-1 cells and NF-κB/AP-1 activation from THP1-XBlue cells was analyzed (Fig. 3A). The truncated YscF proteins induced significantly higher cytokines levels (TNF-α, IL-6, and IL-8) (Fig. 3B to D) and NF-κB/AP-1 activation (Fig. 3A) than full-length YscF. Truncation of amino acids S2 to D15 from the N terminus of YscF resulted in the highest levels of cytokine and NF-κB/AP-1 activation, similar to levels observed with SsaG (Fig. 2).

FIG 4.

Coomassie blue-stained SDS-PAGE and native gels of purified needle proteins and immunoblot of YscF purified from Y. pestis, detected with anti-YscF primary antibody. (A and B) Coomassie blue-stained native gels (A) and SDS-PAGE gels (B) of recombinant needle protein purifications. Lanes: 1, YscF; 2, YscFΔ(S2-S5); 3, YscFΔ(S2-G10); 4,YscFΔ(S2-D15), 5, YscFΔ(S2-A20); 6, YscFΔ(S2-T22). (C) Immunoblot of recombinant needle proteins probed with anti-YscF antibody. (D and E) Coomassie blue-stained native gels (D) and SDS-PAGE gels (E) of sheared needle proteins from Y. pestis KIM8-3002 (lanes A) and Y. pestis ΔyscF KIM-3002.p61 containing plasmids pBAD18-Kan expressing YscFΔ(S2-D15) (lanes B) or YscFΔ(S2-A20) (lanes C). Lane M in panels A to E indicates prestained size markers (Precision Plus markers; Bio-Rad Laboratories, Hercules, CA).

FIG 3.

YscF truncated at 15 amino acids has the highest NF-κB/AP-1 and cytokine activation. (A) THP1-XBlue cells were seeded in wells and treated with PBS, 1 μg/ml of LPS, or 1 μg/ml of recombinant needle protein dissolved in PBS. SEAP levels were measured as a representation of NF-κB/AP-1 activation. (B to D) THP-1 cells were treated with PBS, 1 μg/ml of HKLM, or recombinant needle proteins (1 μg/ml). After 5 h (TNF-α) and 24 h (IL-6 and IL-8), supernatants were collected and tested by ELISA for production of TNF-α (B), IL-8 (C), and IL-6 (D). Error bars represent SEMs (n = 3). Data are representative of at least three experiments. *, P is between 0.05 and 0.01; ***, P is between 0.001 and 0.0001; ****, P < 0.0001.

To confirm that proteins were activating NF-κB/AP-1 (20), needle proteins were digested with proteinase K and added to THP1-XBlue cells. As expected, the proteinase K-treated needle proteins failed to activate NF-κB/AP-1 activation (see Fig. S1A in the supplemental material), suggesting that proteins caused the activation of NF-κB/AP-1. To further demonstrate that needle proteins, not LPS or lipoprotein (LPL) contaminants, were responsible for the cellular response, needle proteins were incubated with lipoprotein lipase to inactivate LPLs or treated with polymyxin B (PMB) to neutralize LPS. Treatment of needle proteins with either LPL or PMB did not affect the ability of the proteins to induce NF-κB/AP-1 activation (see Fig. S1B and C in the supplemental material), showing that the cytokine expression seen was not due to contaminating lipoproteins or LPS.

N-terminal truncations of YscF are responsible for the differences in NF-κB activation of THP1-Xblue cells.

Full-length forms of monomeric needle proteins are known to have a strong tendency to oligomerize and self-associate (36). This needle polymerization can, however, be eliminated in some cases by truncating 5 or more amino acids from the C termini of PrgI, MxiH, and BsaL (36). This modification does not have the same effect on YscF and PscF, as they rapidly self-associate upon expression and purification (37). Consequently, truncation of 5 or 10 amino acids from the C terminus did not eliminate needle polymerization of YscF (data not shown), suggesting that polymers could be the dominant protein species in our protein preparation samples.

To show whether the observed differences between the various mutants were due to differences in proportions of the protein species, purified recombinant and sheared needle proteins were visualized by Coomassie blue staining of SDS-PAGE gels and native gels, followed by immunoblotting with anti-YscF. There were no observed differences in band sizes between full-length and truncated YscF. Additionally, all bands on the gel reacted with an anti-YscF antibody, demonstrating that the protein preparation samples were composed of YscF needle proteins (Fig. 4). These results demonstrate that the differences in NF-κB and AP-1 activation were dependent on the number of amino acids truncated from the N terminus of YscF and not differences in the proportions of the protein species. To further confirm that deleting amino acids from the N terminus was responsible for the observed differences in cytokine expression, THP1-XBlue cells were stimulated with C-terminally truncated YscF needle proteins. These C-terminally truncated YscF mutants activated NF-κB and AP-1 similarly to full-length YscF when added to THP1-XBlue cells (data not shown). Taken together, our results argue that N-terminal truncations of YscF are responsible for the observed differences in NF-κB and AP-1 activation and not differences in proportion of protein species in our protein samples.

Truncated YscF forms functional needles.

The structure of these truncated proteins is not known, and therefore, the effect of these deletions on the function of YscF is uncertain. To address this issue, a ΔyscF Y. pestis strain was transcomplemented with expression plasmids encoding YscF proteins (under the control of the araBAD_pr) lacking amino acids S2 to D15 [YscFΔ(S2-D15)] or S2 to A20 [YscFΔ(S2-A20)]. By comparing growth curves of ΔyscF Y. pestis transcomplemented with wild-type YscF to growth curves of ΔyscF Y. pestis transcomplemented with mutant YscFs, no differences were seen between YscFΔ(S2-D15), YscFΔ(S2-A20), or full-length YscF (data not shown). Similar results were obtained when Yops secretion and translocation profiles were compared (Fig. 5), demonstrating that these truncated YscF proteins formed functional needles.

FIG 5.

Truncated forms of YscF form functional needles. (A) Silver stain analysis of culture supernatant fractions from Y. pestis KIM8-3002 (parent), Y. pestis ΔyscF KIM-3002.p61 containing plasmids pBAD18-Kan (vector), pBAD18-Kan expressing YscF (+YscF), pBAD18-Kan expressing YscFΔ(S2-D15) (+ΔD15YscF), and pBAD18-Kan expressing YscFΔ(S2-A20) (+Δ20YscF) grown for 7 h at 37°C in the presence (+) or absence (−) of 2.5 mM CaCl₂ and 0.1% (wt/vol) l-arabinose. Proteins were separated on a 12.5% SDS-polyacrylamide gel and detected by silver staining. (B) HeLa cells were infected at MOI of 10 with same Y. pestis strains as in panel A. Images were captured 3 h postinfection on an Olympus IX50 inverted microscope fitted with a Nikon D70 digital camera (magnification, ×400) to document cell cytotoxicity.

Wild-type sheared YscF needles are able to trigger NF-κB/AP-1 activation and cytokine (TNF-α, IL-6, and IL-8) expression (20). To verify whether truncated sheared YscF needles could also activate cytokine (TNF-α, IL-6, and IL-8) expression, a ΔyscF Y. pestis strain was transcomplemented with plasmids expressing truncated forms of YscF: YscFΔ(S2-D15) or YscFΔ(S2-A20). When these sheared needles composed of truncated YscF were used to treat THP-1 cells, activation was significantly higher with truncated YscF needles than wild-type needles (Fig. 6A). The higher NF-κB/AP-1 activation translated to higher cytokine (TNF-α, IL-6, and IL-8) expression, similar to what was observed with the recombinant proteins (Fig. 6B to D). The needle formed by N-terminal truncation of amino acids S2 to D15 had the highest activation of NF-κB/AP-1 and cytokine expression. This result with sheared needles paralleled results shown in Fig. 2 and 3 with recombinant YscF needle proteins. Similar to the recombinant proteins, treatment of sheared needles with proteinase K abrogated NF-κB/AP-1 activation of THP1-XBlue cells (see Fig. S1D in the supplemental material). To further confirm that YscF was activating NF-κB/AP-1, the sheared needles were treated with anti-YscF. Antibody to YscF significantly reduced NF-κB/AP-1 (Fig. 6E), demonstrating that neutralizing YscF can abrogate the YscF activation of NF-κB/AP-1. These results demonstrated that needles composed of the truncated forms of YscF could activate NF-κB/AP-1 and cytokine expression in a manner analogous to that of wild-type needles.

FIG 6.

Ysc needles lacking the N terminus of YscF increases NF-κB/AP-1 and cytokine activation. (A) THP1-XBlue cells were seeded in wells and treated with PBS, 1 μg/ml LPS, or 1 μg/ml of sheared needle dissolved in PBS. SEAP levels were measured as representation of NF-κB/AP-1 activation. (B to D) THP-1 cells were treated with PBS, 1 μg/ml of HKLM, or purified needles (1 μg/ml). After 5 h (TNF-α) and 24 h (IL-6 and IL-8), supernatants were collected and tested by ELISA for production of TNF-α (B), IL-8 (C), and IL-6 (D). (E) THP1-Xblue cells were treated with anti-YscF alone, YscF, or YscF incubated with anti-YscF. Error bars represent SEMs (n = 3). Data are representative of at least three experiments. *, P is between 0.05 and 0.01; **, P is between 0.01 and 0.001; ***, P is between 0.001 and 0.0001; ****, P < 0.0001.

Alanine-scanning mutants of YscF reveal a region involved in cytokine expression.

The above-described results showed that needles formed by N-terminal truncation of amino acids S2 to D15 from YscF induced the highest activation of NF-κB/AP-1 and stimulation of cytokine (TNF-α, IL-6, and IL-8) production when used to treat THP-1 cells. This high activation was reduced by truncation of N-terminal amino acids S2 to A20, similar to what was observed with the recombinant proteins. By comparing the aligned sequences (Fig. 1), we identified 3 conserved amino acid residues in this region between amino acids D15 and Q21. This raised the possibility that certain amino acids within this region could be involved in NF-κB/AP-1 activation. The nonalanine codons (encoding residues L16, D17, and V19) in this region were selected for mutagenesis studies. Using site-directed mutagenesis, the codons for these conserved residues (L16, D17, and V19) were replaced with alanine-encoding codons. The individual alanine mutants were tested for NF-κB/AP-1 activation. The results demonstrated that 2 alanine mutants (YscF L16A and YscF V19A) had drastically reduced NF-κB/AP-1 activation, while YscF D17A had no effect on NF-κB/AP-1 activation (Fig. 7). Double mutant (YscF L16A V19A) and triple mutant (YscF L16A D17A V19A) YscF substitutions further reduced NF-κB/AP-1 activation. However, none of these YscF mutants reduced activation to wild-type levels. These results suggested that the region between amino acids D15 and A20 was involved in the higher NF-κB/AP-1 activation of the YscFΔ(S2-D15) mutant than of the YscFΔ(S2-A20) mutant, suggesting, in turn, that the region of YscF between amino acid 15 and amino acid 20 plays a role in NF-κB/AP-1 activation. When these point mutants were examined in the TLR2 and TLR4 HEK 293 reporter cells, all point mutants showed cellular activation patterns similar to what was observed with the THP1-XBlue cells (Fig. 7B and C).

FIG 7.

Alanine-scanning mutants of YscFΔ(S2-D15). THP1-XBlue (A), HEK 293 TLR2 (B), and HEK 293 TLR2 (C) cells were seeded in wells and treated with PBS, 1 μg/ml of LPS, 1 μg/ml of HKLM, or 1 μg/ml of recombinant needle proteins from the various YscFΔ(S2-D15) point mutants dissolved in PBS. SEAP levels were measured as representation of NF-κB/AP-1 activation. Error bars represent SEMs (n = 3). Data are representative of at least three experiments. **, P is between 0.01 and 0.001; ***, P is between 0.001 and 0.0001; ****, P < 0.0001. NS, not significant.

The N terminus of YscF reduces SsaG activation of NF-κB/AP-1 and cytokine expression.

The results shown above demonstrated that the variable N terminus of YscF modulates the interaction of the needle protein with host cells to influence NF-κB/AP-1 activation, likely via TLR2 and TLR4. In order to confirm that the N terminus of YscF interfered with NF-κB/AP-1 activation, the N terminus of YscF was moved onto SsaG, a protein that highly activates NF-κB/AP-1. The YscF-SsaG chimeric protein was constructed by adding the 15 N-terminal amino acids of YscF to SsaG. As shown in Fig. 8A, SsaG activated NF-κB/AP-1 significantly more than YscF. When the N terminus of YscF was added to SsaG, the activation of NF-κB/AP-1 was reduced to the level seen with wild-type YscF (Fig. 8). This showed that the N terminus of YscF modulated the NF-κB/AP-1 activation ability of SsaG considerably, confirming that the N terminus of YscF was capable of decreasing NF-κB/AP-1 activation. To determine whether the reduction in NF-κB/AP-1 activation was achieved by just adding an N terminus to SsaG or if the reduction was due specifically to YscF residues, a frameshift mutation in the plasmid encoding the YscF-SsaG chimera was introduced after the start codon and then the frameshift was fixed to restore the original reading frame of SsaG. These mutations in the plasmid encoding chimeric YscF-SsaG changed the N-terminal amino acids of the new RN-SsaG chimera from MSNFSGFTKGTDIAD to MRYLLWIYERNRYRR. The RN-SsaG chimera had the same level of NF-κB/AP-1 activation as SsaG (Fig. 8A), suggesting that the N terminus of YscF was responsible for the observed decrease in NF-κB/AP-1 activation. As shown above, treatment of chimeric needle proteins with proteinase K, but not LPL or PMB, abrogated NF-κB/AP-1 activation of THP1-Xblue cells (see Fig. S2 in the supplemental material).

FIG 8.

The N terminus of YscF reduces NF-κB/AP-1 and cytokine activation by SsaG. THP1-XBlue cells were seeded in wells and treated with PBS, 1 μg/ml of LPS, or 1 μg/ml of needle proteins: YscF, SsaG, and chimeric YscF-SsaG (A) or YscF, SsaG, and chimeric YscF-SsaG and RN-SsaG (E) dissolved in PBS. SEAP levels were measured as representation of NF-κB/AP-1 activation. (B to D) THP-1 cells were treated with PBS, 1 μg/ml of HKLM, or 1 μg/ml of needle proteins (YscF, SsaG, and chimeric YscF-SsaG). After 5 h (TNF-α) and 24 h (IL-6 and IL-8), supernatants were collected and tested by ELISA for production of TNF-α (B), IL-8 (C), and IL-6 (D). Error bars represent SEMs (n = 3). Data are representative of at least three experiments. ****, P < 0.0001. NS, not significant.

An alternative explanation for the ability of the YscF-SsaG chimera to reduce NF-κB/AP-1 activation is that the addition of the YscF N terminus to SsaG resulted in a misfolded or unstable protein. Therefore, a genetic analysis was devised to test whether these chimeric proteins could form functional needles. First, a ΔssaG Salmonella enterica serovar Typhimurium mutant was constructed and subsequently transcomplemented with YscF-SsaG, RN-SsaG, or SsaG. Because SsaG is required for the intracellular survival of Salmonella (24–26), we tested the functionality of these chimeric proteins by the ability of these mutant Salmonella to survive in macrophages using a gentamicin protection assay (39). Our results showed that the ΔssaG Salmonella mutants transcomplemented with the YscF-SsaG or the RN-SsaG chimeras were able to survive in macrophages similarly to wild-type Salmonella and in contrast to the ΔssaG Salmonella transcomplemented with vector plasmid (Fig. 9). This suggests that the chimeric proteins were able to form functional needles and that the addition of YscF's N terminus onto SsaG had no effect on SsaG's function.

FIG 9.

Addition of the N terminus of YscF to SsaG had no effect on SsaG's function. THP-1 macrophages were infected with S. Typhimurium 14028 (WT) or ΔssaG 14028 (BLS101) transcomplemented with plasmids pBAD18-Kan (+Vector), pBAD18-Kan expressing SsaG (+SsaG), pBAD18-Kan expressing YscF-SsaG (+YscF-SsaG), or pBAD18-Kan expressing RN-SsaG (+RN-SsaG) in the presence of 0.1% l-arabinose. Intracellular survival was determined by comparing the number of bacteria 24 h postinfection. Error bars represent SEMs (n = 3). Data are representative of at least three experiments. **, P is between 0.01 and 0.001.

DISCUSSION

A recent report demonstrated that T3S system needle proteins are pathogen-associated molecular patterns (PAMPs) (20). These needle proteins function as ligands of TLR2 and TLR4 to induce cell activation via an MyD88-dependent pathway leading to expression of proinflammatory cytokines (20). Needle proteins from different species of Gram-negative bacteria, however, induce cell activation to different magnitudes (20) (Fig. 2). In the current study, the hypervariable N termini of these needle proteins were demonstrated to affect the levels of proinflammatory cytokines induced from a human monocyte cell line. These findings may give new insight into the pathogenesis of Y. pestis and other Gram-negative bacteria that use T3S systems to inject bacterial effector proteins to subvert host defenses and promote infection.

T3S needle protein sequences from Y. pestis and other Gram-negative bacteria have been compared and reported to be highly similar (4, 16) (Fig. 1). T3S needle proteins show higher sequence conservation past the first ∼20 to 25 amino acids. Short deletions of residues at the C termini of PrgI, MxiH, and BsaL prevent needle polymerization (16, 17), suggesting that the conserved C termini are involved in needle formation. However, the same observations cannot be made for the N termini of needle proteins, which vary among bacterial species not only in amino acid composition but also in the number of amino acids (Fig. 1). Because the N termini of needle proteins do not appear to be involved in needle assembly (16, 17) and since they are exposed on the needle surface (14, 15), we hypothesized that the differences in cytokine expression (19) could be due to the variations in the N termini of these proteins (40–42). This observation is reinforced by alignments presented in Fig. 2 from reference 15, in which Demers et al. also illustrate the variability of the N termini of needle proteins. Importantly, Demers et al. demonstrate that the N termini of the needle proteins are found on the outside of the assembled needles (15) from Shigella, where they may have more opportunity to influence interaction with TLRs. This topology of the needle structure is in agreement with the work of Loquet et al. (14), who demonstrated that the N terminus of PrgI is also surface localized.

Consistent with the hypothesis that the N terminus of needle proteins could influence activation of NF-κB/AP-1, recombinant needle proteins lacking the variable N termini showed different activation of NF-κB/AP-1 and cytokine expression (TNF-α, IL-6, and IL-8) from those of their wild-type counterparts. These differences in cytokine expression indicate that the variable N terminus could confer a selective advantage for the survival of these pathogenic bacteria. Accordingly, the removal of the N terminus of YscF increased induction of NF-κB/AP-1 activation and cytokine expression compared to those of the full-length protein. Owing to the initial preinflammatory phase required for Y. pestis to survive and replicate (43), our results imply that the N terminus of YscF in Y. pestis may function to interfere with host sensing in an attempt to either evade or suppress early host innate immune responses to prevent inflammation and bacterial clearance. Similar to YscF, the N terminus of PrgI may play a role to dampen the host immune response by blocking inflammation, resulting in a mechanism by which Salmonella organisms control their population density during the initial stages of infection (44). In contrast to the case with Yersinia and Salmonella, induction of immune responses leading to inflammation and subsequent neutrophil recruitment is known to promote invasion and dissemination of Shigella (3, 22, 23). Consequently, removal of the N terminus of MxiH attenuated the high inflammatory response associated with the full-length protein. This suggests that in accordance with the infection goal of Shigella, the N terminus of MxiH may positively add to the proinflammatory environment, possibly by enhancing activation of TLR2 and TLR4. These observations are in agreement with results obtained with flagellin by Smith et al. (40), who showed that the N terminus of flagellin is involved in immune manipulation by some bacteria (Helicobacter, Campylobacter, and Bartonella) to the advantage of the pathogens. Similar to the case with flagella, our data show that the N termini of needle proteins could be involved in manipulating the host response to the advantage of the bacteria.

An exception to the observed differences in cytokine expression between full-length and truncated proteins involved the Salmonella SPI-2 needle protein, SsaG. Truncated SsaG showed levels of NF-κB/AP-1 activation and cytokine expression similar to those of the full-length protein (Fig. 2). This result was not surprising because multiple-sequence alignment of T3S system needle proteins (Fig. 1) showed that SsaG lacked residues corresponding to the 15 N-terminal amino acids of YscF and comparable residues in other needle proteins. As a result, N-terminal deletion SsaG was not expected to have a significant effect on SsaG's NF-κB/AP-1 activation and cytokine expression. Under a hypothesis that the N terminus could modulate TLR2 and/or TLR4 interaction, SsaG would not need an extended N terminus, because SsaG is only expressed when Salmonella is enclosed in the Salmonella-containing vacuole (24–26, 45) and therefore is probably not exposed to TLR2 and TLR4 on the outside of the host cell. These findings give strength to the notion that the variable N termini of T3S system needle proteins do appear to have an immune modulatory function.

Having shown that N-terminally truncated YscF induced higher proinflammatory cytokines than the full-length protein, it would be important to determine the region of YscF that is responsible for this induction of inflammation. To begin to address this issue, a mutational analysis of recombinant and sheared needle proteins was utilized. Deletion of amino acids S2 to S5, S2 to G10, S2 to D15, or S2 to A20 from the N terminus of YscF increased proinflammatory activity (Fig. 3 and 6). The highest activity was observed when amino acids S2 to D15 were deleted from the N terminus of YscF. The level of NF-κB/AP-1 activation and cytokine expression for YscF with an N-terminal truncation of S2 to D15 [YscF Δ(S2-D15)] was similar to that of full-length SsaG. The observed similarity in cytokine expression was striking, as the deletion of S2 to D15 from the N terminus of YscF corresponded to the beginning of SsaG according to our multiple-sequence alignment (Fig. 1). This strengthened the notion that the variable N termini of T3S system needle proteins may not play a role in maintaining a functional needle but may function in manipulating host immune responses. A second possibility is that the differences in NF-κB/AP-1 activation between the full-length and truncated proteins may be due to differences in proportions of protein polymers. However, this seems unlikely given that NF-κB/AP-1 activation did not differ significantly between full-length and truncated SsaG (Fig. 2A), nor did deleting residues from the C terminus of YscF increase NF-κB/AP-1 activation compared to that obtained with full-length YscF. Additionally, gels and blots of purified needle proteins confirm that they are all stable and demonstrate qualitatively similar patterns of polymer distribution. Furthermore, while it is possible to prepare soluble monomers that maintain their native secondary structures when five residues are deleted from the C termini in MxiH, PrgI, and BsaL, YscF and PscF rapidly self-associate upon expression and purification (Fig. 4) (36, 37), corroborating our results showing that deleting residues from YscF does not shift the balance between monomers and polymers.

Interestingly, truncation of amino acids S2 to A20 from the N terminus of YscF [YscFΔ(S2-A20)] yielded reduced cytokine expression compared to that of YscFΔ(S2-D15). These results open the possibility of a unique TLR binding signature between amino acids D15 and A20 in YscF that is blocked by the exposed N terminus. The region (D15 to A20) contains three nonalanine residues that are highly conserved among different bacterial species. As demonstrated in Fig. 7, site-directed mutagenesis of two amino acids (L16A and V19A) showed significantly reduced NF-κB/AP-1 activation in THP1-XBlue cells. Substitution of all 3 amino acids further reduced NF-κB/AP-1 activation to almost wild-type levels; however, mutation of D17A alone had no effect on NF-κB/AP-1 activation. This indicated that amino acids L16 and V19 were important in NF-κB/AP-1 activation by YscF. Similar findings are reported for studies focusing on the PAMP activity of flagellin (40, 46). Those studies demonstrated that mutant flagellin with substitutions of L88A, L94A, I411A, D412A, and L425A either singly or in combination significantly reduced TLR5 recognition by S. Typhimurium FliC (40, 46). Recent studies with the neisserial porin protein PorB provide insight into binding signatures for TLR2 in PorB. PorB is highly conserved between Neisseria species, except for the surface-exposed loops (47–50). Much like T3S needle proteins, PorB (specifically the exposed loops) was found to interact with TLR2. The interaction with TLR2 is dependent on specific amino acids, and the variation between PorB of different Neisseria species creates unique TLR2 binding signatures similar to what our findings suggest. Specific binding signatures were found to be more or less inflammatory through interaction with TLR2 (47–50), which corroborates our earlier report (20). Collectively, our data suggest that some of the proinflammatory activity of YscF resides in the conserved region between amino acids D15 and A20.

Transcomplementation studies were used to analyze the ability of the N-terminal YscF deletions to function in Y. pestis. N-terminal YscF deletions of S2 to D15 or S2 to A20 did not affect needle function (Fig. 5). This observation is consistent with earlier reports that showed that deletion of 12 (19) or 19 (51) amino acids from the N terminus of YscF from Y. enterocolitica and Y. pseudotuberculosis, respectively, did not affect needle assembly and function. Similar to the findings of Lwande and Wedemeyer (51), the YscF deletions did not have a constitutive secretion phenotype, in contrast to results reported for Y. enterocolitica by Allaoui et al. (19). The discrepancy between these findings regarding the secretion phenotype could be due to species-specific differences in yscF from Y. pestis and Y. enterocolitica. Another possible explanation for this discrepancy could be the differences in copy number of yscF used in the separate studies. In the current study, the YscF mutants were expressed from a high-copy-number plasmid (pBAD18-Kan). Allaoui et al. made their truncation on the pYV plasmid. In spite of these explanations, the effect of various mutations in the N terminus of YscF on proinflammatory activation, Yops secretion and translocation is no doubt a complex area, and further work will be required to fully elucidate the role of the N terminus of YscF in secretion and translocation of Yops. Additionally, changes in YscF structure due to the N-terminal deletions cannot be excluded, although the ability of the YscF truncations to transcomplement a yscF strain suggests that native structure is maintained. Furthermore, removal of the N terminus was required to obtain crystals for structural data for several needle proteins and the N terminus is reported to be unstructured and flexible (7, 16, 17).

In this study, the N terminus of YscF was found to modulate NF-κB/AP-1 activation by YscF. To further confirm the modulatory effect of YscF on host immune responses, recombinant chimeric constructs were made to add the 15 N-terminal residues of YscF to the N terminus of SsaG. The chimeric YscF-SsaG construct reduced NF-κB/AP-1 activation and cytokine expression compared to those with SsaG (Fig. 8). In contrast, the addition of random amino acids to the N terminus of SsaG did not reduce NF-κB/AP-1 activation by SsaG (Fig. 7A). This observation suggested that the modulatory effect of YscF on NF-κB/AP-1 activation was sequence specific. This result is in agreement with the findings of Murthy et al. (52), who showed that changes in amino acid sequences of motif N (52) of flagellin abolishes its proinflammatory activity. Moreover, genetic analysis showed that the addition of YscF's N terminus to SsaG had no effect on SsaG's function, evidenced by the ability of a ΔssaG Salmonella mutant transcomplemented with these chimeric needle proteins to survive in macrophages (Fig. 9). Together, these results confirmed that the N terminus of YscF can modulate host immune responses and further confirmed that sequence alterations at the N termini of needle proteins have little or no effect on needle function (39).

This study independently provided a structural functional correlation that reflected a role for the N termini of needle proteins in manipulating host immune responses. Currently, the role that the N terminus of YscF plays during Y. pestis infection is being investigated. These studies will allow us to begin to address the question of the role of N terminus of YscF in the pathogenesis of Yersinia infections as well as other Gram-negative bacteria that utilize a T3S system to subvert the host immune response. Detailed understanding of pattern recognition receptor-needle protein interaction will provide a clearer picture of host-pathogen interactions and permit design of needle proteins as adjuvants or immunomodulatory drugs.