Bacterial growth conditions

N. meningitidis strains HT1125 [35] and its derivatives, H44/76 [45] as well as derivatives harboring IncQ plasmids (Table 1) were routinely grown on GC agar plates at 37 ^oC in a 5% CO₂ atmosphere [46]. E. coli strains were grown in Luria-Bertani (LB) broth liquid cultures (Becton- Dickinson) or on LB plate (LB liquid medium containing 1.5% agar) or in MagicMedia (Invitrogen) at the indicated temperatures. When required, antibiotics were added at the following concentrations: kanamycin at 150 μg/ml, chloramphenicol at 5 μg/ml, erythromycin at 4 μg/ml, and spectinomycin at 75 μg/ml for N. meningitidis; kanamycin at 50 μg/ml, chloramphenicol at 25 μg/ml, and ampicillin at 50 μg/ml for E. coli. The N. meningitidis and E. coli strains used in the present study are listed in Table 2.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. IncQ plasmids introduced into N. meningitidis strains which results were shown in this study.

https://doi.org/10.1371/journal.pone.0237883.t001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Strains used in this study.

https://doi.org/10.1371/journal.pone.0237883.t002

Production of anti-PamA protein rabbit antiserum

All PCR experiments were performed with PrimeSTAR Max GXL DNA polymerase (Takara Bio, Japan). The approximately 3.5-kb PCR fragment containing the pamA gene amplified from N. meningitidis HT1125 genomic DNA with the primer set (NMB1345-21 and NMB1345-2) (S1 Table) was cloned into the pTWV228 vector (Takara Bio, Japan), resulting in pHT922 (S2 Table), which is the ancestral plasmid in the present study. A ΔN-pamA gene, in which the N-terminal hydrophobic region consisting of the first 23 amino acids was removed (S1 Fig), was amplified by PCR with the primer set (NMB1345(pET303)-1(NsiI) and NMB1345(pET303)-2(BH)) (S1 Table) and cloned into the same site of the pET303/CT-His expression plasmid (Invitrogen) to construct pHT934 (S3 Table). Regarding the production of the ΔN-PamA protein for rabbit immunization, a 1 L culture of E. coli strain BL21 (NEB) harboring the plasmid was grown in MagicMedia at 37°C overnight, and the cells were collected. The subsequent purification of the recombinant protein and generation of polyclonal rabbit antibodies to the putative hydrophilic domain of the PamA protein were performed as described previously [46].

Construction of meningococcal mutants

In order to construct the N. meningitidis pamA deletion mutant, the 6.3-kb PCR fragment was amplified with the primers (NMB1345-3 and NMB1345-4) (S1 Table) from pHT922, in which a 1.5-kb DNA fragment containing the coding region of the pamA gene was completely removed, and ligated with a spectinomycin-resistance gene (spc) after phosphorylation to construct pHT930. A 3-kb DNA fragment, in which the pamA structural gene was replaced with the spc gene, was amplified with the primers (NMB1345-21 and NMB1345-2) (S1 Table) from pHT930 and transformed into N. meningitidis strain HT1125, and spectinomycin-resistant (Spc^R) clones were selected, resulting in the ΔpamA mutant HT1822 (Table 2).

The ΔpamA mutant complemented with the pamA⁺ gene at the ggt allele was constructed as follows: A chloramphenicol-resistance gene (cat) was inserted into the SmaI site downstream of the pamA gene of pHT922 to generate pHT923 (S2 Table). A 4.6-kb PCR fragment containing the pamA⁺-cat gene was amplified from pHT923, in which the meningococcal ggt structural gene was cloned in pTWV228 with primers (M13-RV-ggt-5′and M13-47-ggt-3′) (S1 Table) and cloned into the BstXI site of pHT195 (at the middle of the ggt coding region) by In-Fusion cloning (Clontech), resulting in the plasmid pHT924. A 6.3-kb DNA fragment containing the ggt::pamA⁺-cat genes was amplified from pHT924 with primers (ggt-3 and ggt-4) (S1 Table), transformed into N. meningitidis ggt::pamA⁺-cat mutants, and chloramphenicol-resistant (Cml^r) clones were selected, resulting in the pamA deletion mutant HT2224, which were ectopically complemented with the pamA⁺ genes expressed from their own promoter at the ggt allele (Table 1).

Since we have demonstrated that the pilE-cat translational fusion could work as well as the wild-type PilE⁺ N. meningitidis strain [36], an N. meningitidis I12A pilE mutant, in which Ile at position 12 was substituted with Ala, was constructed as described previously. In brief, pHT872 (pilE⁺-cat translational fusion on pGEM-3z) (S2 Table) [36] was site-directed mutagenized with primers (pilE-I12A-1 and pilE I12A-2) (S1 Table) to construct pHT1536 (S2 Table). After confirming its sequence, a 1.5-kb PCR fragment containing the pilE I12A-cat gene was amplified from pHT1536 by a primer set (T7 and SP6) (S1 Table) and transformed into HT1125. Cml^r clones were selected and the mutation was confirmed by direct sequencing as for the pilE I12A-cat N. meningitidis mutant HT2167 (Table 1).

To construct the pilE⁺-ermC or pilE I12A-ermC translational fusion in N. meningitidis strain H44/76 harboring the IncQ plasmid, a 0.9-kb DNA fragment from HT1125 chromosomal DNA containing the structural gene of the wild-type pilE gene and its upstream region was amplified with both primers (pGEM-3z-(SmaI)-2(15mer)-pilE-51 and pilE-52) (S1 Table). A 0.2-kb DNA fragment of the pilE downstream region was also amplified with a primer set (pilE-53 and pilE-18) (S1 Table), and a 0.75-kb DNA fragment of the ermC gene containing the SD sequence was amplified from pHT24 with a primer set (ermC-21 and ermC-22) (S1 Table). The three PCR fragments were cloned into the SmaI site of pGEM-3z to construct pHT1504 (S2 Table). The N. meningitidis I12A pilE mutant was constructed as described above using the primer set (pilE-I12A-1 and pilE I12A-2) (S1 Table) with pHT1504 as the template. After the sequence was confirmed, a 1.5-kb PCR fragment containing the pilE I12A-ermC genes was amplified from pHT1536 (S2 Table) with universal primers (T7 and SP6) and transformed into HT1125. Erm^r clones were selected and the mutation was confirmed by direct sequencing as for the pilE I12A-ermC N. meningitidis mutant HT2166 (Table 1). A 1.5-kb PCR fragment containing the pilE⁺-ermC genes amplified from pHT1504 (S2 Table) or the pilE I12A-ermC genes amplified from pHT1536 (S2 Table) by universal primers (T7 and SP6) was transformed into HT1940. Erm^r clones were selected and the mutation was confirmed by direct sequencing as for the pilE⁺-ermC HT2162 or pilE I12A-ermC HT2166 N. meningitidis mutants in the ΔpamA::spc genetic background (Table 1).

The addition of a FLAG tag to the pilE gene at the 3′-terminus on its chromosomal locus was achieved as follows: a 1.6-kb PCR fragment containing the pilE allele in H44/76 was amplified with a primer set (pilE-11and pilE-12) (S1 Table), and after phosphorylation, it was cloned into pMW119 to construct pHT1419 (S2 Table). A 5.8-kb PCR fragment was amplified with a primer set (pilE-17 and FLAG’(15mer)-pilE-18) (S1 Table) from pHT24, and a 1-kb PCR fragment containing the ermC gene was also amplified with a primer set (pilE-17’(15mer)-M13-47 and FLAG-RV-M) (S1 Table) from pHT24. The two PCR fragments were ligated by In-Fusion Cloning to construct pHT1420, in which the FLAG tag was fused to the pilE structural gene followed by an ermC selection marker (S2 Table). A 2.6-kb PCR fragment containing the pilE^+-FLAG-ermC genes was amplified from pHT1420 with a primer set (pilE-11 and pilE-12) and transformed into HT1940. Erm^r clones were selected and the mutation was confirmed by direct sequencing as for the pilE⁺-FLAG-ermC N. meningitidis mutant HT2095 (Table 1).

The pilE insertional mutant HT2014 was constructed by transforming HT1940 with 500 ng of HT1015 (H44/76 pilE::ermC) [35] chromosomal DNA, followed by the selection of the Erm^r strain HT2014 (Table 1).

The addition of a FLAG tag to the pilF gene at the 3′-terminus on its chromosomal locus was performed as follows: a 1.8-kb PCR fragment containing the pilF allele in the H44/76 N. meningitidis strain was amplified with a primer set (pMW(SmaI)-up(15mer)-pilF-1 and pMW(SmaI)-down(15mer)-pilF-2) (S1 Table) and cloned into the SmaI site of pMW119 by In-Fusion cloning to construct pHT1449 (S1 Table). A 6-kb PCR fragment amplified with a primer set (pilF-3 and FLAG’(15mer)-pilF-4) (S1 Table) from pHT1499 and a 1-kb PCR fragment containing the ermC gene amplified with a primer set (pilF-3’(15mer)-M13-47 and FLAG-RV-M) (S1 Table) from pHT24 were ligated by In-Fusion Cloning to construct pHT1457, in which the FLAG tag was fused to the pilF structural gene followed by an ermC selection marker (S1 Table). A 2.8-kb PCR fragment containing the pilF⁺-FLAG-ermC genes was amplified from pHT1457 with a primer set (pMW(SmaI)-up(15mer)-pilF-1 and pMW(SmaI)-down(15mer)-pilF-2) (S1 Table), and then transformed into HT1125 and HT1822. Erm^r clones were selected and the mutation was confirmed by direct sequencing as for the pilF⁺-FLAG-ermC N. meningitidis mutants HT2218 and HT2219, respectively (Table 1).

A 3.9-kb DNA fragment containing partial pilM-pilN⁺-pilO⁺-pilP⁺ genes in H44/76 N. meningitidis was amplified with a primer set (pMW(SmaI)-up(15mer)-pilMNOP-3 and pMW(SmaI)-down(15mer)-pilMNOP-4) (S1 Table) and cloned into the SmaI site of pMW119 by In-Fusion cloning to construct pHT1450 (S2 Table).

The addition of a FLAG tag to the pilP gene at the 3’ terminus on its chromosomal locus was achieved as follows: a 7.2-kb DNA fragment amplified with a primer set (pilP-3 and FLAG’(15mer)-pilP-2) (S1 Table) from pHT1450 and a 1-kb PCR fragment containing the ermC gene amplified with a primer set (pilP-3’(15mer)-M13-47 and FLAG-RV-M) (S1 Table) from pHT24 were ligated by In-Fusion cloning to construct pHT1454 (S2 Table), in which the FLAG tag was fused to the pilP structural gene at the 3’ terminus followed by an ermC selection marker. A 4.9-kb PCR fragment containing the pilP⁺-FLAG-ermC genes was amplified from pHT1454 with a primer set (pMW(SmaI)-up(15mer)-pilMNOP-3 and pMW(SmaI)-down(15mer)-pilMNOP-4)(S1 Table) and transformed into HT1125 and HT1822. Erm^r clones were selected and the mutation was confirmed by direct sequencing as for the pilP⁺-FLAG-ermC N. meningitidis mutants HT2136 and HT2137, respectively (Table 1).

The addition of the HA tag to the pilM gene at the 3’ terminus on its chromosomal locus was achieved; however, the ORFs of the pilM (1.1 kb), pilN (0.6 kb), and pilO (0.6 kb) genes were very close or overlapped, and thus an ermC selection marker was inserted downstream of the pilO gene. A 7.1-kb PCR fragment containing the partial pilM gene and pMW119 from pHT1450 was amplified with a primer set (pilO-3 and HA’-pilM-2) (S1 Table). A 1.2-kb PCR fragment was amplified with a primer set (HA’(15mer)-pilN-1 and pilO-2) from pHT1450 and a 1-kb ermC gene fragment was also amplified with a primer set (pilO-2’(15mer)-M13-RV and pilO-3(15mer)-M13-47) (S1 Table) from pHT24. The three PCR fragments were ligated by In-Fusion Cloning to construct pHT1453, in which the HA tag was fused to the pilM structural gene with an ermC selection marker (S2 Table). A 2.6-kb PCR fragment containing the pilM^+-HA pilN⁺pilO⁺-ermC-pilP genes was amplified from pHT1453 with a primer set (pMW(SmaI)-up(15mer)-pilMNOP-3 and pMW(SmaI)-down(15mer)-pilMNOP-4) (S1 Table) and transformed into HT1125 and HT1822. Erm^r clones were selected and the mutation was confirmed by direct sequencing as for the pilM⁺-HA-ermC N. meningitidis mutants HT2132 and HT2133, respectively.

The addition of a FLAG tag to the pilX gene at the 3’ terminus on its chromosomal locus was achieved as follows: a 0.5-kb pilX DNA region containing the 400-bp upstream and 160-bp downstream regions of the pilX ochre codon in H44/76 was amplified with a primer set (pMW(SmaI)-up(15mer)-pilX-3 and pMW(SmaI)-down(15mer)-pilX-6) (S1 Table), and then cloned into the SmaI site of pMW119 by In-Fusion cloning to construct pHT1594. A 4.7-kb PCR fragment amplified with a primer set (M13-47′(15mer)-pilX-5 and FLAG′(15mer)-pilX-8) (S1 Table) from pHT1594, and a 1-kb PCR fragment containing the ermC gene amplified with a primer set (FLAG-RV-M and M13-47) (S1 Table) from pHT24 were ligated by In-Fusion Cloning to construct pHT1597, in which the FLAG tag was fused to the pilX structural gene followed by an ermC selection marker (S2 Table). A 1.5-kb PCR fragment containing partial pilX^-FLAG-ermC genes was amplified from pHT1597 with a primer set (pMW(SmaI)-up(15mer)-pilX-3 and pMW(SmaI)-down(15mer)-pilX-6) (S1 Table) and transformed into HT1125 and HT1822. Erm^r clones were selected and the mutation was confirmed by direct sequencing as for the pilX⁺-FLAG-ermC N. meningitidis mutants HT2211 and HT2212, respectively (Table 1).

pamA-lacZ translational fusion N. meningitidis mutants were constructed as follows: a 3-kb PCR fragment containing the lacZ gene was amplified from E. coli strain W3110 [47] chromosomal DNA with the primers (lacZ-21 and M13-RV’(15mer)-lacZ-22) (S1 Table) and a 1-kb ermC gene fragment was amplified with universal primers (M13-RV and M13-47) (S1 Table) from pHT24 [33]. The two PCR fragments were cloned into a 7.2-kb DNA fragment (designated Fragment C) amplified from pHT922 containing the upstream and downstream regions of the pamA gene with the first 408-bp region including the N-terminal membrane spanning region (designated pamA′) with a primer set (M13-47′(15mer)-nmb1345-61 and lacZ′-21(15mer)-nmb1345-62) (S1 Table) by In-Fusion cloning to construct pH1605 (S2 Table and Fig 2B).

pamA′-phoA translational fusion N. meningitidis mutants were constructed as follows: a 1.5-kb PCR fragment containing the phoA gene was amplified from E. coli strain W3110 chromosomal DNA with the primers (phoA-21 and M13-RV’(15mer)-phoA-2) (S1 Table), and a 1-kb ermC gene fragment was amplified with universal primers (M13-RV and M13-47) from pHT24 [33]. The two PCR fragments were cloned into Fragment C by In-Fusion Cloning to construct pH1603 (S2 Table). PCR fragments amplified with nmb1345-5 and nmb1345-21 (S1 Table) from pHT1605 containing the pamA′-lacZ-ermC-downstream region of the pamA gene or from pHT1603 containing the pamA′-phoA-ermC-downstream region of the pamA gene were transformed to HT1125, and erythromycin-resistant clones were selected, as for the pamA′-lacZ::ermC or pamA′-phoA::ermC N. meningitidis mutants on its chromosome HT2215 and HT2217, respectively (Table 1 and Fig 2B).

Assessment of host cell-associated and internalized bacteria

The infection of HBMEC with N. meningitidis was performed as described previously [46]. Results are expressed as means ± the standard deviation (SD), and bacterial numbers were statistically compared using the two-tailed Student’s t-test.

Western blotting

SDS-PAGE and Western blotting were performed as described previously [48].

Observations of meningococci and ezrin accumulation by immunofluorescence staining

HBMEC monolayer preparations, N. meningitidis infections of these layers, and the preparation of immunostaining samples were performed as described previously [46]. Infected HBMEC and attached meningococci were observed using an ECLIPSE E600 microscope (Nikon) with a 100× oil immersion objective.

Construction of IncQ plasmids for N. meningitidis

The mutated M. mazei pylT (tRNA^Pyl) and mutated pylRS genes for mAzZLys (PylRS[Y306A/Y384F]) [49] and pBPa (PylRS A302T/N346T/C348T/W417C]) [50] (MmpBpaRS) were derived from pCDF-Pyl-Fx1 [49]. The four point mutations were generated by three-step mutations with the following three primer sets (A302-T and A302-7-R) (S1 Table), (N346T-C348T-F and N346-C348-R) (S1 Table), and (W417C-F and W417C-R) (S1 Table) using the Mutagenesis Basal Kit (Takara Bio). The plasmid constructed from pCDF-Pyl-Fx1 as described above was named pCDF-Pyl-Fx3 in this study (S3 Table). A 2-kb PCR fragment containing one copy of the pylRS gene (Y306A/Y384F) and two copies of the pylT genes expressed from the E. coli lpp promoter [49] (pylRS (Y306A/Y384F)/T) from pCDF-Pyl-Fx1 or a 2-kb PCR fragment containing one copy of the MmpBPaRS genes and two copies of the pylT genes expressed from the E. coli lpp promoter [49] (MmpBPaRS/T) from pCDF-Pyl-Fx3 were amplified by PCR with a primer set (ptac-12-pylRS-1 and M13-47-reverse-lppT-2) (S1 Table), and then cloned into pTTQ18, an expression vector with a strong tac promoter (P_tac) controlled by the LacI protein [51] using In-Fusion cloning to construct pHT1208 and pHT1261, respectively (S2 Table). The PCR fragment containing the lacI^q-P_tac-pylRS[Y306A/Y384F]/T or lacI^q-P_tac-MmpBPaRS/T genes was amplified by PCR with primers (M13-47 reverse and ptac-12) (S1 Table), and after phosphorylation, the fragment was cloned into the blunted SacI and PstI sites of pHT128, an IncQ plasmid [33], resulting in pHT1212 (lacI^q-P_tac-pylRS[Y306A/Y384F]/T) and pHT1262 (lacI^q-P_tac-MmpBPaRS/T), respectively (Table 1 and S4 Table).

PCR fragments containing the gst gene derived from pGEX-6P1 (GE Healthcare) with an amber codon at position Glu 51 (gst E51) or Phe 52 (gst F52) were constructed as follows: the gst E51 or gst F52 mutation was introduced by site-directed mutagenesis with two sets of primers (gst E51amb-1 and gst E51amb-2 for gst E51) (gst F52amb-1 and gst F52amb-2 for gst F52) (S1 Table) using the PrimeSTAR Mutagenesis Basal Kit. After the sequence was confirmed, a 0.5-kb PCR fragment containing Ptac-gstE51amb or Ptac-gstF52amb was amplified with a primer set (tetA-1-gst-up and gst-down-tetA-2) (S1 Table) and cloned into the SalI site of pHT1212 or pHT1216 to construct pHT1355, pHT1356, pHT1357, and pHT1358, respectively (Table 1).

To construct a series of amber mutations in the pamA gene, the His₆ tag was initially introduced into the 3’ terminus of the native pamA gene as follows: a 7.2-kb PCR fragment was amplified from pHT922 with a primer set (NMB1345-1 and NMB1345-2-(His)₆) (S1 Table), phosphorylated, and then self-ligated to construct pHT1213 (Table 1).

Amber mutation series (pamA K-amb) plasmids (see S1 Fig) were constructed by site-directed mutagenesis of pHT1213 (S2 Table) with primer sets (S1 Table). The pamA K-amb gene on pTWV228 was confirmed by sequencing and the plasmid is listed in S2 Table. A 2.2-kb PCR fragment containing the pamA K-amb gene was amplified with a primer set (tetA-NMB1345-up and tetA-NMB1345-down) (S1 Table) from the pTWV228-based plasmids listed in S2 Table, and cloned into the SalI site of pHT1216 or pHT1262 by In-Fusion cloning to construct IncQ plasmids containing the lacI^q-P_tac-pylRS[Y306A/Y384F]/T and pamA K-amb genes (pHT1212) or lacI^q-P_tac-MmpBPaRS/T and pamA K-amb genes (pHT1262) listed in Table 1.

The addition of StrepTag₂-His₆ synthetic DNA (S3 Fig) to the pamA K278amb gene was achieved as follows: a 6.5-kb PCR fragment amplified from pHT1242 with a primer set (spacer-1(15mer)-nmb1345-12 and His₆′(15mer)-nmb1345-11) (S1 Table) was ligated to StrepTag₂-His₆ synthetic DNA by In-Fusion cloning to construct pHT1293 (S2 Table). A 2.4-kb PCR fragment containing the pamAK278amb-StrepTag₂-His₆ gene was amplified with a primer set (tetA-NMB1345-up and tetA-NMB1345-down) (S1 Table) from pHT1293 and cloned into the SalI site of pHT1262 by In-Fusion cloning, to construct pHT1388 (Table 1).

The introduction of IncQ plasmids into N. meningitidis strains was performed as described previously [33].

UV crosslinking of whole N. meningitidis cells expressing the PamA Lys-amber (pamA K-amb) mutants by pyrrolysine-based amber suppression with pBPa

The N. meningitidis H44/76 ΔpamA::spc mutant (HT1940) (Table 2) harboring IncQ plasmids containing lacI^q -Ptac-MmpBPaRS/T genes and pamA K-amber genes (Table 1) was cultivated on GC agar plates containing 1 mM IPTG and 1 mM pBPa at 37°C in 5% CO₂ for 18 hours. To identify the optimal pamA K-amb mutants for pyrrolyl-based amber suppression followed by photocrosslinking, N. meningitidis transformants grown one-fourth of a GC plate were suspended in 50 μl PBS, transferred into the wells of a 96-well plate/strain, and irradiated with 365 nm UV using a FL-365-SD UV lamp stand (Opticode, Japan) at a distance of 2 cm on ice for 2 hours. After harvesting, UV-irradiated bacteria were resuspended in 400 μl of urea buffer (50 mM Tris-HCl pH7.5, 8 M urea, 1% Triton X-100, and 100 mM NaCl) and then transferred into a 1.5-mL tube. Bacteria were disrupted by sonication, and the soluble fraction was obtained by centrifugation. PamA K-amb-His₆ proteins with expression suppressed by pBPa were purified with 200 μL TALON resin (Clontech). The purified fraction (approx. 250 μl) was precipitated by TCA/acetone at -80 ^oC overnight, and the precipitate was finally dried and suspended in 25 μl 1 × SDS buffer. Aliquots were analyzed by SDS-PAGE.

Identification of endogenous proteins crosslinked with PamA K278(pBPa) in N. meningitidis

To identify endogenous protein(s) crosslinked to PamA K278(pBPa) in N. meningitidis, an synthetic DNA fragment containing twin Strep-Tag (StrepTag₂) [52] was further introduced into pHT1274, resulting in pHT1388 (S3 Fig and Table 2). N. meningitidis HT1940 harboring pHT1388 was grown on approximately 200 GC agar plates containing 1 mM IPTG and 1 mM pBPa at 37°C in 5% CO₂ for 18 hours. The bacteria from 3 GC agar plates were suspended in 0.5 mL PBS, transferred into the wells of a 12-well plate (more than 50 wells in total), and irradiated with 365 nm UV on ice for 2 hours. The irradiated bacteria from 2 wells were transferred into a 15-mL tube and harvested by centrifugation. The collected bacteria were suspended in 5 mL IP Lysis buffer (25 mM Tris-HCl pH7.5, 150 mM NaCl, 1 mM EDTA, 1% NP-40, and 5% glycerol) and disrupted by sonication followed by centrifugation at 12,000 rpm at 4°C for 30 min. The supernatant (soluble fraction) was collected into one tube. Crosslinked proteins were purified by two column chromatography steps using Ni-Sepharose High Performance (GE Healthcare) and Strep-Tactin Sepharose (Iba, Germany). The purified fraction was concentrated approximately 70-fold using a VIVASPIN TURBO 15 (MW 50k) Centrifuge Filter unit (Sartorius), to a final volume of approximately 100 μL. An aliquot of the sample was analyzed by 4–12% NuPAGE (Invitrogen), and immunoblotting with an anti-His₆ monoclonal antibody (mAb) (Wako, Japan). All proteins were also visualized using the CBB staining kit (APRO SCIENCE, Japan) and bands that were also detected by immunoblotting were excised. These bands were subjected to trypsin digestion and peptide extraction. Extracted peptides were dried and resuspended in H₂O and 0.1% (w/w) trifluoroacetic acid for MS. The LC-MS/MS analysis was performed with an Advance nanoLC (Michrom Biosources) coupled to an LTS Mass Spectrometer (Thermo Fisher Scientific). Proteins were identified by running MASCOT (http://www.matrixscience.com) against the NCBI database (NCBInr 20160202). Candidates with molecular masses that were consistent with the difference between the crosslinked complexes and the PamA protein were selected. The accuracy of each interaction between PamA K278(pBPa) and the candidate was further examined by constructing N. meningitidis HT1940 harboring pHT1388 mutants, in which the FLAG tag was added to the candidate gene on its chromosome, as in the pilX-FLAG N. meningitidis strain (Table 1) described above.

Identification of the counterpart of the amino acid residue of PilE crosslinked with PamA K278(pBPa)

A DNA fragment containing ΔN-pamA K278amb was amplified from pHT1242 with a primer set (NMB1345(pGEX)-1(BH) and NMB1345(pGEX)-2(XhoI)) (S1 Table) and cloned into the same sites of pMAL-c2 (NEB) to construct pHT1473 (S3 Table). A 0.5-kb PCR fragment containing the Gln₆-ΔN-PilE gene, in which the first 9 amino acids (PilE leader peptides in N. meningitidis) were replaced with 6 glutamine residues (Gln₆) (see the Results section), was amplified with a primer set ((Gln)₆-pilE-45 and pCDF-1b(PacI)(15mer)-pilE-44) (S1 Table) from H44/76 N. meningitidis chromosomal DNA, and designated as fragment B. To clone the PCR fragment into pCDF-1b-kan (S3 Table) [53], a 3.6-kb DNA fragment was amplified from pCDF-1b-kan with a primer set (pCDF-1b (downstream of PacI) and (Gln)₆(15mer)-pCDF-1b-2) (S1 Table) and ligated to PCR fragment B by In-Fusion Cloning to construct pHT1447 (S3 Table), in which the ΔN-(Gln)₆-PilE protein was expressed from the T7 promoter (P_T7). A 0.6-kb PCR fragment containing the P_T7- ΔN-(Gln)₆-pilE gene was amplified from pHT1447 with a primer set (pSTV28(SmaI-5′)-pCDF-1b-1 and pSTV28(SmaI-3′)-pCDF-1b-2) (S1 Table) and cloned into the SmaI site of pSTV28 (Takara Bio) by In-Fusion Cloning to construct pHT1474 (S3 Table). E. coli BL21(DE3) cells harboring pCDF-Pyl-Fx3, pHT1473, and pHT1474 were cultivated in 100 mL MagicMedia containing 1 mM pBPa at 30°C overnight. The harvested bacteria were suspended in 1.25 mL PBS and divided into 5 aliquots (0.25 mL/tube), transferred into the wells of a 24-well plate/aliquot, and irradiated with 365 nm UV light on ice for 2 hours. UV-irradiated bacteria were collected again in one tube and disrupted by suspension in 1 mL B-PER (Thermo) containing protease inhibitor cocktail (Nacalai Tesque, Japan), followed by centrifugation at 12,000 rpm at 4°C for 30 min. The MBP-ΔN-PamA K278(pBPa) protein crosslinked to Gln₆-ΔN-PilE [as well as uncrosslinked MBP-ΔN-PamA K278(pBPa)] was purified with amylose resin (NEB). The purified fraction was concentrated approximately 70-fold using a VIVASPIN TURBO 15 (MW 50k) Centrifuge Filter unit (Sartorius) to a final volume of approximately 250 μL. An aliquot of the sample was analyzed by 4–12% NuPAGE (Invitrogen), and the bands corresponding to the complex containing MBP-ΔN-PamA K278(pBPa) crosslinked to Gln₆-ΔN-PilE were excised. In addition, an uncrosslinked band corresponding to MBP-ΔN-PamA K278(pBPa) was isolated for a comparative analysis. The band corresponding to the Gln₆-ΔN-PilE protein was also excised. The amino acid residue of Gln₆-ΔN-PilE crosslinked to ΔN-PamA K278(pBPa) was identified by comparing peptides from the three bands as described previously [54].

Measurement of the ratio of pilus components to PilE in purified pili

Meningococcal pili were prepared as described previously [36]. The relative amounts of PilF-FLAG, PilP-FLAG, PilM-FLAG, PilV, and PilX-FLAG to PilE in purified pili were examined by Western blotting with anti-FLAG mAb (Wako, Japan), anti-PilV [36], or anti-PilE rabbit serum [33] with appropriate dilutions using a quantification program in Fusion Capt 17 in Fusion Solo 7S (M&S Instruments). Results are expressed as means ± SD. The adhered and internalized bacterial numbers and the relative amounts of pilus components to PilE in purified pili were compared using the two-tailed Student’s t-test, and P values of <0.05 were considered significant.

Genetic evidence that a hypothetical protein annotated as NMB1345 is important for N. meningitidis internalization into HBMEC

During the course of our research on the infectious abilities of the ST-2032 N. meningitidis strain mutagenized by STM to HBMEC, we found that a transposon mutant, in which the gene annotated as NMB1345 (renamed as PamA in this study) in the N. meningitidis MC58 genome [39] was disrupted and resulted in defective infectious ability in HBMEC (Fig 1A). To confirm the relationship between the mutation and infectious defect in the mutant, we constructed a null mutant (HT1822 ΔpamA::spc) and an ectopically complemented strain with the wild-type pamA⁺ gene at the ggt locus (HT2224) and examined their infectious activities in HBMEC (Fig 1B). While the insertional and null mutants adhered to HBMEC less efficiently than the wild-type strain HT1125, the number of internalized bacteria largely decreased to approximately 1/100 of HT1125 and the defect in internalized bacterial numbers was recovered in HT2224. These results indicate that the mutation in the pamA gene affects meningococcal infection in HBMEC, particularly internalization.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Characterization of ΔpamA N. meningitidis mutant.

(A) Schematic representation of the wild-type, insertion, and deletion mutants in the pamA gene and ectopic complementation of the pamA⁺ gene at the ggt locus in N. meningitidis strains. (B) Effect of ΔpamA mutation on N. meningitidis infection of HBMEC. The adherence (left) and internalization (right) of pamA N. meningitidis mutants to HBMEC, and the effects of complementation of the pamA⁺ gene in the pamA deletion mutant on bacterial infection. Each value is the mean ± standard error of the mean (CFU per 10⁴ HBMEC) from at least four experiments. Open, filled, light gray, and dark gray bars indicate the bacterial number of N. meningitidis wild-type pamA⁺ (HT1125), pamA::Tn-spc (HT1572), and ΔpamA::spc (HT1822), and pamA^- mutants in which the pamA⁺ gene was ectopically complemented (HT1736 and HT2224), respectively (see S1 Table). *P<0.01, #P<0.05, significantly different from the pamA⁺ strain or pamA^- mutants complemented with the pamA⁺ gene. (C) Immunofluorescence microscopy showing the accumulation of ezrin beneath N. meningitidis-infected HBMEC. The HBMEC monolayer was infected with wild-type pamA⁺ (middle-left), ΔpamA (middle-right), and ΔpamA/pamA⁺ (right) N. meningitidis strains. A non-infected HBMEC monolayer is also shown in the left panels. Bacteria and HBMEC were observed by phase-contrast microscopy (upper panels). Ezrin was immunostained with an anti-ezrin mAb and Alexa Fluor 488-conjugated rabbit anti-mouse IgG (green channel, lower panels).

https://doi.org/10.1371/journal.pone.0237883.g001

We also investigated host cell cytoskeleton rearrangements caused by meningococcal infection by monitoring the localization of ezrin, because the accumulation of ezrin beneath meningococci on HBMEC is required for bacterial internalization into host cells [36, 37] (Fig 1C). While ezrin was widely distributed throughout non-infected cells, it was condensed at the site of bacterial attachment in HBMEC infected with wild-type N. meningitidis strain HT1125. In contrast, ezrin condensation was not observed in cells infected with the ΔpamA N. meningitidis mutant HT1822, but was detected in HT2224 (ΔpamA/pamA⁺). These results are consistent with previous findings, and strongly suggest that PamA participates in the meningococcal-elicited bacteria-induced reorganization of the host cell cytoskeleton upon N. meningitidis infection.

Characterization of the biochemical properties of PamA in N. meningitidis

The deduced amino acid sequence of the PamA protein consists of 516 residues (S1 Fig) with a predicted molecular mass of 57 kDa. The hydrophobicity analysis indicated that the PamA protein had a one membrane-spanning region at its N terminus (S1 Fig), suggesting that the PamA protein is localized at the membrane in N. meningitidis. To clarify the localization in N. meningitidis, cellular proteins were biochemically fractionated using a differential detergent solubilization method [45] and analyzed by Western blotting (Fig 2A). LptA, a meningococcal inner membrane (IM) protein [55], was largely detected in the 4% Triton X-100 soluble (IM-enriched) fraction, supporting the proper fractionation in this experiment. Under the same conditions, the PamA protein was mainly enriched in the IM-rich fraction. This result suggests that PamA is located at the IM in N. meningitidis.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Characterization of the biochemical properties of the PamA protein in N. meningitidis.

(A) Western blot analyses of subcellular fractions of the PamA protein in N. meningitidis strain HT1125. A whole-cell lysate periplasmic fraction (P) isolated by CHCl₃ shock, cytoplasmic and periplasmic fraction (CP), IM-enriched fraction, and outer membrane (OM)-enriched fraction were prepared by the sonication method from the N. meningitidis strain, as described previously [45]. Western blotting with anti-PamA (left) and anti-LptA (right) [55] rabbit antisera. LptA is a control for the proper isolation of the IM-rich fraction. Black and gray arrows show PamA and LptA, respectively. (B) Membrane topology of the PamA protein using the pamA′-lacZ and pamA′-phoA fused N. meningitidis strain. The hatched region consists of the 20 amino acid residues corresponding to the putative membrane-spanning region (S1 Fig). The first 408-bp region of the pamA gene was fused to the lacZ or phoA gene, and the gene was incorporated into the pamA locus on the meningococcal chromosome. (C) β-Galactosidase and alkaline phosphatase activities. (D,E) Western blot of spheroplasted N. meningitidis strain H44/76 harboring pHT936 (an IncQ plasmid expressing the gst⁺ gene, Table 1) followed by a treatment with Proteinase K. Aliquots of the sample were analyzed by Western blotting with anti-GST mAb (D) and anti-PamA rabbit serum (E). Black and gray arrows show PamA and GST, respectively.

https://doi.org/10.1371/journal.pone.0237883.g002

We also investigated the membrane topology of the PamA molecule in N. meningitidis by constructing N. meningitidis mutants, in which the first N-terminal 136 amino acid residues of PamA including the membrane-spanning region, termed pamA′, were fused to the lacZ or phoA gene [56] on the chromosome (Fig 2B)(see Materials and Methods). While the wild-type HT1125 and pamA′-lacZ N. meningitidis HT2215 strains (Table 2) did not exhibit β-galactosidase activity, the pamA′-phoA N. meningitidis strain HT2217 (Table 2) showed alkaline phosphatase activity (Fig 2C). Since the alkaline phosphatase activity of PhoA was detected only in the periplasmic space [57], these genetic results suggest that the hydrophobic region of the PamA protein faces the periplasmic side. We also biochemically examined the membrane topology, using a Proteinase K digestion experiment with the spheroplasted N. meningitidis strain expressing glutathione S-transferase (GST) from an IncQ plasmid (Fig 2D and 2E). Under the condition where the cellular protein GST was mostly protected from Proteinase K digestion (Fig 2D), PamA was largely digested (Fig 2E). Collectively, these results suggest that PamA is localized at the IM and faces the periplasmic space in N. meningitidis.

Incorporation of ncAAs into proteins in living N. meningitidis

To attempt the application of genetic code expansion to N. meningitidis for the incorporation of ncAAs with a photocrosslinker into the PamA protein, we used two mutated MmPylRS/T genes to genetically encode ncAAs in N. meningitidis. A previous study reported that a modified MmPylRS gene for E. coli, MmPylRS[Y306A/Y384F] [49] is applicable to N^ε-(m-azidobenzyloxycarbonyl)-L-lysine (mAzZLys) (Fig 3A, left), which is a photocrosslinker excited by ultraviolet (UV) irradiation [58], and MmPylRS([A302T/N346T/C348T/W417C]) [50] (MmpBPaRS) is optimized for p-benzoyl-L-phenylalanine (pBPa) (Fig 3A, right), which is also one of the widely used photocrosslinkers [44].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Incorporation of ncAAs into N. meningitidis proteins monitored by Western blotting.

(A) Chemical structures of ncAAs, mAzZLys (Left) or pBPa (right) used in this study. (B) Estimation of the efficiency of pyrrolysine-based amber suppression with mAzZLys or pBPa for GST E51amb or GST F52amb in N. meningitidis. Bacterial extracts equivalent to an OD₆₀₀ of 0.1 were analyzed by Western blotting with an anti-GST mAb. The horizontal black arrow shows the full-length GST protein. Numbers indicate relative suppression efficiency when wild-type GST expression is defined as 100%. (C) Detection of irradiation- and time-dependent GST dimerization of GST F52 (pBPa) in N. meningitidis. Bacterial extracts equivalent to an OD₆₀₀ of 0.05, irradiated with UV light for the indicated times, were analyzed by Western blotting with an anti-GST mAb. Dark gray and black arrows show the monomer and dimer forms of the GST protein, respectively. The light gray arrow shows an unspecified form of GST generated by UV irradiation.

https://doi.org/10.1371/journal.pone.0237883.g003

We examined the incorporation of mAzZLys and pBPa in N. meningitidis by biochemically monitoring the suppression of the amber codon at position Glu51 or Phe52 of GST (GST E51amb or GST F52amb) by Western blotting (Fig 3B). The expression of both the gst E51amb and gst F52amb mutants was suppressed by pBPa, while the gst F52amb mutant was not suppressed by mAzZLys. The suppression efficiency could be considered as approximately 25% of the wild-type GST protein with pBPa in the gst E51amb mutant, and 37% and 35% with mAzZLys and pBPa in the gst F52amb mutant, respectively. Since the incorporation efficiency of ncAAs is approximately 35–50% in E. coli [59], this result suggested that genetic code expansion by the two modified MmPylRS/T genes was efficiently conducted in N. meningitidis (Fig 3B). These results obtained by biochemical and genetic pyrrolysine-based amber suppression suggest that mAzZLys and pBPa are incorporated into the GST protein in N. meningitidis. The proper incorporation of pBPa into a protein in N. meningitidis, as determined by an MS analysis, was examined with PamA K278 (pBPa)(see a later section, S5 Fig) since the amounts of GST proteins in N. meningitidis were too low to purify.

We further examined the UV crosslinking by using whole bacterial cells (Fig 3C). Longer UV irradiation increased the amount of the dimer form of GST, and the majority of GST F52(pBPa) was fixed as the dimer form for 2 hours because GST is a dimeric enzyme [44, 60, 61]. On the other hand, the dimerization of the GST E51amb(mAzZLys) and GST E51(pBPa) proteins was not observed by photocrosslinking, while GST E51(mTmdZLys) dimerized in E. coli [62], indicating that the processes involved in UV crosslinking with ncAAs do not always function well in N. meningitidis. Collectively, these results demonstrated that physiological transient protein-protein interactions can be detected by crosslinking via the incorporation of photoreactive ncAAs in N. meningitidis and showed that the incorporation of ncAAs by pyrrolysine-based amber suppression could be performed in N. meningitidis.

Incorporation of a genetically encoded photocrosslinking amino acid into the PamA protein to identify the accompanying protein in N. meningitidis

Since it was not possible to predict the three-dimensional structure of PamA by Phyre2 [63], due to the lack of homology to any other protein identified to date, we were unable to speculate which amino acid residues are exposed on the outside of the PamA molecule in N. meningitidis. Therefore, the identities of suitable amino acid residues for the incorporation of photoreactive ncAAs into the PamA protein remained unclear. Thus, in the present study, we focused on the lysine (Lys or K) residue for amber (UAG) substitution because lysine is primarily located on protein surfaces [64, 65]. Forty Lys residues are present in the PamA protein (S1 Fig), and we constructed three pamA amber mutants (designated as pamA K-amb) at K148 (pamA K148amb), K273 (pamA K273amb), and K388 (pamA K388amb) to estimate the proper position(s) in the PamA protein for ncAAs incorporation. N. meningitidis strain H44/76 ΔpamA (HT1940) expressing the MmPylRS[Y306A/Y384F]/T genes and one of the three pamA K-amb genes on an IncQ plasmid was cultivated in the presence of mAzZLys, and whole bacterial extracts were analyzed by Western blotting (S2A Fig). The full length and a sufficient amount of PamA K148amb by incorporation of mAzZLys [(PamA K148(mAzZLys)], PamA K273(mAzZLys), and PamA K388(mAzZLys) were only observed in the presence of mAzZLys, suggesting that mAzZLys is incorporated into the PamA protein by pyrrolysine-based amber suppression in N. meningitidis. We also examined UV-irradiated whole bacterial cells by Western blotting, faint bands appeared to correspond to proteins crosslinked to PamA (mAzZLys), were detected in the PamA K273(mAzZLys) and PamA K388(mAzZLys) samples (S2B Fig). The same result was also obtained for N. meningitidis strain HT1940 expressing the MmpBPaRS/T genes and one of the three pamA K-amb genes on an IncQ plasmid, cultivated in the presence of pBPa (S3 Fig). These results indicate that the Lys residues located from position Lys148 to the carboxy-terminal end (S1 Fig) were localized on the surface of the PamA molecule in N. meningitidis. Moreover, the amber suppression appeared to be more efficiently for PamA K273(pBPa) and PamA K388(pBPa) with pBPa than with mAzZLys (Fig 4 and S3 Fig). Thus, the photocrosslinking experiments by pyrrolysine-based amber suppression described hereafter were conducted with the pBPa and MmpBPaRS/T genes.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. PamA K208(pBPa), PamA K278(pBPa), PamA K314(pBPa), PamA K382(pBPa) and PamA K395(pBPa) proteins crosslinked to unidentified endogenous proteins in N. meningitidis.

UV crosslinking of Pam K208amb, K278amb, K309amb, K341amb, K395amb, and K382amb mutants expressed by pyrrolysine-based amber suppression with pBPa in ΔpamA N. meningitidis. Bacteria that grew on one-fourth of a GC agar plate in the presence of 1 mM pBPa were crosslinked by UV light irradiation for 2 hours and then the PamA K(pBPa)-His₆ protein was purified. An aliquot was analyzed by anti-His₆ mAb. The black arrow shows the full-length PamA K-amb protein expressed by pyrrolysine-based amber suppression with pBPa, and the gray arrow indicates protein complexes crosslinked between PamA K(pBPa) and unidentified endogenous proteins in N. meningitidis.

https://doi.org/10.1371/journal.pone.0237883.g004

To confirm the proper position(s) for the incorporation of pBPa into the PamA protein, we constructed 26 pamA K-amber mutants, in which each Lys residue located from position Lys148 to the carboxy-terminal end in the PamA protein was replaced with the amber codon (S1 Fig, Table 1 and S4 Table). UV-irradiated N. meningitidis strain HT1940 cells expressing the MmpBPaRS/T and pamA K-amber-His₆ genes were analyzed by Western blotting (S3 Fig). Some PamA(pBPa) mutants seemed to be crosslinked to some endogenous protein in N. meningitidis (shown in red in S3 Fig), and the more detailed analysis revealed that the following six pamA K-amber mutants with expression suppressed by pBPa [K208(pBPa), K278(pBPa), K309(pBPa), K341(pBPa), K382(pBPa), and K395(pBPa)] appeared to be crosslinked to some endogenous proteins in N. meningitidis (shown in red in Fig 4) while K309(pBPa) was eliminated. In the present study, we focused only on the proteins crosslinked to PamA K278(pBPa), because the three bands with the highest intensities were found in the PamA K278(pBPa) sample.

To characterize the 150-, 140-, and 85-kDa protein complexes that contained an endogenous protein crosslinked to PamA K278(pBPa), we added a Twin-Strep tag (Strep₂) [66] and His₆ tag at the carboxy-terminal end (S4 Fig) and purified the three corresponding protein complexes from N. meningitidis strain HT1940 harboring pHT1388 grown on approximately 200 plates as described in the Materials and Methods. During the purification process, we also obtained the purified PamA K278 (pBPa) protein expressed in N. meningitidis. The incorporation of pBPa at K278amb of PamA-Strep₂ -His₆, was confirmed by MS analyses (S5 Fig). This result demonstrates that an ncAA was selectively incorporated into a protein in N. meningitidis.

The three corresponding protein complexes were subjected to a gel-based proteomic analysis using MS (Fig 5A and 5B). A compilation of the results obtained using MASCOT (MATRIX SCIENCE, Japan) suggested several candidates, and we selected those with higher scores and numbers of matched peptides for more reliable results in the present study (Table 3). Since the molecular mass of PamA-Strep₂-His₆ is approximately 62 kDa due to the addition of Strep₂-His₆ at the carboxy-terminal end (S4 Fig), the bands corresponding to 150- and 140-kDa were presumed to contain counterparts with a molecular mass of approximately 100 kDa. PamA is listed as the top score of the MASTCOT analysis (Table 3) for the 150 and 140 kDa bands, providing support for the presence of this protein in the two bands. On the other hand, we did not identify any candidate with a deduced molecular mass from 80 to 90 kDa by MASCOT (Table 3). To examine the possibility that the observed molecular mass on SDS-PAGE was different from the calculated molecular mass in the database, we constructed N. meningitidis mutants that expressed the FLAG-tagged proteins of 5 candidates with estimated molecular masses of 75–160 kDa for 150 kDa, and those of 9 candidates with estimated molecular masses of 48–74 kDa (shown in bold in Table 3). Crosslinked samples from the N. meningitidis mutants expressing the FLAG-tagged protein, MmpBPaRS/T and PamA K278(pBPa)-Strep₂-His₆, were examined by Western blotting; however, none of these candidates were crosslinked to PamA K278(pBPa) in N. meningitidis. Considering these results, the two proteins presented by the 150- and 140-kDa bands with the highest intensities on Western blotting apparently correspond to the homo-trimer of the PamA protein (see Discussion and S12 Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Purification and identification of an endogenous protein crosslinked to PamA K278(pBPa)-Strep₂-His₆.

Proteins in each purification step were analyzed by 4–12% SDS-PAGE and staining with the CBB staining kit (A) and Western blotting with anti-His₆ mAb (B). Lane 1, molecular mass standards; lane 2, N. meningitidis crude soluble extract; lane 3, after the Ni-Sepharose column; lane 4, after dialysis with IP Lysis buffer; lane 5, after the Strep-Tactin Sepharose column; lane 6, after concentration with VIVASPIN TURBO 15. Black and red arrows indicate K278(pBPa)-Strep₂-His₆ and the unidentified endogenous proteins crosslinked to PamA K278(pBPa)-Strep₂-His₆, respectively. (C) The PilE-FLAG protein was crosslinked to PamA K278(pBPa) in N. meningitidis. The black arrow shows the full-length PamA K278amb protein expressed by pyrrolysine-based amber suppression with pBPa, and the gray arrow indicates the homotrimer (See text) in N. meningitidis. The red arrow shows the complex of PilE-FLAG crosslinked to PamA K278(pBPa)-Strep₂-His₆ in N. meningitidis. +/- indicates the presence or absence of pBPa, and irradiation or no irradiation with UV light, respectively. (D) The crosslinked band was reacted with anti-PilE rabbit serum. Black and gray arrows show PamA K278(pBPa)-Strep₂-His₆ and the putative homotrimer complex of PamA K278(pBPa), respectively, in N. meningitidis. The red arrow shows the complex of PilE crosslinked with PamA K278(pBPa)-Strep₂-His₆ in N. meningitidis. +/- indicates the presence or absence of the pilE gene, and irradiation or no irradiation with UV light, respectively.

https://doi.org/10.1371/journal.pone.0237883.g005

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Summary of the protein candidates crosslinked to PamA in N. meningitidis identified by mass spectrometry.

https://doi.org/10.1371/journal.pone.0237883.t003

We then characterized the 85-kDa band, which could contain an approximately 20 kDa endogenous protein crosslinked to PamA K278(pBPa). LC-MS/MS followed by MASCOT analyses revealed two possible candidates, a fimbrial protein (17 kDa) and the hypothetical protein NMEN93004_1215 (18 kDa) (Table 3). The BLAST homology search indicated that while we could not find any protein corresponding to NMEN93004_1215 in the databases, the fimbrial protein (accession number gi|496712676) was 95% identical to the protein NMBH4476_0018 annotated as “type IV pilus assembly protein PilA” in N. meningitidis strain H44/76 with the highest E-value (Table 3). Further computational analyses revealed that the “PilA protein” annotated in GenBank was 96% identical to PilE, a major component of meningococcal pili (reviewed in [17]). To confirm that the 85-kDa complex was composed of PilE and K278(pBPa)-Strep₂-His₆, we constructed the pilE-FLAG N. meningitidis strain HT2095 and examined the crosslinked samples by Western blotting with an anti-FLAG mAb (Fig 5C). The analysis revealed that PilE-FLAG was present in the 85-kDa band, with dependence on pBPa and UV irradiation (Fig 5C). We further examined crosslinked samples from the wild-type (HT1940 harboring pHT1388) and pilE-insertional mutant (HT2014 harboring pHT1388) by Western blotting with anti-PilE rabbit serum [33] (Fig 5D). The 85-kDa band was also observed with anti-PilE rabbit serum in the pilE⁺ genetic background only with pBPa and UV irradiation. Considering these results, we conclude that the 85-kDa band contains the PilE protein, indicating that PamA interacts with PilE in N. meningitidis.

Mapping the amino acid residue in PilE that crosslinks to the PamA protein

To further clarify the interaction between PamA and PilE, we attempted to identify the amino acid residue of PilE that crosslinks with PamA K278(pBPa). Since mapping was difficult and tedious [42], particularly for the endogenous proteins, the examination was conducted with recombinant proteins in E. coli.

We examined the interaction of two truncated recombinants, ΔN-PamA K278amb fused to maltose-binding protein (MBP) (MBP-ΔN-PamA K278amb) and ΔN-PilE with six glutamines (Gln₆) in E. coli (S6 Fig). The crosslinked complex between MBP-ΔN-PamA K278(pBPa) and Gln₆-ΔN-PilE was purified from E. coli and investigated using a gel-based proteomic analysis with LC-MS/MS (S7–S9 Figs) [42]. The peptides containing pBPa (IEVGK*[pBPa]LAFSTK), corresponding to aa 639 to 650 of MBP-ΔN-PamA K278(pBPa) (corresponding to aa 274 to 284 of native PamA), from crosslinked and uncrosslinked samples were analyzed by MS/MS. As shown in S9 Fig, the sequence was read from the annotated b and y ions of a crosslinked peptide until isoleucine (Ile) at position 19. This result indicated that the region around Ile19 in Gln₆-ΔN -PilE is the site crosslinked to K278(pBPa) in MBP-ΔN-PamA.

Genetic elucidation of the interaction between PamA and PilE in N. meningitidis and its important for meningococcal internalization into HBMEC

Ile at position 19 in Gln₆-ΔN-PilE corresponded to position 12 in native PilE in N. meningitidis (S10 Fig) because prepilin is processed by the prepilin peptidase PilD during translocation at the IM in N. meningitidis [36, 67]. Multiple alignments showed that the N-terminal region of PilE around Ile12 is highly conserved among neisserial species, including N. gonorrhoeae and N. lactamica, implicating the importance of the region in PilE (S10A Fig) [68].

To further examine the functional interaction between PamA and the region containing the PilE I12 residue in N. meningitidis, the meningococcal pilE I12A-cat mutant HT2167 was constructed (Table 1), and we monitored its infectious ability and the accompanying host cell’s cytoskeleton rearrangement by the localization of ezrin. The pilE I12A N. meningitidis mutant was less efficiently internalized into HBMEC, while the adhesion remained unchanged (Fig 6A), and ezrin did not accumulate beneath the pilE I12A-cat mutant (Fig 6B). Western blotting with UV-irradiated samples purified from HT2167 expressing the MmpBPaRS/T and pamA K278amb genes also confirmed that the interaction between PilE and PamA K278(pBPa) is markedly reduced by the pilE I12A mutation (Fig 6C), providing genetic evidence for the interaction between PamA and PilE in N. meningitidis. Furthermore, since the phenotypes appeared to be mostly identical to that of the ΔpamA N. meningitidis mutant (Figs 1 and 6A), these results indicated that PamA and PilE interact functionally with each other in N. meningitidis and that this interaction is important for meningococcal infection in HBMEC.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. The pilE I12A N. meningitidis mutant also showed reduced internalization into HBMEC, the induction of cytoskeleton rearrangements, and interactions with PamA in N. meningitidis.

(A) A I12A mutation in the pilE gene affects N. meningitidis internalization into HBMEC. Adherence (white) and internalization (gray) of N. meningitidis strains in HBMEC. Each value is the mean ± standard error of the mean (CFU per 10⁴ HBMEC) from at least four experiments. The bars indicate the bacterial numbers of N. meningitidis wild-type pamA⁺ (HT1125, left), ΔpamA::spc (HT1822, middle-left), pilE⁺-cat (HT1744 middle-right), and pilE I12A-cat (HT2167, right), respectively (see S1 Table). *P<0.01, **P<0.02, significantly different from the pamA⁺ strain or pilE⁺-cat strain. (B) Immunofluorescence microscopy showing the accumulation of ezrin beneath N. meningitidis strains. The HBMEC monolayer was infected with wild-type pilE⁺-cat (middle) and pilE I12A-cat (right) N. meningitidis strains. A non-infected HBMEC monolayer is also shown in the left panels. Bacteria and HBMEC were observed by phase-contrast microscopy (upper panels). Ezrin was immunostained with anti-ezrin mAb and Alexa Fluor 488-conjugated rabbit anti-mouse IgG (green channel, lower panels). (C) PilE I12A was crosslinked to PamA K278(pBPa) less efficiently than wild-type PilE in N. meningitidis. Black and red arrows show the crosslinked complex between PamA K278(pBPa) and PilE, and the gray arrow indicates an unidentified band that crossreacted with anti-PilE rabbit serum. +/- indicates the presence or absence of the pilE gene, respectively.

https://doi.org/10.1371/journal.pone.0237883.g006

Characterization of pili in the ΔpamA N. meningitidis mutant

The results described above suggested that PamA affects the functions of meningococcal pili, resulting in a negative impact on meningococcal infection in HBMEC. To further clarify the effects of PamA on pili, we investigated the function of pili in N. meningitidis. Electron microscopy with negative staining showed that the ΔpamA N. meningitidis mutant was similarly piliated to the wild-type strain HT1125 (S11A Fig, left and middle-left). Meningococcal pili are known to play a role not only in the infection of host cells [17, 69], but also in meningococcal natural competence, switching motility, and aggregation [70, 71]. An examination of pilus function in natural competence and aggregation revealed that the pili of the ΔpamA N. meningitidis mutant retain similar functions to those of the wild-type pilE N. meningitidis strain (S11B and S11C Fig). These results suggest that PamA does not participate in pilus formation itself or in functions related to natural competence and aggregation. It is important to note that the pilE I12A-cat N. meningitidis mutant is also as functional as the wild-type pilE-cat N. meningitidis strain while its piliation number its apparently smaller possibly due to translational fusion with the cat gene (S11 Fig and Fig 6).

We examined the effects of PamA on meningococcal pili in more detail. Although the main component of the meningococcal pilus is PilE, its formation requires more than 20 proteins [72–74]. To investigate the effects of the pamA mutation on pili, we initially assessed the pilus protein levels in HT1125 and HT1822 using LC-MS/MS with Tandem Mass Tag (TMT) labeling [38] (S5 Table), and found that the amounts of some pilus components appeared to differ between HT1125 and HT1822. However, accurate comparisons of the protein amounts, particularly that for PilE, by TMT are methodologically difficult because N. meningitidis has 8 non-expressed pilS genes [75], according to the deduced amino acids recorded as Protein in the database. Therefore, we directly monitored the contents of pilus components, particularly for three minor pilins: PilX for aggregation in and adherence to human cultured cells [71, 76, 77], PilV to trigger plasma membrane reorganization [36, 71, 78], and ComP for natural competence [79]. PilF, PilM, and PilP, located on the meningococcal inner membrane for pilus formation [8, 67, 69, 80, 81], were also monitored. The relative amounts of the components to the PilE subunit in purified pili were measured by Western blotting followed by quantification (Fig 7). The relative ratios of PilF, PilP, and PilM to PilE did not significantly differ between the ΔpamA and wild-type N. meningitidis strains. On the other hand, in ΔpamA pili the relative amount of PilV increased, whereas that of PilX decreased (Fig 7A). It is important to note that the amount of ComP was too low in the whole cell extract or purified pili to detect by WB, even when the comP-FLAG N. meningitidis strain or anti-ComP rabbit antiserum was used (data not shown). We also confirmed that the expressions of PilE, PilV, and PilX-FLAG were not affected by the ΔpamA mutation (Fig 7B). Considering these results, the pamA mutation may cause unusual proportions of PilV and PilX to PilE in meningococcal pili, probably by improper sorting of PilE during pili formation, which might affect meningococcal infection in human cultured cells.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Disruption of the interaction between PamA and PilE leads to abnormal proportions of PilV and PilX relative to PilE in meningococcal pili.

(A) Ratio of pilus components to PilE in purified pili. The relative amounts of PilF-FLAG (HT2218 and HT2219), PilP-FLAG (HT2136 and HT2137), PilM-HA (HT2132 and HT2133), PilV (HT1125 and HT1822), and PilX-FLAG (HT2211 and HT2212) to PilE in purified pili were measured by Western blotting with appropriately diluted samples followed by quantification of the band intensity using the 30-kDa band in the molecular marker as an internal golden standard. The values obtained were divided by the value of PilE in the same sample. The significance of differences was examined by the t-test. (B, C, D) The pamA mutation does not affect the expression of the PilE (B), PilV (C), or PilX-FLAG (D) protein in N. meningitidis. Bacterial extracts equivalent to OD₆₀₀ value of 0.0025 (for PilE) or 0.05 (for PilV and PilX-FLAG) were analyzed by Western blotting with anti-PilE and anti-PilV rabbit sera, and an anti-FLAG mAb, respectively. Black arrows indicate PilE, PilV, and PilX-FLAG, respectively.

https://doi.org/10.1371/journal.pone.0237883.g007