Neisseria meningitidis Translation Elongation Factor P and Its Active-Site Arginine Residue Are Essential for Cell Viability

Translation elongation factor P (EF-P), a ubiquitous protein over the entire range of bacterial species, rescues ribosomal stalling at consecutive prolines in proteins. In Escherichia coli and Salmonella enterica, the post-translational β-lysyl modification of Lys34 of EF-P is important for the EF-P activity. The β-lysyl EF-P modification pathway is conserved among only 26–28% of bacteria. Recently, it was found that the Shewanella oneidensis and Pseudomonas aeruginosa EF-P proteins, containing an Arg residue at position 32, are modified with rhamnose, which is a novel post-translational modification. In these bacteria, EF-P and its Arg modification are both dispensable for cell viability, similar to the E. coli and S. enterica EF-P proteins and their Lys34 modification. However, in the present study, we found that EF-P and Arg32 are essential for the viability of the human pathogen, Neisseria meningitidis. We therefore analyzed the modification of Arg32 in the N. meningitidis EF-P protein, and identified the same rhamnosyl modification as in the S. oneidensis and P. aeruginosa EF-P proteins. N. meningitidis also has the orthologue of the rhamnosyl modification enzyme (EarP) from S. oneidensis and P. aeruginosa. Therefore, EarP should be a promising target for antibacterial drug development specifically against N. meningitidis. The pair of genes encoding N. meningitidis EF-P and EarP suppressed the slow-growth phenotype of the EF-P-deficient mutant of E. coli, indicating that the activity of N. meningitidis rhamnosyl–EF-P for rescuing the stalled ribosomes at proline stretches is similar to that of E. coli β-lysyl–EF-P. The possible reasons for the unique requirement of rhamnosyl–EF-P for N. meningitidis cells are that more proline stretch-containing proteins are essential and/or the basal ribosomal activity to synthesize proline stretch-containing proteins in the absence of EF-P is lower in this bacterium than in others.


Introduction
The ribosome connects amino acids together to synthesize a protein in the order specified by the mRNA sequence.During this translation process, multiple proline stretches with two or more consecutive prolines in the amino acid sequence retard peptide bond formation [1] and cause ribosome stalling [2].Translation elongation factor P (EF-P) alleviates ribosome stalling at proline stretches [3,4,5,6,7,8,9,10], by binding between the peptidyl (P) site and the tRNA exit (E) site of the ribosome [11,12].EF-P was discovered as a protein that stimulates the ribosomal peptidyltransferase activity [13,14,15], and is almost universally conserved among bacteria [16].
EF-P is composed of domains 1, 2, and 3, and the overall structure assumes an L shape, which mimics that of tRNA [18,27,28].In contrast, EpmA is a paralogue of lysyl-tRNA synthetase (LysRS).Therefore, the β-lysyl modification of EF-P Lys34 by EpmA may be regarded as molecular mimicry, in that Lys34 corresponds to the 3 0 -end adenosine (A76) of tRNA, and the mechanisms of substrate recognition and aminoacylation catalysis by EpmA resemble those of an aminoacyl-tRNA synthetase [18,29].The (R)-β-lysyl group of the post-translationally modified Lys34, at the tip of the L-shaped EF-P, may contact the ribosomal peptidyltransferase center [12].
Lys34 is conserved in the EF-Ps among about 80% of bacteria.However, the β-lysyl modification enzymes EpmA and EpmB are conserved in only 26-28% of bacteria, including E. coli and its phylogenetically related γ-proteobacteria (e.g., Enterobacter aerogenes, Salmonella enterica, Vibrio cholerae, Shigella flexneri, Haemophilus influenzae, and Yersinia pestis).Therefore, the other types of post-translational lysine modifications, if any, of the EF-Ps from a large number of bacteria are still not known.In this regard, initiation factor 5A (e/aIF5A) is the eukaryotic/archaeal orthologue of EF-P, and its highly-conserved lysine residue (Lys50) is posttranslationally modified to hypusine by deoxyhypusine synthase (DHS) and deoxyhypusine hydroxylase (DOHH) [30,31,32].Moreover, the deoxyhypusine modification is important for translation elongation and restoration of stalled ribosomes [33,34].
In bacteria, including E. coli, the EF-P deletion is not lethal [35,36,37], although the genome encodes over one thousand proline stretches in its proteins.This is probably because even a protein containing a strong pausing consecutive proline sequence, such as PPPP, can be expressed at the basal level: i.e., at several percent relative to a protein with no proline stretch [9].Similarly, the deficiency in the enzyme activity for the known Lys or Arg modification is not lethal.Consequently, EF-P and its Lys/Arg modification are important, but not essential, to alleviate ribosome stalling at proline stretches.
In the present study, we studied the EF-P protein from Neisseria meningitidis, a member of the β-proteobacteria and a leading cause of bacterial meningitis and septicemia worldwide, which is therefore a potential target for drug development.N. meningitidis has EF-P containing Arg32 and its putative modification enzyme.Remarkably, the number of proline stretches encoded in the N. meningitidis genome is much smaller than those in the genomes of other bacteria, including E. coli, S. oneidensis and P. aeruginosa.However, it was unknown whether N. meningitidis EF-P, or EF-P(Nm), is essential for viability, and whether EF-P(Nm) is posttranslationally modified in the same manner as in S. oneidensis and P. aeruginosa.
Therefore, we analyzed the modification of Arg32 in EF-P(Nm), and identified it as the same rhamnosylation as those of the S. oneidensis and P. aeruginosa EF-P proteins.We successfully deleted the N. meningitidis gene encoding the EF-P rhamnosyl modification enzyme, EarP.However, our attempt to disrupt the N. meningitidis gene encoding EF-P failed, indicating that EF-P is essential for cell viability.We confirmed that, in contrast to most bacteria, both EF-P(Nm) and Arg32 are crucial for the viability of N. meningitidis.

N. meningitidis EF-P is essential for cell viability
We first tried to disrupt the efp gene, encoding EF-P, in the N. meningitidis genome, but could not obtain any erythromycin-resistant (Erm r ) colonies with the Δefp::ermC allele (data not shown).This result suggested that the efp gene is essential for N. meningitidis viability.To further examine this possibility, N. meningitidis cells with the endogenous efp gene in the chromosome were transformed with pHT261 (S1 Table ), derived from the broad-host-range IncQ plasmid and harboring a second efp gene, which is designated hereafter as pHT969 (Fig 1A).These meningococcal transformants were further transformed with a PCR fragment containing the efp-flanking regions and the ermC gene, in order to disrupt the efp gene in the chromosome.Numerous colonies of the erythromycin-resistant mutant were obtained for N. meningitidis cells harboring pHT969 (the wild-type efp-containing plasmid).In contrast, for N. meningitidis cells harboring pHT261 (the empty vector plasmid), very few colonies of the erythromycin-resistant mutant(s) were obtained, and they lacked the ermC gene in the efp locus.These results indicated that the N. meningitidis efp gene is essential for cell viability (Table 1).
In parallel, we performed a complementary experiment to assess whether the efp gene is actually essential for N. meningitidis viability.First, N. meningitidis cells were transformed with the IncQ plasmid pHT1139 (S1 Table ), containing an IPTG-inducible copy of the efp gene under the control of the tac promoter.Then, under conditions with the induced expression of the efp gene, we deleted the efp gene from the N. meningitidis H44/76 genome, by integrating an erythromycin resistance gene (ermC) or the efp gene with the Arg32 codon replaced by an opal (TGA) stop codon.The growth characteristics of the N. meningitidis cells containing the inducible efp gene, with and without the inducer, are shown in Fig 1B and 1C, respectively.Without IPTG, the N. meningitidis HT1913/pHT1139 and HT1914/pHT1139 cells barely grew, and the very small number of colonies should be ascribed to the leaky expression of the In addition, meningococcal transformants with a plasmid harboring the efp(R32K) or efp (R32A) mutant, with Arg32 replaced by either Lys32 or Ala32, respectively, were also constructed.As a result, almost no colonies of the efp-null mutant were obtained in the presence of the plasmid harboring the efp(R32K) or efp(R32A) mutant gene (Table 1).Thus Arg32 is indispensable for the EF-P activity in N. meningitidis.Conversely, the efp gene disruption is not lethal in other bacteria, such as E. coli MG1655 [37], E. coli W3110 [38], S. enterica [17], Agrobacterium tumefaciens [24], P. aeruginosa [35,36,39], and Bacillus subtilis [40].This is the first report that the EF-P function is essential for cell viability.

N. meningitidis EF-P is post-translationally modified at Arg32
With the finding that EF-P is essential for N. meningitidis viability, we next attempted to examine whether there is any difference in the post-translational modification of the EF-P proteins between N. meningitidis and S. oneidensis/P.aeruginosa.First, we analyzed whether EF-P(Nm) is post-translationally modified.As the N. meningitidis cells can only be cultured on solid media, we prepared 100 plates to harvest a sufficient amount of the N. meningitidis HT1125 cells.The endogenous EF-P was purified from the cells in three column chromatography steps, as described in the "Materials and methods" (Fig 2A and S1 Fig).The molecular mass of the endogenous EF-P(Nm) was estimated to be 30 kDa by SDS-PAGE (Fig 2A and 2B), and was slightly larger than that of the EF-P from E. coli (EF-P(Ec)) [18,41].MALDI-TOF MS and ESI-MS analyses indicated that the molecular masses of the endogenous EF-P(Nm) were 21,034.74Da (Fig 2C) and 21,040.00Da (data not shown), respectively, which are higher by 147-153 Da than that of the recombinant EF-P(Nm) (obsd: 20,887.39,calcd: 20,879.75)(Fig 2D, S2 Table ).A peptide mass fingerprinting (PMF) analysis of the endogenous EF-P(Nm) was performed to identify the peptide segment with the post-translational modification (Fig 3A and 3B).The endogenous EF-P(Nm) generated peptides with the molecular masses of 808.51 Da (with AspN and trypsin digestion) and 808.53 Da (with AspN and API digestion), which are +146 Da higher than that of the recombinant EF-P(Nm) peptide "GGRSSAK" (calcd: 662.36 [M+H] + , obsd: 662.45 [M+H] + ) (Fig 3B).The endogenous EF-P(Nm) peptide was introduced into the efp allele within the N. meningitidis H44/76/pHT1139 genome, to obtain the N. meningitidis strains HT1913/pHT1139 (B, left) and HT1914/pHT1139 (C, left), respectively.In these strains, the efp gene expression can be controlled by IPTG, and EF-P can be inducibly produced in the presence of IPTG.(B, C, right) Growth of the N. meningitidis HT1913/pHT1139 and HT1914/pHT1139 cells, with and without IPTG.Both of the N. meningitidis cells lack the efp gene in the genome, but contain an inducible copy of the efp gene in the IncQ plasmid.doi:10.1371/journal.pone.0147907.g001The post-translational modification at Arg32 of N. meningitidis EF-P is rhamnosylation Arg32 was apparently modified with a monosaccharide, because the modified arginine decomposed during MALDI-TOF MS using α-cyano-4-hydroxycinnamic acid (CHCA) as the matrix.
The +146 Da monosaccharide may be either a rhamnosyl or fucosyl modification [35].S. oneidensis and P. aeruginosa only have the biosynthesis pathway for dTDP-rhamnose, and lack that for GDP-fucose, and their EF-P proteins are ramnosylated [35,36].However, N. meningitidis might have the fucose pathway in addition to the rhamnose pathway, as a few of its genes appear to be homologous (E values of 3e-12 and 4e-5 by the Protein BLAST, 40) to the fcl gene, encoding the key fucose pathway enzyme in E. coli.Therefore, we performed an HPLC analysis using 11 monosaccharides as standards, and found that the +146 Da modification is rhamnose (Fig 4A).Furthermore, a quantitative analysis of the sugar and amino acid components revealed that the endogenous EF-P(Nm) peptide contains Arg, Ser, Gly, Ala, Lys, and rhamnose, at concentration ratios of 1: 2: 2: 1: 1: 1, respectively (Fig 4B ).Taken together, these data demonstrated that the guanidino group of Arg32 in EF-P(Nm) is linked with rhamnose in the endogenous EF-P(Nm) (Fig 4C ), which is the same post-translational modification as in the S. oneidensis and P. aeruginosa EF-P proteins [35].
The N. meningitidis HT1125 genome encodes the EF-P rhamnosylation enzyme EarP We found that EF-P(Nm) can be modified with rhamnose by crude extracts of N. meningitidis cells (S3 Fig) .In the cases of the S. oneidensis and P. aeruginosa EF-P proteins, EarP performs the rhamnosyl modification at Arg32, using dTDP-L-rhamnose as a substrate [35,36].The earP gene encoding EarP (NMB0935a, accession code: YP_008920709; NMH_0797, accession code: EFV64284) is located next to the efp gene (NMB0937, accession code: AAF41343; NMH_0798, accession code: EFV64285) in the genomes of N. meningitidis strains MC58 and H44/76, while the genome sequence of the N. meningitidis HT1125 strain is not yet available.The gene encoding EarP was cloned from N. meningitidis HT1125, and it shared 96% amino acid sequence identity with those from the H44/76 and MC58 strains (S4 Fig).
The EarP protein from N. meningitidis modifies EF-P(Nm) with rhamnose We tested whether the N. meningitidis EarP, or EarP(Nm), could rhamnosylate EF-P(Nm) (Fig 5).The plasmid pET-NmED was constructed to express EF-P(Nm) and EarP(Nm) (S1  The rhamnosylated EF-P(Nm), but not the unmodified EF-P(Nm), restores the growth rate of EF-P-deleted E. coli cells to the wild-type level To examine whether the rhamnosylated EF-P(Nm) functions in E. coli cells, we used the Keio collection of an E. coli K12 deletion mutant, the Δefp strain JW4107 (BW25113 Δefp::kan).The Δefp mutant grew more slowly than the parent strain BW25113, designated hereafter as the wild type (S6 Fig) .As described previously [18], the Δefp mutant cells transformed with the plasmid vector pMW119, harboring the E. coli efp, epmA, and epmB genes (pMW-EcEGY), grew as fast as the wild-type cells, while the Δefp mutant cells transformed with the empty vector, pMW119, grew as slowly as the parent Δefp mutant cells (S6 Fig) .Interestingly, the plasmid pMW-NmE, containing the N. meningitidis efp gene, actually slowed the cell growth, as compared with the growth of cells transformed with the empty vector (S6 Fig) .In contrast, the Δefp mutant cells transformed with the plasmid pMW-NmED, containing both the N. meningitidis efp and earP genes, grew as fast as the wild-type cells (S6 Fig) .As EF-P(Nm) has Arg32 at this position, instead of Lys, EF-P(Nm) cannot be β-lysylated by E. coli EpmA and EpmB.Thus, the rhamnosylated EF-P(Nm), rather than its unmodified version, is required for the complementation of the EF-P deficiency of E. coli cells.Therefore, the unmodified EF-P(Nm) is non-functional, but the rhamnosyl-EF-P(Nm) functions, in place of the endogenous β-lysyl-EF-P(Ec), on the E. coli ribosome.

Rhamnosyl-EF-P(Nm) rescues ribosomes stalled at proline stretches in proteins
To examine whether rhamnosyl-EF-P(Nm) rescues stalled ribosomes at proline stretches in Δefp E. coli cells, we used the E. coli flagellar regulator Flk, consisting of 331 amino acid residues, and E. coli GntX, consisting of 227 amino acid residues, as model proteins containing proline stretches, such as the PPP and PPG motifs ( Proline stretch-containing proteins encoded by the N. meningitidis genome reason for the requirement of the EF-P(Nm) function in N. meningitidis cells is that the synthesis of proline stretch-containing proteins in the absence of EF-P may be much less efficient than that in E. coli cells.

N. meningitidis EarP is important for cell viability
In the same manner as for the efp gene, we attempted to disrupt the earP gene, encoding EarP, in the N. meningitidis genome (Fig 7A).N. meningitidis HT1125 cells were transformed with a fragment containing the earP-flanking regions and the spectinomycin resistance gene (spc), in order to disrupt the earP gene in the N meningitidis chromosome, and we obtained many spectinomycin-resistant (Spc r ) colonies of strain HT1907, with the ΔearP::spc allele (Fig 7B , right).Nevertheless, the growth of the N. meningitidis ΔearP cells was much slower than that of the N. meningitidis HT1125 cells (Fig 7B , left).As described above, no colonies of the efp null mutant were obtained in the presence of the plasmid harboring the efp(R32K) or efp(R32A) mutant gene (Table 1).Taken together, we confirmed that the Arg32 residue in EF-P is essential, but the post-translational rhamnosyl modification of EF-P(Arg32) is not essential, for the viability of N. meningitidis cells.Therefore, EF-P(Nm) with the unmodified Arg32 can function to some extent on the N. meningitidis ribosome, although the activity is much lower than that of the rhamnosyl-EF-P(Nm).In contrast, the unmodified EF-P(Nm) did not function on the E. coli ing that the unmodified EF-P(Nm) is not only non-functional but also inhibitory for the E. coli ribosome.Consequently, the N. meningitidis ribosome is different from the E. coli ribosome, in that the unmodified EF-P(Nm) is not inhibitory but minimally functional.It is therefore important to clarify the difference in the precise mechanisms of proline stretch translation between the E. coli ribosome and the N. meningitidis/S.oneidensis ribosomes.

The divergent post-translational modifications of EF-P may be due to evolutionary convergence
Although the β-lysyl and rhamnosyl modification pathways are phylogenetically unrelated to each other, both post-translational modifications are analogous, in that the modified Lys34/ Arg32 residue might thus be extended to reach the ribosomal peptidyltransferase center [35].
We propose that the long side chain at the Lys34/Arg32 position of EF-P is required for EF-P to rescue stalled ribosomes efficiently.Here, we suggest that such diverse post-translational modifications, including the β-lysylation/rhamnosylation of EF-P (and the hypusination of eIF5A) (Fig 8), are typical examples of "convergent evolution".Evolutionary convergence creates analogous structures that have similar forms or functions, which were not present in the common ancestor.Previously, we proposed that the molecular mimicry between the lysyl modification of EF-P(Ec)(Lys34) by EpmA and the aminoacylation of tRNA(A76) by aaRS resulted from convergent evolution [18].Likewise, such evolutionary convergence might have occurred in the post-translational EF-P/eIF5A modifications during evolution.
The post-translational rhamnosyl modification of EF-P(Arg32) in N. meningitidis is a new antibiotic target N. meningitidis is a Gram-negative diplococcus pathogen that colonizes the nasopharynx.It can spread into the bloodstream, where it causes septicaemia and furthermore induces meningitis when it reaches the cerebrospinal fluid.Although some of the factors involved in its pathogenesis have been identified (reviewed in [44]), many problems still remain to be resolved: e.g., the development of an appropriate antibiotic therapy for systemic meningococcal disease is urgently required.The present in vivo analyses using the EF-P-deficient N. meningitidis mutant revealed that EF-P(Nm) is important for N. meningitidis survival.Therefore, EF-P(Nm) may be a target for antibiotic drug development.In general, the three-dimensional structures of EF-Ps are mostly composed of convex surfaces, and it is therefore difficult to design chemical compounds that specifically bind to EF-P and inhibit it efficiently.However, the present study revealed that the post-translational modification of EF-P(Nm) is important, but not essential, for the growth of N. meningitidis, which is quite unusual among bacteria.Consequently, since the enzymatic catalytic site may have one or more pockets for substrate binding, and may generally be suitable for drug design and development (i.e., druggable), the rhamnosylation enzyme EarP(Nm) is an attractive target for antibiotic drug development toward the treatment of meningitis.

The post-translational rhamnosyl modification of EF-P(Arg32) in other pathogenic bacteria
The deficiency in EF-P and/or its post-translational modification reportedly attenuate the virulence and infectivity of pathogenic bacteria [17,24,26,35,45,46].Besides N. meningitidis, the genomes of certain pathogenic β-proteobacteria and γ-proteobacteria encode the conserved EF-P(Arg32) and the rhamnosyl modification enzyme EarP (S4 Table ).Therefore, the posttranslational rhamnosyl modification might be equally important in other clinically relevant species, such as Neisseria gonorrhoeae, Bordetella pertussis, Burkholderia pseudomallei, and Burkholderia cepacia, because all of the EF-P proteins from these bacteria contain Arg32.As a limited number of pathogenic bacteria have the EF-P bearing Arg32 and the corresponding rhamnosylation enzyme homologue, the EF-P rhamnosylation pathway should be a target for new species-specific antibacterial agents against these pathogenic bacteria.

Materials and Methods
Biochemical and molecular biological procedures were performed using commercially available materials, enzymes, and chemicals.The polyclonal antibody against EF-P(Ec) was purchased from Keari Bio (Osaka, Japan).

Purification of the endogenous EF-P from N. meningitidis cells
All steps were performed at 4°C.Frozen N. meningitidis HT1125 cells (13 g) were suspended in B-PER bacterial protein extraction reagent (Takara), and disrupted by freeze-thawing.The crude cell extract (50 ml, 489 mg total protein) was centrifuged at 10,000 × g for 20 min.After the resulting supernatant was dialyzed overnight against 50 mM potassium phosphate buffer (pH 7.4), containing 1 mM DTT and 0.1 mM PMSF (buffer A), the solution was applied to a column of DEAE-Sephacel (50 ml, GE Healthcare) equilibrated with buffer A. The column was washed and then developed with a 0 to 0.4 M NaCl gradient.Active fractions of the eluate were identified with a polyclonal antibody against EF-P(Ec), bearing cross-reactivity against EF-P (Nm) (Fig 2B ), collected, and dialyzed against buffer A. The DEAE-Sephacel fraction (30.3 mg total protein) was applied to a HiTrap Q HP column (GE Healthcare) equilibrated with buffer A. The column was washed, and the proteins were eluted by a linear gradient of 0 to 0.5 M NaCl.Active fractions of the eluate were identified with the anti-EF-P(Ec) polyclonal antibody, collected, and dialyzed against buffer A. To the HiTrap Q fraction (3.49mg total protein), (NH) 2 SO 4 was added to a final concentration of 1 M.This solution was loaded on a HiTrap Butyl HP column (GE Healthcare), equilibrated with buffer A containing 1 M (NH) 2 SO 4 .After the column was washed with buffer A containing 1 M (NH) 2 SO 4 , the proteins were eluted with a linear gradient of 1 to 0 M (NH) 2 SO 4 .The eluted EF-P(Nm) proteins were collected, dialyzed against 20 mM potassium phosphate buffer (pH 7.4) containing 0.15 M NaCl and 10 mM βmercaptoethanol, flash cooled with liquid nitrogen, and stored at -80°C (1.06 mg total protein) until use.
cloned into the NdeI and BamHI sites of the pET23 and pET28 vectors to construct the plasmids pET-NmE1 and pET-NmE2, respectively (S1 Table ).E. coli BL21-Gold(DE3) cells were transformed with the plasmid pET-NmE1, containing a non-tagged N. meningitidis efp gene.The cells harboring pET-NmE1 were grown in LB broth (Miller) medium to an OD 600 of 0.6, and then protein expression was induced with 1 mM IPTG at 37°C for 4 hr.The cells were harvested, sonicated, and centrifuged to remove the cell debris.The supernatant, containing the recombinant EF-P(Nm) protein, was dialyzed against 20 mM potassium phosphate buffer (pH  ).Expression, purification, and MS analyses of the EarP(Nm)-modified recombinant EF-P(Nm) were performed as described below.
Post-translational modification assays E. coli BL21-Gold(DE3) cells were transformed with the plasmid pET-NmE2, containing the efp gene encoding EF-P with an N-terminal hexahistidine tag (His 6 , MGSSHHHHHHSS GLVPRGSH), and the recombinant EF-P(Nm) protein was overexpressed as described above.
The crude extract (100 μl, 1.5 mg/ml protein) was prepared from the recombinant EF-P(Nm)producing cells, mixed with the crude cell extract (40 μl, 4.78 mg/ml protein) from N. meningitidis HT1125 cells, and incubated at 37°C overnight.The His 6 -tagged EF-P(Nm) was purified by batch chromatography on Ni-Sepharose (GE Healthcare), as follows.Ni-Sepharose (25 μl) was added to this solution, which was incubated on a rotary shaker at 4°C for 2 hr.The gel was washed three times with 50 mM potassium phosphate buffer (pH 7.4), containing 150 mM NaCl and 10 mM β-mercaptoethanol (buffer C), and EF-P(Nm) was eluted with buffer C containing 350 mM imidazole.The purified EF-P(Nm) was dialyzed against 20 mM potassium phosphate buffer (pH 7.4), containing 150 mM NaCl and 10 mM β-mercaptoethanol, treated with thrombin (0.1 mg protein per unit) at 4°C overnight, and analyzed by MALDI-TOF MS.

Western blot analysis
The crude extract prepared from N. meningitidis cells and the fractions from each column chromatography step were resolved by 10-20 % SDS-PAGE, and transferred to an Immobilon P membrane (Millipore).EF-P(Nm) was recognized using a polyclonal antibody against EF-P (Ec) and a horseradish peroxidase (HRP)-conjugated anti-rabbit IgG antibody from donkey (GE Healthcare).Western blot analyses were performed as described previously [49,50], with some modifications.

Mass spectrometry analyses
The molecular masses of the N. meningitidis endogenous and recombinant EF-P proteins were determined with the aid of a Voyager-DE STR MALDI-TOF mass spectrometer (Applied Biosystems) and a TOF/TOF5800 system (AB SCIEX).The endogenous and recombinant EF-P (Nm) proteins were subjected to SDS-PAGE, followed by staining with Coomassie Brilliant Blue or SimplyBlue SafeStain (Life Technologies).The protein bands (about 30 kDa) were excised and digested in the gel with endoproteinase AspN (Sequencing Grade, Basel, Switzerland).The digests were separated on a column of Inertsil ODS-3 (1 × 100mm; GL Sciences Inc., Tokyo) with a model 1100 series liquid chromatography system (Agilent Technologies, Waldbronn, Germany), using solvents A and B, which were 0.085% (v/v) aqueous trifluoroacetic acid (TFA) and 0.075% (v/v) TFA in 80% (v/v) acetonitrile, respectively.The peptides were eluted at a flow rate of 20 μl/min, using a linear gradient of 0-60% solvent B. The two peptides selected from the peptide map of these digests were subjected to Edman degradation, using a Procise cLC protein sequencing system (Applied Biosystems).The peptide from the endogenous EF-P(Nm), with the sequence DPMVVQKTEYIkggXssaxv (X designates a modified arginine residue), was further digested with trypsin or Achromobacter protease I (API; a gift from Dr. Masaki, Ibaraki University [51]).The digest was subjected to MALDI-TOF MS on an Ultraflex mass spectrometer (Bruker Daltonics, Bremen, Germany) in the reflector mode, using 2, 5-dihydroxybenzoic acid (DHB) as the matrix.The API digest was further purified on a porous graphitic carbon column (Hypercarb, 1.0 × 100 mm, 3 μm, Thermo Fisher Scientific Inc., San Jose, CA) with a model 1100 series liquid chromatography system (Agilent Technologies) at a flow rate of 20 μl/min, using a linear gradient of 1-50% solvent B. The purified peptide was subjected to MALDI-TOF MS and MS/MS analyses, using an Ultraflex system.

Sugar components and amino acid analysis
The Hypercarb column purified peptide was subjected to a sugar components analysis and an amino acid analysis.The sugar components analysis was performed after hydrolysis with 4 M TFA for 3 hr at 100°C.The sample was N-acetylated, labeled with a fluorescent label (ABEE, 4-aminobenzoic acid ethyl ester), and then analyzed using a Honenpak C18 column (4.6 mm i. d. × 75 mm) from J-Oil Mills Inc., developed with 0.2 M potassium borate buffer (pH 8.9) containing 7% acetonitrile, along with a standard kit from J-Oil Mills Inc. containing 11 monosaccharides [52].For amino acid analysis, the peptide was hydrolyzed in the vapor gas phase in constantly boiling conc.HCl at 110°C for 24 hr.The hydrolysate was analyzed by the pre-label method, using 6-aminoquinolyl-N-hydroxysuccinimidyl carbamate for fluorescence detection, ion-pair chromatography with a reversed-phase column, and an ultra-high-pressure high-performance liquid chromatography (UHPLC) system [53].
To construct the N. meningitidis mutant with the deletion of the efp gene, a 1.6-kbDNA fragment containing the efp gene (0.5 kb) and its upstream (0.6 kb) and downstream (0.5 kb) regions was amplified with the primers efp-1 (5 0 -TGAAAGGCTGCATCTGATGCCT TCGCCGCA-3 0 ) and efp-2 (5 0 -CGGACTGTCTGTTTGCCCTTTCCCATCACG-3 0 ) by Pri-meStar Max DNA polymerase, using N. meningitidis HT1125 genomic DNA as the template.The fragment was cloned into the SmaI site of pMW119 (Nippon Gene), to construct the plasmid pMW-NmE2 (5.8 kb).The 5.3 kb DNA fragment from pMW-NmE2, which does not contain the efp coding region, was amplified again by PrimeSTAR Max DNA polymerase (Takara), and ligated to an erythromycin resistance gene (ermC) to obtain pMW-NmE2-Erm.The linearized DNA fragment (2.1 kb) containing the ermC gene and the efp flanking regions (Δefp::ermC) was amplified from pMW-NmE2-Erm, and transformed into HT1125 in the presence or absence of plasmids harboring the wild-type N. meningitidis efp, efp(R32K), and efp (R32A) genes.Transformation was performed as described previously, with some modifications: 1 μl (50 ng) of the PCR product from pMW-NmE2-Erm was added to 1 ml N. meningitidis suspension in GC broth [47], corresponding to 0.2 OD 600 , and rotated at 37°C for 3 hr.A 100 μl portion of the suspension was spread onto GC agar medium [47], containing 5 μg/ml chloramphenicol and 4 μg/ml erythromycin, and incubated overnight at 37°C in 5% CO 2 .The number of erythromycin-resistant (Erm r ) colonies was counted.At the same time, to examine the efp allele, 16 Erm r colonies were selected and colony PCR was performed with the primers efp-1 and efp-2, which only amplify the chromosomal efp allele, using Tks Gflex DNA polymerase (Takara Bio).The amplified DNA was fractionated on a 0.7% agarose gel.The lengths of the wild-type efp and Δefp::ermC alleles were 2 kb and 2.5 kb, respectively.In this study, Erm r colonies appeared in all of the N. meningitidis transformation experiments, probably because the Δefp::ermC gene was inserted into unidentified allele(s) other than efp due to unexpected recombination.Therefore, the number of Δefp::ermC alleles (n) in 16 Erm r colonies was counted, and the ratio n/16 was multiplied by the total number of Erm r colonies, in order to estimate the total number of Δefp::ermC mutants.

Examination of the essentiality of the inducible efp in N. meningitidis
The N. meningitidis strain with an inducible efp gene was constructed as follows.The 1.1-kb BamHI-KpnI DNA fragment of pMW-NmE was subcloned into the same sites of pTTQ19 [55], to construct the plasmid pHT1095.To reduce the expression level, the translational start codon (ATG) in the efp gene was replaced with TTG by site-directed mutagenesis, to construct the plasmid pTTQ-Ptac-Δ200-TTG-Nmefp.The pTTQ-Ptac-Δ200-TTG-Nmefp was digested with ScaI and AlwNI, and the AlwNI-ScaI DNA fragment was blunt-ended and cloned into the blunted ScaI and PstI sites of the IncQ plasmid pGSS33 [47], to construct pHT1139.The pHT1139 plasmid was introduced into N. meningitidis H44/76 cells, to construct the transformant H44/76/pHT1139.
The endogenous efp gene in the N. meningitidis genome was inactivated by the introduction of the Δefp::ermC mutation, as follows.The linearized DNA fragment (2.1 kb), containing the ermC gene and the efp flanking regions (Δefp::ermC), was amplified from pMW-NmE2-Erm, and transformed into the N. meningitidis H44/76/pHT1139 cells.Transformation was performed as described previously, with some modifications: 1 μl (50 ng) of the PCR product amplified from pMW-NmE2-Erm was added to a 1 ml suspension of the N. meningitidis H44/ 76/pHT1139 cells in GC broth [47], corresponding to 0.2 OD 600 , rotated at 37°C for 8 hr, spread onto GC agar medium [47], containing 5 μg/ml chloramphenicol and 4 μg/ml erythromycin, and incubated overnight at 37°C in 5% CO 2 , to obtain N. meningitidis HT1913/ pHT1139.To examine the efp allele in the HT1913 genome, 16 Erm r colonies were selected and colony PCR was performed with the primers efp-1 and efp-2, which amplify the chromosomal efp allele only, using Tks Gflex DNA polymerase (Takara Bio).The amplified DNA was fractionated on a 0.7% agarose gel.The lengths of the wild-type efp and Δefp::ermC alleles were 2 kb and 2.5 kb, respectively.
The genomic N. meningitidis efp gene was also inactivated by the introduction of the efp (R32opal) mutation, as follows.pMW-NmE2 was mutagenized with PrimeStar Max DNA polymerase to obtain the plasmid pHT1094, harboring N. meningitidis efp(R32opal) (6.3kb).The 5.3-kb DNA fragment was amplified 150 bp downstream of the stop codon of the efp gene, with the primers efp-25 (TTTGTCGGGATTGCGTTCACGGTT) and efp-26 (CGTTTTTA GACATCCATTTTGACGAAA) by PrimeStar Max DNA polymerase, and ligated to the 0.8 kb DNA fragment of the erythromycin resistance gene (ermC), to construct the plasmid pHT1098.A 2.3-kb DNA fragment, containing the efp(R32opal) gene (0.5 kb), and the ermC gene (0.8 kb) and its upstream (0.8 kb) and downstream (0.6 kb) regions, was transformed into N. meningitidis H44/76/pHT1139 cells.The resultant Erm r colonies were selected and named N. meningitidis HT1914/pHT1139, and the efp(R32opal) mutation was confirmed by direct PCR sequencing.The essentiality of the N. meningitidis efp gene was examined as follows.N. meningitidis HT1913/pHT1139 and HT1914/pHT1139 were grown on GC agar, containing 0.5 mM IPTG, at 37°C in a 5% CO 2 atmosphere overnight.The colonies were collected and suspended in 1 ml PBS, at an OD 600 of 0.01.The bacteria were harvested by centrifugation and resuspended in 1 ml PBS.This procedure was repeated five times.A 100 μl portion of the bacterial suspension was plated onto GC agar with or without 0.5 mM IPTG, and the plates were observed after an incubation at 37°C in a 5% CO 2 atmosphere for 20 hrs.
Analysis of the earP null mutant of N. meningitidis A 2.5-kb DNA fragment, containing the earP (1.2 kb) and efp (0.5 kb) genes and their upstream (0.5 kb) and downstream (0.2 kb) regions, was amplified with the primers miaA-6 (5 0 -GGCG GTCGGCCCGAGCAGGGCAA-3 0 ) and efp-down-2 (5 0 -CGGAGAAATGCTTGAAAACC AATC-3 0 ) by PrimeStar Max DNA polymerase (Takara Bio, Japan), using N. meningitidis HT1125 genomic DNA as the template, and cloned into the SmaI site of pMW119 (Nippon Gene), to construct the plasmid pHT1088 (6.7 kb).Since the ATG start codon of the efp gene is located only 43 bp downstream of the stop codon of earP gene, we speculated that the earP and efp genes could be co-transcribed.To avoid the polar effect, the earP null mutant was constructed as follows.The 4.5-kb DNA fragment of pHT1088, which lacks the region encoding the earP and efp genes, was amplified by PrimeStar Max DNA polymerase.In addition, a 1-kb DNA fragment of pMW-NmE in which the efp gene was fused to the E. coli lac promoter, and a 1-kb DNA fragment of the spectinomycin resistance gene (spc) were also amplified, using Pri-meStar Max DNA polymerase.These three fragments were fused by In Fusion cloning (Clontech) to construct pHT1089, containing the ΔearP::spc-Plac-efp genes.To construct the N. meningitidis ΔearP::spc-Plac-efp mutant strain HT1907, the linearized DNA fragment (2.7 kb), containing the ΔearP::spc-Plac-efp genes, was amplified from pHT1089, and transformed into N. meningitidis HT1125 cells.Transformation was performed as described previously [47].Spc r colonies were selected, and the ΔearP::spc-Plac-efp allele in the N. meningitidis genome was confirmed by PCR.

Expression of a proline stretch-containing protein in EF-P-deficient E. coli cells
The E. coli Flk and GntX expression plasmids (pCA24N-derived ASKA library plasmid) were obtained from the National Bio-Resource Project (NBRP), Japan [55].The DNA fragment encoding EarP(Nm) was PCR-amplified from N. meningitidis HT1125 genomic DNA by Pri-meStar Max DNA polymerase (Takara), and cloned into pMW-NmE to construct the plasmid pMW-NmED (S1 Table ).The DNA fragment encoding E. coli EF-P, EpmA, and EpmB was PCR-amplified from the plasmid pACTK-EGY [18] by PrimeStar Max DNA polymerase (Takara), and cloned into pMW119 to construct the plasmid pMW-EcEGY.The plasmids pMW119, pMW-NmE, pMW-NmED, and pMW-EcEGY were co-transformed with the Flk plasmid into E. coli BW25113 and Δefp cells.The cells harboring these plasmids were grown in LB broth (Miller) medium to an OD 600 of 0.6, and then protein expression was induced with 1 mM IPTG at 37°C for 6 hr.The cells were harvested, and the expressed Flk protein with the His 6 -tag (MRGSHHHHHHTDPALRP) and the EF-P protein were analyzed by SDS-PAGE and western blotting, using an anti-His 6 antibody and a polyclonal antibody against EF-P(Ec).

Fig 1 .
Fig 1. Strategies for efp deletion from the N. meningitidis genome.(A) The efp allele in the N. meningitidis genome can be disrupted, but only in the presence of a plasmid containing the wild-type N. meningitidis efp gene.(B, C) The IncQ plasmid pHT1139, containing Ptac-TTG-efp-lacI q , was transformed into N. meningitidis H44/76 cells.Subsequently, the DNA fragment bearing the erythromycin resistance gene (ermC) or the earP-efp(R32opal) gene was

doi: 10 .
1371/journal.pone.0147907.t001further investigated by MALDI-TOF MS/MS analysis (Fig 3C), which confirmed that Arg32 is modified with a +146 Da moiety.Therefore, the post-translational modification of EF-P(Nm) occurs at the same position as Arg32 in the S. oneidensis and P. aeruginosa EF-P proteins and Lys34 in EF-P(Ec) (S2 Fig).

Fig 2 .
Fig 2. Purification and MS analysis of N. meningitidis EF-P(Nm).Proteins in each purification step were analyzed by 10-20% SDS-PAGE and stained with SimplyBlue SafeStain.(A) Lane 1, molecular mass standards; lane 2, N. meningitidis crude cell extract; lane 3, after DEAE-Sephacel column; lane 4, after HiTrap Q HP column; lane 5, endogenous EF-P(Nm) purified on a HiTrap Butyl HP column; lane 6, MagicMark molecular mass standards (Life Technologies).(B) The polyclonal antibody against EF-P(Ec) cross-reacts with EF-P(Nm).The EF-P proteins in the N. meningitidis cell extracts and the column chromatography fractions were monitored by western blotting with the antibody.Lane 1, N. meningitidis crude cell extract; lane 2, MagicMark molecular mass standards.The arrow designates EF-P(Nm).(C) MALDI-TOF MS and SDS-PAGE of the endogenous EF-P(Nm), shown in the left and right panels, respectively.(D) MALDI-TOF MS and SDS-PAGE of the recombinant EF-P(Nm), shown in the left and right panels, respectively.doi:10.1371/journal.pone.0147907.g002

Fig 4 .
Fig 4. Analysis of the sugar and amino acid components in the modified peptide of N. meningitidis endogenous EF-P.(A) Analysis of the sugar composition in the endogenous EF-P(Nm) peptide, GGR*SSAK.The peptide was hydrolyzed to free amino acids and other components.The samples were derivatized with ABEE, and were subjected to the UHPLC analysis, using 11 monosaccharides as standards (std).A blank sample including deionized water was also loaded (blank).Gal, galactose; Man, mannose; Glc, glucose; Ara, arabinose; Rib, ribose; ManNAc, N-acetylmannosamine; Xyl, xylose; GlcNAc, Nacetylglucosamine; Fuc, fucose; Rha, rhamnose; GalNAc, N-acetylgalactosamine.(B) Quantitative analysis of the amino acid and rhamnose components in the EF-P(Nm) peptide.(C) Rhamnosyl-Arg32 and a 3D structural model of N. meningitidis EF-P.doi:10.1371/journal.pone.0147907.g004

7 . 4 )
, containing 150 mM NaCl and 10 mM β-mercaptoethanol, treated with thrombin (0.1 mg protein per unit) at 4°C overnight, and subjected to MALDI-TOF MS or ESI-MS analyses.The purified EF-P(Nm) protein was flash-cooled with liquid nitrogen, and stored at −80°C.The gene encoding the EarP(Nm) homologue (accession code: LC059993) was PCR-amplified from N. meningitidis HT1125 genomic DNA by ExTaq DNA polymerase (Takara), and cloned into the BamHI and HindIII sites of pET-NmEFP2, and the NdeI and HindIII sites of pET28, to construct the plasmids pET-NmED and pET-NmD, respectively (S1 Table

Table 1 .
The N. meningitidis efp gene is essential for cell viability.containing the ermC gene and the efp flanking regions was transformed into N. meningitidis HT1125, but the disruption of the efp gene failed in the absence of the wild-type efp-containing plasmid.Experiments were performed 5 times and the averages of the results are shown, together with the standard deviation for the wild-type efp, in this table.