Analysis of Bacterial Phosphorylcholine-Related Genes Reveals an Association between Type-Specific Biosynthesis Pathways and Biomolecules Targeted for Phosphorylcholine Modification

Abstract

Many bacterial surface proteins and carbohydrates are modified with phosphorylcholine (ChoP), which contributes to host mimicry and can also promote colonization and survival in the host. However, the ChoP biosynthetic pathways that are used in bacterial species that express ChoP have not been systematically studied. For example, the well-studied Lic-1 pathway is absent in some ChoP-expressing bacteria, such as Neisseria meningitidis and Neisseria gonorrhoeae. This raises a question as to the origin of the ChoP used for macromolecule biosynthesis in these species. In the current study, we used in silico analyses to identify the potential pathways involved in ChoP biosynthesis in genomes of the 26 bacterial species reported to express a ChoP-modified biomolecule. We used the four known ChoP biosynthetic pathways and a ChoP transferase as search terms to probe for their presence in these genomes. We found that the Lic-1 pathway is primarily associated with organisms producing ChoP-modified carbohydrates, such as lipooligosaccharide. Pilin phosphorylcholine transferase A (PptA) homologs were detected in all bacteria that express ChoP-modified proteins. Additionally, ChoP biosynthesis pathways, such as phospholipid N-methyltransferase (PmtA), phosphatidylcholine synthase (Pcs), or the acylation-dependent phosphatidylcholine biosynthesis pathway, which generate phosphatidylcholine, were also identified in species that produce ChoP-modified proteins. Thus, a major finding of this study is the association of a particular ChoP biosynthetic pathway with a cognate, target ChoP-modified surface factor; i.e., protein versus carbohydrate. This survey failed to identify a known biosynthetic pathway for some species that express ChoP, indicating that a novel ChoP biosynthetic pathway(s) may remain to be identified. IMPORTANCE The modification of bacterial surface virulence factors with phosphorylcholine (ChoP) plays an important role in bacterial virulence and pathogenesis. However, the ChoP biosynthetic pathways in bacteria have not been fully understood. In this study, we used in silico analysis to identify potential ChoP biosynthetic pathways in bacteria that express ChoP-modified biomolecules and found the association between a specific ChoP biosynthesis pathway and the cognate target ChoP-modified surface factor.

Keywords: ChoP; bacterial virulence; phosphatidylcholine; phosphoethanolamine; phosphorylcholine.

FIG 1

Characterized pathways for biosynthesis of candidate ChoP donor molecules for macromolecule modification by ChoP. (a) Lic-1 pathway. LicB takes up choline from the environment, which is then converted to CDP-choline by LicA and LicC. LicD transfers the activated CDP-ChoP to a glycan structure, such as LOS or teichoic acid. (b) PmtA pathway. Phosphatidylethanolamine is methylated at three positions by PmtA to form phosphatidylcholine. (c) Pcs pathway. Pcs enzymes catalyze the condensation of choline with CDP-diacylglycerol to phosphatidylcholine (PC), releasing a CMP molecule. (d) Acylation-dependent PC biosynthesis pathway. Abbreviations: SAHC, S-adenosylhomocysteine; SAM, S-adenosylmethionine; PPi, pyrophosphate; Lplt, lysophospholipid transporter; Aas, acyltransferase-acyl carrier protein synthase.

FIG 2

Enzyme homologs identified in bacteria containing ChoP-carbohydrates. The arrangements of the genes and proteins in the Lic-1 pathway are indicated, with LicA, LicB, LicC, and LicD depicted in blue, purple, green, and brown, respectively. Organisms and strain name are shown to the left. Homologs that were found are indicated next to the stain name with accession numbers. The arrow represents the gene orientation. The percent sequence identity with Lic-1 pathway proteins in H. influenzae is displayed below each gene box.

FIG 3

Enzyme homologs identified in bacteria containing ChoP-modified protein. The known PmtA of R. sphaeroides (61), PmtA of S. meliloti (62), Pcs of S. meliloti (62), and characterized PptA (13, 23) with accession numbers are depicted in red, yellow, pink, and green, respectively. Organisms and strain names are shown to the left. Homologs that were found are indicated next to the stain name with accession numbers. The arrow represents the gene orientation. The percent sequence identity with R. sphaeroides PmtA, S. meliloti PmtA, S. meliloti Pcs, and N. meningitidis PptA is displayed below each gene box.

FIG 4

Enzyme homologs identified in bacteria containing undefined ChoP-modified structures. The arrangements of genes and proteins in the Lic-1 pathway are illustrated, with LicA, LicB, LicC, and LicD depicted in blue, purple, green, and brown, respectively. The known PmtA of R. sphaeroides (61), PmtA of S. meliloti (62), Pcs of S. meliloti (62), characterized N. meningitidis PptA (13, 23), and P. aeruginosa EftM (29), with their respective accession numbers, are shown in red, yellow, pink, green, and orange. Organisms and strain names are indicated to the left, and homologs that were identified are noted next to the stain name, along with accession numbers. Gene orientation is represented by arrows, and the percent sequence identity is shown below each gene box.

FIG 5

Unrooted phylogenetic tree of PptA and PptA-like ORFs. The protein sequences of PptA-like ORFs were used to construct a neighbor-joining tree with characterized PptA (13, 23), EptA (49), EptB (55), PmrC (50, 51), LptA (52), Lpt3 (53, 54), and Lpt6 (48). N. meningitidis PptA is depicted in pink. E. coli EptB, E. coli EptA, N. meningitidis Lpt3, N. meningitidis Lpt6, N. meningitidis LptA, and A. baumannii PmrC are depicted in green. Distances between sequences are expressed as 0.2 change per amino acid residue. Ng, N. gonorrhoeae; Ab, A. baumannii; Aggr.a, A. actinomycetemcomitans. Accession numbers of the sequences of the proteins are shown.

FIG 6

Unrooted phylogenetic tree of PmtA and PmtA-like ORFs. The protein sequences of PmtA-like ORFs were used to construct a neighbor-joining tree with known PmtA of R. sphaeroides (61), S. meliloti (62), PmtA, PmtX1, PmtX3, and PmtX4 of B. japonicum (64), and PmtA of X. campestris (27). R. sphaeroides PmtA and B. japonicum PmtAX1 are shown in green. S. meliloti PmtA, B. japonicum PmtA, PmtAX4, and PmtAX3, and X. campestris PmtA are highlighted in blue. R. sphaeroides PmtA-like ORFs are abbreviated as Rs ORFs. S. meliloti PmtA-like ORFs are referred to as Sm ORFs. Distances between sequences are expressed as 0.2 change per amino acid residue. Rs, R. sphaeroides; Sm, S. meliloti; Nm, N. meningitidis; Micr, Micrococcus spp.; Lact, Lactococcus spp.; Cory, C. jeikeium. Accession numbers of the sequences of the proteins are shown.

FIG 7

Unrooted phylogenetic tree of Pcs and Pcs-like ORFs. The protein sequences of Pcs-like ORFs were used to construct a neighbor-joining tree with known Pcs of S. meliloti (62) (orange boxes) and PssA and PgsA of E. coli (65) (green boxes). Distances between sequences are expressed as 0.1 change per amino acid residue. Accession numbers of the sequences of the proteins are shown.

FIG 8

Phylogenetic analysis of a number of known and predicted acyltransferases involved in PC biosynthesis. The protein sequences of Aas, Xc_0188, and Xc_0238-like ORFs were used to construct a neighbor-joining tree with known Aas of E. coli (26), Xc_0188 of X. campestris (27), and Xc_0238 of X. campestris (27). They are depicted in dark purple, dark red, and dark yellow, respectively. The tree shows three different lyso-PC acyltransferase families, represented by shaded clusters with the following colors: purple for the Aas family, pink for the Xc_0188 family, and yellow for the Xc_0238 family. Sp, S. pneumoniae; St, S. mitis; So, S. oralis. Distances between sequences are expressed as 0.1 change per amino acid residue. Accession numbers of the sequences of the proteins are shown.

FIG 9

Phylogenetic analysis of a number of known and predicted Lplt transporters involved in PC biosynthesis. The protein sequences of Lplt-like ORFs were used to construct a neighbor-joining tree with known Lplt of E. coli (26). The shaded cluster represents the Lplt family. Distances between sequences are expressed as 0.1 change per amino acid residue. Accession numbers of the sequences of the proteins are shown.