Identification and Functional Characterization of Peptides With Antimicrobial Activity From the Syphilis Spirochete, Treponema pallidum

Abstract

The etiological agent of syphilis, Treponema pallidum ssp. pallidum, is a highly invasive "stealth" pathogen that can evade the host immune response and persist within the host for decades. This obligate human pathogen is adept at establishing infection and surviving at sites within the host that have a multitude of competing microbes, sometimes including pathogens. One survival strategy employed by bacteria found at polymicrobial sites is elimination of competing microorganisms by production of antimicrobial peptides (AMPs). Antimicrobial peptides are low molecular weight proteins (miniproteins) that function directly via inhibition and killing of microbes and/or indirectly via modulation of the host immune response, which can facilitate immune evasion. In the current study, we used bioinformatics to show that approximately 7% of the T. pallidum proteome is comprised of miniproteins of 150 amino acids or less with unknown functions. To investigate the possibility that AMP production is an unrecognized defense strategy used by T. pallidum during infection, we developed a bioinformatics pipeline to analyze the complement of T. pallidum miniproteins of unknown function for the identification of potential AMPs. This analysis identified 45 T. pallidum AMP candidates; of these, Tp0451a and Tp0749 were subjected to further bioinformatic analyses to identify AMP critical core regions (AMPCCRs). Four potential AMPCCRs from the two predicted AMPs were identified and peptides corresponding to these AMPCCRs were experimentally confirmed to exhibit bacteriostatic and bactericidal activity against a panel of biologically relevant Gram-positive and Gram-negative bacteria. Immunomodulation assays performed under inflammatory conditions demonstrated that one of the AMPCCRs was also capable of differentially regulating expression of two pro-inflammatory chemokines [monocyte chemoattractant protein-1 (MCP-1) and interleukin-8 (IL-8)]. These findings demonstrate proof-of-concept for our developed AMP identification pipeline and are consistent with the novel concept that T. pallidum expresses AMPs to defend against competing microbes and modulate the host immune response.

Keywords: Treponema pallidum; antimicrobial peptides; bactericidal; bacteriostatic; syphilis.

FIGURE 1

Bioinformatics pipeline for the identification of potential AMPs and AMPCCRs in the proteome of T. pallidum. (A) Proteome Analysis: The whole proteome (from NCBI genome database, July 2021 annotation) of T. pallidum was searched for proteins with no assigned function and 150 amino acids or less. (B) AMP Prediction: Complete amino acid sequences of 68 T. pallidum proteins (≤150 amino acids) of unknown function were submitted to three AMP prediction servers which allowed for ranking from most likely AMP to least likely AMP. Additional leader peptide, physicochemical, expression, and proteome clustering analyses facilitated the identification of two potential AMPs (Tp0451a and Tp0749) for further analyses. (C) AMPCCR prediction: A combination of critical core region (CCR) mapping, structural, homology, and cell penetrating peptide analyses were then performed to further confirm the initial AMP predictions generated by the three prediction servers and to identify putative antimicrobial peptide critical core regions (AMPCCRs) within the identified predicted AMPs, Tp0451a and Tp0749.

FIGURE 2

Distribution of the number of positive AMP predictions and corresponding mean probability score ranges for T. pallidum miniproteins of unknown function. Pie chart indicating the number of positive AMP predictions following submission of the full-length amino acid sequences of 68 T. pallidum miniproteins (≤150 amino acids) to three AMP prediction servers (eight AMP prediction algorithms total). The arrow begins at the 0/8 positive AMP prediction class and finishes at the 8/8 positive AMP prediction class. The corresponding mean probability score range for all proteins from each positive AMP prediction class is also shown.

FIGURE 3

Physicochemical properties of T. pallidum miniproteins of unknown function. Physicochemical properties known to be important for AMP function were calculated for the 68 T. pallidum proteins (≤150 amino acids) of unknown function using the APD3 online AMP calculator and predictor tool. Left: Scatter plots depicting (A) net charges, (B) hydrophobic residue content, and (C) cysteine residue content (orange circles represent the mean values, +/- standard error) of all proteins from each positive AMP prediction class. Right: Bar graphs showing (A) net charge, (B) hydrophobic amino acid content, and (C) cysteine residue content of all 68 miniproteins (≤150 amino acids) from predicted AMP ranking 1–68. No. +ve AMP Predictions equates to the number of AMP prediction programs that assigned the miniprotein as an AMP.

FIGURE 4

Proteome clustering of T. pallidum (Nichols strain) miniproteins of unknown function. Each of the T. pallidum (Nichols strain) proteins (≤150 amino acids) of unknown function were arranged from the lowest (Tp0004) to the highest (Tp1032) locus tag number within the proteome (Tp0001–Tp1041) and clusters comprised of at least two miniproteins separated by five or less intervening proteins were identified. (A) Schematic depicting the spatial arrangement of all miniproteins (≤50 amino acids) of unknown function within the T. pallidum (Nichols strain) proteome. The location of each protein within the proteome is represented by a vertical black line. (B) Schematic showing the spatial arrangement of the 30 top-ranking predicted AMPs within the T. pallidum (Nichols strain) proteome. The location of proteins corresponding to rankings 1–10, 11–20, and 21–30 are shown in red, blue, and green, respectively. Asterisks denote the location of 20 of the top 30-ranking predicted AMPs that are found to be located within 13 miniprotein clusters.

FIGURE 5

In silico identification of potential T. pallidum AMPCCRs. The critical core regions of two candidate AMPs, (A) Tp0451a and (B) Tp0749, were predicted using our bioinformatics pipeline. The first step of the pipeline involved CCR mapping (A,B, top): four prediction servers [AMPA (one algorithm), CAMP (three algorithms), AntiBP (three algorithms), and AntiBP2 (one algorithm)] were used to identify the amino acid boundaries of potential antimicrobial active regions (critical core regions, CCRs). High probability/scoring regions predicted by at least three of the four servers are shown with their corresponding probabilities (AMPA and CAMP algorithms) or scores (AntiBP and AntiBP2 algorithms). Hydrophobic residues: green; Positively-charged residues: red; Cysteines: blue. In the second step of the pipeline, secondary structure analyses and modeling were performed (A and B panels, middle): secondary structure analyses of the full-length proteins were performed using Jpred 4 (H: alpha helix; E: beta strand; dashed line: coiled) and PSIPRED (pink highlight: alpha helix; orange highlight: beta strand; gray highlight: coiled). HeliQuest was used to generate helical wheel diagrams for potential alpha helices (yellow: hydrophobic residues; purple: serine or threonine; blue: positively charged residues: gray: glycine or alanine). Structure modeling using Modeller generated a confident model for the central region of Tp0451a (residues E36-I71), but confident models were not generated for the N- or C-terminal regions. Structure modeling using a combination of PEP-FOLD-2, Swiss-Model, Molsoft ICM, and Modeller generated models for the N- and C-terminal regions of Tp0749 (residues P14-S48 and I52-K65, respectively). Together, these findings allowed for the identification of two potential critical core regions within the N-terminus (Tp0451a_N and Tp0749_N) and C-terminus (Tp0451a_C and Tp0749_C) of Tp0451a and Tp0749 (A and B panels, bottom).

FIGURE 6

Structure modeling of the candidate T. pallidum AMP, Tp0749. (A, Left): Table showing the scoring functions of the Tp0749 N-terminal model (residues P14-S48) generated by a combination of PEP-FOLD-2, Swiss-Model, Molsoft ICM, and Modeller. (Middle) Model ribbon structure of Tp0749 residues P14-S48 and rotated view showing amphipathicity. (Right) Helical wheel schematic of Tp0749 (P14-S48) generated using HeliQuest. Dashed line separates the hydrophilic/polar and hydrophobic/non-polar faces of the predicted alpha helix. (B) Ribbon and surface/charge distribution images of the Tp0749 (P14-S48) model showing one face of the alpha helix rich in positively-charged/polar residues (blue) and the opposing face rich in hydrophobic/non-polar residues (white). Red: negatively-charged/polar residues. (C) PROMALS3D was used to generate a structure-based comparative model of the Tp0749 N-terminal alpha helix model (P14-S48) using the structure of the known AMP, human cathelicidin LL-37 (PDB:5NMN) as a template. RMSD—0.31 Å over 19 Cα atoms and 3/6 positively charged residues involved in binding target cell membrane lipids are conserved (indicated by rectangles; triangles show non-conserved residues). (D) Left: Table showing the scoring functions of the Tp0749 C-terminal model (residues I52-K65) generated by using a combination of PEP-FOLD-2, Swiss-Model, Molsoft ICM, and Modeller. (Middle) Model ribbon structure of Tp0749 residues I52-K65 and rotated view showing partial amphipathicity (red: negatively charged/polar residue side chains; blue: positively charged/polar residue side chains; yellow: cysteine residue side chain). (Right) Helical wheel schematic of Tp0749 (I52-K65) generated using HeliQuest. Dashed line separates the hydrophilic/polar and partially hydrophobic/non-polar faces of the predicted alpha helix. (E) Surface and charge distribution views of Tp0749 (I52-K65) (red: negatively charged/polar residues; blue: positively charged/polar residues; white: hydrophobic/non-polar residues).

FIGURE 7

Analysis of the immunomodulatory capacities of the T. pallidum AMPCCRs. Human THP-1 cells were differentiated to macrophages and then stimulated with or without IL-32γ. Cells were then immediately exposed to either LL-37, sLL-37, Tp0751_p5, Tp0451a_C, Tp0451a_N, Tp0749_C, or Tp0749_N and analyzed for expression of (A) MCP-1 and (B) IL-8. Each data point is representative of cells from one well of a 12-well plate. Data shown is representative of three independent experiments. A Dunnett’s multiple comparisons test was used for normally distributed data and a Dunn’s multiple comparisons test was used for data that was not normally distributed. For statistical analyses, mean values from each peptide were compared to the mean of the unstimulated control (IL-32γ only stimulation); p-values (Dunnett’s multiple comparisons test) indicating statistically significant differences observed in three independent experiments are indicated.