Isolation and functional analysis of phage-displayed antibody fragments targeting the staphylococcal superantigen-like proteins

Abstract

Staphylococcus aureus produces numerous virulence factors that manipulate the immune system, helping the bacteria avoid phagocytosis. In this study, we are investigating three immune evasion molecules called the staphylococcal superantigen-like proteins 1, 5, and 10 (SSL1, SSL5, and SSL10). All three SSLs inhibit vital host immune processes and contribute to S. aureus immune evasion. This study aimed to identify single-chain variable fragment (scFvs) antibodies from synthetic antibody phage libraries, which can recognize either of the three SSLs and could block the interaction between the SSLs and their respective human targets. The antibodies were isolated after three rounds of panning against SSL1, SSL5, and SSL10, and their ability to bind to the SSLs was studied using a time-resolved fluorescence-based immunoassay. We successfully obtained altogether 44 unique clones displaying binding activity to either SSL1, SSL5, or SSL10. The capability of the SSL-recognizing scFvs to inhibit the SSLs' function was tested in an MMP9 enzymatic activity assay, a P-selectin glycoprotein ligand 1 competitive binding assay, and an IgG1-mediated phagocytosis assay. We could show that one scFv was able to inhibit SSL1 and maintain MMP9 activity in a concentration-dependent manner. Finally, the structure of this inhibiting scFv was modeled and used to create putative scFv-SSL1-complex models by protein-protein docking. The complex models were subjected to a 100-ns molecular dynamics simulation to assess the possible binding mode of the antibody.

Keywords: Staphylococcus aureus; antivirulence strategy; phage display; scFv; staphylococcal superantigen-like protein.

Figure 1

Crystal structure of SSL1 (PDB ID: 4O1N). The N‐terminal oligosaccharide‐binding (OB) domain is colored slate blue, and the C‐terminal β‐grasp domain is green. The SSL1 N‐terminal β1‐3 domain and the C‐terminal α4β9 domain are colored violet and yellow, respectively.

Figure 2

(a–c) Enrichment of the phage antibody library against. (a) SSL1, (b) SSL5, or (c) SSL10 determined by phage immunoassay after each panning round. The bars represent an average of triplicate fluorescent signals obtained and their respective standard deviations.

Figure 3

Binding profiles for the 29 SSL‐binding scFv‐ALP (alkaline phosphatase fusion scFv) chosen for further activity testing. The binding of the scFvs to the biotinylated antigens was detected through the fusion ALP activity. scFv‐ALP, alkaline phosphatase fusion proteins.

Figure 4

Effect of SSL5‐binding scFvs on relative MMP9 activity after inhibition by 0.125 μg/mL SSL5. The gray bar represents the enzymatic activity of MMP9 with no SSL present and the white bar its activity after the addition of SSL5. The rest of the bars represent MMP9 activity after the addition of SSL5 that was pretreated with a concentration range (in μg/mL) of SSL5‐binding scFvs. MMP9, matrix metallopeptidase 9; scFvs, selective single chain variable fragment.

Figure 5

Relative MMP9 activity after SSL1 inhibition by 0.25 μg/mL SSL1. The gray bar represents MMP9 enzymatic activity with no SSL1 present and the white bar the activity after the addition of SSL1. The rest of the bars represent MMP9 activity after the addition of SSL1 that was pretreated with a concentration range (in μg/mL) of SSL1‐binding scFvs. MMP9, matrix metallopeptidase 9; scFvs, selective single chain variable fragment.

Figure 6

MMP9‐mediated collagen degradation visualized on an SDS‐PAGE. All the wells contain collagen. Activated MMP9 and MMP9 inhibited with SSL1 were used as positive and negative controls, respectively. Noninhibiting scFv‐45‐ALP and SSL5 that scFv‐93‐ALP does not recognize were used as additional controls. scFv‐93‐ALP was added in a concentration series from 3.7 to 30 μg/mL. MMP9, matrix metallopeptidase 9; scFvs, selective single chain variable fragment; SDS‐PAGE, sodium dodecyl sulphate‐polyacrylamide gel electrophoresis.

Figure 7

Gel filtration chromatogram. 280 µg/mL scFv‐93‐ALP was preincubated with 500 µg/mL HIS‐SSL1 in a total volume of 500 µL for 30 min at RT before running on a gel filtration column. For comparison, 280 µg/mL scFv‐93‐ALP and 500 µg/mL HIS‐SSL1 were run separately. ALP, alkaline phosphatase; RT, room temperature.

Figure 8

ScFvs' effect on P‐selectin binding to PSGL‐1 on human neutrophils in the presence of SSL5. P‐selectin binding to PSGL‐1 was measured after the addition of SSL5 (1.0 μg/mL) and scFvs (50 μg/mL). The control of uninhibited PSGL‐1 was measured and the signals for the rest of the measurements were adjusted in relation to its fluorescent signal. A low fluorescent signal indicates a low level of PSGL‐1 bound P‐selectin. scFvs, selective single chain variable fragment.

Figure 9

A pairwise alignment of the SSL1‐binding scFv‐99 and SSL1‐binding and inhibiting scFv‐93. Black arrows indicate conservative residue variation and red arrows nonconservative variation. Most of the variation between these antibodies is located in the CDR‐H2 loop (from residue 50 to 65; Kabat numbering; Kabat et al., 1991). The linker joining the variable chains together is marked with a dashed line. Sequence alignment was conducted with Clustal Omega (1.2.4) (Figure A4) and modified with the ESPript online tool (Madeira et al., ; Robert & Gouet, 2014).

Figure 10

A predicted binding complex of SSL1 with scFv‐93. A crystal structure of SSL1 (PDB ID: 4O1N) was docked to the homology model of scFv‐93 and the docking complex was subjected to a 100‐ns MD simulation. The last frame of the simulation is presented here. SSL1 is presented as a light blue surface/sticks and the N‐terminal β1‐3 domain residues are colored slate blue. The CDR‐loops of scFv‐93 are colored as follows: L1 magenta, L2 cyan, L3 orange, H1 brown, H2 yellow, and H3 green. Interacting residues are labeled and presented as sticks (right). Polar contacts are shown as black dashed lines. MD, molecular dynamics.

Figure A1

(a–c) Screening immunoassay results for SSL‐binding scFv‐ALP clones after three rounds of panning. The binding of the scFvs to the biotinylated antigens was detected through the fusion ALP activity. The clones selected for larger‐scale production and further activity testing are denoted with a red asterisk. ALP, alkaline phosphatase; scFvs, selective single chain variable fragment.

Figure A2

IgG1 mediated phagocytosis. Uninhibited IgG1 is shown as a green bar, SSL10 inhibited IgG1 is shown as an orange bar and the scFv pretreated SSL10 with IgG1 are shown as blue bars. No SSL10 inhibition of the antibody fragments could be detected.

Figure A3

Multiple sequence alignment between the three reviewed SSLs created with Clustal Omega.

Figure A4

Pairwise sequence alignment of scFv‐93 and scFv‐99 created with Clustal Omega.

Figure A5

A predicted binding complex of SSL1 with scFv‐93. A crystal structure of SSL1 (PDB ID: 4O1N) was docked to a homology modeled scFv‐93 and the complex was subjected to a 100‐ns MD simulation. The last frame of the simulation is presented here. SSL1 is presented as light blue surface/sticks and the C‐terminal α4β9 domain residues are colored slate blue. The CDR‐loops of scFv‐93 are colored as follows: L1 magenta, L2 cyan, L3 orange, H1 brown, H2 yellow, and H3 green. Interacting residues are labeled and presented as sticks (right). Polar contacts are shown as black dashed lines. MD, molecular dynamics.

Figure A6

Interacting residues between the scFv‐93 (column) and SSL1 β1‐3 (row) interface. The residues that are in contact during the MD trajectory are shown as squares with different intensities of blue, depending on the frequency of the interaction (light blue for sporadic contacts and the most frequently occurring in black). The data are from the last 50 ns of the simulation collected from every ns of the trajectory. HC, heavy chain; LC, light chain; MD, molecular dynamics.