Characterization of a glycan-binding complex of minor pilins completes the analysis of Streptococcus sanguinis type 4 pili subunits

Significance Type 4 pili (T4P)—important in bacterial pathogens—are filaments composed of one well-characterized major pilin, and several minor pilins whose roles are often poorly understood. Streptococcus sanguinis T4P are composed of five pilins, which makes it a good model to determine the role of each pilin subunit in detail. Here, we characterize PilA and PilC, showing how they interact and how they function. Together with our previous findings, this provides an integrated view of the role of the five pilin subunits in S. sanguinis T4P. PilE1/PilE2 are the major pilins forming the backbone of the filament, while the three minor pilins (PilA, PilB, and PilC) form a tip-located complex promoting adhesion to various host receptors.

equilibrated with binding buffer. The resin was mixed with the cell lysate and the column was allowed to drain. The column was then washed with binding buffer (50 mM HEPES pH 7.5, 150 mM NaCl, 20 mM imidazole) several times, before being eluted with elution buffer (50 mM HEPES pH 7.5, 150 mM NaCl, 300 mM imidazole).
Strep-tagged proteins were affinity-purified on an ÄKTA Purifier using 5 ml StrepTrap HP columns (GE Healthcare) according to manufacturer's instructions. The columns were washed with binding buffer (50 mM HEPES pH 7.5, 150 mM NaCl) and protein was eluted with elution buffer (50 mM HEPES pH 7.5, 150 mM NaCl, 2.5 mM desthiobiotin). Both the 6His-and Strep-tagged proteins were then further purified by SEC on an ÄKTA Purifier using a Superdex 200 16/600 GL column (GE Healthcare), and simultaneously buffer-exchanged into 50 mM HEPES pH 7.5, 150 mM NaCl.
Protein concentration was quantified spectrophotometrically on a NanoDrop Lite (Thermo Fisher Scientific).
To purify SeMet-labelled PilA and PilC for phasing, the corresponding pET-28b derivatives were transformed in E. coli B834(DE3). A single colony was picked and grown O/N in LB with kanamycin. After back dilution, transformants were grown at 37°C in selective liquid LB, until OD600 reached 0.6-0.7. The cells were pelleted at 8,000 g for 5 min and washed twice with 2 ml of CDM containing no Met. The pellets were then washed with 2 ml of CDM supplemented with 20 mg/ml L-Met (Sigma) and used to inoculate, at 1/200 dilution, 20 ml of CDM supplemented with 20 mg/ml Met.
These cultures were grown O/N at 37°C. Cells were pelleted and washed three times with CDM. Then, the pellets were re-suspended in 20 ml of CDM, supplemented with 20 mg/ml SeMet, and used to inoculate 1 l of CDM supplemented with SeMet. Cells were grown at 37°C until OD600 reached 0.5-0.7. The cultures were cooled down to 16°C, and protein expression was induced by adding 1 mM IPTG and 4 ml of 36 % glucose (w/v). The cultures were further supplemented with another 4 ml of 36 % glucose 2.5 h later. The next day, the cells were harvested, and SeMet-labelled proteins were purified as above.
To purify 15 N and/or 13 C-labelled PilA for NMR analysis, the corresponding pET-28b derivative was transformed in E. coli BL21(DE3). A single colony was used to inoculate a starter culture, which was back-diluted 1/100 into 1 l kanamycinsupplemented LB the following morning. The cultures were grown at 37°C until the OD600 reached 0.9. The cells were harvested by centrifugation at 6,000 g for 10 min at 4°C. The pelleted cells were resuspended in 490 ml M9 salts (3.37 mM Na2HPO4. 2.2 mM KH2PO4, 0.855 mM NaCl, pH 7.2), containing 1 ml 1 M MgSO4 and 250 µl 0.2 M CaCl2. The cell suspension was supplemented with 5 mg vitamin B1, 50 µg/ml kanamycin, 0.5 g 15 NH4Cl and 2 g unlabelled D-glucose (or 13 C-labelled glucose if double isotopic labelling was required) dissolved in 10 ml H2O and filter-sterilised.
The cells were rested at 16°C for 20 min before protein expression was induced O/N with 0.5 mM IPTG. The cells were harvested, lysed, and purified as described above.

Pull-down assays
Pull-down assays were carried out using Dynabeads His-tag Isolation and Pulldown (Invitrogen). The 6His-tagged soluble pilin domains were used as bait, while Streptagged soluble pilins were used as prey. For each pull-down reaction, 25 µl of magnetic beads and 1.5 nmol of protein were used (23 µg for PilA, 75 µg for PilC, and 60 µg PilCΔpilin). The pull-down assays were performed three times for each combination. Prior to the pull-down assays, the bait and prey proteins were mixed in 1 ml of binding buffer (50 mM HEPES pH 7.6, 150 mM NaCl, 10 mM imidazole, 0.1 % Tween-20) and incubated on ice for 1 h. Meanwhile, the magnetic beads were incubated with 700 µl blocking buffer (50 mM HEPES pH 7.6, 150 mM NaCl, 10 mM imidazole, 0.1 % Tween-20, 5 % skim milk) on a rotating wheel at 4°C. After 1 h, the blocking reaction was placed on a magnet for 10 sec to capture the beads, and the flow-through was discarded. The beads were then rinsed twice with 700 µl binding buffer and mixed with 1 ml of bait-prey reaction mixture. Following a 20-min incubation on a rotating wheel at 4°C, the beads were captured on the side of the tube using a magnet, and the flow-through was removed. Beads were washed 10 times by vortexing them with 500 µl washing buffer (50 mM HEPES pH 7.6, 150 mM NaCl, 20 mM imidazole, 0.1 % Tween-20) for 10 sec. After the final wash, the beads were incubated with 100 µl elution buffer (50 mM HEPES pH 7.6, 300 mM NaCl, 500 mM imidazole, 0.1 % Tween-20) for 5 min on a rotating wheel at 4°C. The beads were captured on the side of the tube using a magnet, and the flow-through was carefully transferred to a fresh tube. For each pull-down reaction, input, flow-through, and elution samples were analysed. The samples were mixed with 2x Laemmli Buffer (Bio-Rad), boiled for 5 min at 100°C, and subsequently analysed by immunoblotting.

SDS-PAGE and immunoblotting
SDS-PAGE was carried out using 1x Tris/Glycine/SDS Buffer (Bio-Rad) in a Mini-Protean Tetra cell system (Bio-Rad) for 1 h at 200 V. The Precision Plus Protein All Blue Prestained Protein Standards (Bio-Rad) was used as molecular weight marker and loaded alongside the protein samples. Gels were either stained with Bio-Safe Coomassie (Bio-Rad) and imaged with a Gel Doc EZ Imager (Bio-Rad) or transferred to a membrane and analysed by immunoblotting.
Immunoblotting was done as follows. After proteins were separated by SDS-PAGE, they were transferred onto Amersham Hybond ECL nitrocellulose membrane (GE Healthcare). The wet transfer was carried out for 1 h at 100 V in ice-cold buffer (39 mM glycine, 48 mM Tris base, 0.037 % SDS, 20 % isopropanol). The blotted membranes were blocked for 1 h at room temperature, while shaking, in PBS supplemented with 0.1 % Tween-20 (PBST) containing 5 % (w/v) skim milk powder (VWR). The membranes were then incubated for 1 h with primary antibodies diluted 1/3,000 in 5 % milk PBST. The specific antibodies generated in rabbits against PilE1, PilE2, PilA, PilB, PilC, were previously described (2,3). Following three 10-min washes with PBST, the membranes were incubated for 1 h with an anti-rabbit secondary antibody conjugated to horseradish peroxidase (GE Healthcare), diluted 1/10,000 in PBST. The membranes were washed again three times for 10 min in PBST, dried and developed with Amersham ECL Prime Western Blotting Detection Reagent (GE Healthcare). Protein bands were detected using a ChemiDoc Imaging System (Bio-Rad).

SEC-MALS
SEC-MALS was performed using an ÄKTA Prime system with a S200 10/300 GL column (GE Healthcare) and a MALS detector (Wyatt). The column and the system

NMR assignment and chemical shift perturbations
A sample containing 13 C, 15 N labelled PilA at 1 mM in NMR buffer (10 mM Na2HPO4/NaH2PO4 pH 7, 150 mM NaCl, 5 % D2O) was used for the TROSY-based assignment experiments. Experiments were processed using MddNMR (4) for reconstruction after Non-Uniform Sampling, and NMRPipe (5). Peak picking and assignments were performed in SPARKY (6).

Protein stability assays
To perform long-term protein stability tests, 6His-PilA and 6His-PilC were purified and mixed at 400 µM concentration. The single proteins and the complex -three independent aliquots for each -were kept at 4°C for four weeks. Samples were taken once a week, mixed with 2x Laemmli buffer and analysed by SDS-PAGE/Coomassie staining to reveal protein degradation.
To perform trypsin sensitivity assays, 6His-PilA and 6His-PilC were purified and one hundred µl aliquots of PilA, PilC and the PilA-PilC complex were prepared at 80 µM concentration. The PilA-PilC complex was incubated at 4°C O/N. The next day, the three aliquots were incubated with trypsin (Sigma) at 1/1,000 dilution for 60 min on ice. Samples of 10 µl were taken at 1, 5, 10, 20, 40 and 60 min. The samples were immediately mixed with 2x Laemmli buffer, boiled at 100°C and analysed by SDS-PAGE/Coomassie staining. The trypsin sensitivity assays were performed three times with freshly purified proteins.

Glycan microarrays
The binding specificities of 6His-PilC, 6His-PilCΔpilin and 6His-PilC SK36 -purified and concentrated to 1 mg/ml in 10 mM HEPES pH 7.5, 150 mM NaCl, 5 mM CaCl2were analysed using a NGL-based microarray system (7). The list of glycan probes is given in the Dataset S1. Details of the preparation of the glycan probes and the generation of the microarrays are listed in Table S4. The microarray analyses were performed essentially as described (8). In brief, after blocking of the slides for 1 h with HBS buffer (10 mM HEPES pH 7.4, 150 mM NaCl) containing 1 % (w/v) BSA (Sigma), 0.02 % (w/v) Casein (Pierce), and 10 mM CaCl2, the 6His-tagged PilC proteins were analysed under two conditions. In condition A, the microarrays were overlaid with the 6His-PilC proteins for 90 min as precomplexed protein-antibody complexes. These were prepared by preincubating the His-tagged PilC with mouse monoclonal anti-poly-histidine and biotinylated anti-mouse IgG antibodies (both from Sigma) at a ratio of 1:1.5:1.5 (by weight) and diluted in the blocking solution to provide a final PilC concentration of 50 µg/ml. In condition B, the microarrays were first overlaid with the 6His-PilC proteins at 100 µg/ml. This was followed by incubation with mouse monoclonal anti-His and biotinylated anti-mouse IgG antibodies (both at 10 µg/ml). In both conditions, binding was detected with Alexa Fluor-647-labelled streptavidin (Molecular Probes) at 1 µg/ml for 30 min. All steps were carried out at ambient temperature except for the precomplexation step which was carried out on ice. Imaging and data analysis are described in Table S4.

Bioinformatics and modelling
Protein sequences were routinely analysed using DNA Strider (9). Prediction of protein domains was done by interrogating the InterPro database with InterProScan (10). Molecular visualisation of 3D structures was done using PyMOL (Schrödinger), which was used for generating the figures in this manuscript. The DALI server was used for comparing protein structures in 3D (11). Protein 3D structures were downloaded from the RCSB PDB server. The 3d-SS (12) server was used to superpose 3D protein structures with the STAMP algorithm (13). PDBePISA (14) was used for the exploration of macromolecular interfaces. Modelling was done using AlphaFold (15) and AlphaFold-Multimer (16).      in Dataset S1. For easy comparison of the sialyl glycan probes bound by the two proteins, a "condensed matrix" focused on sialyl glycan probes has been added in Dataset S1. * Signals with large error bars due to artefacts on the array slides.

Fig. S7. Glycan microarray analysis of the glycan-binding activity of PilCΔpilin.
PilCΔpilin is the purified PilC protein without pilin module. The protein was analysed with and without precomplexation with detection antibodies, yielding similar results.

Assay protocol
Microarray analyses were performed essentially as described (7), for modifications of the protocol see "Glycan microarrays" section in the main text.

Glycan description for defined glycans
A broad-spectrum screening microarray containing 672 sequencedefined oligosaccharide probes was used. The probe names, corresponding sequences, and IDs in the international glycan structure repository GlyTouCan (https://glytoucan.org/) are displayed in Dataset S1. These probes represent a subset of a recently generated large screening microarray containing around 900 glycan probes (in-house designation "Array sets 42-56", which will be published elsewhere). The NGL probes are from the collection assembled by the Glycosciences Laboratory (https://glycosciences.med.ic.ac.uk/glycanLibraryList.html). Glycan description for undefined glycans Not relevant.

Printing surface
Description of surface Nitrocellulose-coated glass microarray slides.
Dispensing mechanism Non-contact liquid delivery with four dispensing tips.
Glycan deposition Approximately 0.33 nl was printed per spot. Lipid-linked glycan probes were printed at 2 and 5 fmol per spot in duplicate.

Printing conditions
The printing solutions were all aqueous based. Printing was performed at ambient temperature and relative humidity of 58 %. In addition fro the lipid-linked glycan probes, the "liposome" printing solutions contained 100 pmol/µl of DHPC and cholesterol as lipid carriers (both from Sigma). The concentrations of the lipid-linked glycan probes were 5 and 15 pmol/µl for the 2 and 5 fmol per spot levels, respectively. The printing solutions also contained Cy3 NHS ester (GE Healthcare) at 20 ng/ml (26 fmol/µl) as a marker to monitor the printing process.

Glycan microarray with "map"
Array layout Each array slide contained 16-pad subarrays. Each pad was set up for printing 64 probes maximum, each at 2 levels in duplicate (four spots for one probe in a row); 256 spots (16x16) in total in each pad. The 672 lipid-linked probes in the screening arrays were printed on multiple subarrays for parallel binding analyses.
Glycan identification and quality control The quality control of the glycan microarrays was routinely carried out with a panel of (i) biotinylated plant lectins (Vector Laboratories) including Ricinus communis agglutinin I, Aleuria aurantia lectin, Concanavalin A, and wheat germ agglutinin, (ii) anti-carbohydrate antibodies, and (iii) commercial bacterial adhesins and toxins. Binding results will be shared via the GlyGen glycan array repository currently under development as part of the NIH-funded GlyGen iniatiative (https://www.glygen.org/), which has entered its final testing phase. In the meantime, datasets can be seen in the shared Google folder https://drive.google.com/drive/folders/1hMr-bWX4k3XxBd8FIB8cHbPkHBsxNzkj?usp=sharing. The sialylated glycan probes included in the present glycan array analyses have been extensively validated in previous studies with influenza viruses (20), and a number of viral adhesive proteins including VP1 proteins of polyomaviruses -simian virus 40 (21) and human JC polyomavirus (22) -and the fiber knob of human adenovirus 52 (23).

Detector and data processing
Scanning hardware GenePix 4300A from Molecular Devices (UK).

Data processing
The gpr files were entered into an in-house microarray database using software designed by Mark Stoll for data processing (http://www.beilstein-institut.de/en/publications/proceedings/glyco-2009). No particular normalisation method or statistical analysis was used for the results of the screening arrays.

Data presentation
The microarray binding results are presented as histogram charts in Fig. 7A, Fig. S6 and Fig. S7. The results table with binding scores and relative binding intensities shown as "matrix" are in Dataset S1. This dataset also displays a focused matrix of the bound sialyl glycan probes with their sequences.

Data interpretation
No software or algorithms were used to interpret processed data.

Conclusions
PilC, PilCΔpilin and PilC SK36 bound to sialyl glycans and sulphated GAGs (mainly heparin) NGL probes in the screening array. PilCΔpilin showed much stronger binding to heparin NGL probes relative to sialyl probes. For the full-length proteins, PilC and PilC SK36 , very similar binding patterns were observed to a broad range of sialylated glycans, mainly sialyl α2-3-linked and α2-9-linked polysialic. Among the differences there is the probe NeuAcα-6GalNAc-AO bound strongly by PilC SK36 and weakly by PilC.
Dataset S1. Results of the different glycan array experiments, with detailed list of probes on the array. This includes experiments performed with PilC, PilCΔpilin and PilC SK36 .