Binding of the kringle‐2 domain of human plasminogen to streptococcal PAM‐type M‐protein causes dissociation of PAM dimers

Abstract The direct binding of human plasminogen (hPg), via its kringle‐2 domain (K2hPg), to streptococcal M‐protein (PAM), largely contributes to the pathogenesis of Pattern D Group A Streptococcus pyogenes (GAS). However, the mechanism of complex formation is unknown. In a system consisting of a Class II PAM from Pattern D GAS isolate NS88.2 (PAMNS88.2), with one K2hPg binding a‐repeat in its A‐domain, we employed biophysical techniques to analyze the mechanism of the K2hPg/PAMNS88.2 interaction. We show that apo‐PAMNS88.2 is a coiled‐coil homodimer (M.Wt. ~80 kDa) at 4°C–25°C, and is monomeric (M.Wt. ~40 kDa) at 37°C, demonstrating a temperature‐dependent dissociation of PAMNS88.2 over a narrow temperature range. PAMNS88.2 displayed a single tight binding site for K2hPg at 4°C, which progressively increased at 25°C through 37°C. We isolated the K2hPg/PAMNS88.2 complexes at 4°C, 25°C, and 37°C and found molecular weights of ~50 kDa at each temperature, corresponding to a 1:1 (m:m) K2hPg/PAMNS88.2 monomer complex. hPg activation experiments by streptokinase demonstrated that the hPg/PAMNS88.2 monomer complexes are fully functional. The data show that PAM dimers dissociate into functional monomers at physiological temperatures or when presented with the active hPg module (K2hPg) showing that PAM is a functional monomer at 37°C.

expressed by GAS that covalently binds to the cell wall of Grampositive bacteria and extends beyond the capsular cell surface as a hair-like projection. This protein is the basis of emm gene serotyping for the >250 strains of GAS that have been identified (Fischetti, 1989;Phillips et al., 1981). Functionally, M-protein is a versatile virulence factor that participates in several resistance mechanisms aimed at the elimination of GAS. These extended surface M-protein appendages initially mediate strong attachment of bacteria to keratinocytes and epithelial cells to initiate infection (Courtney et al., 1992;Cue et al., 1998;Ellen & Gibbons, 1972). Some serotypically distinct Mproteins, e.g., M1-, M3 -, and M6-proteins, expressed by the emm1, emm3, and emm6 genes, respectively, directly bind to extracellular matrix components (ECM), such as fibronectin, to promote bacterial colonization (Cue et al., 2001;Schmidt et al., 1993). In addition to the mediation of bacterial adhesion to host cells, blockage of the complement pathways (Agrahari et al., 2013;Buffalo et al., 2016;Prasad et al., 2015) and upregulation of the fibrinolytic system on the GAS surface (Glinton et al., 2017) are other major functions of some M-proteins.

Plasminogen-binding Group A streptococcal M-protein (PAM) is
a specialized M-protein found on skin-tropic Pattern D GAS strains (Bessen & Lizano, 2010) that directly binds to human plasminogen (hPg) with nM-scale affinity, and, accordingly, accumulates hPg on the GAS cell surface. There are several sequence variations of PAM-type M-proteins on different strains of Pattern D GAS (Qiu et al., 2018), but all bind tightly to hPg and stimulate hPg activation by the GAS-secreted streptokinase (SK2b) that is coinherited with PAMs on Pattern D strains (Zhang et al., 2012). These steps result in activation of hPg bound to GAS cells in this manner and thus generate a functional protease, plasmin (hPm), on the GAS surface. These proteolytically competent cells degrade the fibrin formed as a host response to infection that encapsulates invading GAS cells, thereby first promoting the release of the bacteria and, ultimately, dissemination of these pathogens by digesting cellular tight junction proteins and degrading the extracellular matrix (Loof et al., 2014).
Structurally, all known PAMs are domain-assembled proteins composed consecutively of a 41-residue NH 2 -terminal signal peptide, a hypervariable region (HVR), A-, B-, C-, and D-domains, a Pro/ Gly-rich region, a LPXTG motif recognized by sortase A, a COOHterminal transmembrane anchor, and a very short C-terminal extracellular region (Fischetti et al., 1988;Smeesters et al., 2010). In Pattern D strains, PAM subdomains, viz., the a1 and/or a2 repeats of its A-domain, specifically bind to the kringle-2 module of hPg (K2 hPg ; Berge & Sjobring, 1993;Rios-Steiner et al., 2001). According to the amino acid sequence and the number of tandem repeats in the PAM A-domain, we previously categorized PAMs into three classes: Class I and Class III (both containing a1-and a2-repeats) and Class II (a2-repeat only; Qiu et al., 2018). Each class of PAM tightly associates with hPg in the low-nM range. We hence established a PAM structure-function model of solution phase PAM to elucidate the significant participation of A-domain α-helices in hPg-binding (Qiu et al., 2020).
The C-domain of PAM contains the most abundant α-helices among all domains. These regions of PAM drive parallel coiled-coil dimerization via hydrophobic clustering that is mediated by Leu-and Val-containing heptad repeats (Fischetti et al., 1988(Fischetti et al., , 1990Qiu et al., 2018). This non-covalent dimerization mode is stable at ≤25°C but is disrupted by a small increase in temperature in the physiological range (Qiu et al., 2019). At 37°C, most PAM dimers dissociate, leading to a large portion of monomers, irrespective of the PAM classifications (Qiu et al., 2019). This suggests that the dimer is not highly stable. Nonetheless, it is thus far unknown whether K2 hPg -binding is another event leading to the dissociation of PAM dimers. This investigation addresses this uncertainty.  (McKay et al., 2004) begins at the first amino acid residue in the HVR (D 1 of the protein after cleavage of a 41-residue signal peptide, and terminates at the last residue of the Pro/Gly-rich region (Q 344 ), slightly upstream of the L 345 PSTG sortase A (srtA) cleavage motif region were excluded from construction of pam expression plasmids (Chandrahas et al., 2015;Glinton et al., 2017;Qiu et al., 2018Qiu et al., , 2019Zhang et al., 2012). AGL55, cloned from the PAM NS88.2 gene, encompasses A 54 -Y 108 of PAM NS88.2 . This construct contains 14 COOH-terminal residues of the HVR, the entire hPg binding a-repeat (D 68 -E 85 ), and 24 NH 2terminal residues of the downstream B-domain (Qiu et al., 2018(Qiu et al., , 2019.

| Protein expression and purification
Recombinant AGL55 and PAM NS88.2 contained NH 2 -and COOHterminal His 6 tags, respectively, for ease in purification by Ni + -based affinity chromatography. Detailed procedures with respect to expression in Escherichia coli BL21 (DE3) cells and purification of PAM proteins and peptides have been reported (Qiu et al., 2018;Yuan et al., 2017).
K2 hPg (E 164 E-C 166 -C 243 of mature hPg) was cloned from the hpg gene, expressed in Pichia pastoris, and purified on Lys-Sepharose (Nilsen et al., 1999). There were several additional residues at the N-and C-termini of K2 hPg which were derived from the nature of the processing of the expression plasmid in P. pastoris cells.

| Isothermal titration calorimetry (ITC)
All titrations were conducted in a VP-ITC Microcalorimeter (Malvern Panalytical, UK) at the designated temperatures. The proteins were dissolved in 50 mM sodium phosphate/100 mM NaCl, pH 7.4. The During the entire titration, the syringe stirred at a rate of 329 rpm.
The differential power (DP) was initially set at 10 μcal/sec. When the actual DP stabilized in the range of 10 ± 1 μcal/s, the titration automatically started and the data were recorded in real-time.

| Analytical ultracentrifugation (AUC)
Sedimentation velocity (SV) and sedimentation equilibrium (SE) experiments were conducted using a Beckman Optima XL-I analytical ultracentrifuge in the absorbance optics mode. SV experiments were conducted for 17-44 h depending on the temperature (4, 15, 25°C, or 37°C) with a rotor speed of 170,700 g using 1.2 cm two-channel centerpieces. Prior to the loading step, all proteins were dissolved in 50 mM sodium phosphate/100 mM NaCl, pH 7.4 (sample buffer).
PAM NS88.2 was diluted to A 230 nm of ~0.15, 0.3, and 0.6. In each assembled cell, 420 μl sample buffer, was injected as the reference into the left channel, and 400 μl of the protein solution was loaded into the right channel. Scans were recorded every three mins and 300-700 scans were collected for each SV experiment. The data were analyzed by Sedfit (version 14.1) using a continuous c(s) distribution model (Schuck et al., 2014). The values of viscosity and density of the buffer solutions were obtained using Sednterp (http://www.rasmb. org/sednterp). The normalized c(s) vs. S 20 ,w plots were generated using public domain software, GUSSI version 1.4.2 (https://www. utsou thwes tern.edu/labs/mbr.software).
For AUC-SE experiments, PAM NS88.2 (A 280nm ~0.125) and purified PAM NS88.2 /K2 hPg (A 280 nm ~0.3) were loaded into the sample well of a two-sector centerpiece. The sample channel contained 150 μL protein while the reference channel was loaded with 160 μl sample buffer. All experiments were conducted at 4°C, 25°C and 37°C. Scans were recorded hourly at 26,100 g and 32,200 g until equilibrium was achieved, as evaluated by the lack of change of the concentration gradient with time. The equilibrium data were analyzed for molecular weight using Optima XL-A/XL-I software (Beckman Coulter).

| Circular dichroism (CD)
Far-UV CD spectral measurements in the wavelength range of 195-250 nm were conducted using a Jasco J-815 circular dichroism spectropolarimeter and a 1 mm path-length cuvette. The buffer used was 50 mM sodium phosphate, pH 7.4, at 4°C, 25°C, and 37°C. To observe the influence of K2 hPg on the 2°structure of PAM NS88.2 and AGL55, the CD spectra of these proteins/peptides were mixed with excess K2 hPg and the spectra were collected. The spectrum of K2 hPg , alone, which is non-helical, was similarly obtained and subtracted from that of the complex. An average of three replicate scans, from which an average buffer scan was subtracted, were collected. For the AGL55 experiments, the concentration of AGL55 was 10 µM, and that of K2 hPg , when present, was 12.5 µM. For the PAM NS88.2 experiments, the concentration of PAM NS88.2 was 0.3 µM and that of K2 hPg , when present, was 1.0 µM.
Mean residue ellipticity ([θ]) of the proteins/peptides were calculated using: [θ] = (θ obs × MRW)/(l × c), where θ obs is the observed signal in mdeg, MRW is the mean residue weight in g/mol, l is the path length in mm, and c is the protein concentration in mg/ ml (Greenfield, 2006). The fractional α-helical content (f H ) was estimated from: [θ] 222 nm = −30,300f H −2,340, in which f H refers to the fraction of helices in the protein (Chen et al., 1972). Very little interference was obtained from K2 hPg at this wavelength since K2 hPg does not contain α-helical regions (Wang et al., 2010).

| Size exclusion chromatography/multi detection system
An Agilent 1260 II HPLC system (Agilent Technologies), which consists of a variable wavelength detector, a quaternary pump, an online vacuum degasser, and an autosampler, was used in tandem with a multiangle light scattering (MALS) detector (DAWN-HELEOS; Wyatt Technology), and a differential refractive index detector (Optilab ® T-Rex; Wyatt Technology). The mobile phase was 50 mM Na-phosphate/100 mM NaCl, pH 7.4 (PBS). Prior to injection, the protein samples were passed through a membrane filter (pore size 0.1 μm; Supelco Analytical). The sample volume injected into the channel was 100 μl, and the flow rate was set at 0.5 ml/min. Data acquisition and evaluation were accomplished using Astra 7 software (Wyatt Technology Corporation).
The stoichiometry of binding between the AGL55 and K2 hPg in the two-component complex was determined using the Protein Conjugate Analysis Method in Astra 7 software (Wyatt Technology).
This method, based on the unique specific refractive index increments (dη/dc) and extinction coefficients of each protein enables accurate determination of the components of the complex along molecular mass distributions.

| hPg activation assays
The assays were performed at 25°C, and 37°C in 96-well microti-  Figure 1A) contains one a-repeat and we expect that this peptide will contain one binding site for K2 hPg ( Figure 1B).
To initially determine whether a K2 hPg /AGL55 complex was formed at 25°C, a size exclusion chromatography (SEC) system was employed, which consisted of a multi-detection system (ultraviolet, refractive index, multi-angle light scattering) (Rebolj et al., 2012). From the deconvoluted fractogram ( At all temperatures, the binding stoichiometry is 1:1 (m:m).
The interaction of PAM and hPg is dependent on an α-helix within the a-repeats of PAM for optimal organization of the amino acid side chains required for binding (Qiu et al., 2020;Rios-Steiner et al., 2001;Wang et al., 2010). Thus, we examined by circular dichroism (CD) the 2°structure of AGL55 at 4°C, 25°C, and 37°C to assess whether an α-helical structure is maintained at these temperatures. The far-UV CD ( Figure 3A) shows that, while helical structures are present at 4°C and 25°C, there is a loss of the α-helix of AGL55 at 37°C, in agreement with other studies on a different M-protein (Nilson et al., 1995). Upon binding to K2 hPg , the α-helical nature of AGL55 ( Figure 3B-D) is enhanced at all three temperatures, and that the presence of K2 hPg does not interfere with these spectra ( Figure 3E). Thus, K2 hPg can induce a conformational change in AGL55 that stabilizes the complex at each temperature studied.

| Binding of K2 hPg to PAM NS88.2 results in dissociation of PAM dimers into monomers
We next turned our attention to the nature of the interaction of K2 hPg with PAM NS88.2 . To assess the molecular species of F I G U R E 1 Schematic depiction of the proteins. (A) PAM NS88.2 . The genomic DNA of PAM NS88.2 was cloned from GAS isolate NS88.2, inserted into the expression plasmid, pET-28a, and expressed in E. coli BL21 cells. The protein was purified by Ni + -based affinity chromatography. AGL55 was constructed as [MGSS(H) 6 GB1-(thrombin cleavage peptide-LVPR'GS)-AGL55]. After expression in E. coli and purification using Ni + -based affinity chromatography, the peptide was cleaved by thrombin (at R'G) and repurified by the same column as the column pass-through. GS-AGL55 was the remaining peptide. (B) K2 hPg , consisting of residues C 166 -C 243 , preceded by EE from hPg, was cloned from the hPg gene, inserted into expression plasmid pPIC9K, expressed in P. pastoris cells, and purified by Lys-Sepharose affinity chromatography. The (YVEFS) residues at the N-terminus, as well as the (AA) at the C-terminus, originated from the poly-linker sites of the expression plasmid  (Cedervall et al., 1995;Nilson et al., 1995) and PAM NS88.2 (Qiu et al., 2019) that show that a small temperature increase within a narrow range of nondenaturing temperatures destabilizes the dimeric M-protein and a similar protein, Protein H. In concert with this, the S°2 0,w values of dimeric PAM NS88.2 show single distributions of 2.6-2.8S at 4°C, 15°C, and 25°C, and was reduced to 2.1S at 37°C, also with a single S value distribution ( Figure 4D). This is consistent with a molecular weight reduction to a monomer at 37°C. To assess whether the dissociation of the PAM dimers is related to the helical nature of PAM, we performed far UV CD scans of PAM at these temperatures ( Figure 4e). Indeed, a large reduction of αhelix is seen for PAM NS88.2 at 37°C (from originally ~65% at 4°C and 25°C to ~30% at 37°C) as reflected by a less negative mean residue ellipticity value at 222 nm, when compared to the ellipticity values at 4°C and 25°C.
Given this conclusion of a weakly associated dimer for PAM NS88.2 , experiments were performed to determine whether K2 hPg binding to PAM NS88.2 was also capable of affecting the quaternary structure of PAM. When a mixture of 12 µM K2 hPg /6 µM PAM NS88.2 (based on the monomer M. Wt. of PAM NS88.2 ) is subjected to SV experiments at each of these temperatures, two S°2 0,w distributions are present. Figure 5A illustrates the S°2 0,w data for the free K2 hPg with an S°2 0,w of 1.6 S at 4°C, 15°C, 25°C, and 37°C. Figure 5B shows the SV data obtained for the 2/1 (m/m) mixture. Here, two S°2 0,w clusters are observed. Cluster 1 is the free K2 hPg in the mixture at all three temperatures. In Cluster 2 the increases in S°2 0,w values from ~2.6S (from Figure 4D) to ~3.3S at 4°C and 25°C and from ~2.1S (from Figure 4D) to ~2.7S at 37°C is due to complex formation between K2 hPg and PAM NS88.2 . The data obtained at 37°C demonstrate that the monomeric form of PAM NS88.2 , with a global reduction in α-helix, will interact with K2 hPg with strong affinity. However, whether the S°2 0,w values at 4°C and 25°C are monomers or dimers complexed with K2 hPg is not settled to this point. In these latter cases, the increase in S°2 0,w values could indicate that the dimer binds two K2 hPg peptides or that a monomer binds a single K2 hPg accompanied by a less rigid coil, each of which would lead to increased S°2 0,w values upon F I G U R E 2 Complex formation of AGL55 with K2 hPg . (A) RI fractograms representing the K2 hPg /AGL55 complex, together with the M. Wt. versus elution volume; black for the K2 hPg /AGL55 complex, blue for the K2 hPg constituent of the complex, and red for the AGL55 constituent of the complex. The protein complex (100 µl) complex was injected onto a Wyatt -030S5 column (7.6 mm × 300 mm, 5 µM, 300 Å), equilibrated, and eluted with PBS, pH 7.4, at a flow rate of 0.5 ml/min. The M. Wt. of the complex of 16.2 kDa is consistent with a 1:1 complex of K2 hPg (M. Wt. of 10.2 kDa) and AGL55 (M. Wt. 6.6 kDa). (B-D). K2 hPg was titrated into AGL55 at 4°C (b), 25°C (C), and 37°C (D). The heat changes, measured by ITC, accompanying each injection are plotted against the molar ratio of K2 hPg to AGL55. The data points are in black and the best-fit lines for these points are shown in red. The experimental data were best-fit using the hetero-association model of K2hPg and AGL55 in SEDPHAT, which provided values of N, Kd, and ΔH for the K2 hPg /AGL55 interaction  ITC was employed to measure the stoichiometry and dissociation constants of K2 hPg to PAM NS88.2 . To more fully interpret the binding data, we also relied on biophysical data. For example, it was important to understand that the final product of the interaction was a 1:1 (m:m) complex of K2 hPg and PAM NS88.2 monomer. ITC titration data were employed at 4°C and the data are shown in Figure 7A. The molar ratio in the plot was based on the molecular weight of the PAM monomer.
The curve is best-fit in Sedphat using the simple hetero-association model of K2 hPg and PAM monomer ( Figure 7A). The best interpretation of the data is that K2 hPg first binds to the PAM dimer, which results in dissociation of the dimer and binding to the monomer. The fact that one simple titration curve is seen indicates that the binding of K2 hPg occurs with a similar binding constant to the dimer and monomer. The resulting Kd of 0.7 nM indicates a very tight binding of K2 to PAM.
Similarly, at 37°C, PAM NS88.2 exists as a monomer and is also best-fit by the simple hetero-association model of binding ( Figure 7C). Again, the titration curve is relatively simple and is deconvoluted to a Kd of 181 nM. The ΔH of −13.7 kcal/mol at 4°C is elevated to −37 kcal/mol at 37°C and this exothermic reaction is the foundation for the elevation of the Kd at this higher temperature. At 25°C, the titration curve is more complex (Figure 7b). The major final inflection of the titration curve is B to A monomer, as is evident from the biophysical data which we fit a hetero-association model of A + B. The Kd of the single binding site increases to 2.7 nM. The initial phase of the titration (not fit) has several possible explanations, the simplest of which is that there is initial tight binding to dimer, which dissociates to a slightly weaker binding monomer (with an accompanying heat of dissociation). While this is somewhat unclear, these experiments have revealed the nature of the binding of K2 hPg to the PAM NS88.2 monomer, which is the final product showing the lack of α-helices in this domain. The concentration of AGL55 was 10 mM and the K2 hPg was 12.5 mM when present. The buffer was 50 mM Na-phosphate, pH 7.4 at each temperature of the binding. The results that we obtain for the strong hPg-type ligand binding to PAM-type M-proteins at 37°C are opposite to the very weak binding of fibrinogen, IgG, and albumin binding to M1-monomers at 37°C (Cedervall et al., 1995), thus bringing into question the biological relevance of binding of these ligands to M1 at the temperature of 37°C at which GAS infections occur.  The data of Figure 8A-D show that SK2b which is coinherited by GAS strains (Pattern D) that express PAM (Zhang et al., 2012) is inefficient in generating plasmin (hPm) in the absence of PAM.

| DISCUSS ION
Most molecular models for the structure of surface fibrous Mproteins of S. pyogenes are widely accepted to be centered on a coiled-coil α-helical rich dimer of ~60 nm in length covalently bound at their C-termini to the cell wall by a transpeptidation step catalyzed by sortase A. This is based on the presence of characteristic seven residue heptad repeats, a very early example being that of the coiled-coil of α-tropomyosin (Cohen & Parry, 1990;Parry, 1975).
However, based on the data presented herein, this model requires further consideration as it applies to its biological relevance.
Secretion of M-protein from the cytoplasm of GAS is via the single functional membrane microdomain, the ExPortal, which contains the SecA translocon, anionic lipids, and accessory proteins, e.g., membrane-associated chaperones (e.g., Htra), and sortases (Rosch & Caparon, 2004). Retention of M-protein, perhaps by the interaction of its positively charged cytoplasmic tail with anionic phospholipids in the ExPortal, allows time for unfolded M-protein from the cytoplasm to properly mature and for enzymes such as sortase A to function to anchor M-protein to the cell wall peptidoglycan. Thus, the ExPortal couples protein secretion with maturation. Regarding its subcellular distribution, M-protein first appears in the septum and then becomes distributed over the entire cell surface. It thus appears that information contained in the signal sequence directs the distribution of this protein to its subcellular location (Carlsson et al., 2006).
While M-protein secretion appears to be a highly organized se- Lastly, Pattern D M-protein monomers are functional with regard to stimulation of hPg activation by the coinherited SK2b, which does not activate hPg in solution. This step allows hPm to form and remain bound to the cell surface. The bound hPm is resistant to inactivation by its natural inhibitor, α2-antiplasmin, and thus persists on the bacterial surface to aid in its dissemination.
In conclusion, coiled-coil dimeric models of M-protein have been proposed in the past, but these only appear to apply to M-proteins F I G U R E 8 Stimulation of the SK2binduced activation of hPg by PAM NS88.2 at different temperatures. An unprocessed activation titration of hPg by SK2b at (A) 25°C and (B) 37°C. A typical assay mixture contained hPg (200 nM), various concentrations of PAM NS88.2 (indicated on the graphs) and S2251 (0.25 mM) in 10 mM Na-Hepes/150 mM NaCl, pH 7.4. The reaction was accelerated by adding SK2b (5 nM). The release of p-nitroaniline from S2251 from the continuously generated hPm was monitored at 405 nm for up to 120 min. (C) and (D) Initial velocities were calculated from plots of A 405nm vs t 2 and plotted as a function of the PAM NS88.2 concentrations in solution at temperatures up to ~25°C. While such models are highly useful for understanding the protein chemistry that allows for a-helical coiled-coils to form, it is unlikely that these models are appropriate for an understanding of the biology of M-proteins on bacterial surfaces, which appear to be monomers with reduced helical content.

E TH I C S S TATEM ENT
None required.

ACK N OWLED G M ENTS
We thank Dr. Giselle Jacobson of the Biophysics Instrumentation Core Facility, University of Notre Dame, for her support in the CD experiment. This work was supported by Grant HL013423 from the N.I.H.

CO N FLI C T O F I NTE R E S T
None declared.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data are provided in full in this article.