Force Spectroscopy of the Plasmodium falciparum Vaccine Candidate Circumsporozoite Protein Suggests a Mechanically Pliable Repeat Region*

The most effective vaccine candidate of malaria is based on the Plasmodium falciparum circumsporozoite protein (CSP), a major surface protein implicated in the structural strength, motility, and immune evasion properties of the infective sporozoites. It is suspected that reversible conformational changes of CSP are required for infection of the mammalian host, but the detailed structure and dynamic properties of CSP remain incompletely understood, limiting our understanding of its function in the infection. Here, we report the structural and mechanical properties of the CSP studied using single-molecule force spectroscopy on several constructs, one including the central region of CSP, which is rich in NANP amino acid repeats (CSPrep), and a second consisting of a near full-length sequence without the signal and anchor hydrophobic domains (CSPΔHP). Our results show that the CSPrep is heterogeneous, with 40% of molecules requiring virtually no mechanical force to unfold (<10 piconewtons (pN)), suggesting that these molecules are mechanically compliant and perhaps act as entropic springs, whereas the remaining 60% are partially structured with low mechanical resistance (∼70 pN). CSPΔHP having multiple force peaks suggests specifically folded domains, with two major populations possibly indicating the open and collapsed forms. Our findings suggest that the overall low mechanical resistance of the repeat region, exposed on the outer surface of the sporozoites, combined with the flexible full-length conformations of CSP, may provide the sporozoites not only with immune evasion properties, but also with lubricating capacity required during its navigation through the mosquito and vertebrate host tissues. We anticipate that these findings would further assist in the design and development of future malarial vaccines.

Malaria remains a major cause of morbidity and mortality throughout the world. According to the World Health Organization (WHO) report, there were an estimated 214 million new cases of malaria and 438 thousand deaths due to malaria alone in 2015 (1). The failure to control the disease effectively stems from pesticide resistance of mosquitoes, drug resistance of malaria parasites, and the lack of an effective vaccine against malaria. At present, the most successful malaria vaccine on trial is RTS,S, which makes use of domains of the circumsporozoite protein (CSP), 3 which is the most abundant surface protein of the infective sporozoites of Plasmodium species (2)(3)(4)(5).
Sporozoites are the infective stage for the vertebrate host and are therefore the appropriate target for a malaria vaccine. The CSP, expressed exclusively in the sporozoite stage, was shown to be the target of protective antibodies (6,7). Early studies on CSP established the presence of a centrally located immunogenic repeat region in the protein from all species of Plasmodium (7,8). It was demonstrated that the central repeat domain of the CSP was the variant and immunodominant domain, with about 90% of the protective antibodies reacting against the repeat domain, with only a small response observed for the flanking regions (9). Simian malaria such as Plasmodium knowlesi (10) and Plasmodium cynomolgi (11,12), as well as the human malaria parasite Plasmodium vivax (13,14), exhibited huge diversity in the repeat sequences of CSP among different strains of these Plasmodium species, but so far all Plasmodium falciparum strains tested show the presence of NANP repeat sequences, indicating its value as a vaccine candidate epitope (3,15).
Gene knock-out studies in P. falciparum have demonstrated that CSP is required for the structural development of sporozoites (16). A recent study demonstrated that the repeat region is also critically required for sporozoite formation and maturation (17). Other than the structural requirement of CSP for sporozoites, in the human host, it is functionally involved in hepatocyte invasion by the parasite, as antibodies directed against CSP prevented sporozoite invasion into hepatocytes (18 -20). The gliding motility of the sporozoites is also vital for infectivity (20 -22). Sporozoites are introduced subcutaneously, and an invasive sporozoite has to squeeze in through the endothelial cell layer before finally infecting a hepatocyte after traversing through Kupffer cells and several hepatocytes (23)(24)(25). These observations raise some important questions: 1) How does the sporozoite protect itself against the mechanical and frictional forces operative during such journeys involving * This work was supported by the Department of Atomic Energy (DAE) and Tata Institute of Fundamental Research (TIFR). The authors declare that they have no conflicts of interest with the contents of this article. □ S This article contains supplemental Figs. S1-S3 and supplemental Table S1. 1 To whom correspondence may be addressed. E-mail: sharma@tifr.res.in. 2 To whom correspondence may be addressed. E-mail: koti@tifr.res.in. several cell penetrations? 2) Does CSP, which forms a dense coat on the surface of the sporozoites, provide some mechanical buffer for such forces? CSP repeats are immunodominant and form immunoevasive cross-reactive domains that elicit a Tcell-independent immune response (8,26). The immunodominant nature of the CSP repeat region would predict that these would be the majorly exposed domains on the sporozoite.
Although the CSP has been studied for over three decades, and is the only parasite protein used in the RTS,S vaccine (the only malaria vaccine under trial so far), very little is known regarding CSP structure. The study of structural and dynamical properties of CSP has been hampered due to the lack of crystal structure, and so far only a structure of (NANP) 3 from the central repeat region has been assessed through NMR studies (27,28). The C-terminal structure is better predicted as it contains the thrombospondin type 1 repeat (TSR) domain, whereas no structure could be obtained for the N-terminal domain (29,30). CSP is proposed to have a rod-like structure anchored at the C terminus (31). Recent studies have also suggested that CSP undergoes reversible conformational changes, between an "open (or non-adhesive)" state and a "collapsed (or adhesive)" state, before the sporozoite invades a hepatocyte (18,32). It was proposed that mechanical force might regulate the conformational change from "collapsed" to "open," although the force required for such a force-induced change remains to be measured. How divergent can such conformations of repeat region or the terminal domains of CSP be? With the idea that a measurement of the strengths of the repeat region and the internally folded domains of CSP would provide an insight into the dispersion of such structures, we prepared tagged P. falciparum CSP (sandwiched between (I27) 3 units) and studied the protein using single-molecule force spectroscopy (SMFS) (33). We observed that the NANP repeat region (CSP rep ) displayed flexibility and heterogeneity in its structure, whereas the near fulllength version without the terminal hydrophobic residues (CSP ⌬HP ) exhibited distinctly folded domains with mechanically weak interdomain interactions.

Results
Polyprotein Construction of Malarial CSP-CSP from P. falciparum is a 397-amino acid protein containing an N-terminal signal peptide, a C-terminal TSR domain and GPI anchor peptide, and an NANP repeat region in the middle (Fig. 1). The NANP repeat region used in our study consists of 172 residues (i.e. 101-272 in the protein sequence), mostly composed of NANP repeats (asparagine, 47%; proline, 25.6%; alanine, 22%; aspartate, 3%; valine, 2.4%). Such proline-rich regions, for example, PEVK segments that are rich in proline, glutamate, valine, and lysine of the muscle protein titin, are found to have unique mechanical properties suited for their environment (34,35). In this study, we cloned the full-length CSP (CSP FL ), the CSP rep , the CSP ⌬HP , and the N-and C-terminal deletion constructs (Fig. 1B). The polyproteins used in this study were constructed using these CSP variants and the I27 domain from the I-band of the giant muscle protein titin (36,37). Each CSP is flanked at both termini by three I27 domains to construct the polyproteins (I27) 3 -CSP-(I27) 3 , where the I27 domain would provide a molecular fingerprint in the SMFS experiments (37,38). The details of the polyprotein construction, expression, and purification are given under "Experimental Procedures." The expressed CSP FL was largely aggregated, and a small amount of soluble protein could be subjected to the SMFS study, but the N-and C-terminal deletion proteins were not soluble at all. The CSP rep and the CSP ⌬HP polyproteins were largely soluble, and the purified proteins were analyzed by SDS-PAGE and Western blotting techniques using anti-(NANP) 5 antibodies (Fig. 2) and were used in SMFS studies.
SMFS Studies on (I27) 3 -CSP rep -(I27) 3 Polyproteins-The polyproteins in phosphate-buffered saline (pH 7.4) were used for SMFS pulling studies. A pulling velocity of 1000 nm/s was kept constant for all the experiments. A representative forceversus-extension (FX) trace of (I27) 3 -CSP rep -(I27) 3 polyprotein (given by 60% population) is shown in Fig. 3A. The FX trace has a sawtooth pattern of eight force peaks with peak force in the range 100 -250 pN. The force peaks were fitted with a wormlike chain (WLC) model of polymer elasticity, and the increment in contour length (⌬L c ) followed by the force peaks was estimated (39). In the FX trace, the ⌬L c followed by the force peaks 2-7 is ϳ27 nm, and their peak force is ϳ200 pN, which are characteristic properties of the I27 domain (36 -38). The last force peak in the FX trace is due to the detachment of the molecule from the cantilever tip or from the substrate. Hence, six I27 domains in the polyprotein serve as a molecular fingerprint of single molecules, and observing more than four force peaks of I27 in FX traces ensures that the sandwiched CSP domains have been subjected to the pulling force (37,38). Therefore, FX traces containing at least four I27 unfolding force peaks were considered for data analysis. Once the I27 signatures have been identified in the FX trace, the remaining features are assigned to the mechanical unraveling of the CSP protein.
In Fig. 3A, we assigned the force peak in the beginning of the FX trace, preceding the I27 sawtooth pattern, to the mechanical unraveling of CSP rep . The ⌬L c of the peak is 36 nm, and its rupture force is ϳ100 pN. This force is much lower than that of the I27, and hence, it precedes the I27 sawtooth pattern in the FX trace. Surprisingly, the ⌬L c of CSP rep is much lower than the expected value from the unraveling of NANP repeats ϳ76 nm (200 aa ϫ 0.38 nm/aa) (40). In Fig. 3A, the ⌬L c of 36 nm means that ϳ20 NANP repeats are compliant and unraveled without any resistance, and the remaining ϳ23 NANP repeats have resisted stretching force of ϳ100 pN before their rupture, giving rise to a 36-nm elongation after the force peak. The distributions of the rupture forces (73 Ϯ 35 pN) and the ⌬L c (44 Ϯ 20 nm) obtained from 55 molecules are given in Fig. 3, B and C. We also made a scatter plot, looking at the correlation between the unfolding force and ⌬L c of CSP rep (Fig. 3D). We found no significant correlation (correlation coefficient, r ϭ Ϫ0.21, and the p value associated with this correlation is 0.12), ruling out any dependence between the unfolding force and ⌬L c . These unfolding force and ⌬L c histograms of CSP rep are much broader than those of the I27 domain (rupture force ϭ 205 Ϯ 23 pN and ⌬L c ϭ 27.1 Ϯ 0.4 nm; see supplemental Fig. S1). A typically folded domain with a well defined structure such as I27 would give a narrow distribution of ⌬L c . The broader width of ⌬L c for CSP rep suggests conformational heterogeneity in NANP repeat structures as compared with any well folded globular proteins such as I27. It also means that the size of the mechanically resistant structures varies because some of the NANP repeats of CSP rep have conformations that require 10 -180 pN force to unravel, whereas other NANP repeats are mechanically compliant and unravel prior to the resisting NANPs (Fig. 3A). This composition varies from molecule to molecule, giving rise to a broader ⌬L c (Fig. 3B). Furthermore, we performed Monte Carlo simulations to see whether the observed ⌬L c distribution of CSP rep in Fig. 3B is compatible  F1 and F2). MWM, molecular weight markers. B, Western blot of (I27) 3 -CSP rep -(I27) 3 using anti-(NANP) 5 monoclonal antibody (lanes F1 and F2). C, Coomassie Brilliant Blue-stained gel of the tagged (I27) 3 -CSP ⌬HP -(I27) 3 F1 and F2). D, Western blot of (I27) 3 -CSP ⌬HP -(I27) 3 using anti-(NANP) 5 F1 and F2).

monoclonal antibody (lanes
with a distributed composition of NANP repeats in a mechanically resistant form (see supplemental Fig. S2). Based on the experimentally observed 60% population of CSP rep (43 NANP repeats), which exhibited mechanical resistance, we assumed a probability of 0.6 for a given NANP repeat to be in mechanical resistant conformation. The observed broad distribution of 17-35 repeats in the resistant conformation from the Monte Carlo simulations is in support of the experimentally observed broad ⌬L c distribution of CSP rep (Fig. 3B and supplemental Fig. S2).
Heterogeneity of CSP rep Conformations-The type of FX trace shown in Fig. 3A is only given by 60% of the molecules. The remaining 40% of the molecules have FX traces as shown in Fig.  3E, where CSP rep did not give any discernible force peak in the FX trace. These FX traces have only force peaks of I27 (regular sawtooth pattern), with a long spacer preceding the I27 sawtooth pattern. We measured the spacer length (L c ) to be 104 Ϯ 13 nm, as shown in Fig. 3F. This long spacer suggests that CSP rep has been unraveled without any mechanical resistance. We compared the elongation in CSP rep (i.e. spacer length) with the corresponding spacer length of the I27 heptamer, (I27) 7 , shown in Fig. 4. In the case of (I27) 7 , it is 45 Ϯ 16 nm as compared with CSP rep , which shows 104 Ϯ 13 nm. The difference in the spacer length between these two cases is ϳ60 nm, which is coming from the mechanical unraveling of the CSP rep . Overall, the CSP rep mechanical properties suggest that the structure of 43 NANP repeats is conformationally heterogeneous wherein 40% of cases do not require force (or Ͻ10 pN) to mechanically unravel them, and in 60% of cases, some of the 43 NANP repeats have a structure that is mechanically resistant, but nevertheless unfolds at weak forces.
Complexity in the Mechanical Properties of (I27) 3 -CSP ⌬HP -(I27) 3 -We tried pulling experiments on the full-length CSP polyprotein (I27) 3 -CSP FL -(I27) 3 . This protein often aggregated after purification, and we could not get single-molecule FX traces with 4 -6 force peaks of the I27 domain. We often got FX

Mechanical Properties of Circumsporozoite Protein (CSP)
FEBRUARY 10, 2017 • VOLUME 292 • NUMBER 6 traces with forces higher than 200 pN and without a clear I27 fingerprint, which indicated aggregation (see supplemental Fig.  S3). The aggregation of CSP FL might be due to the hydrophobic regions at the N and C termini (Fig. 1). We have also made CSP constructs with N-terminal deletion (residues 91-397) and C-terminal deletion (residues 1-290) sandwiched with (I27) 3 units for pulling experiments, but the proteins were insoluble, and hence could not be used for the SMFS study.
The data from polyprotein (I27) 3 -CSP ⌬HP -(I27) 3 in reducing and oxidizing conditions is shown in Fig. 5. For CSP ⌬HP in reducing conditions (buffer containing 5 mM ␤-mercaptoethanol), we often observed multiple force peaks preceding the I27 sawtooth pattern in FX traces (Fig. 5A). Of the 75 recorded FX traces, there are 12 traces with one force peak, 30 traces with two force peaks, 24 traces with three force peaks, and 9 traces with four force peaks of CSP ⌬HP . The corresponding histograms of the ⌬L c of the first force peak with respect to the first I27 force peak is shown in Fig. 5B. We observed a bimodal distribution for ⌬L c , with maxima at ϳ89 nm (25%) and ϳ128 nm (75%). The CSP ⌬HP protein consists of CSP, excluding the terminal hydrophobic sequences (i.e. 369 amino acids as shown in Fig. 1). If we assume that the entire protein is folded and the termini are closer, we would expect a contour length of about ϳ140 nm as per calculation (369 aa ϫ 0.38 nm/aa). It is possible that the 89-nm contour length reflects a partially unfolded protein, whereas the 128-nm contour length suggests that these protein molecules had to be opened up from a relatively compact folded state. Also, the observed contour lengths of CSP ⌬HP are much longer than those observed for CSP rep . The multiple force peaks observed for CSP ⌬HP suggest that the mechanical resistance might come from regions other than NANP repeats (i.e. N-terminal domain, C-terminal domain, or TSR domain). The structure of the TSR domain is known, and it is expected to give ϳ22 nm of ⌬L c upon unraveling (29,30). As shown in Fig.  5, the first two CSP ⌬HP force peaks have ⌬L c values in this range, and it is likely that one of these force peaks is due to the   3 proteins with and without the reducing agent ␤ME (lanes 1 and 2) showed no significant difference in their mobility. The ⌬L c and unfolding force data of the protein in oxidizing conditions (i.e. without ␤ME) are shown in E and F. E, ⌬L c between the first force peak of CSP ⌬HP and the first I27 force peak is found to be a bimodal distribution with Gaussians at 90 Ϯ 8 nm (average Ϯ S.D.), n ϭ 18, and 125 Ϯ 10 nm (average Ϯ S.D.), n ϭ 57. F, distribution of the unfolding forces of all force peaks of CSP ⌬HP . The measured unfolding force is 88 Ϯ 52 pN (average Ϯ S.D.), n ϭ 177 (see "Results" for details).
TSR domain unraveling. It must be noted, however, that we cannot separate TSR domain unfolding events from others in the FX traces. We have also performed experiments in oxidizing conditions (i.e. without the reducing agent in the buffer) to see whether the disulfide bonds of the TSR domain have any effect on the mechanical properties (Fig. 1A) (29,30). The mobility of the CSP ⌬HP polyprotein is similar in reducing and oxidizing conditions as seen on SDS-PAGE (Fig. 5D), consistent with the earlier studies (41). The SMFS experiments on the CSP ⌬HP polyprotein in oxidizing conditions showed that irrespective of the reducing condition, the two populations of proteins with two sets of contour lengths remain virtually unchanged and demonstrate no significant changes in their mechanical properties (Fig. 5E). Interestingly, although the forces required for stretching the proteins remain broadly the same, we observe that the oxidized disulfide-bonded CSP ⌬HP polyprotein can be unfolded with greater ease (Fig. 5F).
Furthermore, it is known that CSP may undergo reversible conformational changes between adhesive (collapsed) and nonadhesive (open) conformations during the sporozoite's journey from mosquito midgut to mammalian liver (18,32). In the collapsed conformation, it was proposed that there could be an intermolecular interaction between the N-terminal and C-terminal domains of CSP, which needs to be broken to change its conformation to the open state to interact with and invade hepatocytes (32). It was speculated that the conformational change might be due to a mechanical signal. Our experimental results are consistent with these two proposed populations. Based on our observation of a bimodal distribution for ⌬L c of CSP ⌬HP , we can say that the protein might be in both collapsed and open conformations where the collapsed conformation would require a rupture force and gives a long ⌬L c , whereas the open conformation is already elongated (i.e. the end-to-end distance of the conformation is large) and gives a short ⌬L c upon stretching. Our studies showed that about 75% of the proteins are present in the less unfolded state, whereas 25% of the proteins possess the straight structure. This distribution of the population is independent of the reducing agent, confirming that the covalent disulfide bonds in the TSR domains are unlikely to play a role in the unfolding of the molecules observed under our pulling experiments (Fig. 5), which explore largely the non-covalent interactions Overall, our single-molecule measurements give direct evidence for the existence of two populations, possibly the open and collapsed conformations of CSP.

Discussion
What Is the CSP rep Conformation?-Our SMFS experimental data suggest that CSP rep is conformationally heterogeneous: some of the NANP repeats in a given protein are mechanically resistant, whereas the other repeats are mechanically compliant. The conformations that give observable force peaks might have short stretches of NANP repeats in the resisting structures (Fig. 6A). X-ray structure of NANP and NMR structure of (NANP) 3 suggested a super-helical structure for the NANP repeats (28,42). It is possible that the super-helical structure, as suggested by previous NMR and X-ray data, might resist stretching forces before unraveling. So does the super-helical structure persist over all the NANP repeats? It seems that this is not the case; otherwise, all the CSP rep molecules would have a definite well defined unique conformation, and all the FX traces would have looked similar. On the other hand, the conformations that do not possess long range persistent structure would not give any observable force peak within the instrumental resolution, and they could be unraveled at low forces (Ͻ10 pN) (Fig. 6A). Also, the stabilizing force in the NANP super-helical structures is not the strong H-bonding as in the case of ␤-sheets and other helices in protein structures (31,42). As proline is a cyclic amino acid, it does not participate in the H-bonding to stabilize helical structures. Hence, the helicity of the NANP repeats does not persist (or extend) for longer lengths within the 43 repeats of CSP rep but lasts only over a few repeats, as observed in the NMR and X-ray studies. However, whenever the super-helical structure is formed, it might require force to unravel it, as observed in our experiments.
Interestingly, there are other proline-rich regions in proteins such as the PEVK region of giant muscle protein titin, where the proline content is 25%, just as in NANP (34,35). PEVK regions are not known to have any structure and provide muscle-passive elasticity, and they do not possess any mechanical resistance and unravel at very low forces (34,35). The fingerprint of PEVK regions is akin to the 40% of FX traces of CSP rep in our study, where they unraveled at very low forces. This further supports our observation of the heterogeneity of conformations in CSP rep . Unlike PEVK, where the entire region is devoid of structure and is mechanically compliant, CSP rep has a mixture of super-helical structure and unstructured regions distributed over the repeat region. The CSP repeats are unique in nature because they all contain prolines at regular intervals in different Plasmodium species and strains (supplemental Table S1). Proline-rich sequences are also known as helix and ␤-sheet breakers, and overall they provide conformationally rigid structure to protein, i.e. polyproline II (PPII) structure (43)(44)(45). However, the structural properties of these repeat sequences in proteins have not been studied in detail as they are difficult to crystallize.
Role of Mechanical Flexibility of CSP rep in Sporozoite Motility-The conformational heterogeneity and the sparsely populated super-helical structure along the CSP rep might have important roles in the motility of the sporozoite. It has been known that CSP covers sporozoite's surface densely and that it is essential for sporozoite's assembly (16). As the sporozoite navigates through the mechanically rigid environment of salivary glands in the mosquito, and through endothelial cells of host vasculature, Kupffer cells, and hepatocytes in the host liver, the densely populated CSP molecules on the sporozoite surface might respond through the mechanical response of CSP rep . When the parasite encounters a rigid/stiff environment during its navigation, the local forces experienced by the CSP molecules due to the parasite squeezing through the harsh environment would stretch the CSP rep and hence may release the "local mechanical stress" by unraveling the super-helical structures and attain their conformations quickly when the stress (or local stretching force) on the molecules is released. Overall, our studies at the single-molecule level suggest that the flexible structure of the NANP repeats may play a key role to sustain the mechanical stress/forces experienced by the densely populated Mechanical Properties of Circumsporozoite Protein (CSP) FEBRUARY (18,32). Such a reversible conformational change, from a collapsed to open state, is thought to mask some epitopes in the N-and C-terminal domains until the sporozoite interacts with the appropriate host/vector tissues (32). Our results showing the heterogeneity of NANP repeats suggest that they might also contribute to the regulation of the collapsed and open conformations of CSP (Fig. 6B). Also, the interdomain or intermolecular interactions between N-terminal and C-terminal domains may stabilize their conformations (32). The disulfide-containing TSR domain in CSP has been reported to be strikingly divergent in sequence and structure from other TSR domains, including those in other Plasmodium surface proteins (30). The thin and elongated prototypical TSR domain was found to be shortened on its long axis and widened on the other axis, as the N terminus of the CSP TSR domain was located on the same end of the domain as the C terminus, overall constituting a well formed compact structure that did not bind to heparin (30). Our results indicate that under oxidized conditions, where the TSR domain would be well folded, the overall CSP gets to be more unstructured and easier to unfold. Because the TSR domain sequence is extremely conserved in all Plasmodium species, it would appear that the consequent enhanced flexibility of the CSP under non-reducing conditions might be a widespread property of CSP across all species. How the TSR is oriented on the P. falciparum sporozoite surface, between the NANP repeats and C-terminal GPI anchor, remains to be resolved. Our studies show that the conformations of CSP are flexible, and they suggest that weak stretching mechanical forces could be used to change from collapsed to open conformation of CSP on the sporozoite surface (Fig. 6C).
For motility as well as invasion, the sporozoites require a large degree of flexibility. Because CSP constitutes a large fraction of the sporozoite membrane proteins, and because our molecular data suggest the requirement of very low forces in stretching the CSP rep , we speculate that CSP could provide the lubrication to the sporozoites during motility and cell invasion. During this navigation, if the force is high enough locally, then the individual CSP molecules might even be dislodged or pulled out of the membrane as shown by previous studies of shedding CSP from the sporozoite during its movement in liver tissue (24). This would only change the CSP concentration minimally on the parasite surface as the CSP protein is densely populated. It is known from previous studies that newly synthesized CSP is added or replenished from the apical end of sporozoite and that the shedding occurs at the distal end during its motility (24,46).
In conclusion, our single-molecule force spectroscopy studies on the NANP repeats and near full-length CSP of P. falciparum indicate the heterogeneity of conformations, where some molecules are entirely mechanically compliant, whereas others have both mechanically compliant and mechanically resistant conformations. In the larger propensity of collapsed structures of CSP, the flexible repeat regions would be exposed on the surface and would thus be available for generating the immunodominant antibody response. These mechanical properties of CSP repeats combined with their high density on the sporozoite surface might provide the lubrication required for its navigation through host tissues, whereas the open conformation of CSP molecules would be able to facilitate tissue invasion through interactions with the host receptors. mined using the equipartition theorem (49). All the experiments were performed at room temperature at a constant pulling speed of 1000 nm/s. The concentration of the polyproteins used in the experiment varied between 2 and 10 M. For pulling studies, 50 -70 l of protein solution in phosphate-buffered saline (pH 7.4) with 5 mM ␤-mercaptoethanol were added onto the gold-coated coverslip and then left for 10 min before starting pulling experiments. The pulling experiments in oxidizing conditions were performed without any reducing agent in the buffer. The proteins were non-specifically adsorbed onto the coverslip. The proteins were adsorbed onto the cantilever tip by pushing the tip onto the coverslip with a force of 1-2 nN for 0.1-1 s. On retraction, the proteins were stretched if they adsorbed onto the coverslip as well as the tip surface. The single molecules were identified by the number of I27 domains (i.e. between 1 and 6) unfolding in the sawtooth pattern with their characteristic unfolding contour length (ϳ27 nm) and unfolding forces (ϳ200 pN) in the FX traces. Only those FX traces containing at least four I27 force peaks were considered for data analysis as this guarantees the stretching of CSP sandwiched by (I27) 3 on both sides.
Data Analysis-The following equation of the WLC model of polymer elasticity was used to fit the FX traces (39) where F, p, L, k B , and T denote force, persistence length, contour length, Boltzmann constant, and absolute temperature, respectively.
Author Contributions-S. R. K. A. and S. S. conceived and coordinated the study and wrote the paper. A. P. P. designed, performed, and analyzed the experiments. All authors reviewed the results and approved the final version of the manuscript.