Serine 106 preserves the tertiary structure, function, and stability of a cyclophilin from Staphylococcus aureus

Abstract SaCyp, a staphylococcal cyclophilin involved in both protein folding and pathogenesis, has a Ser residue at position 106 and a Trp residue at position 136. While Ser 106 of SaCyp aligned with a cyclosporin A (CsA) binding Ala residue, its Trp 136 aligned with a Trp or a Phe residue of most other cyclophilins. To demonstrate the exact roles of Ser 106 and Trp 136 in SaCyp, we have elaborately studied rCyp[S106A] and rCyp[W136A], two-point mutants of a recombinant SaCyp (rCyp) harboring an Ala substitution at positions 106 and 136, respectively. Of the mutants, rCyp[W136A] showed the rCyp-like CsA binding affinity and peptidyl-prolyl cis-trans isomerase (PPIase) activity. Conversely, the PPIase activity, CsA binding affinity, stability, tertiary structure, surface hydrophobicity, and Trp accessibility of rCyp[S106A] notably differed from those of rCyp. The computational experiments also reveal that the structure, dimension, and fluctuation of SaCyp are not identical to those of SaCyp[S106A]. Furthermore, Ser at position 106 of SaCyp, compared to Ala at the same position, formed a higher number of non-covalent bonds with CsA. Collectively, Ser 106 is an indispensable residue for SaCyp that keeps its tertiary structure, function, and stability intact. Communicated by Ramaswamy H. Sarma

Staphylococcus aureus, a pathogenic bacterium with pronounced impacts on human health, expresses one cyclophilin (SaCyp) that is neither stress-induced nor required for its growth (Anderson et al., 2006;Zhang & Lin, 2009). However, it has distinct influences on the virulence properties and hemolysis activity of S. aureus (Wiemels et al., 2016). This protein was shown to be necessary for efficient folding of a staphylococcal virulence factor, nuclease, and the survival ability in mice (Keogh et al., 2018;Lin et al., 2018). Conversely, the PPIase function of SaCyp seemed to be unconnected to its virulence, because a point mutant of this cyclophilin that eliminated its PPIase function did not alter either the hemolytic property or the pathogenicity of S. aureus (Keogh et al., 2018). Together, this staphylococcal cyclophilin might be a prospective target for the screening and generation of new compounds that would specifically curb the infectivity of or fully eliminate the S. aureus strains including the ones that are multidrug-resistant. The latter strains are implicated with severe health problems and remain a matter of concern globally (Assis et al., 2017;Foster, 2017).
Studies on a recombinant SaCyp (rCyp) earlier revealed that it is a single-domain protein and exists as the monomers in the aqueous solution (Polley et al., 2017;Seal et al., 2019;. Additionally, rCyp showed PPIase activity that was specifically abolished by CsA. The CsA binding affinity of rCyp was noted about 1 mM. Furthermore, guanidine hydrochloride-or urea-induced unfolding of rCyp proceeded via the synthesis of an intermediate. The CsA-bound rCyp (rCyp-CsA) also showed similar pattern of unfolding in the presence of above denaturants (Polley et al., 2017;Seal et al., 2019;. The unfolding reactions were reversible though there was the formation of an intermediate. There was also a CsA-mediated rise of the thermodynamic stability of rCyp, suggesting that this protein may be useful in the screening of the CsA analogs (Seal et al., 2019;. SaCyp possesses a region that does not exist in the wellstudied cyclophilins but occurs in some putative cyclophilins (G€ othel & Marahiel, 1999;Gupta & Tuteja, 2011;Ke et al., 1993;Keogh et al., 2018;Peterson et al., 2000;Polley et al., 2017;Seal et al., 2019;€ Unal & Steinert, 2014;Weber et al., 1991). This uncommon segment in SaCyp that harbors 28 amino acid residues ranges from residue 121 to 148. Of the residues in the unusual region, seven residues are conserved among the putative cyclophilins ( Figure 1). The data from the limited proteolysis experiments revealed that the above segment is buried inside the structure of rCyp (Seal et al., 2019). The deletion of this unusual region significantly reduced PPIase activity and the thermodynamic stability of this enzyme . On the contrary, the deletion has strikingly increased the shape and the CsA binding activity of rCyp. In addition, the structure, and hydrophobic surface area of rCyp were considerably affected due to the above deletion .
Structural investigations previously reported that cyclophilins harbor a b-barrel conformation formed by several a-helices, b-strands, and loops (G€ othel & Marahiel, 1999;€ Unal & Steinert, 2014). Additionally, thirteen amino acid residues from the characterized cyclophilins, involved in the generation of non-covalent bonds with different residues from CsA, were found to be conserved (G€ othel & Marahiel, 1999;€ Unal & Steinert, 2014). Modeling studies earlier deciphered that SaCyp also has a similar cyclophilin-like b-barrel structure (Polley et al., 2017;Seal et al., 2019). Of the conserved residues involved in CsA binding, twelve residues were also present in SaCyp (Polley et al., 2017;Seal et al., 2019). The exception lies at the 106 position of SaCyp that carries a Ser residue. Ser 106 of SaCyp showed alignment with a CsAbinding Ala residue of the orthologous cyclophilins (Polley et al., 2017;Seal et al., 2020) (Figure 1). The data suggest that Ser 106 of SaCyp may also be similarly involved in the binding of CsA. In addition, this residue may maintain the structure, PPIase activity, and stability of SaCyp as well. Like Ser 106, Trp 136 in the atypical region of SaCyp may too play a pivotal role as it has aligned with either the Trp or Phe residue of many putative cyclophilins (Figure 1). Thus far, the ways Ser 106 or Trp 136 contributes to SaCyp have not been demonstrated though it may expedite the designing or screening of new anti-staphylococcal agents in the future. Herein, rCyp[S106A] and rCyp[W136A], constructed respectively by replacing the Ser 106 and Trp 136 of rCyp with an Ala residue, have been investigated using various tools. The results have pointed out that the Ser 106 is essential for maintaining the tertiary structure, function, and thermodynamic stability, whereas Trp 136 only has a trivial role in the function of rCyp.

Modeling and docking
To carry out various computational studies, the tertiary structural models of SaCyp, SaCyp[S106A], and SaCyp[W136A] were generated by Swiss-Model server (ExPasy.org) using their respective amino acid sequences as previously described . SaCyp[S106A] carries Ser to Ala change at position 106, whereas SaCyp[W136A] harbors Trp to Ala substitution of SaCyp at position 136. Analyses by Verify3D (Eisenberg et al., 1997) indicate that all of the models are good quality models. The visualization and superimposition of models were performed by UCSF Chimera 1.15 (Pettersen et al., 2004) and Pymol 1.7.4.5 edu (pymol.org), respectively.

Basic analyses of DNA and proteins
Several molecular biological experiments such as PCR, agarose gel electrophoresis, DNA digestion with restriction endonuclease, sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE), DNA estimation, DNA sequencing, competent E. coli cell preparation, plasmid DNA isolation, transformation of DNA, protein quantification, staining and de-staining of SDS-polyacrylamide gels were carried out as reported (Sambrook & Russell, 2001

Purification of proteins
Proteins rCyp[S106A] and rCyp[W136A] were purified from SAU1433 and SAU1435 cells using affinity chromatography Figure 1. Sequence alignment of Cyclophilins. Multiple sequence alignment between the amino acid sequences of SaCyp and its different orthologs, performed using the freely available software, Clustal Omega (www.ebi.ac.uk) as described . The upward arrows indicate residues implicated in CsA binding, while the downward arrows point to the residues that align with the Ser106 and Trp136 of SaCyp, respectively.  Ke et al., 1993); HsCyp, cyclphilin expressed by Homo sapiens (PDB ID: 3CYS; Weber et al., 1991); PfCyp, cyclophilin expressed by Plasmodium falciparum (PDB ID: 1QNG; Peterson et al., 2000). mostly as stated before . Briefly, the IPTG induced cells, resuspended in the 20 mM imidazole containing buffer A [20 mM Tris-HCl (pH 8.0), 500 mM NaCl, 5% glycerol, and 10 mg/ml PMSF], were sonicated to make the cell extracts. The debris-free cell extract was mixed with a Ni-NTA agarose carrying solution followed by its pouring into a purification column. Once buffer A left the column, we thoroughly washed it with buffer A having 40 mM imidazole. Lastly, we eluted each protein from the column using 200 mM imidazole carrying buffer A. The eluted proteins were dialyzed overnight against a buffer made with 300 mM NaCl, 20 mM Tris-HCl (pH 8.0), and 5% glycerol. Apart from rCyp[S106A] and rCyp[W136A], rCyp protein was also purified as mentioned earlier (Seal et al., 2019). The molar concentration of each protein was calculated considering its amount in dialysis buffer and theoretical molecular mass as described .

PPIase activity and CsA binding affinity
We separately determined the PPIase activity of rCyp, rCyp[S106A], and rCyp[W136A] using ribonuclease T1 (RNase T1) refolding assay as earlier demonstrated (Polley et al., 2017;Seal et al., 2019;. In brief, a denatured form of RNase T1 was allowed to undergo refolding by diluting it in the presence and absence of wild-type or mutant rCyp. The Trp fluorescence intensity values of RNase T1 at 323 nm (excitation wavelength 295 nm) were instantly measured by a fluorometer. The band-pass at excitation was 2.5 nm, whereas that at the emission wavelength was 5 nm. The fluorescence results, obtained in the presence and absence of native or mutant rCyp, were fit to a standard non-linear equation (Jana & Sau, 2012;Seal et al., 2019) to acquire the corresponding first-order rate constant values. Lastly, the PPIase activity (k cat /K m ) value of rCyp or rCyp derivative was calculated using a regular equation (Eq. (1)).
where k a , and k p are the first-order rate constants in the absence and presence of rCyp or mutant, respectively.

Conversely, [E] indicates the concentration of rCyp or variant.
Assuming that CsA and rCyp or mutant interacts at the stoichiometry of 1:1, the associated binding affinity (K d ) was determined as earlier reported ( Seal et al., 2019;Jana & Sau, 2012]. In short, the intrinsic Trp fluorescence spectra of rCyp, rCyp[S106A] and rCyp [W136A] in the presence of 0-4 mM CsA were separately recorded by a fluorometer. The values of Trp fluorescence intensity at k max (343 nm ) were serially derived from the spectra and corrected by subtracting the accompanying buffer fluorescence and by amending for volume alteration. To obtain K d values for the drug-protein association, the fluorescence results were fitted to the following Eq. (2) from GraphPad Prism (GraphPad Software Inc.).
Here, Y, B max , and [X] designate the change in fluorescence amount at any [CsA], the highest fluorescence change of rCyp or rCyp variant upon saturation with CsA, and [CsA], respectively.

Understanding protein structure
To have hints about the structures of native and mutant rCyp, the Trp fluorescence spectra (k ex ¼ 295 nm and k em ¼ 300-400 nm), ANS fluorescence spectra (k ex ¼ 360 nm and k em ¼ 400-600 nm), far-UV circular dichroism (CD) spectra (200-260 nm), and near-UV CD spectra (250-320 nm) of these proteins were individually recorded as mentioned before (Jana & Sau, 2012;Seal et al., 2019;. The protein concentration in each spectroscopic study (except near-UV CD spectroscopy) was 10 mM. In the near-UV CD experiment, [protein] was kept 25 mM. The [ANS] was undertaken 100 mM. The path lengths for the far-UV CD and near-UV CD experiments were 1 mm and 5 mm, respectively. To correct the CD and fluorescence spectra, buffer readings were considered as stated .
To gather knowledge about the solvent exposure of the Trp residues in rCyp and rCyp[S106A], quenching of Trp fluorescence of these proteins was independently studied using 0-200 mM acrylamide (Jana & Sau, 2012;Seal et al., 2020). To determine the Stern-Volmer constant (K SV ) values, the quenching data were analyzed as earlier mentioned (Jana & Sau, 2012;Seal et al., 2020).

Protein unfolding-refolding
The urea-induced unfolding of rCyp and rCyp[S106A] was inspected by both Trp fluorescence spectroscopy and ANS fluorescence spectroscopy as previously reported (Seal et al., 2019;. In sum, rCyp and rCyp[S106A] aliquots (10 mM each) were separately incubated with 0-7 M urea for $16-18 h at 4 C as described (Seal et al., 2019;. The fluorescence measurement of each aliquot was performed as stated above. Buffer corrections were made by deducting the values of urea-containing buffer solutions from that of the same buffer-carrying protein. Finally, the unfolding curves of the proteins were generated by plotting their fluorescence data against the matching urea concentrations as stated (Seal et al., 2019;. To check if the unfolding reactions were reversible, the Trp fluorescence spectra of wild-type, denatured, and renatured forms of rCyp[S106A] and rCyp (10 mM each) were independently collected as mentioned (Seal et al., 2019;).

Determination of unfolding mechanism and stability of protein
To discern the unfolding mechanism of rCyp[S106A] and rCyp, we fit the related unfolding curves to the reported two-state (N$U) equation or three-state (N $I$U) equation (Pace & Shaw, 2000;Sancho, 2013;Seal et al., 2019). To have hints on the thermodynamic stability of proteins, the associated parameters were calculated from the two-state equation (Eq. (3)) as presented below.
The terms [C], Y, R, Y N , T, and Y U indicate urea concentration, the collected spectroscopic data of protein at any [urea], the universal gas constant, the spectroscopic data obtained from entirely folded protein, the absolute temperature in Kelvin, and the spectroscopic data emerging from the fully unfolded protein, respectively. The [urea] at the center (C m ) of unfolding transition (N$U) was determined by dividing the free energy change at 0 M urea (DG W ) with the cooperativity parameter (m).
The difference in free energy change (DDG) between rCyp[S106A] and rCyp, relating to the N$U transition, was obtained from the following equation (Eq. (4)) as described (Pace & Shaw, 2000).
Here, the terms ‹m› and DC m designate the average of m values and the difference in C m values, acquired from the N$U unfolding procedure of rCyp and rCyp [S106A].
To know about the production of intermediates during unfolding, the fraction of protein molecules in the unfolded state (f U ) was estimated from the related fluorescence values using the following equation (Eq. (5)) as stated (Pace & Shaw, 2000).
where X U , X, and X N indicate the fluorescence data of protein in the unfolded state, the fluorescence data of rCyp or derivative at any [urea], and the fluorescence data of protein in the native state, respectively. The value of X N or X U was determined from the straight line generated using the fluorescence data of protein at very small or high urea concentrations.

MD Simulation
The modeled structures of SaCyp and SaCyp[S106A], constructed above, were separately subjected to molecular dynamics (MD) simulation using AMBER14 program as described (Case et al., 2005). After parameterization of the above proteins by ff14SB and GAFF force-fields, the solvation of each system was performed using TIP3P water box. As the default geometry seems to be different from actual minima, SaCyp and SaCyp[S106A] were minimized by the typical AMBER procedure (Case et al., 2005) using 500 steps of steepest-descent and 500 steps of conjugate-gradients. To constrain all bonds including hydrogen bonds, each system was heated for 50 ps at 300 K and simultaneously keeping the SHAKE algorithm on. After equilibration for 1 ns, each system was run for 10 ns on an NPT ensemble at 300 K temperature and 1 atm pressure having a step size of 2 fs. The values of radius of gyration (Rg), root mean square fluctuation (RMSF), and root mean square deviation (RMSD) of the backbone atoms (Ca, C, N, O) of each protein were determined by analyzing the corresponding MD trajectories using the CPPTRAJ module of AMBER package. The final trajectories were visualized using VMD and PyMol (Roe & Cheatham III, 2013).

Structural models of SaCyp mutants
To obtain clues about the roles of Ser 106 and Trp 136 on the structure of SaCyp, model structures of SaCyp ( Figure  2A), SaCyp[S106A] ( Figure 2B), and SaCyp[W136A] ( Figure 2C) were built up as stated above. The putative structures were found to have a similar cyclophilin-like b-barrel conformation (G€ othel & Marahiel, 1999) harboring three a-helices (a1-a3), eight b-strands (b1-b8), and twelve loops (L1-L12). The amino acid composition of each a-helix, b-strand, and loop (except L7 or L8) of SaCyp (Table S1) has matched to those of corresponding structural elements of SaCyp[S106A] or SaCyp[W136A]. Ser 106 and Trp 136 of SaCyp are located in the loops L7 and L8, respectively. The introduction of Ala residue at position 106 or 136 did not interrupt L7 or L8. The barrel structure in each model is formed by an antiparallel sheet in which b-strands are arranged as b8-b1-b2-b7-b5-b6-b4-b3. There are two helices (a1 and a3) on the opposite sides of the barrel. As noted in several cyclophilins (G€ othel & Marahiel, 1999), many residues (Table S2) from three b-strands (b3, b4, and b6), one a-helix (a2), and two loops (L7 and L9) have together formed a putative active site in one face of SaCyp and mutants.
To check whether there is any structural distortion due to Ala substitution at position 106 or 136 of SaCyp, the model of this protein was separately superimposed to those of SaCyp[S106A] and SaCyp[W136A]. The root mean square deviation (RMSD) values for the SaCyp-SaCyp[S106A] and SaCyp-SaCyp[W136A] sets are 0.056 Å ( Figure 2D) and 0.021 Å ( Figure 2E), respectively. A similar superimposition study between the structures SaCyp[W136A] and SaCyp[S106A] yielded an RMSD value of 0.034 Å ( Figure 2F). Collectively, the structure of no protein completely matched that of others. Additionally, the structure of SaCyp[S106A], compared to the structure of SaCyp[W136A], was more deviated from that of SaCyp. A careful observation of the superimposed structures revealed that most of the structural discrepancies are in the atypical regions of the protein. So, Ser 106 and Trp 136 may play some role to maintain the structure of SaCyp.

Docking of CsA with SaCyp mutants
Previously, a recombinant SaCyp exhibited binding to CsA with a considerably high affinity (Polley et al., 2017;Seal et al., 2019;. To find out the residues involved in the interaction between SaCyp and CsA, the structural model of this cyclophilin was docked with the reported structure of CsA (Fesik et al., 1991) using HDOCK server (Yan et al., 2020). In the docked complex, CsA has shown binding to the putative active site of SaCyp. A thorough analysis of the docked complex has revealed that fifteen amino acid residues of SaCyp have formed several hydrogen bonds, and van der Waals interactions with eight CsA residues such as 4-Methyl-4-[(E)-2-Butenyl]-4-N-Dimethyl-Threonine (MeBmt1), a-aminobutyric acid (Abu2), Sarcosine (Sar3), N-methyl leucine4 (MeLeu4), Val5, MeLeu9, MeLeu10, and MeLeu11 (Table 1). Of the CsA binding SaCyp residues, Arg 59, Phe 64, Met 65, Gln 67, Gly 75, Ala 104, Asn 105, Gln 114, Phe 116, Trp 152, Leu 153, and His 157 are the conserved residues as these have been observed in both the characterized and putative cyclophilins (G€ othel & Marahiel, 1999;Polley et al., 2017;Seal et al., 2020) (Figure 1). The above twelve conserved residues from a well-studied cyclophilin (PfCyP19) (Peterson et al., 2000) similarly interacted with the identical sets of CsA residues (Table 1). Of the rest CsA binding SaCyp residues, Ser 106 of SaCyp aligns with PfCyP19 residue Ala 110 (Figure 1), a residue reported to be conserved previously (G€ othel & Marahiel, 1999;Peterson et al., 2000). The structural data reveals that the side chain carbon atom (C B ) and the peptidyl carbonyl-oxygen atom (O) of Ala 110 of PfCyP19 (Table 1 and Figure 3A) have produced van der Waals interactions with the carbonyl O of Abu2 and the side chain carbon atom (C H ) of MeBmt1, respectively (Peterson et al., 2000). Conversely, Ser 106 of SaCyp has formed five non-covalent bonds with CsA residues MeBmt1, MeLeu4, Sar3, and Abu2 (Table 1 and Figure 3B). Of the non-covalent bonds, two bonds are hydrogen bonds and the rest are van der Waals interactions. Both carbonyl oxygen and peptidyl nitrogen atoms of Abu2 have produced hydrogen bonds with the hydroxyl O G atom of Ser 106. The side-chain hydroxyl oxygen atom (O G ) of Ser 106 has also made van der Waals interactions with the peptidyl nitrogen atom (N) of Sar3, and the C N atom attached to the peptidyl nitrogen of MeLeu4. Another van der Waals interaction has been formed between the peptidyl carbonyl-oxygen atom (O) of Ser 106 and the side chain carbon atom (C H ) of MeBmt1. Two remaining CsA binding SaCyp residues whose orthologs in PfCyP19 did not show interaction with CsA residues are Ile 61 and Met 76 (Table 1). While Met 76 has generated three van der Waals interactions with CsA residues MeLeu4, Sar3, and Abu2, Ile 61 has produced a van der Waals interaction with MeLeu10 of CsA ( Figure S1).
To understand whether the Ala substitution at positions 106 and 136 of SaCyp affects its bond formation with CsA, we have also individually docked the structural models of SaCyp[S106A] and SaCyp[W136A] with the structure of the above cyclophilin inhibitor. The results have revealed that SaCyp[W136A] residues, like the SaCyp residues, identically form non-covalent bonds with the CsA residues ( Table 1). All of the SaCyp[S106A] residues except Ala 106 also have identically interacted with CsA (Table 1). Careful observation has demonstrated that Ala 106 of SaCyp[S106A], unlike SaCyp or PfCyP19, makes three van der Waals interactions with CsA residues MeBmt1, Abu2, and MeLeu4 ( Figure 3C). Jointly, the CsA binding affinity of SaCyp[S106A] may be different from that of SaCyp.

Purification and functional studies of rCyp mutants
To validate the suppositions put forward by the above computational studies, rCyp and its mutants rCyp[S106A] and rCyp[W136A] were purified to uniformity by Ni-NTA To experimentally check if the mutations have affected the CsA binding activity of SaCyp, we individually recorded the Trp fluorescence spectra of 0-4 lM CsA-equilibrated rCyp, rCyp[S106A], and rCyp[W136A] as mentioned above. The spectra reveal the gradual increase of Trp fluorescence intensity of all three proteins upon increasing the [CsA] from 0 to $2.5 lM (Figures 4B-D). Conversely, there is very little change of fluorescence at $2.5 to 4 lM CsA. Further, the plots developed from the fluorescence data of rCyp, rCyp[S106A] and rCyp[W136A] distinctly show that the increase of fluorescence intensity of rCyp or rCyp[W136A] occurs at relatively less concentrations of CsA ( Figure 4E). However, there is saturation in fluorescence increase for each protein at $3-4 lM CsA. The K d values of rCyp-CsA, rCyp[S106A]-CsA, and rCyp[W136A]-CsA interactions, estimated from the above plots, are presented in Table 2.
Comparative analyses indicate that the K d value of rCyp[S106A]-CsA interaction is significantly higher than that of other protein-CsA interactions (all p values <0.031). Alternately, the K d value of rCyp[W136A]-CsA interaction, compared to that of rCyp-CsA interaction, was not R59 ( R62(b5), F67(L7), M68(b6), Q70(b6), A108(L9), N109(L9), F120(b8), L129(a2), H133(L11) a The data obtained from the X-ray crystal structure of Plasmodium falciparum Cyp-CsA complex (Peterson et al., 2000). All of the amino acid residues (except the underlined amino acid residues) are conserved residues. 1.21 ± 0.02 0.036 ± 0.002 ----- The K d and K SV values were determined using the results presented in Figures 4E and 5D, respectively. Using the RNase T1 refolding assay data ( Figure 4F) and the emission k max values of Trp fluorescence ( Figure 6A), the k cat /K m values and the values of different thermodynamic parameters were calculated as described in Materials and methods. The '-' indicates not determined.
significantly altered (p ¼ 0.19). Together, the experimental data not only supports the computational data ( Figure 3) but also suggests that the CsA binding affinity of rCyp[S106A] is less than that of rCyp and rCyp [W136A].
To detect if the rCyp[S106A] and rCyp[W136A] have different PPIase activity, we individually performed RNase T1 refolding assay in the absence and presence of these mutants and rCyp as described (Polley et al., 2017). The obtained data suggest that the mutant rCyp[S106A] refolds RNase T1 more rapidly in comparison with rCyp or rCyp[W136A] ( Figure 4F). The k cat /K m values of rCyp, rCyp[S106A] and rCyp[W136A], estimated from the assay data, are given in Table 2. Further analysis indicates that the k cat /K m value of rCyp[S106A] is significantly higher than that of rCyp or rCyp[W136A] (all p values <0.029). Conversely, there is little difference between the k cat /K m values of rCyp[W136A] and rCyp (p ¼ 0.2). Collectively, the Ser to Ala change at position 106 has considerably enhanced the enzymatic activity of rCyp. However, the Ala substitution at position 136 did not affect the PPIase activity of rCyp.

Structure and surface hydrophobicity of rCyp[S106A]
The PPIase or the CsA-binding activity of rCyp[W136A] is not significantly different from those of rCyp (Table 2), indicating that the structures of these two proteins might be identical. On the contrary, both the enzymatic function and the inhibitor binding affinity were markedly altered due to Ala substitution at the position of 106 of rCyp ( Table 2). The data opines that the structures of rCyp[S106A] and rCyp may not be fully identical. To check this hypothesis, we studied both rCyp and rCyp[S106A] by different spectroscopic probes as stated in Materials and methods. The far-UV CD spectrum of rCyp[S106A] is nearly identical to that of rCyp ( Figure 5A), suggesting the presence of a similar extent of a-helices in both proteins. A computational analysis of the above spectra revealed that the contents of a-helix, b-strands, b-turn, and coil in rCyp[S106A] are nearly identical to those in rCyp (Table S3). In sum, the secondary structure of rCyp was little affected due to mutation at position 106.
The near-UV CD spectra of rCyp and rCyp[S106A], unlike their far-UV CD spectra, are completely different from each other ( Figure 5B). However, both spectra yielded a peak at $284-287 nm. We have observed that the ellipticity value of rCyp is about 6 folds less than that of rCyp[S106A] at 285 nm (Table S3). Collectively, the three-dimensional structure of rCyp might be different from that of rCyp[S106A].
The intrinsic Trp fluorescence spectra of rCyp and rCyp[S106A], like their near-UV spectra, also did not match with each other ( Figure 5C). Both spectra, however, have yielded a peak at 343 nm. The fluorescence intensity of rCyp[S106A], compared to that of rCyp, was lost about 1.4 times at 343 nm (Table S3). The results suggest that the Trp residues in the rCyp mutant may be relatively more surfaceexposed. To verify the hypothesis, we investigated the Trp  Figure 5D). The values of the Stern-Volmer constant (K SV ) of rCyp and rCyp[S106A], determined from the above plots, are presented in Table 2. Analysis revealed that the K SV value of rCyp is significantly less than that of rCyp[S106A] (p ¼ 0.0007). The data suggest that one or both the Trp residues of rCyp[S106A], compared to those of rCyp, may have higher solvent accessibility.
The ANS fluorescence spectra of the two proteins are also not identical ( Figure 5E). There is a 1.2 fold increase of the ANS fluorescence intensity of the mutant protein compared to that of rCyp at 470 nm (Table S3). Thus, the Ala substitution at position 106 partly affected the surface hydrophobicity of rCyp as well.

Unfolding and refolding of rCyp[S106A]
To know whether rCyp[S106A] folds like rCyp (Seal et al., 2019; 2020), we studied the urea-induced denaturation of  these proteins using two spectroscopic probes. The obtained spectra revealed the gradual decrease of Trp fluorescence intensities of both proteins upon the increase of [urea] from 0 to 3.5 M (Figures 6A, B). The associated emission k max values were also increased from 343 to 350 nm when urea concentrations were raised from 0 to 3.5 M. However, there was little increase of k max or intensity values when [urea] concentration was further increased to 7 M. The plotting of the normalized fluorescence intensity or the k max values against [urea] yielded monophasic unfolding curve for both rCyp and rCyp[S106A] ( Figure 6C). Careful observation revealed that the curves of rCyp[S106A] and rCyp have transitions at $1-3.25 and $1.5-3.5 M urea, respectively. The data indicate that the mutant rCyp may be relatively more susceptible to urea. The ANS fluorescence intensity values of rCyp and rCyp[S106A] were also gradually reduced upon enhancing the [urea] from 0 to 5.5 M (Figures 6D, E). There was a very little decrease of the fluorescence intensity upon further increase of [urea] to 7 M. Unlike the Trp fluorescence datamade unfolding curves ( Figure 6C), the curves, created using the normalized ANS fluorescence intensity data (at 470 nm) of native and rCyp mutant, are roughly biphasic in nature ( Figure 6F).
To learn whether the process of urea-induced unfolding of both rCyp and rCyp[S106A] is reversible, the three different forms, namely the native, unfolded, and possibly refolded forms of these proteins were separately prepared followed by the recording of their Trp fluorescence spectra as stated before . The spectra of the possibly refolded forms of all proteins merged with the spectra of their corresponding native forms ( Figure 7A), indicating that the urea-induced unfolding of both proteins occurred by a reversible pathway.

Denaturation pathway of proteins
To precisely know whether rCyp[S106A], like rCyp (Seal et al., 2019;Seal et al., 2020], was also denatured via the generation of an intermediate, the unfolding curves ( Figure 6) of this mutant and rCyp were fit to the standard equations as demonstrated above (Pace & Shaw, 2000;Sancho, 2013;Seal et al., 2019;. The unfolding curves, prepared from the Trp fluorescence data of rCyp[S106A] and rCyp ( Figure 6C), showed excellent fitting to a two-state equation, whereas those made using their ANS fluorescence data ( Figure 6F) exhibited perfect fitting to a three-state equation. Jointly, the urea-mediated denaturation of both native and mutant rCyp most possibly proceeded through the synthesis of an intermediate. To validate the above supposition, the fraction unfolded molecules of rCyp and rCyp[S106A], determined from their Trp fluorescence ( Figure 6C) and ANS fluorescence ( Figure 6F) data, were plotted against the relevant [urea]. Neither the rCyp-specific nor the rCyp[S106A]-specific fraction unfolded curves from two spectroscopic studies showed any superimposition, particularly at the transition regions ( Figure 7B). Together, the data have not only supported the previously demonstrated unfolding mechanism of rCyp (Seal et al., 2019; but also indicated the production of at least one intermediate during the unfolding of rCyp[S106A] in the presence of urea.

Stability of rCyp[S106A]
The thermodynamic stability of a mutant protein often differs from that of its native form (Case et al., 2005;Jana & Sau, 2012;Mandal et al., 2018;Roe & Cheatham III, 2013;Seal et al., 2020). As the unfolding of rCyp[S106A] started at comparatively less urea concentrations (Figure 6), the Ala substitution at position 106 of rCyp may affect its stability. To validate this assumption, the C m and DG W values of both rCyp and rCyp[S106A] (Table 2), yielded from the two-state equation-fitted k max data ( Figure 6C), were compared as described . Of the values, the C m value of rCyp[S106A], compared to that of rCyp, was found considerably less (p ¼ 0.006). The C m values of rCyp, obtained from the two-state equation-fitted Trp fluorescence intensity data ( Figure 6C) or the three-state equation-fitted ANS fluorescence data ( Figure 6F), were also considerably higher than the analogous C m values of rCyp[S106A] (all p values <0.022; Table S4). In addition, the DDG value or the change in free energy between native and mutant rCyp, calculated using the emission k max data ( Figure 6C), was 0.87 ± 0.29 kcal mol À1 (Table 2). Jointly, Ser 106 residue is critical for preserving the thermodynamic stability of rCyp.

MD Simulation of SaCyp[S106A]
MD simulation, a computational tool, earlier showed the mutation-induced structural changes in many proteins including PPIase enzymes (Arjomand et al., 2016;Chen et al., 2013;Li & Zheng, 2011;Mandal et al., 2018;Odera et al., 2018;Polley et al., 2015;Sinha et al., 2020). To verify the structural change of rCyp ( Figure 5), we have carried out MD simulations on the energy-minimized structural models of both SaCyp[S106A] (Movie S1) and SaCyp (Movie S2) for 10 ns. The generated RMSD data reveal that SaCyp[S106A] curve diverges from the SaCyp curve at $6.5 ns and reaches from $1.4 Å to about $3.5 Å at 10 ns ( Figure 8A). Conversely, SaCyp trajectory has shown roughly a steady equilibrium with minute fluctuations during the entire simulation period. The results together suggest the conformational change of SaCyp due to the Ala substitution at position 106. In addition, SaCyp[S106A] backbone is comparatively less stable and has a relatively higher structural flexibility.
The Rg curves of SaCyp[S106A] and SaCyp also did not fully match with each other ( Figure 8B). While SaCyp curve remained nearly steady, SaCyp[S106A] curve shows about 1 Å change during the simulation period. In sum, the mutant protein that looks relatively more flexible may have a different dimension as well.
The RMSF data also show that most SaCyp[S106A] residues, compared to the corresponding SaCyp residues, have higher fluctuations ( Figure 8C). We have also observed that the RMSF values of nine CsA binding residues of SaCyp[S106A] are about 0.4-1.23 Å higher than those of Figure 6. Unfolding of cyclophilins. The Trp fluorescence spectra and the ANS fluorescence spectra from different aliquots of urea-treated rCyp (A and D) and rCyp[S106A] (B and E), were recorded as described in the Materials and methods. The Trp fluorescence intensity (at 343 nm) and the linked k max values of rCyp and rCyp[S106A], derived from panels (A) and (B), were normalized and plotted versus related [urea] (C). The ANS fluorescence intensity data (at 470 nm) of those proteins (D and E), were also collected, analogously normalized, and plotted versus [urea] (F). analogous SaCyp residues (Table 3). On the other hand, the RMSF values of three other CsA binding residues of SaCyp are $ 0.3-0.45 Å higher than those of analogous SaCyp[S106A] residues. In sum, the relatively greater fluctuations and flexibility in SaCyp[S106A] have possibly affected its CsA binding activity (Table 2).

Discussion
The change of an amino acid residue of a protein to another residue often affects its structure, function, and stability (Keogh et al., 2018;Mandal et al., 2018;Polley et al., 2015;Sinha et al., 2020). The current investigations have shown that the substitution of Trp 136 with an Ala residue little affected the PPIase activity or the CsA binding affinity of rCyp ( Figure 4 and Table 2). Conversely, the replacement of Ser with an Ala residue at position 106 of rCyp has differently altered its CsA-binding affinity, PPIase activity, tertiary structure, Trp accessibility, surface-hydrophobicity, and thermodynamic stability (Figures 2, 4-6, and Table 2). While the enzymatic activity, Trp accessibility, and surface hydrophobicity of rCyp were notably enhanced, the CsA-binding affinity, tertiary structure, and stability of rCyp were considerably affected due to the above mutation. Interestingly, this mutation hardly changed the secondary structure, and unfolding mechanism of rCyp ( Figures 5, 6). The present MD simulation experiment also suggested the mutation-mediated structural change in SaCyp (Figure 8 and Table 3). Therefore, the Ser 106 of SaCyp might be a crucial residue that is employed to preserve its three-dimensional structure, function, and stability. Contrarily, Trp 136, though aligned with Trp/Phe residues of many SaCyp orthologs, looks dispensable for this protein.
Some mutations in SaCyp and other cyclophilins earlier differently changed the PPIase activity and CsA binding affinity (Fejzo et al., 1994;G€ othel et al., 1996;Hoffmann et al., 1995;Ke et al., 1993;Liu et al., 1991;Seal et al., 2020;Skagia et al., 2017;Spitzfaden et al., 1994). The change of structure was also noted in a few cyclophilin mutants that altered the PPIase activity and CsA binding affinity in an unequal way (Fejzo et al., 1994;Ke et al., 1993;Seal et al., 2020;Skagia et al., 2017). The above data indicate that the decrease of CsA-binding affinity in rCyp[S106A] may be due to its altered tertiary structure (Figures 4,5,and Table 2). In addition, the formation of a lesser number of non-covalent bonds between Ala 106 and CsA may also contribute to the reduced CsA-binding affinity in rCyp[S106A] (Figure 3). On the other hand, the increase of PPIase activity in rCyp[S106A] is quite interesting (Figure 4 and Table 2) as the Ala residues of various cyclophilins that aligned with Ala 106 of SaCyp[S106A] (data not shown) do not bind the peptide substrate (G€ othel & Marahiel, 1999;Seal et al., 2020). In the present study, to determine the PPIase activity of rCyp or rCyp[S106A], a protein substrate (i.e. RNase T1) was used as it is more closer to the physiological substrate(s) of cyclophilins (G€ othel & Marahiel, 1999). Currently, it is not known whether RNase T1 interacts with Ala 106/Ser 106 of rCyp[S106A]/rCyp or the interaction between RNase T1 and Ala 106 of rCyp[S106A] is relatively superior. The higher PPIase activity in rCyp[S106A] may also be due to the consequence of its structural alteration. The way structural change increased the PPIase activity of rCyp[S106A] is, however, not known with certainty.
The side chain of Ser residue carries a hydroxymethyl group whereas that of Ala residue harbors a methyl group. While Ala was equally found both in the surface and interior of a protein structure, Ser prefers to remain on the surface of a protein. Ser was also found inside a protein particularly when its side chain forms a hydrogen bond. In addition, Ala is usually involved in the formation of a-helix, whereas Ser tends to disrupt it. Analyses of the SaCyp and SaCyp[S106A] structural models revealed that both Ser 106 and Ala 106 reside in a loop region of the respective proteins ( Figure S3). Additionally, Ala 106 forms a hydrogen bond with a Thr 110 residue in SaCyp[S106A] ( Figure S3B), whereas Ser 106 forms hydrogen bonds with both Thr 110 and Asn 105 in SaCyp ( Figure S3A). Therefore, the loss of hydrogen bond(s) most possibly changed the tertiary structure of SaCyp[S106A] or rCyp[S106A] (Figures 2, 5). The structural alteration (particularly tertiary structure) was also noted in some proteins harboring Ser to Ala change (Lendrihas et al., 2010;Poyner et al., 2002).
SaCyp and its close orthologs have a larger substrate/ inhibitor binding site due to the presence of additional 28 amino acid residues in their C-terminal halves . A few members of this family (including SaCyp) also harbor a Ser residue instead of an Ala residue at position 106 ( Figure 1). Our docking study indicates that no residue from the extra region participates in the interaction between SaCyp and CsA ( Table 1). The docking data was partly supported by our experimental data as the substitution of one of the additional residues (Trp 136) with an Ala residue did not affect the CsA binding affinity of SaCyp (Table 2). As the deletion of the entire extra region of SaCyp previously enhanced its CsA binding affinity , we rule out the direct interaction between this region and CsA. However, our in silico results have proposed that fifteen residues, located outside of the above atypical region of SaCyp, are involved in binding with CsA (Table 1 and Figure 3B). We have noted that 80% of the above 15 residues are employed by other cyclophilins for binding CsA (G€ othel & Marahiel, 1999;Ke et al., 1993;Peterson et al., 2000;€ Unal & Steinert, 2014;Weber et al., 1991). Of the remaining SaCyp residues, the exact role of Ile 61 (a conserved residue) or Met 76 in CsA binding is not currently known with certainty. Conversely, our genetic experiment has confirmed the participation of Ser 106 in CsA binding (Table 2). Collectively, the present docking data that may be nearly 87% trustworthy has for the first time provided invaluable clues about the mechanism of interaction between CsA and SaCyp, a representative of a new type of cyclophilins .
A protein usually requires a stable three-dimensional structure to exert its desired function. This structural stability is attributed mostly to the numerous non-covalent interactions/bonds formed, like the hydrophobic interactions, van der Waals interactions, hydrogen bonds, and ionic bonds (Jaenicke, 2000;Pace et al., 1996;Privalov & Gill, 1988). Many of these non-covalent interactions/bonds that occur in cyclophilins also appeared to be crucial for their stability (Boudko et al., 2012;Ke et al., 1991;Kozlov et al., 2010;Seal et al., 2020). As the Ala substitution at position 106 affected the rCyp structure (Figure 5), the non-covalent bonds/interactions present in the native and mutant rCyp might not be alike. To confirm this hypothesis, the structural model of SaCyp[S106A] and SaCyp were separately analyzed by the iCn3D server (Wang et al., 2020). The obtained data have shown that these proteins have different number of hydrogen bonds (data not shown). Collectively, the dissimilar quantity of hydrogen bonds in these two proteins has presumably diminished the stability of rCyp[S106A].