Functional Mechanism of C-Terminal Tail in the Enzymatic Role of Porcine Testicular Carbonyl Reductase: A Combined Experiment and Molecular Dynamics Simulation Study of the C-Terminal Tail in the Enzymatic Role of PTCR

Porcine testicular carbonyl reductase, PTCR which is one of the short chain dehydrogenases/reductases (SDR) superfamily catalyzes the NADPH-dependent reduction of carbonyl compounds including steroids and prostaglandins. Previously we reported C- terminal tail of PTCR was deleted due to a nonsynonymous single nucleotide variation (nsSNV). Here we identified from kinetic studies that the enzymatic properties for 5α-dihydrotestosterone (5α-DHT) were different between wild-type and C-terminal-deleted PTCRs. Compared to wild-type PTCR, C-terminal-deleted PTCR has much higher reduction rate. To investigate structural difference between wild-type and C-terminal-deleted PTCRs upon 5α-DHT binding, we performed molecular dynamics simulations for two complexes. Using trajectories, molecular interactions including hydrogen bonding patterns, distance between 5α-DHT and catalytic Tyr193, and interaction energies are analyzed and compared. During the MD simulation time, the dynamic behavior of C-terminal tail in wild-type PTCR is also examined using essential dynamics analysis. The results of our simulations reveal that the binding conformation of 5α-DHT in C-terminal-deleted PTCR is more favorable for reduction reaction in PTCR, which shows strong agreement with kinetic data. These structural findings provide valuable information to understand substrate specificity of PTCR and further kinetic properties of enzymes belonging to the SDR superfamily.


Introduction
Porcine testicular carbonyl reductase, PTCR (also known as 20b-hydroxysteroid dehydrogenase, 20b-HSD) belongs to the short-chain dehydrogenases/reductases (SDR) superfamily. The SDR catalyze crucial steps of activation and inactivation of steroids, vitamins, prostaglandins, and other bioactive molecules by oxidation and reduction of hydroxyl and carbonyl groups, respectively [1]. Almost all SDRs have Tyr-Lys-Ser as catalytic triad and their functional units are homotetramers or homodimers [1][2][3][4][5][6][7][8]. However, interestingly, PTCR is the first known monomeric structure among the SDR superfamily and it catalyzes the NADPH-dependent reduction of ketones on androgens, aldehydes, progestins, and prostaglandins as well as various xenobiotics [9,10]. The high activity of enzyme is shown by the reduction of 20-carbonyl groups of C 21 -steroids, such as conversion of 17ahydroxyprogesterone to 17a, 20b-dihydroxy-4-pregnen-3-one, which is present in pig testes during the neonatal period [11,12]. Purified PTCR has enzymatic activities to 3a-hydroxysteroid, 3bhydroxysteroid and 5a-dihydrotestosterone (5a-DHT) as substrates [13]. Meantime, the endogenous PTCR proteins were identified as two bands, 31 kDa and 30 kDa proteins from porcine testis through western blot analysis even though the reason for detection of a minor 30 kDa protein remained unclear [14]. Recently, our report described that the minor 30 kDa protein, detected in porcine testis, is a C-terminal-deleted PTCR, expressed from the PTCR gene having a paralogous sequence variant (T), probably generated by gene duplication [15]. It also suggested that the nonsynonymous variation (G.T), generated by gene duplication, leads to a deletion of the C-terminal region (E281 to A288) of PTCR, strongly indicating that the pig testicular genome endogenously produces both wild-type and C-terminaldeleted PTCR, 31 kDa and 30 kDa proteins, respectively [15]. Additionally, the C-terminal fragment from E281 to A288, in the PTCR structure resolved at high resolution, is reported to be in the vicinity of the active site [16] and is absent in human carbonyl reductase despite the high homology (about 85%) between PTCR and human carbonyl reductase [9,16]. Accordingly, it has been suggested that the PTCR-unique C terminus might provide a clue to the differential kinetics between porcine and human carbonyl reductases [9], and the structural importance of the C-terminal tail also has been discussed with regard to catalysis of the PTCR enzyme [17].
In the present study, we found that WT and C-terminal-deleted PTCRs have shown different kinetics in NADPH-dependent carbonyl reductase activity. To find structure related mechanistic explanation for the different enzymatic properties between WT and C-terminal-deleted PTCRs, molecular dynamics (MD) simulations were performed. Using trajectories, molecular interactions including hydrogen bonding patterns, distance between and catalytic Tyr193 and interaction energies are analyzed and compared. During the MD simulation time, the dynamic behavior of C-terminal tail in WT PTCR is also examined using essential dynamics analysis. Our simulation results provide detailed information which can explain the different binding mode of 5a-DHT in WT and C-terminal-deleted PTCRs and these explanations are in the agreement with the kinetic experimental data.

Recombinant protein purification
Two types of pPROEX HTb-PTCR clones were already constructed in previous study [1] and further used for the IPTGinduced expression of each of the clones in E. coli BL21, in order to produce the his-tagged fusion proteins, PTCR(WT) and PTCR(DCterm). Subsequently, the IPTG-induced proteins were subjected to the affinity chromatography using Ni-NTA agarose, according to manufacturer's manual (Peptron, Daejon, Korea). Briefly, basal buffer for protein purification was prepared by 50 mM sodium phosphate buffer (pH 8.0) and 500 mM NaCl, and imidazole (Sigma, USA) was added by the concentrations required for the affinity chromatography. The overexpressed cells was precipitated by centrifugation and suspended by binding buffer, the basal buffer including 5 mM imidazole. The genomic DNAs of collected cells were fragmented by SONICS Vibracell VCX 750 Ultrasonic Cell Disruptor (Newtown, CT), which was done twice by conditions as following; 5 min by 2 sec interval of on/off and 35% amplitude during ice cooling. The supernatant to obtain water-soluble protein was collected from the centrifugation at 12,0006 g for 30 min at 4uC, afterward the supernatant was loaded to a column including the Ni-NTA agarose, washed by the above binding buffer and then subjected to the stepwise elution using the basal buffer including various imidazole concentrations. The purified recombinant proteins were concentrated by Ultrafree-0.5 Centrifugal Filter Device (Millipore, Bedford, MA) and then were dissolved with 50% glycerol and 50 mM Sodium Phosphate Buffer (pH 6.4) for long-term storage at 220uC. Finally, concentrations of the recombinant proteins, PTCR(WT) and PTCR(DCterm), were determined by Bradford Protein assay kit (Bio-Rad, Richmond, CA) as about 4.57 mg/ml and 1.26 mg/ml, respectively.

Measurement of NADPH-dependent carbonyl reductase activity
Reaction mixtures consisted of 60 mM sodium phosphate (pH 6.5), the purified PTCR(WT) or PTCR(DCterm) proteins, 0.1 mM NADPH, and various amounts of substrates as indicated in Figure 1 and 2 were incubated in a quartz cuvette at 37uC. The assays of NADPH-dependent reduction activity for various substrates were carried out spectrophotometrically by monitoring the change in absorbance at 340 nm with time.

Statistical analysis
To determine kinetic parameters with a Michaelis-Menten plot, the data were analyzed by nonlinear regression using GraphPad

Preparation of PTCR complex structure
The WT PTCR and C-terminal (E281 to A288)-deleted PTCR that each protein in complex with NADPH and 5a-DHT were already modeled through molecular docking simulation in our previous study [15]. The binding conformations of 5a-DHT at the active site of WT and C-terminal-deleted PTCRs were predicted using GOLD program [18,19]. The substrate binding site was defined by 20 Å around the center of the site between C-terminal tail and NADPH. During the calculation, the number of GA runs was set to 30 and the predicted poses were ranked by GOLD fitness score. All other parameters were used as their defaults.

Molecular dynamics simulations
Two modeled structures were then used as a starting structure for molecular dynamics simulation. The MD simulations were carried out with CHARMM27 force field using GROMACS 4.5.3 package [20,21]. The topology files for cofactors and substrates were generated using SwissParam [22]. As a validation purpose, we calculated the topologies of cofactors and substrates using ParamChem [23][24][25] and MATCH [26] and compared with the topologies obtained from SwissParam. At first, hydrogen atoms were added and all the ionizable residues in the protein were protonated at pH 7. A cubic water box of 1.5 nm from the surface of the protein was generated and solvated with TIP3P water model [27] to perform the simulations in an aqueous environment. The Na + counter-ions were added by replacing water molecules to neutralize the system. The systems were subjected to energy minimization using steepest descent algorithm to improve the model quality to convergence on maximum force lower than 2000 kJ/mol. The energy minimized systems were then equilibrated in three steps. In the first step, NVT equilibration was conducted for 100 ps at 300 K. A constant temperature was controlled by V-rescale thermostat [28]. After NVT, 100 ps NPT ensemble was applied at 1 bar of pressure followed by 20 ns production run under the same ensemble. During this process, 300 K and 1 bar was kept using the same thermostat and Parrinello-Rahman barostat [29,30]. For the equilibration process, the protein backbone was restrained and the solvent molecules with counter-ions were allowed to move. Bonds between heavy atoms and corresponding hydrogen atoms were restrained to their equilibrium bond lengths using the LINCS algorithm [31,32] and the geometry of water molecules was constrained using SETTLE algorithm [33]. Short-range interactions were considered by using the cut-off value of 1.2 nm and long-range electrostatic interaction was calculated using particle mesh Ewald (PME) method [34,35]. A grid spacing of 0.12 nm was applied for fast Fourier transform calculations. All simulations were performed under periodic boundary conditions to avoid edge effects. The time step of the simulation was 2 fs and the coordinate data were stored to the file every 1 ps. We repeated the simulation of each system four times (Rep 1 to Rep 4) under the same conditions. All analysis was done using VMD, Discovery studio (DS) v3.1, and GROMACS. The representative structure which is the closest conformation to the average structure was selected from each simulation and used for the analysis. Essential dynamics analysis [36,37] also called principal component analysis was performed. The covariance matrices were diagonalized by constructing projections of trajectories on the eigenvectors. Movements of all C a atoms of the protein in the essential subspace were projected according to the most significant eigenvectors.

Results and Discussion
The C-Terminal Region Contributes Significantly to the Enzymatic Properties of PTCR Previously, the PTCR had been identified as two differential variants with 31 kDa or 30 kDa in porcine testis [9]. Recently, our approach integrating RNA-Seq and molecular biological analyses supported that the minor 30 kDa protein, detected in porcine testis, is a C-terminal (E281 to A288)-deleted PTCR, expressed from the PTCR gene having a paralogous sequence variant [15]. In addition, the C-terminal region (E281 to A288) was suggested to contribute to different kinetic properties between the NADPHdependent carbonyl reductase activities of WT and C-terminaldeleted PTCRs [15]. Thus, these indicate that the C-terminal region (E281 to A288) of PTCR has a significant effect on catalytic turnover rate (k cat ) with a slight one on substrate binding affinity (K m ).
To identify their kinetic characteristics, the his-tagged PTCR(WT) and PTCR(DCterm), recombinant PTCR proteins with or without the C-terminal region, were purified by affinity chromatography using a Ni-NTA resin. After stepwise elution with imidazole, the results of SDS-PAGE revealed that several fractions contain the high purity proteins such as his-tagged (about 1 kD) PTCR (WT) and PTCR (DCterm), which have different protein sizes of about 32 kD and 31 kDa, respectively ( Figure 1A). Subsequently, the highly purified proteins were subjected to the investigation of NADPH-dependent carbonyl reductase activity using 5a-DHT as a substrate. As shown in the Michaelis-Menten plot ( Figure 1B), the rates of 5a-DHT reduction by PTCR (WT) and PTCR (DCterm) were significantly different: PTCR(DCterm) showed much higher reduction rate than PTCR(WT) did. In addition, the kinetic parameters, determined from the Michaelis-Menten plot ( Figure 1B), revealed significant increases in V max (p,0.01), k cat (p,0.01), K m (p,0.05), and k cat /K m (p,0.01) for the NADPH-dependent 5a-DHT reduction by the PTCR(DCterm), compared to the PTCR(WT) ( Table 1). Approximately a 5-fold increase was observed in the 5a-DHT reduction rate (V max or k cat ) by the PTCR(DCterm), with a slight increase in the K m for this reaction, compared to the PTCR(WT). Therefore, the catalytic efficiency (k cat /K m ) for the 5a-DHT reduction increased by V max when comparing the PTCR(DCterm) to the PTCR(WT), primarily as a result of a markedly increased V max (or k cat ) despite of a slightly increased K m for 5a-DHT reduction by the PTCR(DCterm) ( Table 1). Besides, the comparison of their NADPH-dependent carbonyl reductase activities were assessed using other carbonyl compounds such as methylglyoxal, 9,10phenanthrenequinone and hydrindantin as substrates. The PTCR(DCterm) showed significantly higher reduction rates than the PTCR(WT) when using the methylglyoxal and 9,10-phenanthrenequinone as substrates, like using the 5a-DHT as a substrate, whereas it showed no difference when using the hydrindantin as a substrate (Figure 2), which suggests that the C-terminal region (E281 to A288) has a significant effect on substrate specificity of PTCR. Altogether, our results indicate that the C-terminal region (E281 to A288) contributes to the different enzymatic properties, such as the catalytic turnover rate, the substrate binding affinity and even the substrate specificity, between PTCR(WT) and PTCR(DCterm). Furthermore, these imply that the PTCR may have evolved its C-terminal region by generating a paralogous sequence variant through gene duplication.

MD Simulation of WT and C-terminal-deleted PTCRs in
Complex with Its Substrate, 5a-DHT To investigate structural difference between WT and Cterminal-deleted PTCRs upon 5a-DHT binding, 20 ns MD simulations were performed using the docked conformations as initial structures. The detail information about the system used in MD simulations is listed in Table 2. During the simulation time, the root-mean-square deviation (RMSD) for the C a atoms of protein and potential energy were measured to examine the stability of the system. The RMSD values for two systems were calculated with respect to their initial configurations, as a function Each value indicates mean 6 SEM (n = 3). Kinetic parameters were determined using data in Figure 1B

Binding Mode of 5a-DHT at the Active Site of WT and Cterminal-deleted PTCRs
The binding modes of 5a-DHT in WT and C-terminal-deleted PTCRs are analyzed and compared using their representative structures to understand structural differences induced by the substrate binding. Superimposition of WT PTCR with Cterminal-deleted PTCR showed that there was no significant difference with C a RMSDs of 0.94 Å . The catalytic triad (Ser139, Tyr193, and Lys197) and cofactor (NADPH) are well conserved in the active site in both structures (Figure 4). In case of the NADPH binding, it reveals similar conformations in the both structures, which is consistent with typical cofactor binding in the SDRs to facilitate hydride transfer reaction catalyzed by enzyme [2,[4][5][6][7]38,39]. The C4 atom of the nicotinamide ring, specifically 4-pro-S hydride of NADPH, is facing towards to the carbonyl group of 5a-DHT. Hydrogen bonds between NADPH and active site residues such as Asn13, Lys14, Ile16, Gly17 Arg37, Arg41, Ile63, Asn89, Tyr193, Lys197, Val230, Thr232, and Met234 were observed and the residues including Ile92, Val137, Ser138, Thr140, Cys226, Pro227, Gly228, and Trp229 were involved in hydrophobic interactions (data not shown).
On the other hand, 5a-DHT binding has shown slight changes after 20 ns MD simulation compared with our previous results [15]. In the initial docked structure, 5a-DHT in WT PTCR formed hydrogen bonds with Tyr193, Met234, and Pro284 but these interactions were broken during the simulation because the carbonyl group of 5a-DHT is gradually moved away from these residues and then the hydroxyl group forms hydrogen bonds with Lys238 and Asn287 ( Figure 5A). Also, it binds to the active site of WT PTCR through hydrophobic interaction with the residues such as Leu96, Trp229, Met234, Gly235, Pro284, Trp285, Val286, and NADPH ( Figure 5A). In case of the binding conformation of 5a-DHT in C-terminal-deleted PTCR, the residues Gln95, Thr140, Glu141, Trp229, Met234, Gly235, and NADPH have formed hydrophobic interaction with 5a-DHT ( Figure 5B). Unlike WT PTCR, carbonyl group of 5a-DHT in C-  terminal-deleted PTCR makes strong hydrogen bonds with hydrogen atoms of Ser139 and Tyr193 compared with the initial docked conformation. These interactions might contribute to stabilize 5a-DHT binding in the active site of C-terminal-deleted PTCR. Since the hydroxyl group of Tyr193 has been regarded as the proton donor in an electrophilic attack on the carbonyl group of substrate in a reduction reaction [2], hydrogen bond interaction with catalytic Tyr, Tyr193 is especially important for the initiation of the reaction. Although 5a-DHT in WT PTCR binds in a similar fashion with that in C-terminal-deleted PTCR, hydrogen bonds with catalytic residues are only found in C-terminal-deleted PTCR. In addition to, 5a-DHT binding in WT PTCR seems that it somewhat faces toward the C-terminal tail due to hydrophobic interaction with C-terminal tail such as Pro284, Trp285, and Val286.
Comparison of Distance between 5a-DHT and Tyr193, Mediating Catalytic Reaction in PTCR As discussed above, C4 position of the nicotinamide ring in NADPH and C3 position of 5a-DHT is a significant factor for hydride transfer reaction. In other words, the carbonyl group of 5a-DHT should be positioned close to the both C4 atom of NADPH and hydroxyl group of Tyr193 to take place the reduction reaction. In our simulation, the distances from 5a-DHT to NADPH or to Tyr193 are relatively short in C-terminaldeleted PTCR compared with WT PTCR (Figure 5A and 5B). The C3 of 5a-DHT in each representative structure is at a distance of 0.48 nm and 0.34 nm from C4 of NADPH in WT PTCR and C-terminal-deleted PTCR, respectively. The distance between C3 of 5a-DHT and the hydroxyl group of Tyr193 is 0.42 nm in WT PTCR and 0.29 nm in C-terminal-deleted PTCR. These distances were monitored throughout the simulation time. First, the distance between carbon C4N in the nicotinamide ring of NADPH and carbon C3 in carbonyl group of 5a-DHT is similar in both simulations until 14 ns but since then it is increased in WT PTCR, while the distance is maintained at 0.4 nm in C-terminal-deleted PTCR ( Figure 5C). The average distance during last 5 ns is 0.51 nm and 0.39 nm in WT and Cterminal-deleted PTCRs. Second, the distance between the hydroxyl group of Tyr193 and C3 in the carbonyl group of 5a-DHT is also revealed comparable pattern with above one. The distance in WT PTCR is gradually increased up to 0.8 nm from 10 ns, whereas it is kept around 0.4 nm in C-terminal-deleted PTCR until the end of the simulation ( Figure 5D). The average distance for last 5 ns is 0.49 nm and 0.37 nm in WT and Cterminal-deleted PTCR, respectively.
The comparisons of two major distances show that the distances in C-terminal-deleted PTCR are relatively shorter than that of WT during the 20 ns MD simulation. Moreover, this indicates that the binding of 5a-DHT in C-terminal-deleted PTCR makes the protein more favorable condition for initiating the reaction.

Comparison of WT and C-terminal-deleted PTCRs from an Energetic Perspective
From the each simulation, coulomb and Leonard-Jones potentials between the protein and 5a-DHT were calculated during the entire simulation time. The coulomb potentials are slightly low in C-terminal-deleted PTCR with average potentials of 257.62 kJ/mol while 243.38 kJ/mol for WT ( Figure S1). On the contrary, Leonard-Jones potentials are relatively low in WT PTCR with average value of 2123.10 kJ/mol and 2100.72 kJ/ mol for WT and C-terminal-deleted PTCRs, respectively ( Figure  S1). The calculation is also completed using four repetitive simulations called Rep1 to Rep4 (Table 3). Similarly, the results show that strong coulomb potentials in C-terminal-deleted PTCR while strong Leonard-Jones potentials in WT PTCR. We have also calculated interaction energy estimated as the sum of the van der Waals energy and electrostatic energy using DS. All individual representative structures from the repetitive simulations were used in the calculation and each value was summarized (Table S1). Although the average of interaction energies is not significantly different in both systems with 240.23 kJ/mol and 234.51 kJ/mol for WT and C-terminal-deleted PTCRs, this study reveals that WT PTCR has more stable interactions in terms of van der Waals energy, whereas C-terminal-deleted PTCR has somewhat more stable electrostatic energy. Taken together, these differences might be resulted from that 5a-DHT in WT PTCR forms more hydrophobic interactions with the C-terminal residues while Cterminal-deleted PTCR has more hydrogen bond interaction than WT during the simulation time ( Figure S2).

The Binding Conformation of 5a-DHT in C-terminaldeleted PTCR Is More Favorable for Catalytic Reaction
To further explore the influence of C-terminal deletion on 5a-DHT binding, essential dynamics analysis was performed using MD simulation trajectories of both systems. By calculating eigenvalues and eigenvectors obtained from covariance matrix based on 5a-DHT or protein C a atoms, we identified configurations which showed the largest correlated displacements throughout the simulation time. From the comparison of the extreme  These are drawn as diagrams to clearly show the difference in binding conformations between two structures. The protein and C-terminal tail of WT PTCR is illustrated as U-shape (gray) and rectangle (cyan), respectively. The residues involved in the interaction are labeled and red bars indicate hydrogen bonds with 5a-DHT. doi:10.1371/journal.pone.0090712.g008 configurations, we observed that b-face of 5a-DHT was rotated in the opposite direction in both WT and C-terminal-deleted PTCRs ( Figure 6A and 6B). To examine the rotational movement of the bface of 5a-DHT, the distance between the methyl group at C10 position of 5a-DHT and C a atom of Gly194, facing to the substrate, was considered. Since b-face of 5a-DHT is directed towards the active site for catalytic reaction, the distance should be retained short during the simulation time. The distance between methyl group of 5a-DHT and Gly194 is relatively short in Cterminal-deleted PTCR compared to the WT ( Figure 6C). In the simulation of WT PTCR, b-face of 5a-DHT has been rotated and far from Tyr193 about 15 ns due to hydrogen bond interaction with Glu141 and then finally placed on the opposite side of the catalytic site. In case of C-terminal-deleted PTCR, b-face of 5a-DHT was not headed for the active site in the initial stages of the simulation but it slowly turned toward to the catalytic site and was kept properly until the end of the simulation. The comparison of the distance revealed that C-terminal-deleted PTCR retains the right direction of 5a-DHT longer than WT PTCR. Furthermore, superimposition of extreme configurations calculated based on all C a atoms in the protein indicates that hydroxyl tail of 5a-DHT is alienated from active site due to interaction with the residues in Cterminal region in WT PTCR (Figure 7). Changes in distances were considered to confirm the structural correlation between 5a-DHT binding and C-terminal tail. As represented in the figure, C a atoms of Gln95 and Pro284 which were placed in spatially parallel with center of 5a-DHT were used in the distance calculation ( Figure 7A and 7B). In WT PTCR, the changes in the distance from Gln95 to 5a-DHT and Pro284 are shown a similar pattern with each other, while the distance between Pro284 and 5a-DHT is maintained at constant value over the simulation ( Figure 7C). This indicates that C-terminal tail has formed stable interactions with 5a-DHT and it might restrict the substrate to approach to the catalytic residues. However, there is no significant change in the distance between Gln95 and 5a-DHT in C-terminal-deleted PTCR ( Figure 7D). From the essential dynamics analyses we demonstrate that the rotational displacement for b-face of 5a-DHT and structural correlation between 5a-DHT binding and Cterminal tail in WT. These observations are supported by distance calculations and the structural comparisons using extreme configurations.

Conclusions
In the present study, the kinetic studies using 5a-DHT revealed that deletion of C-terminal tail had significant effects on enzymatic properties of PTCR. While K m for 5a-DHT was slightly lower in WT than that of C-terminal-deleted PTCR, reduction rate and catalytic efficiency for 5a-DHT were remarkably increased in Cterminal-deleted PTCR compared to the WT. To find structural differences upon 5a-DHT binding in WT and C-terminal-deleted PTCRs, 20 ns MD simulations for the both complexes were performed and the reasonable explanations for these differences have been discussed in this work. In case of 5a-DHT binding in WT PTCR, it mostly depended on hydrophobic interactions with active site residues and hydrogen bonds with Lys238 and Asn287 were also found. On the contrary, C-terminal-deleted PTCR shows strong hydrogen bond interactions with catalytic residues such as Ser139 and Tyr193. In addition, the distance from 5a-DHT to NADPH or to Tyr193, which is the important distance for hydride transfer reaction, was relatively short in C-terminaldeleted PTCR than that of WT during the whole simulation time.
From the structural comparisons based on the distance and interaction energy calculations as well as essential dynamics analyses, our results strongly suggest that 5a-DHT binding in Cterminal-deleted PTCR is structurally more favorable for catalytic reaction compared to the WT. Also, these explanations are in accordance with our experimental result that C-terminal-deleted PTCR has higher catalytic efficiency for 5a-DHT than that of WT. On the basis of the overall results obtained from molecular modeling studies, we propose the schematic diagram to easily understand the difference of 5a-DHT binding between WT and C-terminal-deleted PTCRs (Figure 8). Since molecular dynamics studies for explaining enzymatic properties of PTCR have not studied much until recently, we believe that our findings can provide very meaningful information to understand substrate specificity of PTCR and further kinetic properties of enzymes belonging to the SDR superfamily. Figure S1 Interaction energy between the protein and 5a-DHT. Table S1 Interaction energies between the protein and 5a-DHT. The energies were calculated using representative structure of each repetitive simulation and given in kJ/mol. (DOCX)