Erix Wiliam Hernández-Rodríguez
Yüklə 3,6 Mb.
|
- Bu səhifədəki naviqasiya:
- [-5.42 -1.86] ** 1U19 [a] -36.12
- [-0.09 -0.06] ***
- III. Spectral Analysis
- [-0.24 -0.07] *** ([12.91 45.04]) ***
- [-0.36 -0.24] *** ([43.32 64.75]) ***
- [-0.39 -0.21] *** ([37.46 69.60]) ***
- REFERENCES Merin, S.; Auerbach, E. Surv Ophthalmol. 1976
RESULTS AND DISCUSSION I. Structural Analysis. The large structural similarity between bWT and hWT (not only in the RBP, but for entire structure), besides the biochemical-information wealth for bovine rhodopsin, give an appropriate reference point for evaluating our results from hWT and mutants. All amino acid residues matching for bWT and hWT, in each stratus and within a sphere of 4.5 Å starting from retinal (6-s-cis, 11-cis for both WTs and mutants), were selected; it was found a sequence identity of 100% and a RMSD of 0.5 Å for all atoms between WT structures; the high similarity was also for water molecules at the retinal binding site. Hence, the RBPs of both WTs structures are very similar. A. 3D-Models Evaluation. Structures were reliable and consistent according to the quality values of 3D models, which are similar or better than the values obtained from the starting structure (Table 1). A better compatibility of the 3D environment for every residue in the sequence, taking into account the 3D-1D scores, and a best packing quality with more positive Z-scores (second generation packing quality) was obtained for all models regarding the crystalline structure. All 3D models had a value near to 90 % for the most favoured regions and higher than the crystal value, even when this template structure was resolved at 2.2 Å. The homology models and the chain-B crystal were usual structures, considering the G-Factor above -0.5. On the other hand, the backbone RMSD values calculated between models and the crystalline structure were pretty low; so, the general structure of 3D models was similar to the crystal; probably bWT was resolved at a lower resolution than the chain-B crystal with the methodology used being also appropriate for the hWT and mutants structures with a very similar amino-acid residue sequence. The RBP core and Cys110-Cys187 disulfide bond were conserved for all systems. This evaluation shows the high quality of the 3D models built in this work.
B. Similarity between RBPs. Conformations obtained from the classical MD trajectories with the constrained protein within membrane (Figure 1) had a RBP very similar to those from the structures picked from QM/MM-MD trajectories and optimized by combined QM/MM method. 
Figure 1. Complex rhodopsin/membrane/water/ions. The residue geometries, 11-cis-retinal conformations and significant hydrogen bonds related to spectra were conserved nicely between these retinal binding sites; RMSD values, using VMD for RBP comparison, averaged 1.60 for one snapshot with restrained protein within membrane (RbWT) against (vs) every snapshot of four structures from QM/MM MD and optimized by QM/MM (bWT) (Figure 2a), 1.60 (RhWT vs hWT) (Figure 2b), 1.38 (RhM207R vs hM207R) (Figure 2c) and 1.80 (RhS186W vs hS186W) (Figure 2d). These calculations included all heavy atoms. It is assumed that these rhodopsin structures achieved by the second minimization-equilibration cycle, constraining the protein, are the nearest to the X-ray starting point and according to the reality for those calculated within membrane, its spectral properties and others can be inferred from those structures from the structures picked from QM/MM MD and optimized by the combined QM/MM method because the high similarity between its retinal binding sites (Figure 2). 
Figure 2. High similarity is displayed by the overlay of RBPs selected from QM/MM MD trajectories with a representative RBP selected from the classical MD with the rhodopsin constrained within the membrane, for every system (see text): a) bWT, b) hWT, c) hM207R, d) hS186W. Note the perturbed hydrogen-bonding network at the retinal binding sites of both mutants. Significant conformational deviations from the starting points were reached for some systems at the end of the classical MD with the unconstrained protein (FMD) within membrane. It was expected as extensive deviations from crystal structure have been found in membrane-rhodopsin simulations (99). A large conformational change was observed for Glu181 in bWT system, which moved its carboxylic group away from retinal (Figure 3 a), a no observed conformation in previous structures (Figure 2a), as well as, for Arg207 in hM207R system, which moved its guanidino group away from the cyclohexenil ring of retinal (Figure 3b), acquiring a different conformation with regard to previous structures, where the guanidino group pointed to the ring (Figure 2c). The geometrical properties for the RBP of these structures were calculated on picked snapshots from this MD and optimized, subsequently, with the combined QM/MM approach; its spectral properties were calculated using the methods for UV-Visible spectra calculations above-mentioned. Additionally, the hydrogen-bonding network at the retinal binding site appeared disturbed for both hM207R (Figures 2c, 3b) and hS186W (Figure 2d) mutants. 
Figure 3. Large conformational deviation is showed for Arg207 during classical simulation with protein constrained within membrane: (a) overlay of RBPs (hM207R) selected from both QM/MM MD and classical MD (protein restricted in membrane) trajectories (1) with RBPs selected from the classical MD with the rhodopsin unconstrained within the membrane (2); (b) overlay of RBPs (bWT) selected from the classical MD (protein unrestricted in membrane), it is displayed a significant conformational deviation for Glu181 regarding its more probable conformation (Fig. 2a, 2b). C. Stability. The energy values during classical MDs (Charts 1a and 1b) and QM/MM MD (Chart 1c) or the RMSD values while classical MDs ran with constrained protein (Chart 2a) or not (Chart 2b) within membrane, showed very little variation; a constant stability and geometry was reached for the entire protein with all systems during simulations; so, the structures obtained are not artifact. Similar result was found for the RMSD of the QM/MM MD and thus not shown; besides some atoms were frozen in this calculation. Although, WTs and mutants had this behavior and the QM/MM MD showed similar stability for all systems (Chart 1c) [QM/MM MD was performed for sampling chromophore conformations and from residues around it for later QM/MM optimizations and spectra calculations rather than to analyze stability for the complete protein (i.e., some atoms far away from active region were kept frozen for this simulation, like those belonging to different structural points between WTs and mutants)], a destabilizing effect caused by the mutations M207R or S186W on rhodopsin structure within membrane was observed during classical MDs, these simulations gave a more realistic vision on rhodopsin stability in its natural environment. 
Chart 1. Comparison of the energies during the dynamics with rhodopsins constrained (a) or not (b) within membrane and the QM/MM MD. Both mutations lead to a less stable rhodopsin (see text). 
Chart 2. Comparison of the RMSD (Å) during the dynamics with rhodopsins constrained (a) or not (b) within membrane, the expected deviations during the second dynamic (b) were achieved at the end (see text). Both mutated rhodopsins within membrane (Figure 1) had a higher energy (less stable) than bWT and hWT (more stable) systems during the classical MD with the protein constrained avoiding large structural deviation (Chart 1a) and a very similar result was obtained while a final MD (unconstrained protein) ran looking for large conformational deviation from the starting points (Chart 1b). Therefore, structural differences related to the stability between normal and mutated systems are suggested; two aspects could be considered: (a) protonation state of Cys316 was unprotonated for mutants (protonated in WTs) giving an additional interaction (sidechain hydrogen bond) with Lys67, which is not present in normal structures; this interaction in mutants could change regional conformations and/or the protein-membrane interaction varying the energy system in regard to WTs during the MDs in membrane; even without affecting the RBP geometry, no effect has been reported for this residue in the dark state, absorption spectra or photoisomerization. Cys316 is implicated in protein-G activation (54) and was kept frozen during QM/MM MD calculations. (b) The hydrogen-bonding network at the retinal binding site has been altered in both mutants M207R (Figure 2c) and S186W (Figure 2d, it is broken) regarding bWT (Figure 2a) or hWT (Figure 2b); this network is very important for to stabilize thermally to the protein and to prevent the thermal isomerization, according to previous studies on bovine rhodopsin (100). Hence, the rupture of the hydrogen-bonding network at the RBP caused by M207R and S186W mutations in human rhodopsin can decrease the thermal rhodopsin stability and affect the photoisomerization process; being probably, the most important contribution to the mutant instability displayed in this work, in good agreement with recent experimental and computational studies on WT rhodopsins; even when, the folding and retinal binding in these mutants can occur somehow (33-35). The stability in function of pH was analyzed using PROPKA 2.0 for homology models (MODELLER) and homology models optimized by MM (NAMD) (Chart 3). Expected results from stability was obtained for all homology models (Chart 3a), at physiological pH, a substantial stability reduction was observed for S186W mutant; which had the bigger damage to the hydrogen-bonding network (Figure 2d) at the retinal binding site. The M207R stability was slightly lower than hWT stability; it was consistent with previous molecular analysis and data, a recent study reported that the folding for M207R mutant is more favored than for S186W, using FoldX algorithm (37). However, a different behavior was found for optimized homology models (Chart 3b), the stability was increased for the 3D models as expected because all structures were energy minimized appropriately; curiously, the highest stability vs pH was for M207R mutant and no an important stability decrease was found for mutants regarding WT structures at physiological pH; suggesting that similar conformations to WT rhodopsins could be possible for mutants; studies pointing to a possible folding with retinal binding (33-35). Recent advances propose to use small molecules in order to stabilize proteins tending to misfold (101, 102) and a very recent study considers that researches on mechanisms taking place for the Class II mutants associated to RP could be especially valuable and to help in guiding mutant-specific and personalized treatment of the disease from a structural, functional, and energetic point of view, as patients carrying mutations that destabilize rhodopsin could be treated using these small molecules (37). Our results showed that M207R and S186W mutations can affect the stability of human rhodopsin in some grade and to suggest that more stable structures could be reached with actions changing the mutant conformations, which cloud be possible with stabilizing molecules, considering the advances made recently. 
Chart 3. (a) The lowest stability vs pH was observed for hS186W mutant and it was lightly low for hM207R before a better MM optimization; (b) An acceptable stability for both mutants can be achieved by changes of conformations due to energy minimization. On the other hand, the knowledge on mutation mechanisms further the stability is crucial and valuable for substitutions near chromophore, i.e., M207R or S186W, because abnormal processes related to optical spectra and/or photoisomerization could also impair rhodopsin functions, even when a stable conformation enabling the folding, retinal binding and absorption spectra takes place through a natural way or/and using artificial molecules. The following geometrical, electronic and spectral analyses were made on the structures obtained from the QM/MM MD and subsequently optimized by the combined QM/MM approach, the more reliable and nearest structures to the starting points with properties attributable to the structures from classical MDs with the constrained protein in membrane describing rhodopsin properties within its natural environment. D. Planarity of the β-Ionone Ring of Retinal. The β-ionone ring orientation was 6-s-cis at C6-C7 bond; the negative values of the C5-C6-C7-C8 dihedral angle (φC5-C6-C7-C8), for all systems like it is reported for chain-B of the 1U19 crystal (-31.86°), were: bWT (-48.06°±1.49), hWT (-43.35°±1.44), hM207R (-36.86°±1.52), hS186W (-44.61°±3.86). In good agreement with previous predictions made by an ONIOM-EE scheme (-44.00°), using a B3LYP/6-31G*:AMBER level of theory (103) or by classical MD (-52.00°) and QM/MM MD (-42.00°) (104). 6-s-cis conformation was found in all X-ray studies on bovine rhodopsin (26), being the conformation with energetic preference in this protein (16). An important geometrical parameter of retinal is the C4-C5-C6-C7 dihedral twist angle (φωC4-C5-C6-C7) of the β-ionone ring related to its planarity; a strong deviation with a negative value was observed for hM207R system (-178.88°±0.32) with regard to crystalline structure (170.63°) and the other 3D models, bWT (178.66°±0.70), hWT (177.36°±0.43), hS186W (177.01°±0.74). The cyclohexenil ring is affected by the alien guanidino group from Arg207 because the M207R substitution (Figure 2c). This deviation of hM207R tending to a vanishing twist angle but toward other sense, can disrupt the π-system; which is strongly correlated when the planar arrangement of the six double-bonds is possible (40) as in the other studied systems, considering this angle. In hM207R mutant, less double-bonds could contribute to the excitation, leading to higher vertical excitation energies with respect to systems conserving similar values for this dihedral angle to those reported for WT structures. A hypsochromic shift versus the normal planar structure (positive values for this φω near 171°) has been reported when the φωC4-C5-C6-C7 goes away from the planarity (40). E. C11-C12 Double-Bond in the Dark State. The conformation for φC10-C11-C12-C13 was 11-cis for all structures; both bWT and hWT showed values of ~12.00° in accord with theoretical studies reproducing similar values for the dark state as: -11.00° [bovine structure 1F88, Chain-A, ONIOM-EE (B3LYP/6-31G*:AMBER)] (103) and -17.00° [bovine structure 1HZX, Chain A, classical MD] or -16.00° [bovine structure 1HZX, Chain A, QM/MM MD] (104), these MD studies had large standard deviations for φC10-C11-C12-C13. The obtained values for WTs are reasonable with the starting structure (1U19); large deviations were found from X-ray studies for retinal (26), QM/MM calculations reach values more reliable for the chromophore in the protein environment (16). Although, appropriate φC10-C11-C12-C13 values were obtained for WT structures in the dark state, both hM207R and hS186W displayed a value of ~ 9.00° significantly less negative than WTs (Table 2); it is showed results for hWT vs mutants comparisons, similar statistical signification (p<0.001) for comparisons between bWT and mutants are not shown. The protein environment in WT structures would favor a stronger steric strain on 11-cis retinal and to ensure a better pretwisted φC10-C11-C12-C13 because more negative values were observed, allowing an appropriate selection at C11-C12 double-bond for the photoisomerization; this environment effect is important because an inversion of the bond-length pattern takes place in the excited-state configuration of the RPSB and the electronic structure is unselective toward the rotation of the double-bonds of the polyene chain (105), being decisive the protein environment for a selective isomerization. Retinal binding sites of hM207R and hS186W mutants induced a weaker strain and to lead to a less pretwisted φC10-C11-C12-C13, because significant less negative values were observed for mutants. These substitutions could disturb the appropriate dark state geometry for a selective and efficient photoisomerization at C11-C12 double-bond or to favor less this process with regard to WT structures. Additionally, the φωC4-C5-C6-C7 is also affected in hM207R; the ring position is sufficiently conserved during normal isomerization in WT systems, its restrained position should be kept during the reaction (103, 104). On the other hand, SBL is unprotonated in hS186W; hence the quantum photoisomerization yield can be decreased for this mutant as it was reported that bovine rhodopsin with an unprotonated SBL has a quantum yield for the isomerization reaction lower than that of WT rhodopsin with a RPSB (106).
F. Bond Length Alternation (BLA). Single/bond BLA, calculated as the sum of all formal single-bond lengths minus the sum of all formal double-bond lengths along the conjugated carbon chain between the C6 and N atoms like in a previous study (104), plays an important role in spectral tuning within the protein environment (16, 40, 104). A substantial higher average BLA for every system was observed for the hM207R and hS186W mutants (Table 3), it is showed results for hWT vs mutants comparisons, similar statistical signification (p<0.001) for bWT vs mutants comparisons are not shown, besides a significant difference was also found between mutants, being the average BLA from hS186W higher than that of hM20R system [CI (0.01 0.10), p<0.05*]. The BLA values for the crystalline bovine structure are very well represented by other studies from pure crystal or treated with QM/MM methods (16, 26), those data are in fully agreement with our results for bWT system.
The average BLA for every system and bond length showed a behavior in accord with the previous BLA calculated, displaying the highest BLA values for mutants and additional details (Chart 4): only, BLA of hM207R was higher than BLA of WTs at C8-C9 single-bond. Starting from C10-C11 single-bond, BLA of hM207R was lower than that of hS186W but higher than that of WTs, at C9-C10, both mutants BLA were similar each other and higher than BLA of WT systems. The mutant hS186W showed the highest BLA values, starting from C9-C10 double-bond until C15-N16 double-bond (SBL). The results of both BLA calculated were very similar for both WT structures. 
Chart 4. An increased BLA is displayed due to both mutations, the values were very similar for both WTs and hS186W showed a higher BLA than hM207R mutant, the results suggested different mechanisms between mutants.*Bond length expressed in Å. Although, the analysis for every system and bond length also supports a higher BLA for mutants, some deviations are observed for this value among mutants, pointing to possible differences in the electronic structure. It is known, that the Highest Occupiest Molecular Orbital (HOMO) is better localized on ionone ring region with higher BLA values, besides the electrostatic counterion influence on BLA and the positive correlation of BLA with the dipole moment (µ) (104), the vertical excitation energies (16), and positive charge along the polyene chain (the higher the positive charge, the higher the BLA) (27). Since, the stronger BLA was for mutants studied, an electron transfer decreased is expected from the ionone ring region to the nitrogen atom of the SBL of both hM207R and hS186W in the ground state (S0) of the 11-cis retinal in the dark state, leading to a more localized HOMO in the cyclohexenil ring and an increased positive charge on the conjugated carbon chain, as well as, significant changes of µ and the absorption spectra regarding WT structures; although different mechanisms can take place for each mutant. II. Electronic and Electrostatic Analysis. Electrostatic and electronic environments of the RBP influence on the spectral tuning (27, 40), the applied combined QM/MM approach included both electrostatic and electronic polarization contributions on the QM wave function caused by the surrounding MM charges, as described in Methods Section. The QM region included the system 11-cis retinal attached to Lys296 moiety (via SBL), two water molecules and the Glu113 moiety; hence, a lower µ of the whole QM region in the S0 would correspond to a more polarized system chromophore-Lys296 moiety, leading to a higher µ of the 11-cis retinal or a more polar chromophore in the S0, because the Glu113 (as carboxylic or carboxylate, depending on the system) and β-ionone ring are placed at opposed ends. A. Electronic Polarization, Electrostatic Influence and Charge Transfer. Mutants showed a significant decrease of µ in S0 (µS0) (p<0.001) for the entire QM region regarding WT structures (Table 4), pointing to highest µS0 of 11-cis retinal for hM207R and hS186W systems; a more polarized chromophore in hM207R is possible due to the electronic polarization that the positive charge from guanidino group of Arg207, pointing to the β-ionone ring, would cause because its proximity of ~2.0 Å to both methyl groups of the ring attached to C1 and C5 atoms. The electron transfer from these methyl groups to the nitrogen atom of the PRSB would be decreased in the hM207R mutant due to the increase of positive charge around methyl groups made by guanidino group, even when the Glu122 is deprotonated in this system, which had its carboxylate group far from the ring and compensated through hydrogen bonds with surrounding amino acid residues, the influence from its negative charge is decreased. Additionally, the electronic density showed a large electrostatic influence caused by the counterion Glu113 (carboxylate form) on the system 11-cis retinal-Lys296 moiety in the hM207R mutant (Figure 4c); hence a better location of the positive charge on the PRBS is possible; even when the saline-bridge length (2.73 ű0.02) was similar to that of hWT (2.71 ű0.01) or bWT (2.87 ű0.01), the second oxygen from Glu113 was more fence in this mutant for all snapshots; it also prevents the spontaneous electrons flow to the nitrogen of the positive PRSB. Theoretical data supports that this charge must be shifted against the negatively charged counterion and to relate this shift to the increased dipole moment of the S1 state relative to that of S0 (107). Obviously in this mutant, both the additional electronic polarization by the guanidino group and the stronger electrostatic influence keep the electrons on the cyclohexenil ring rather than on the polyene chain leading to a more localized HOMO in the ionone ring region (Figure 4a) and to increase the positive charge in the conjugated carbon chain and its location by the counterion; so, the electrons of the QM region are more paced at the opposed ends (cyclohexenil ring and counterion Glu113), explaining its decreased µS0 and to suggest a more polarized 11-cis retinal with regard to WT structures. These finding are in accord with the results for BLA, which were higher in mutants; moreover consistent with studies on bovine-rhodopsin mutant reporting an electron transfer decreased from the methyl group, attached to C1 atom, to the nitrogen atom of PRBS in E122Q mutant as negative charge around it is halved with the mutation, explaining that the polyene chain possesses more positive charge, resulting in increased BLA (16, 27). 
Figure 4. Locations of HOMO (a) or LUMO (b) and Electronic Density (c) from a representative snapshot, using MRCI for every system were consistent with BLA and µ, supporting changes in HOMO location or electrostatic and electronic influences due to both mutations (see text). Similar behavior of µS0 and also a more polarized 11-cis retinal (QM-region µS0 is significantly lower than that of WTs) was found for hS186W, answering to a different mechanism from hM207R; now, a QM region end has the Glu113 moiety (carboxylic form with a partial negative charge on oxygens) and the other one, the ionone ring region, besides the SBL is unprotonated. Interestingly, an increased BLA as above-mentioned in this mutant regarding normal structures and a more localized HOMO on the β-ionone ring (Figure 4a) suggest an increased positive charge on the polyene chain. Nevertheless, neither a notable electrostatic influence from Glu113 is showed (Figure 4c), nor an additional electronic polarization from a positive charge on methyl groups take place in this system. This condition reveals that the natural or spontaneous tendency of electrons in the system 11-cis retinal-SBL (unprotonated)-Lys296 within the protein environment is to stay on the β-ionone ring, when there is not a positive charge located on the SBL (PRBS) attracting the electrons to move toward it (to the opposed end); so, it is not necessary an additional electrostatic influence from Glu113 for obtaining a significant increased positive charge on the conjugated carbon chain in this mutant. It seems that, the electronic environment conditioned by amino acid residues with its appropriate protonation states around the ionone ring portion of 11-cis retinal in the Dark State from WT structures, which were conserved in hS186W mutant, favors this tendency normally; in this mutated case a better localized HOMO on the β-ionone ring is possible since it is not present the positive electrostatic influence of the PRBS in WTs. In fact, the best location on the ionone ring region of the HOMO was observed for hS186W; additionally, its Lowest Unoccupiest Molecular Orbital (LUMO) was the best placed on the SB moiety (Figures 4a and 4b), leading to the highest HOMO-LUMO distance. This result is in fully agreement with the highest BLA values calculated for this mutant.
On the other hand, a substantial difference was observed between WTs and mutants for the transition dipole moment TDM of S1 (TMDS1) and among mutants. In hM207R, TMDS1 was significantly lower than that of hWT (Table 4) or hS186W [CI (-9.37 -7.16), p<0.001***]. This TMDS1 decreased for the entire QM region indicates that the µ of the 11-cis retinal in S1 (µS1) is increased and a more restricted motion of the excited electrons for this system in S1, which is possible because the electronic polarization by the positive charged near retinal ring and the larger electrostatic repulsion due to counterion Glu113 (Figure 4c) preventing the electrons motion through the conjugate carbon chain to the nitrogen atom of the PRSB and to induce a higher µS1 of the 11-cis retinal than the observed for WTs and hS186W. TMDS1 for hS186W was also significantly lower than that of hWT, but the difference was not substantial, considering the CI information, nor important regarding bWT; so, the electrons were less restricted to move through the polyene chain to the nitrogen atom of the RSB, there is no a repulsive electrostatic force exerted from a counterion as Glu113 is uncharged and the saline bridge is broken in hS186W mutant, the length from the nearest oxygen from Glu113 uncharged to the SBL uncharged and its nitrogen atom was of 6.64 ű0.11, consistent with the Glu113 protonated and SB unprotonated foretold using the calculated pKa by PROPKA 2.0 for this system and considering the physiological pH. Studies on rhodopsin bovine report the massive flow of negative charge from the conjugated carbon chain towards the positive nitrogen atom following excitation in a model without counterion (108, 109). Dipole-moment changes were compatible with those for the BLA and related to strong absorption-spectra deviations observed for both hM207R and hS186W structures. III. Spectral Analysis. Excited states (from S1 to S10) of WT and mutants were calculated using DFT/MRCI and TDDFT, as described in details in the Methods Section. The S1 and S2 energies are summarized for all structures (Table 5) and to show that the S1 vertical-excitation energies and the λmax calculated in this work for WT structures, using both spectra-calculation methods, are in a very good agreement with the experimental data, for bovine or human rhodopsin, and previous theoretical approaches having calculated the optical spectra of bovine rhodopsin (E=2.487 eV or 57.4 kcal/mol, λmax=498 nm) (11, 16, 40, 103, 109).
Additionally, a significant difference was also found between mutants, the higher S1 energy and the stronger hypsochromic shifts of λmax were for hS186W instead of hM207R (p≤0.01), (Charts 5b and 5c, DFT/MRCI). Despite both DFT/MRCI and TDDFT were used for calculation, only, the results from DFT/MRCI and the first excitation are further discussed here. TDDFT results were very similar to those from DFT/MRCI and to lead to the same conclusions, thus are not shown. Our calculations made for absorption spectra are consistent with previous experimental or theoretical data and with other results obtained in this study (geometrical or electronic and electrostatic analysis) and a quite good precision is showed (Chart 5b); its reliability was strengthened because very similar results were achieved with different structures for every system, which were calculated and picked from a large region in the configuration space of the 11-cis retinal and its binding site, generated by the corresponding MDs; so, the agreements were not casual. 
Chart 5. (a) Increased vertical excitation energies and (b,c) a significant blue-shift of the λmax for both mutations were calculated in all optimized snapshots by QM/MM and selected from the QM/MM MD trajectories; (d) the significant increased energy and blue-shifted light absorption remained for hS186W mutant only when large conformational deviations from the starting points were achieved for Arg207, supporting the influence of M207R mutation on the optical spectra. A. hM207R. The optical-spectra pattern calculated for the mutant hM20R is in accord with the blue-shifted λmax reported by an experimental study; in which was also observed a strong hypsochromic shift of λmax (380 nm) and the possibility for an unprotonated SBL was outlined based on this value obtained from the detected UV-Visible spectra for this mutant (33). An equal spectral pattern to WT rhodopsin with an unprotonated SBL for this system, if an unprotonated SBL would be the cause only, could be unlikely due to different retinal binding pocket between this mutant and WT for the 207 position with a strong positive charge in the first one. Λmax of 380 nm is reported for WT rhodopsin with unprotonated SBL (10), the hS186W mutant showed an unprotonated SBL with a strong blue-shifted λmax but it was not 380 nm exactly (both RBP are not similar). On the other hand, we also found a strong blue-shift for hM207R mutant toward a value of ~3.0 eV using DFT/MRCI, showing a similar spectral behavior and pKaSBL values pointing to a PRBS: (8.54) (homology model) or 8.60 (optimized homology model) and the same tendency during the simulations was observed for this value; 9.24 (QM/MM-MD snapshot); 8.59, 9.24, 9.33 (MD snapshot from lipid membrane at 500 ps, 1ns and 3 ns, respectively) and 8.68 (QM/MM optimization snapshot), its corresponding pKaGlu113values were 3.88; 4.61; 3.21; 4.64, 3.21, 3.35 and 4.46; taking into account the pH near 7.2, values for WTs systems and hS186W mutant [i.e., for bWT, 3.82 (pKaGlu113) and 8.79 (pKaSBL) or for hWT, 4.25 (pKaGlu113) and 8.61 (pKaSBL), for the unprotonated SBL in hS186W, 8.31 (pKaGlu113) and 5.88 (pKaSBL)] and additionally, as described above, the distance between the carboxylic group from the Glu113 and the nitrogen atom from the SBL is similar to those for WTs systems, the Glu113-RPSB saline bridge is formed for hM207R mutant in this work. The blue-shift vs WTs is explained by the electronic polarization caused by the guanidino group from Arg207 and the stronger electrostatic influence from the Glu113, which is conceivable because the additional and explained electronic polarization caused by the M207R substitution with a more localized HOMO on the ring and positive charged on the SBL moiety, as well as, an increased electrostatic repulsion caused by the stronger influence of Glu113 than that of WTs would lead to the destabilization of the S1 state and an increased vertical-excitation energies. It is reported the contribution of the counterion for the blue-shift in the normal bovine rhodopsin by means of this mechanism (109). In the mutant hM207R, the electrostatic influence is stronger considering the dipole-moment calculated and the electronic density showing a larger counterion effect (Figure 4) and the increased BLA found in this mutant regarding WT structures (Chart 4); it is known that the BLA is positively correlated with the vertical excitation energies and its variation is induced electrostatically and not sterically (110). On the other hand, the charge transfer associated to the first excitation, which carries a portion of the positive charge from the SBL moiety to the allyl moiety inducing a large dipole-moment change, as described above, must work more in the hM207R mutant against the stronger electric field of the counterion due to its larger influence in this case in regard to WTs. Studies support that an important contribution to the electrostatic environment of the retinal chromophore in rhodopsins (WT) comes from negatively charged groups in the RBP, causing a significant hypsochromic shift in the S1 excitation energy because the charge transfer works against the electric field of the counterion and a resulting blue-shift is obtained (40). B. hS186W. The largest blue-shift was achieved for the mutant hS186W; nevertheless, the conformations of amino acid residues near ionone ring region causing electronic polarization are not very different to those in WTs and no repulsive electrostatic influence from a counterion is possible as Glu113 is uncharged in this system. Since, the SBL is unprotonated, a missing positive charge on SBL carries to a loss of electrostatic forces moving electron to the SB fragment. Electrons would stay on the ring due to its spontaneous tendency in the retinal within the protein environment, giving the best localized HOMO on the ring, as above-mentioned, and consistent with the strongest BLA calculated for this system (Chart 4); probably, the best defined HOMO-LUMO location (HOMO and LUMO toward the β-ionone ring and the SB moiety respectively, Figures 4a and 4b), would be the best contribution to the vertical-excitation energies for this mutant, more than a S1 destabilization due to electronic polarization, electrostatic repulsion or charge transfer, its S1 energies were the highest calculated for all structures studied in this work. This mechanism could take place in normal blue-shift intermediate of rhodopsin activation with an unprotonated SBL. A
C. Conformational Deviations. For completeness, the MD with the protein unconstrained with membrane was used in order to look for large structural RBP deviations far from the starting point or real structure for evaluating its influence on spectral absorption. At the end of this simulation, it was found significant conformational deviations in the retinal binding sites from bWT (Figure 3a) and hM207R (Figure 3b), explained previously. The calculations of the UV-Visible spectra using MRCI on structures (snapshots) picked from this classical MD and subsequently optimized with the combined QM/MM approach were: bWT [2.48 eV±0.05 (500 nm±9.29)], hWT [2.57 eV±0.07 (483 nm±13.08)], hM207R [2.45 eV±0.03 (507 nm±5.86)], hS186W [3.75 eV±0.32 (332 nm±30.35)], the SD showed a high variability in the results due to features of this MD giving systems more distant from the proper structures. Only, hS186W mutant had a significant blue-shift with regard to WTs, [CI (51.98 250.02), p<0.05] for the comparison between hWT and hS186W (Chart 5d, average). This kept behavior responds to a lasting structural change, not conformational only, caused by the S186W mutation leading to a missing counterion and an unprotonated SBL; the previous mechanism remains even after this simulation for obtaining conformations far from the reality, which were not found in other structures of this work, including those from the QM/MM MD. On the other hand, this theoretical experiment reveals the strong influence of the electronic polarization (working with the repulsive electrostatic influence from counterion) due to the guanidino group from Arg207 in hM207R mutant leading to a more localized HOMO on the β-ionone ring, an increased dipole-moment of the 11-cis retinal (Table 4, a decreased QM-region µ) and a stronger BLA (Table 3, Chart 4) resulting in a more increased intake of S1 energy (Table 5, Charts 5a, 5b and 5c) in regard to WT structures since when the conformational deviation (Figure 3b) leads to the guanidino group far from ionone ring region, the blue-shift caused by the mutation disappears (Chart 5d). In addition, it was found, an irrelevant role of Glu181 residue on the absorption spectra, in agreement with a recent study (53), considering the large conformational deviation observed in different structures (snapshots) from the system bWT unrestricted within membrane for the sidechain of Glu181 orienting its carboxylic group toward a contrary side of the 11-cis retinal (Figure 3a) and the calculated vertical-excitation energies for this orientation, which were very similar to those for the more probable conformation of the residue (Figure 2a), (Charts 5c and 5d). D. Energy excess in mutants. Since, a substantial increase of the average vertical-excitation energy for the first excitation in both mutants above the necessary and calculated for human rhodopsin in this work (57.7 kcal/mol) using DFT/MRCI, the experimental values is 57.4 kcal/mol (103), lead to an energy excess of 9.92 kcal/mol in hM207R mutant and of 12.68 kcal/mol in hS186W mutant; this energy excess would cause harmful collateral reactions in the retinal binding site of both mutants related with the pathogeny of the RP at a very molecular level; an increased energy intake and a decreased chemical work take place in these mutants, considering that the energy in vertebrate rod is on order of 40-50 kcal/mol for rhodopsin activation by light, only (111), the energy storage due to 11-cis/all-trans isomerization in rhodopsin is on order of 32-35 kcal/mol, being ~50 % of the photon energy (103), and the retinal binding-site structure in the Dark State for hM207R and hS186W mutants is not optimal for the isomerization, which could impair this reaction (i.e. deviations for φωC4-C5-C6-C7 in hM207R or for φC10-C11-C12-C13 in both mutants and the unprotonated SBL in hS186W as above-mentioned). In fact, less than 23 kcal/mol remains from the difference between excitation energy and storage energy in WT rhodopsins with a normal photoisomerization reaction. Both mutants studied in this work had a larger surplus energy than WT structures. CONCLUSIONS We present an evaluation on stability, geometries, electronic or electrostatic influences and vertical excitation energies in the dark state of the human-rhodopsin mutants associated to RP carrying M207R or S186W substitutions, regarding bovine and human rhodopsins, by molecular dynamic simulations and combined QM/MM calculations, mainly. The calculations explored in large regions of the conformational space for all structures of every system, wild type or mutant, displayed high precision, a very good agreement with previous data from experimental or theoretical studies and to enhance the knowledge on mutant spectral patterns and its relation with geometric or electric changes. Our investigation results support that (a) both M207R and S186W mutations can decrease the human-rhodopsin stability, although the folding, retinal binding and absorption spectra can take place through a natural way and/or favored by artificial stabilizing molecules, considering the advances made recently; (b) geometrical deviations and abnormal mechanisms perturbing the absorption spectra in the dark state due to these substitutions must be also considered for a better understanding on the mutation effects beyond stability; these also impair rhodopsin functions; (c) both mutations lead to less optimal dark-state configurations for the photoisomerization, (d) a strong blue-shift of the λmax was induced by both mutations; (e) studied mutations had a larger surplus energy with regard to wild-type rhodopsins, which could lead to additional abnormal reactions due to an inappropriate energy dissipation in mutants since an increased energy intake takes place in mutants with a conceivable deficient photoisomerization process due to conformational retinal deviation in both mutants and the unprotonated SBL caused by S186W mutation; (f) both the high blue-shift and surplus energy calculated for mutations emphasize the idea of preventing the eyes exposition of patient suffering RP to light tending to the spectral blue region as well as to find a viable solution for decreasing harmful collateral reactions proposed here and its possible noxious species, that otherwise would spread the consequences to other important molecules and tissues in retina; (g) both the highest blue-shift and surplus energy were calculated for S186W mutation; (h) a positive correlation was observed between BLA and vertical excitation energies; (i) the higher S1 vertical-excitation energies observed were consistent with the average BLA calculated and the electric changes of both mutants pointing to a possible S1 destabilization due to the electronic polarization made by Arg207 and a stronger repulsive electrostatic influence from the counterion Glu113, which would have the main contribution for explaining the increased vertical-excitation energy for M207R mutation as well as an increased electronic polarization induced by the surrounding environment to the β-ionone ring when an unprotonated SBL is possible due to S186W mutation leading to the highest averaged BLA, the best localized HOMO on the ring and the highest vertical-excitation energies, which were observed for S186W mutation; (j) it seems that the spontaneous electrons tendency is to stay on the ionone ring region when retinal is linked to an unprotonated SBL, this behavior could be considered for the basic mechanism of spectral tuning in the dark state and normal blue-shift intermediate of rhodopsin activation with an unprotonated SBL; (k) Glu181 emerged as irrelevant for the optical spectra; (l) somehow, the stability in function of pH could be improved by means of conformational changes in mutants, suggesting action points for potential drugs; (m) even when some conformations were less probable for M207R mutation, these could reveal a potential target for prospective therapy implicated with molecules used for stabilizing or changing the protein conformation, since a normal absorption was achieved for some less probable or unnatural conformations in this mutant; (n) The present study also pretend to encourage new experimental studies and combined QM/MM approaches on mechanisms of RP-associated mutations related to stability, spectral shift or photoisomerization connecting the genetical etiology to the pathogeny at a very molecular level . Our results provide for the first time an atomistic and reliable insight on the mechanisms of both M207R and S186W mutations associated to RP, which can open new strategies for the treatment. Acknowledgments. The Charité-Universitätsmedizin Berlin served as excellent host for an essential part of this work as well as the Max Planck Institute für Kohlenforschung of Mülheim an der Ruhr. The Autonomous University of Madrid provided part of the computational facilities altogether with the Max Planck Institute für Kohlenforschung. We thank Sandra M. Blois for supporting this work within the framework of the cooperation between the Charité-Universitätsmedizin Berlin and the Havana Higher Institute of Medical Sciences in Cuba as well as Elke Gnielka for helping this work. Rachel Crespo, Alejandro Gil and Reynier Suardíaz are acknowledged for helpful discussions. REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
