What Kinds of Ferryl Species Exist for Compound II of

May 1, 2009 - Published online 1 May 2009. Published in print 4 June 2009. +. Altmetric Logo Icon More Article Metrics. CURRENT ISSUELATEST NEWS...
0 downloads 0 Views 1MB Size
Download PDF
7912

J. Phys. Chem. B 2009, 113, 7912–7917

What Kinds of Ferryl Species Exist for Compound II of Chloroperoxidase? A Dialog of Theory with Experiment Wenzhen Lai, Hui Chen, and Sason Shaik* Institute of Chemistry and The Lise Meitner-MinerVa Center for Computational Quantum Chemistry, The Hebrew UniVersity of Jerusalem, GiVat Ram Campus, 91904 Jerusalem, Israel ReceiVed: March 14, 2009

QM/MM calculations of Mo¨ssbauer parameters lead to assignments of the three isomeric species of CPO-II, “major”, “minor”, and unassigned “6% species”, which were recently observed in experiment and posed two puzzles (Stone, K. L.; Hoffart, L. M.; Behan, R. K.; Krebs, C.; Green, M. T. J. Am. Chem. Soc. 2006, 128, 6147). Both the minor and major species were found to be iron(IV)-hydroxides, thus accounting for the observed ratio of their relative yield that is pH-independent. The difference between the minor and major species is a single water molecule that acts as a H-bond acceptor from the ferryl in the minor species (2b) and it is essential to get a good match of the calculated Mo¨ssbauer parameters to the experimentally observed ones for the minor species. The major species (2c-2e, 2e-NW) may or may not have a water molecule. The calculations reveal also two candidates for the unassigned 6% species, which are a Por+•FeIIIOH species 2e-Fe(III), without or with a water molecule, or the corresponding aqua complex Por+•FeIIIOH2 3c formed by adding an additional proton to the system. These species have ∆EQ parameters of the same magnitude but with opposite signs: negative (-2.30 mm/s) for the two 2e-Fe(III) species and positive (2.39 mm/s) for 3c. The above assignments were further consolidated by an extended correlation (Figure 2) between the iron spin density and the ∆EQ parameters of the species calculated in the present study and by relating ∆EQ to the d-electronic configuration on iron. A bonding model of the FeO(H) moiety (Figure 3) was used to account for the variation of the spin density and provided further support for the correlation in Figure 2 and the assignment. Experimental determination of the sign of the quadruple parameter will finally confirm the identity of this species. In addition, since 3c possesses an additional proton, its identity can be revealed by pHdependent yield. All in all, the present paper shows that QM/MM calculations can conduct a useful dialogue with experiment in this complex field. 1. Introduction Chloroperoxidase (CPO) is a versatile heme enzyme that, besides its native halogenation and dehalogenation reactions, exhibits reactivities typical of other enzymes such as peroxidases, catalases, and cytochromes P450 (P450s).1 In fact, CPO is the only thiolate-ligated heme enzyme, whose oxoiron(IV)porphyrin cation-radical active species, termed Compound I (CPO-I) and shown in Scheme 1, could be characterized by spectroscopic methods,1,2 thus providing key insight into the behavior of P450s where this species is elusive.3 Furthermore, the investigation of the 1e-reduced species, termed Compound II (CPO-II), showed that this species is in fact protonated (Scheme 1). Thus, initial extended X-ray absorption fine structure (EXAFS) study4 of CPO-II at pH 6.5 revealed that instead of the expected short FeIVdO, CPO-II possessed a long bond of 1.82 Å, corresponding to a protonated ferryl with a single FeIV-OH bond.4 Subsequent resonance Raman characterization of the Fe-O stretching frequency provided direct evidence for the existence of a FeIVOH species.5 These findings allowed Green et al.4,5 to make an important generalization about Nature’s choice of thiolate ligation as a major promoter of the uncanny ability of P450s to perform C-H hydroxylation of * To whom correspondence should be addressed. E-mail: sason@ yfaat.ch.huji.ac.il.

nonactivated C-H bonds. Ever since this generalization, CPOII has become a research focus in bioinorganic and mechanistic chemistry. Still, since other heme enzymes possess also unprotonated Compound II species, with a short FedO bond, Green et al. deemed it necessary to seek for alternative ways, using Mo¨ssbauer spectroscopy, to determine the protonation state of CPO-II species. A recent such study6 revealed three species; one (6% yield) could not be assigned, whereas the other two were assigned based on comparison of their experimentally determined Mo¨ssbauer parameters with density functional theory (DFT) computed values. Thus, the “major species” was assigned as FeIVOH based on its quadrupole splitting value, ∆EQ ) 2.06 ( 0.03 mm/s, which fitted very well the model DFT calculations. The “minor species” was tentatively assigned as FeIVdO, even though its ∆EQ ) 1.59 ( 0.05 mm/s value did not match as well as the calculated one of 1.0 mm/s. As pointed out by the authors, the most surprising result was the yield ratio (70:30) of the two species was independent of the pH.6 How could the ratio be pH independent if one form is protonated and the other is not? In addition, the 6% species, with ∆EQ ) 2.30 mm/s, could not be assigned;6 what is the identity of the unidentified third species? To answer these questions, we explored CPO-I and a variety of CPO-II species (Scheme 1) by hybrid DFT(UB3LYP)/MM calculations. On the basis of experience, it is important to run the Mo¨ssbauer calculations in the protein environment, including

10.1021/jp902288q CCC: $40.75  2009 American Chemical Society Published on Web 05/01/2009

Ferryl Species for Compound II of Chloroperoxidase

J. Phys. Chem. B, Vol. 113, No. 22, 2009 7913 TABLE 1: QM/MM Calculated and Experimental Mo¨ssbauer Spectroscopic Parameters for CPO-I and CPO-II

SCHEME 1: CPO-I and CPO-II

species CPO-II

hydrogen bonding interactions with the active species.7 Thus, our CPO model involves the distal residues (Glu183 and His105) and the water molecule that is liberated during the formation of CPO-I.8 As shall be seen, these calculations enabled us to assign all the CPO-II species observed in the experiment,6 resolve the pH-independence of the relative yield of the major and minor species, and make a new prediction about the nature of the unassigned 6% species in CPO6 with direct relevance to P450.

entry theor

expth

CPO-I

theor expti

2. Computational Methods 9

The QM/MM calculations were preformed using ChemShell interfaced with Turbomole10 and DL_POLY.11 The hybrid B3LYP functional12 was used throughout this study for the QM part, and the CHARMM22 force field13 was used for the MM part. An electronic embedding scheme14 was applied. To treat the QM/MM boundary, hydrogen link atoms with the charge shift model15 were used. Geometry optimizations were preformed with B1, the LACVP16 basis set. Single-point calculation performed with B2, a Wachters’ all electron basis set17 augmented with diffuse18 d and polarization19 f functions on iron (8s7p4d1f) and 6-31++G(d,p) on the other atoms. To ascertain the basis set convergence, we also employed a larger basis set termed B3 for some key species. B3 is comprised of def2-TZVP basis20 (Turbomole notation) for all atoms. Details of preparation of the enzyme for QM/MM calculations are provided in the Supporting Information document. The Mo¨ssbauer parameters were evaluated with the program ORCA21 using single-point B3LYP calculation at the corresponding QM/MM-optimized geometries. While, MM point charges were also included to probe the effect of the protein environment. In these calculation, iron was described by the triply polarized core properties basis set CP(PPP),22 and the other atoms were described by the SV(P) basis set23 with the inner s functions left uncontracted. 3. Results: Identification of the CPO-II Species In brief, we calculated both CPO-I and CPO-II species, in various protonation states of the species and of Glu183, which participate in the O-O activation, as well as with and without an intervening water molecule, which is formed during the O-O activation.8a The entire data is collected in Table 1 and Figure 1, where the CPO-I are labeled as 1a and 1b, while the CPO-II species are indicated as 2 with additional indicators of the protonation state and water environment. Consider first, 2a, the unprotonated CPO-II species with a PorFeIVdO structure, shown in Figure 1 and in entry 1 in Table 1; the species possesses Fe-O/Fe-S bond lengths of 1.66/2.58 Å, ∆EQ of 1.02 mm/s, and isomer shift (δ) of 0.12 mm/s. Clearly therefore, based on the experimental ∆EQ value of 1.59 mm/s for the “minor ferryl species” (entry 13, Table 1),6 the large deviation of the computed ∆EQ of 1.02 mm/s (entry 1, Table 1), rules out 2a as the minor species of CPO-II. Having ruled out 2a as the minor species, we turned to study the cases where the ferryl oxo is protonated. As such, we protonated the nearest proximal residue Glu183, which is the

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

∆EQ (mm/s) δ (mm/s) 2aa,b 1.02 0.12 2bb,c 1.62 0.08 2cb,c 2.18 0.08 2db,c 2.13 0.09 b,c 2e 2.13 0.08 b,c,d 2e-NW1 2.06 0.10 2.00 0.10 2e-NW2b,c,d 2e-Fe(III)b,e -2.29 0.34 2e-NW-Fe(III)b,d,e -2.31 0.33 2.49 0.12 3ac,f 3bc,f 1.95 0.10 b,g 3c 2.39 0.34 minor 1.59 ( 0.05 0.11 ( 0.03 major 2.06 ( 0.03 0.10 ( 0.03 6% species 2.30 0.29 1aa,b 0.92 0.12 1.08 0.09 1ba,f 1.02 0.15

η 0.24 0.29 0.09 0.18 0.08 0.10 0.11 0.19 0.22 0.03 0.17 0.01

0.29 0.22

a Iron(IV)oxo. b Glu183 is unprotonated. c Iron(IV)-hydroxide. d The water molecule is removed. e Iron(III)-hydroxide. f Glu183 is protonated. g Protonated iron(III)-hydroxide. h Reference 6. i Reference 2a.

only residue that can possibly interact with the ferryl FedO unit through hydrogen bond (H-bond). To ascertain the effect of protonated Glu183 on the Mo¨ssbauer parameters, we calculated the corresponding CPO-I species with a deprotonated Glu183 (1a, Figure 1) and a protonated Glu183 (1b, Figure 1). It is apparent that the calculated ∆EQ for 1a and 1b (in entries 16 and 17 (Table 1)) are both in good agreement with experiment (entry 18).2a Thus, CPO-I remains intact when Glu183 is protonated, and its Mo¨ssbauer parameters are reliably computed vis-a`-vis experimental data. Injecting an additional electron to 1b and optimizing the geometry of the reduced species resulted in its automatic protonation and led to CPO-II 2b (Figure 1) wherein the proton of Glu183 moved spontaneously via the intervening water molecule and protonated the ferryl oxo to yield the FeIVOH moiety. Thus, in line with Green et al.4 the ferryl basicity in CPO-II is strong enough to abstract one proton from protonated Glu183. As shown in Table 1, unlike the case of 2a, the calculated Mo¨ssbauer parameters for 2b (entry 2), ∆EQ ) 1.62 mm/s and δ ) 0.08 mm/s, are in excellent match with experimental values observed for the minor species of CPO-II (entry 13).6 This result further strengthens the above conclusion that the minor species that possesses ∆EQ ) 1.5 mm/s cannot be the unprotonated CPO-II and is the protonated form 2b. The structure of 2b in Figure 1 indicates that the minor species observed in the experiment6 should be characterized by the very tight H-bond network, Glu183-water-HO-Fe, which results in Fe-O/Fe-S bond lengths of 1.73/2.46 Å, that is, shorter/longer than the EXAFS results4 for CPO-II. To ascertain the minimal structural requirement for 2b to qualify as the minor species, we removed the water in 2b and kept the geometry of all the other parts fixed. The water removal resulted in a mismatching value of ∆EQ ) 1.9 mm/s (2b′ in Table S2 in the Supporting Information). Therefore, we ascertained in this manner that the presence of water in 2b is essential for the agreement of the calculated and experimental ∆EQ values. The H-bond network in 2b suggested that there should be other conformations in which the H-bonding pattern with the active species is different from that in 2b. By adjusting the H-bond network with subsequent geometric optimizations, we located three new conformations of protonated CPO-II, termed 2c, 2d, and 2e (Figure 1). Thus, while in 2b the water accepts a H-bond from the FeOH moiety, in 2c and 2d the water acts as a H-bond donor to FeOH unit, whereas in 2e Glu183

7914

J. Phys. Chem. B, Vol. 113, No. 22, 2009

Lai et al.

Figure 1. Key DFT(UB3LYP/B1)/MM optimized CPO-I and CPO-II species. Bond distances are in Å. Most H atoms are omitted for clarity.

accepts a H-bond directly from FeOH. The calculated Mo¨ssbauer parameters for 2c-2e in Table 1 (entries 3-5) are all in good agreement with the experimental datum of the major species of CPO-II (entry 14).6 Similarly, the Fe-O/Fe-S bond lengths of 2c (1.81/2.41), 2d (1.80/2.41), and 2e (1.80/2.45) in Figure 1 are all in good agreement with the experimental EXAFS values of 1.82/2.39 Å for CPO-II.4 Interestingly, removal of the water molecule from 2d-2e created structures 2e-NW1 and 2e-NW2 (NW stands for no water) having direct Glu183 · · · HO-Fe H-bonds as in 2e, and ∆EQ ) 2.00-2.06 mm/s and δ ) 0.10 mm/s values, still in good match with experimental datum for the major species.6 Thus, our calculation shows that the major species of CPO-II is also an iron(IV)-hydroxide entity, albeit now either as a H-bonding acceptor from the intervening water molecule, as in 2c and 2d, or as an acceptor from Glu183, as in 2e and 2eNW; the latter lacking a water molecule. Clearly therefore, the assignments of both minor and major species as protonated CPO-II entities explain perfectly the pH-independence of their relative yields.6 There remains to identify the unassigned third species (6% of total intensity) with ∆EQ ) 2.30 mm/s and δ ) 0.29 mm/s.6 Previous QM/MM calculations for P450s indicate the Por+•FeIIIOH electromer of Compound II is very close to the PorFeIVOH species in energy, both at the B3LYP/MM and Multireference-CI/MM levels.24 Indeed, here too we obtained the corresponding Por+•FeIIIOH species of CPO-II (with the exception of the one corresponding to 2b; see Table S3 in the Supporting Information where this FeIII state is too high in energy to locate). The so-located Por+•FeIIIOH electromers 2eFe(III) and 2e-NW-Fe(III), entries 8 and 9 in Table 1, have ∆EQ values of ca. -2.3 mm/s, similar to the corresponding calculated value for the corresponding HRP-II (FeIII) species.25 Furthermore, this ∆EQ value is identical in absolute magnitude to the value reported for the unassigned “6% species”.6 Therefore, the “third species” might be tentatively assigned as the Por+•FeIIIOH electromer of CPO-II, but this assignment would require experimental determination of the sign of the respective ∆EQ value, which was not done in the original study for this species.6 We also note that the absence of the water

from 2e-Fe(III) hardly affects the ∆EQ and δ values (see 2eNW-Fe(III) in entry 9, Table 1). For completeness sake, we introduced an additional proton to the system (which may be there depending on the pH), and explored also the effect of protonating Glu183 on the FeIV-OH CPO-II species (see Table S5 in the Supporting Information for all the species). The so-formed lowest energy species 3a-3c are shown in Figure 1 (see also Table 1, entries 10-12) with 3a being the most stable and 3b the least (4.0 and 7.8 kcal/mol higher, respectively). Thus, in 3a the OH moiety of CPO-II is a H-bond acceptor, in 3b it is a H-bond donor, whereas 3c is an aqua complex, generated by proton transform from Glu183 and having a Por•+FeIII-OH2 structure, which is the protonated form of 2e-Fe(III) (in entry 8, Table 1). The calculated isomer shift for 3a and 3b are 0.12 and 0.10 mm/s, respectively, and similar to those of 2c-2e. By contrast, the calculated isomer shift for 3c is 0.34 mm/s, which is much larger than 3a and 3b, but similar to the same parameter in 2e-Fe(III). The ∆EQ parameter of 3a is on the higher side of the experimental data (entries 10 vs 14, Table 1) and could probably be ruled out as a major species candidate. On the other hand, based only on the calculated δ and ∆EQ parameters, 3b could have also been assigned (in addition to 2c-2e) as a candidate for the major species observed in the Mo¨ssbauer experiment.6 However, since 3b has an excess proton relative to the minor species 2b, the assignment of 3b as the major species can be ruled out, because the extra proton is in discord with the observed pHindependence of “major/minor” relative yields.6 In conclusion, the most likely candidates for the major species are 2c-2e, while 2b is the minor species. The aqua complex 3c has δ and ∆EQ parameters (entry 12, Table 1) very close to those observed for the unassigned 6% species (see entry 15, Table 1; ∆EQ ) 2.30 mm/s and δ ) 0.29 mm/s).6 As such, we now have two candidates for the 6% species, 3c or 2e-Fe(III). Note that these species have opposite signs of the ∆EQ parameter; 3c has a positively signed parameter, while 2e-Fe(III) has a negatively signed one. The choice between them will ultimately be made by experimental determination of sign of the ∆EQ of the third species. Another helpful probe is the fact that 3c has an extra proton compared

Ferryl Species for Compound II of Chloroperoxidase

Figure 2. Correlations between the calculated quadrupole splitting ∆EQ of CPO-I and CPO-II species and spin distributions on (a) iron and (b) oxygen.

with 2e-Fe(III). Therefore, theory predicts that if 3c is indeed the 6% species, its yield should be sensitive to the pH; the original report6 did not specify such dependence however possibly because such sensitivity was not observed in which case 2e-Fe(III) remains the only candidate. It is noteworthy that this third species is likely to exist also in P450BM3 Compound II system, where it remained also unassigned.26 4. Discussion As shown above, 2b is the only species among those computed that possesses ∆EQ and δ parameters consistent with those observed experimentally for the minor species, while 2c-2e (2e-NW) qualify as the major species. These assignments, of the major and minor species, are in good accord with the observed Mo¨ssbauer parameters as well as with the pHdependency of the relative yields of the two species. Both 2eFe(III) and 3c qualify as the 6% species; this however depends on the sign of ∆EQ and importantly on the pH, as 3c has an additional proton over 2c-2e (and over 2e-Fe(III)). These are all testable predictions. Electronic Structure Considerations. We shall come back to the assignment issue in the next subsection, while at this point we wish to comprehend the sensitivity of the ∆EQ parameter as revealed in Table 1 and to try to relate this sensitivity to the electronic structure of the species. To put these electronic structures in perspective, we depict in Scheme 2 the orbital occupation diagrams for all the species in Table 1, using the d-block orbitals and porphyrin’s a2u orbital.25 The electronic structures in Scheme 2 reveal that 2a-3b are all PorFeIV)O and/or PorFeIV-OH species with two singly occupied π*(FeO) orbitals (πyz,xz*). On the other hand, 2eFe(III) and 2e-NW-Fe(III) are Por+•FeIII-OH species, while 3c is a Por+•FeIII-OH2 species. With two/three singly occupied d-type orbitals, these complexes will possess significant spin

J. Phys. Chem. B, Vol. 113, No. 22, 2009 7915 densities on the Fe-O moiety, and this might strongly affect the value of ∆EQ. Figure 2 shows plots of the computed ∆EQ parameters for all the species of the study (with the exception of 3c, which suffers from more spin contamination compared to others). Thus, as can be seen from Figure 2a the calculated quadrupole splitting ∆EQ for the various CPO-I and CPO-II species, is sensitive to the spin density on iron, F(Fe); the larger the iron spin density, the larger gets the ∆EQ value. Interestingly, one can see from Figure 2a, that the removal of the water molecule from 2b to yield 2b′ increases the Fe spin density by 0.14 units and as a result this raises the ∆EQ value by 0.28 mm/s, thereby causing a mismatch with the experimental datum for the minor species. Clearly therefore, the strong H-bonding situation between the FeOH moiety and the water molecule in 2b lowers the spin density on iron and gauges the quadrupole splitting parameter for the minor species. What is the reason for that? To get more insight, consider the plot of ∆EQ with F(O) (ferryl FeO) in Figure 2b. It is seen that the correlation exists but in an inverse sense; the larger the F(O) and the closer it approaches 1.0, the smaller gets the ∆EQ value. These inverse correlations show that the increase of the spin density on iron comes at the expense of its decrease on O and thereby affect the ∆EQ value. This complementary change in the spin densities on Fe and O can be rationalized using a valence bond (VB) modeling of Fe-O(H) π-bonding, using covalent and ionic structures, as shown in Figure 3. This model is essentially the same as the one used to explain the effect of H-bonding on the nature of the Fe-O bond in HRP-I.25 Figure 3, part a, shows the covalent and ionic structures of the bond, along with their spin density distribution, which are shown explicitly underneath the structure. The iron spin density is also indicated by subscript of the VB structures; thus 3Φ1.0 is the covalent structure, which possesses a spin density of 1.0 on Fe, whereas 3Φ2.0 is the corresponding ionic structure with Fe spin density of 2.0. The H in parentheses means that we are discussing a range of species involving the gamut from FedO ferryls all the way to the Fe-OH ferryl-hydroxo species with and without H-bonding situations. This range is discussed in Figure 3b,c, using VB mixing diagrams based on the two VB structures. Let us start with bare ferryls (FedO), such as 1a, 1b, and 2a in Figure 3b. The π(FedO) bond of bare ferryls is dominated by the covalent structure, which is depicted lower than the corresponding ionic structure. λ is the mixing coefficient of the ionic structure, and if the energy gap between the structures is large λ will vary inversely to the energy gap and lead to the perturbational expression of the bond wave function, which is shown underneath the diagram. Thus, in bare ferryls we would expect that the wave function will be dominated by the covalent structure 3Φ1.0 with small mixing of the ionic structure, thus having a spin density distribution of 1 + λ2/(1 + λ2) on Fe and 1 - λ2/(1 + λ2) on O. As the H-bonding to the oxo of the ferryl becomes stronger, this stabilizes the ionic structure, increases λ2, and thereby transferring spin density from O to Fe. As seen from Figure 2, this trend is manifested in 1a, 2a, and 1b, wherein the protonation of Glu183 increases the spin density on Fe by ∼0.2 e and decreases it by the same amount on O. As the H-bonding gets stronger, the ionic structure approaches the covalent and at the limit they become degenerate (λ ) 1), leading to a spin density of 1.5 on Fe and 0.5 on O; 1b is approaching this limit. Further increase of the H-bond will continue to stabilize the ionic structure below the covalent and will result in protonation of FedO to produce the FeOH ferryls.

7916

J. Phys. Chem. B, Vol. 113, No. 22, 2009

Lai et al.

Figure 3. A VB description of the leading to the π-bonding wave function of ferryls: The contributing VB structures to the π-bonding in Fe-O(H), and their spin distributions are shown in (a), where the subscript near the VB structures indicates the spin density on Fe, while the superscript signifies the triplet state of the species. The VB diagram for FedO ferryls is shown in (b), while the VB diagram for Fe-OH ferryls is depicted in (c). ∆E is the energy gap between the structures and λ is the mixing coefficient of the higher VB structure in each case. Note that the overlap of the VB structures is omitted.

SCHEME 2: Schematic Representation of the Electronic Structures of CPO-I and CPO-II Species

This is shown in Figure 3c, where now the ionic structure is the lowest between the two. The π-bond wave function in Figure 3c, which is a resonance hybrid of the two structures, gives rise to spin densities given approximately as 2 - λ2/(1 + λ2) on Fe and λ2/(1 + λ2) on O. Here too, the H-bonding situation will modulate the energy gap between the structures and as its strength increases λ decreases and approaches asymptotically zero, in which case the spin density on Fe is 2.0 and on O it is 0.27 As can be seen from Figure 2, as the Fe-OH ferryl become increasingly engaged in stronger H-bonding the Fe spin density increases and that on O decreases, reaching the limit in 3a, where an extra proton was added to the system. A good example is the case of 2b versus 2b′. Here in 2b the intervening water that acts as a H-bonding acceptor from the FeOH ferryl pulls the H+ and destabilizes the ionic structure, thus decreasing the bond ionicity and hence also the Fe spin density. As the intervening water is removed in 2b′, the ionic structure is stabilized, the bond ionicity increases, and the Fe spin density also increases by 0.2 with a concomitant decrease in the O spin density. It is apparent that the interactions of the ferryl with the protein environment modulate the FeO(H) bond ionicity and change the spin density on Fe and O in a complementary manner as observed in Figure 2a versus Figure 2b.27 To relate the spin density to the value of quadrupole splitting parameter, ∆EQ, we note that in the covalent structure, 3Φ1.0, the iron has approximately an atomic configuration consisting of dx2-y22dxz2dyz1, while in the ionic structure 3Φ2.0 the approximate atomic configuration is dx2-y22dxz1dyz1. The simplest consideration of the local valence electron contribution (anisotropic electron distribution in the iron valence shell for dx2-y22dxz2dyz1 and dx2-y22dxz1dyz1) to the electric field gradient tensor that determines ∆EQ (∆EQ ) (e2QVzz/2)(1 + η2/3)1/2), shows that the main tensor component Vzz (where |Vzz| g |Vyy|

g |Vxx|) for 3Φ1.0 is negative (-4/7), while the main tensor component Vzz for 3Φ2.0 is positive (4/7). On the basis of this consideration, the ionic structure 3Φ2.0 in the bond wave function in Figure 3 will contribute positively to ∆EQ while 3 Φ1.0 will make a negative contribution. This analysis, combined with the VB bonding model, provides a reasonable physical rationale for the computational trends in the ∆EQ and its correlation with the Fe spin density, as well as with the protonation state and H-bonding situation of the species, and hence our assignments based on the variation in this spectroscopic parameter are deemed reliable. We may therefore conclude that increasing the Fe spin density (by mixing more of the ionic structure) via protonation or H-bonding to the Fe-O(H) bond will result in a larger ∆EQ parameter and vice versa. Inspection of Table 1 and Figure 2 shows that this is indeed the case. Thus, the assignments of 2b as the minor species and 2c-2e as the major species with either 2e-Fe(III) or 3c as the 6% species, rest now not only on computations but also on a model of bonding.28 Energetic Considerations. Having established the identity of the observed species based on the Mo¨ssbauer parameters, this is the place to consider the question of the relative yields of the species, which is a computationally more difficult issue than the spectroscopic assignment. Experimentally, the relative yield of 70:30 of the two species requires almost identical relative free energy, but we cannot calculate accurate enough free energies with current capabilities of B3LYP/MM calculations. Nevertheless, we can draw some educated judgment based on our computational results. Thus, the QM/MM energies of the 2b-2e species, which share a water molecule in H-bonding with FeOH and Glu183, show that 2b is the lowest, while energies of the 2c-2e species all lie 5 kcal/mol higher (see Table S2 in the Supporting Information). However, as shown

Ferryl Species for Compound II of Chloroperoxidase in Figure 1, the H-bond network between Glu183, the oxo unit, and the water molecule in 2b is obviously stronger and more rigid than those in all the other three species. Therefore, entropic factors would favor to some extent the other three species relative to 2b. Additionally, if the major species lacks the intervening water molecule, as in 2c-2e, while 2b that possesses a water molecule is the minor species, then 2b will be disfavored by entropic factors, which can amount to 5 kcal/mol, relative to 2c-2e. Indeed, with 2b being the minor and 2c-2e the major species one can predict a temperature-dependent ratio of the major and minor species of CPO-II. In fact, the ∆EQ parameter of the minor species, generated by cryoreduction of CPO-I, exhibited temperature dependence,6 thus revealing an effect of molecular relaxation, which could derive from the presence/ absence of the water molecule. 5. Conclusion The above QM/MM calculations of Mo¨ssbauer parameters allowed us to assign the three isomeric species of CPO-II, which were observed in experiment.6 Thus, both the minor and major species are iron(IV)-hydroxides, and therefore their ratio is pHindependent but is likely to be temperature dependent. This confirms Green’s conclusion that the ferryl basicity in CPO-II is strong enough to abstract the proton from the distal Glu183 residue. The difference between the minor and major species is a single water molecule that acts as a H-bond acceptor from the ferryl in 2b vis-a`-vis 2c-2e where it acts as a H-bond donor or is completely absent (as in 2e-NW species). The presence of this water molecule in 2b is essential to get a good match of the calculated Mo¨ssbauer parameters to the experimentally observed ones6 for the minor species. The QM/MM calculations further reveals two candidates for the unassigned 6% species, the third species observed in experiment,6 should be a Por+•Fe(III)OH(H) species 2e-Fe(III) without or with a water molecule or the aqua complex 3c formed by adding an additional proton to the system. These species have ∆EQ parameters of the same magnitude but with opposite signs; negative for the two 2e-Fe(III) species and positive for 3c. Experimental determination of the sign of the quadruple parameter will finally nail the identity of this species. In addition, since 3c possesses an additional proton its potential candidacy can be probed by pH-dependent yield. The above assignments were further consolidated by an extended correlation between the iron spin density and the ∆EQ parameters of all the species calculated in the present study and by relating ∆EQ to the d-electronic configuration on iron. A bonding model of the FeO(H) moiety (Figure 3) was used to account for the variation of the spin density and provided further support for the correlation in Figure 2, as well as for the assignments of the species. All in all, the present paper shows that QM/MM calculations can be a useful partner to experiment in this complex field. Acknowledgment. S.S. is supported by an ISF grant (16/ 06). H.C. thanks the Golda Meir fellowship fund. Supporting Information Available: Computational procedures and full set of computational results. This material is available free of charge via the Internet at http://pubs.acs.org.

J. Phys. Chem. B, Vol. 113, No. 22, 2009 7917 References and Notes (1) Dawson, J. H.; Sono, M. Chem. ReV. 1987, 87, 1255. (2) (a) Rutter, R.; Hager, L. P.; Dhonau, H.; Hendrich, M.; Valentine, M.; Debrunner, P. Biochemistry 1984, 23, 6809. (b) Palcic, M. M.; Rutter, R.; Araiso, T.; Hager, L. P.; Dunford, H. B. Biochem. Biophys. Res. Commun. 1980, 94, 1123. (c) Egawa, T.; Proshlyakov, D. A.; Miki, H.; Makino, R.; Ogura, T.; Kitagawa, T.; Ishimura, Y. J. Biol. Inorg. Chem. 2001, 6, 46. (d) Hosten, C. M.; Sullivan, A. M.; Palaniappan, V.; Fitzgerald, M. M.; Terner, J. J. Biol. Chem. 1994, 269, 13966. (e) Kim, S. H.; Perera, R.; Hager, L. P.; Dawson, J. H.; Hoffman, B. M. J. Am. Chem. Soc. 2006, 128, 5598. (3) (a) Denisov, I. G.; Makris, T. M.; Sligar, S. G.; Schlichting, I. Chem. ReV. 2005, 105, 2253. (b) Sono, M.; Roach, M. P.; Coulter, E. D.; Dawson, J. H. Chem. ReV. 1996, 96, 2841. (c) Poulos, T. L. Nat. Prod. Rep. 2007, 24, 504. (4) Green, M. T.; Dawson, J. H.; Gray, H. B. Science 2004, 304, 1653. (5) Stone, K. L.; Behan, R. K.; Green, M. T. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 12307. (6) Stone, K. L.; Hoffart, L. M.; Behan, R. K.; Krebs, C.; Green, M. T. J. Am. Chem. Soc. 2006, 128, 6147. (7) Derat, E.; Cohen, S.; Shaik, S.; Altun, A.; Thiel, W. J. Am. Chem. Soc. 2005, 127, 13611. (8) (a) Chen, H.; Hajime, H.; Derat, E.; Schlichting, I.; Shaik, S. J. Phys. Chem. B 2008, 112, 9490. (b) Ku¨hnel, K.; Derat, E.; Terner, J.; Shaik, S.; Schlichting, I. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 99. (9) Sherwood, P.; et al. J. Mol. Struct. (THEOCHEM) 2003, 632, 1. (10) Ahlrichs, R.; Ba¨r, M.; Ha¨ser, M.; Horn, H.; Ko¨lmel, C. Chem. Phys. Lett. 1989, 162, 165. (11) Smith, W.; Forester, T. R. J. Mol. Graph. 1996, 14, 136. (12) (a) Becke, A. D. Phys. ReV. A 1988, 38, 3098. (b) Lee, C.; Yang, W. T.; Parr, R. G. Phys. ReV. B 1988, 37, 785. (c) Becke, A. D. J. Chem. Phys. 1993, 98, 5648. (d) Becke, A. D. J. Chem. Phys. 1993, 98, 1372. (13) Mackerell, A. D., Jr.; et al. J. Phys. Chem. B 1998, 102, 3586. (14) Bakowies, D.; Thiel, W. J. Phys. Chem. 1996, 100, 10580. (15) de Vries, A. H.; Sherwood, P.; Collins, S. J.; Rigby, A. M.; Rigutto, M.; Kramer, G. J. J. Phys. Chem. B 1999, 103, 6133. (16) (a) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 299. (b) Friesner, R. A.; Murphy, R. B.; Beachy, M. D.; Ringnalda, M. N.; Pollard, W. T.; Dunietz, B. D.; Cao, Y. X. J. Phys. Chem. A 1999, 103, 1913. (17) Wachters, A. J. H. J. Chem. Phys. 1970, 52, 1033. (18) Hay, P. J. J. Chem. Phys. 1977, 66, 4377. (19) (a) Bauschlicher, C. W., Jr.; Langhoff, S. R.; Partridge, H.; Barnes, L. A. J. Chem. Phys. 1989, 91, 2399. (b) Stewart, R. F. J. Chem. Phys. 1970, 52, 1033. (20) Scha¨fer, A.; Huber, C.; Ahlrichs, R. J. Chem. Phys. 1994, 100, 5829. (21) Neese, F.; ORCA, version 2.6, revision 35; University of Bonn: Bonn, Germany, 2006. (22) Neese, F. Inorg. Chim. Acta 2002, 337, 181. (23) Scha¨fer, A.; Horn, H.; Ahlrichs, R. J. Chem. Phys. 1992, 97, 2571. (24) (a) Scho¨neboom, J. C.; Cohen, S.; Lin, H.; Shaik, S.; Thiel, W. J. Am. Chem. Soc. 2004, 126, 4017. (b) Altun, A.; Kumar, D.; Neese, F.; Thiel, W. J. Phys. Chem. A 2008, 112, 12904. (25) Derat, E.; Shaik, S. J. Am. Chem. Soc. 2006, 128, 8185. (26) Behan, R. K.; Hoffart, L. M.; Stone, K. L.; Krebs, C.; Green, M. T. J. Am. Chem. Soc. 2006, 128, 11471. (27) It should be noted that small spin polarization occurs between ferryl FeO center and porphyrin ring in CPO-II, leading to total spin of the FeO exceeding 2 (see Table S5 in the Supporting Information). This would cause small shifts if one wanted to estimate the weights of the VB component from spin density. (28) Since we only considered the local valence contribution to the electric field gradient but omitted the other terms such as core-polarization and bond contributions,29 our consideration is not meant to address the absolute quantitative value of ∆EQ. However, due to the similarity in electronic structure of our studied species, wherein terms that are non valence terms are less species dependent, our approach is capable of explaining the ∆EQ difference between the species, based on Figure 3. (29) (a) Neese, F. J. Inorg. Biochem. 2006, 100, 716. (b) Serres, R. G.; Grapperhaus, C. A.; Bothe, E.; Bill, E.; Weyhermu¨ller, T.; Neese, F.; Wieghardt, K. J. Am. Chem. Soc. 2004, 126, 5138.

JP902288Q