Conformational differences between the methoxy groups of Q_A and Q_B site ubisemiquinones in bacterial reaction centers: a key role for methoxy group orientation in modulating ubiquinone redox potential

Abstract

Ubiquinone is an almost universal, membrane-associated redox mediator. Its ability to accept either one or two electrons allows it to function in critical roles in biological electron transport. The redox properties of ubiquinone in vivo are determined by its environment in the binding sites of proteins and by the dihedral angle of each methoxy group relative to the ring plane. This is an attribute unique to ubiquinone among natural quinones and could account for its widespread function with many different redox complexes. In this work, we use the photosynthetic reaction center as a model system for understanding the role of methoxy conformations in determining the redox potential of the ubiquinone/semiquinone couple. Despite the abundance of X-ray crystal structures for the reaction center, quinone site resolution has thus far been too low to provide a reliable measure of the methoxy dihedral angles of the primary and secondary quinones, Q_A and Q_B. We performed 2D ESEEM (HYSCORE) on isolated reaction centers with ubiquinones ¹³C-labeled at the headgroup methyl and methoxy substituents, and have measured the ¹³C isotropic and anisotropic components of the hyperfine tensors. Hyperfine couplings were compared to those derived by DFT calculations as a function of methoxy torsional angle allowing estimation of the methoxy dihedral angles for the semiquinones in the Q_A and Q_B sites. Based on this analysis, the orientation of the 2-methoxy groups are distinct in the two sites, with Q_B more out of plane by 20-30°. This corresponds to an ≈50 meV larger electron affinity for the Q_B quinone, indicating a substantial contribution to the experimental difference in redox potentials (60-75 mV) of the two quinones. The methods developed here can be readily extended to ubiquinone-binding sites in other protein complexes.

The reaction center (RC) of the photosynthetic bacterium, Rhodobacter (Rb.) sphaeroides, is an integral membrane protein that separates charge upon photoactivation with ~100% quantum yield.^1,2 Light activation causes transfer of an electron from a pair of bacteriochlorophylls to a ubiquinone-10 (UQ-10) occupying the Q_A site on the other side of the membrane, forming the semiquinone anion radical, SQ_A. SQ_A can then transfer its electron to another, chemically identical, UQ-10 in the Q_B site, forming SQ_B.^3,4 Passage of the electron from SQ_A to Q_B is essential to allow further photochemical turnover of the RC, and is energetically favorable in spite of the chemical identity of Q_A and Q_B. The redox potential of Q_B is 60-75 mV more positive than Q_A.^5-9 The origin of this redox potential difference is not yet known, but only quinones with methoxy groups are able to serve as Q_A and Q_B simultaneously.^10,11 Furthermore, experiments with mono-methoxy ubiquinone-4 analogs show that simultaneous function as Q_A and Q_B specifically requires only the 2-methoxy group¹⁰ suggesting that this group is somehow involved in generating the necessary difference in redox potential between Q_A and Q_B. The methoxy groups not only have an impact on the redox potential of the quinone and resultant catalytic activity,^12,13 but may also contribute to interactions necessary for correct and adequate binding, especially for Q_B.¹⁰ Thus, a role for the methoxy groups is indicated that suggests protein-imposed constraints on the dihedral angle of the methoxy groups of UQ in their respective sites. This has clear implications for the much wider involvement of ubiquinones in key membrane redox function.

The plethora of X-ray structures now available does not provide an unequivocal description of the two quinone sites in the RC, which are positioned almost symmetrically about a central Fe(His)₄ complex. The H-bond distances and the torsional angles of the two methoxy group substituents of the ubiquinone ring display a significant spread in different structures, indicating mobility or mosaicity (reviewed in ref 4). However, for Q_B in the proximal position relative to the Fe(His)₄ center, the majority of structures yield a consensus with both methoxy groups strongly out of plane. This is superficially similar to the average conformation for Q_A, except 180° out of phase.⁴ Alternative approaches to obtaining these geometric parameters are spectroscopy and computation. Vibrational spectroscopy, and FTIR spectroscopy, in particular, provide significant and independent insights to the quinone binding interactions,^4,14 while EPR methods have been applied with great effect to the semiquinone states.^15,16 An advantage of IR and Raman over magnetic resonance spectroscopies is that the former, by difference spectroscopy, can see all redox states of the quinones not just the radical species. The use of light-induced (or redox potential-induced) difference spectra has allowed resolution of the difference between quinone and SQ states in RCs (for reviews, see refs 13,17). The theoretical and experimental bases for the IR spectra of ubiquinones, in solution, have been systematically explored.^18-23 The observation of two C=O stretches indicates asymmetry of substitution, arising from the conformations of the two methoxy groups, since an in-plane methoxy donates electrons into the ring while those out of plane are electron withdrawing. In addition, the influence of methoxy group conformations on transition energies and electron affinity have been quantum-chemically calculated for model quinones and ubiquinones.^19,21,24-26

Q_A^––Q_A and Q_B^––Q_B IR difference spectra, including spectra of UQ-10 site-specifically ¹³C-labelled at the C2 and C3 positions (see Scheme 1 for numbering convention) and reconstituted to either the Q_A or the Q_B binding site, were used to study the inequivalence of the methoxy groups.²⁷ The spectra did not show a shift in frequency of (ring)C-O vibrations at either the Q_A or Q_B binding sites, as compared with unbound UQ-10. On this basis, it was concluded that the methoxy groups are similarly oriented out of plane in both sites and do not contribute significantly to the differences of redox energies between Q_A and Q_B.²⁷ This is at odds with the functional observations described above, and one can suggest that the necessary redox potential difference is quite small^5-9 and might not result in significant IR spectroscopic changes.

Scheme 1.

Considering specifically the SQ states in the Q_A and Q_B sites, one can anticipate that orientations of the methoxy substituents can be inferred from analysis of the hyperfine tensor of ¹³C labelled methoxy groups as described recently for the SQ in the high-affinity site of cytochrome bo₃ ubiquinol oxidase from E. coli.²⁸ At present, no information about ¹³C hyperfine couplings in methoxy groups of SQ_A and SQ_B in RCs is available.

In this work, we report the experimental ¹³C hyperfine couplings for the methyl and two methoxy groups of ¹³C site-specifically labeled SQ_A and SQ_B in RCs of Rb. sphaeroides, determined using 2D ESEEM (HYSCORE). Analysis of the couplings to characterize the distribution of the unpaired spin density and conformations of methoxy groups and their influence on the redox properties of the two sites is performed with the aid of DFT calculations of hyperfine tensors and electron affinity values.

EXPERIMENTAL SECTION

Samples

Headgroup ¹³C-methyl labeled ubiquinone was biosynthesized in a strain of E. coli ²⁹ that is auxotrophic for eight amino acids, including methionine, which is the methyl donor in ubiquinone synthesis. Growth of this strain in the presence of ¹³C-methyl methionine results in ¹³C labeling of the methoxy and methyl carbons of the ubiquinone headgroup (Scheme 1). The product of biosynthesis in E. coli is UQ-8, rather than UQ-10, but no functional differences exist between them.¹¹

After growth, ubiquinone was extracted in organic solvents and purified by TLC.²⁸

Reaction centers used in this study were isolated from a strain of Rb. sphaeroides expressing RCs with a histidine-tag on the M subunit.³⁰ Cells were grown using ¹⁵N-labeled ammonium sulfate (Cambridge Isotopes), to prevent peak overlap and the strong cross-suppression effects of ¹⁴N on the ¹³C modulation.³¹ In order to isolate SQ EPR signals, the native, high spin Fe²⁺ must be replaced by diamagnetic Zn²⁺. Procedures for biochemical metal exchange, along with the methods of bacterial cell growth and RC isolation, were as previously described.³² Quinones were extracted from RCs by the method of Okamura et al.,³³ and were replaced with the ¹³C-methyl labeled ubiquinones. The SQ_A radical was generated by dithionite reduction in semianaerobic conditions (with continuous Argon flow over the sample). SQ_B was generated by exposing the RCs to a single 532 nm Nd-YAG laser pulse in the presence of ferrocytochrome c (to quickly rereduce the bacteriochlorophyll dimer after charge separation).³² All samples were frozen promptly in liquid nitrogen. ¹³C-labeling of the methyls does not influence the SQ linewidth, indicating that it is still dominated by the g-tensor anisotropy. The ¹H, ¹⁵N ESEEM for UQ-8 SQs was identical with the spectra of natural UQ-10 SQs thus confirming similar H-bond patterns after replacement.

EPR and ESEEM measurements

The CW EPR measurements were performed on an X-band Varian EPR-E122 spectrometer. The pulsed EPR experiments were carried out using an X-band Bruker ELEXSYS E580 spectrometer equipped with an Oxford CF 935 cryostat. All measurements were made at 80 K. The 2D, four-pulse experiment (π/2-τ-π/2-t₁-π-t₂-π/2-τ-echo, also called HYSCORE),³⁴ was employed with appropriate phase-cycling schemes to eliminate unwanted features from the experimental echo envelopes. The intensity of the echo after the fourth pulse was measured with t₂ and t₁ varied and constant τ. The length of a π/2 pulse was 16 ns and a π pulse 32 ns. HYSCORE data were collected in the form of 2D time-domain patterns containing 256×256 points with steps of 20 or 32 ns. Spectral processing of ESEEM patterns, including subtraction of the relaxation decay (fitting by polynomials of 3-6 degree), apodization (Hamming window), zero filling, and fast Fourier transformation (FT), were performed using Bruker WIN-EPR V2.22 Rev.10. Processed data were then imported into Matlab R2010a via the EasySpin package³⁵ to either be simulated by EasySpin, or be analyzed by a homemade script for fitting data in (ν₁)² vs. (ν₂)² coordinates. After plotting the HYSCORE as (ν₁)² vs. (ν₂)², ridges were fit via a linear regression with each point on the ridge weighted according to its HYSCORE intensity (see below).

Computational Methods

All density functional calculations were performed using Gaussian 09.³⁶ All calculations, including geometry optimization, conformational analysis and hyperfine coupling, were performed using the B3LYP functional and the EPR-II basis set. Specific details concerning hyperfine coupling calculations and the SQ_A and SQ_B site models are as previously described.^37-39 In addition to these, two new models of ubisemiquinone were used, termed SQ_M1 and SQ_M2. These are shown in Figure S1 of Supporting Information and were used to model the semiquinone hydrogen bonding to four water molecules (SQ_M1,) and one water molecule (SQ_M2). Conformational analysis using the SQ_M2 model was achieved by varying the C_mO_mC2C1 dihedral angle from 0° to 180° in 20° steps while optimizing all other parameters.

Powder ¹³C ESEEM spectra

The high-resolution pulsed EPR techniques, such as ESEEM and ENDOR, make use of the paramagnetic properties of the SQ intermediate, its interactions with nearby magnetic nuclei of the protein, the aqueous solvent, and the quinone molecule itself. 1D and 2D ESEEM, can be used to explore the fine-tuning of the environment, and the geometry of substituents and electronic structure of the SQ, via the isotropic and anisotropic hyperfine interactions with magnetic nuclei (¹³C in this work).¹⁵ ESEEM measures frequencies of nuclear transitions from nuclei interacting with an S=1/2 electron spin of the SQ. There are only two transitions with frequencies ν_α and ν_β for ¹³C with nuclear spin I=1/2, corresponding to two different states m_s=±1/2 of the SQ electron spin in a constant applied magnetic field. The value of the frequencies depends on the vector sum of the applied magnetic field and local magnetic field induced at the nucleus by the isotropic and anisotropic hyperfine interactions with the electron spin. In this work we used X-band EPR with microwave frequency ~9.7 GHz and magnetic field ~350 mT. The X-band EPR spectrum of the SQ in frozen solutions is a single line with the width ~0.8-1.0 mT with unresolved hyperfine structure. This width is comparable to the excitation width of the EPR spectrum by microwave pulses. In this case, the pulses can be considered as a giving a complete excitation of the powder EPR spectrum and thus the ESEEM spectra obtained are the powder-type spectra of nuclear frequencies with all different orientations of applied magnetic field relative to the principal axes of ¹³C hyperfine tensor(s). The frequencies of ν_α and ν_β transitions vary between

$$\begin{array}{cc}\hfill {\nu }_{\alpha \left(\beta \right)\perp }& =\mid {\nu }_{\mathrm{C}}+(-){\mathrm{A}}_{\perp }∕2\mid \text{and}\hfill \\ \hfill {\nu }_{\alpha \left(\beta \right)\parallel }& =\mid {\nu }_{\mathrm{C}}+(-){\mathrm{A}}_{\parallel }∕2\mid \hfill \end{array}$$ (1)

corresponding to the perpendicular and parallel orientations of the magnetic field and the unique axis of the axial hyperfine tensor (ν_C is the Zeeman frequency of ¹³C in the applied magnetic field, A_⊥=|a–T| and A_∥ = |a+2T|, a is the isotropic hyperfine constant, and the T-components of the anisotropic hyperfine tensor are (-T, -T, 2T). The principal values of the rhombic hyperfine tensor can be defined as follows: (–T(1 + δ), –T(1 – δ), 2T) with 0 ≤ δ ≤ 1, where δ is a rhombic parameter.

In this work we used HYSCORE because it provides better resolution of the extended lines of low intensity. The HYSCORE experiment creates off-diagonal cross-peaks (ν_α, ν_β) and (ν_β, ν_α) from each I=1/2 nucleus in the 2D spectrum. Powder HYSCORE spectra of I = ½ nuclei reveal, in the form of cross-ridges, the interdependence of ν_α and ν_β in the same orientations (see Figure 1). The two coordinates of the arbitrary point at the cross-ridge, described in the first-order by the equation

$${\nu }_{\alpha \left(\beta \right)}=\mid {\nu }_{\mathrm{c}}+(-)\mathrm{A}∕2\mid$$ (2)

can be used for the first-order estimate of the corresponding hyperfine coupling constant A:

$${\nu }_{\alpha }\text{-}{\nu }_{\beta }=\mathrm{A}$$ (3)

Figure 1.

Contour (top) and stacked (bottom) HYSCORE spectra of SQ_A (left) and SQ_B (right) in ¹⁵N uniformly labeled RCs [magnetic field 345.2 mT (Q_A) and 345.1 mT (Q_B), time between first and second pulses (τ) 136 ns, microwave frequency 9.686 GHz (Q_A) and 9.684 GHz (Q_B)].

Analysis of the ridges in (ν_α)² vs. (ν_β)² coordinates allows for direct, simultaneous determination of the isotropic a and anisotropic T components of the hyperfine tensor as described below in “Results”.^40,41

RESULTS

The ¹³C hyperfine couplings in SQ_A and SQ_B were probed by HYSCORE experiments. Figures 1 shows representative HYSCORE spectra for SQ_A and SQ_B in the frequency interval from 0 to 7 MHz for both axes. The spectra of the SQ in both sites contain the lines from ¹⁵N and ¹³C nuclei. Here we focus on the analysis of the ¹³C lines. The ¹³C cross-features are located along the antidiagonal, symmetrically around the diagonal point (ν_C,ν_C) where ν_C is ~3.7 MHz, in the applied magnetic field. The locations of the cross-peaks are significantly different for the two SQs.

The spectrum of SQ_A (Figure 1) exhibits a peak 1_C located around diagonal point (ν_C,ν_C) and two pairs of cross-peaks 2_C and 3_C. They are positioned symmetrically around the diagonal peak, along the antidiagonal, with maxima at (4.3, 3.1) MHz (2_C) and (5.5, 1.9) MHz (3_C), which correspond to first-order estimated hyperfine couplings 1.2 and 3.6 MHz, respectively. The width of these cross-ridges allows us to conclude that the dominant contributions to these couplings come from the isotropic constant. On the other hand, the lineshape of the 1_C peak suggests that the isotropic constant is close to zero in this case.

In contrast, the spectrum of SQ_B (Figure 1) consists of three pairs of cross-peaks 1_C – 3_C. Cross-peaks 2_C and 3_C are partially overlapped, but their maxima are well separated. Coordinates of the maxima at (4.5, 3.0) MHz (1_C), (5.7, 1.7) MHz (2_C), and (6.0, 1.5) MHz (3_C) define the couplings 1.5, 4.0 and 4.5 MHz, respectively. The spectrum also contains a sharp peak of low intensity at the diagonal point (ν_C,ν_C). We assign this line to weakly coupled, natural abundance (1.1%) ¹³C nuclei present in the protein surrounding of SQ_B. In the SQ_A spectrum a similar line is masked by the significantly more intense 1_C diagonal feature from one site-specifically labeled carbon.

More complete information about separate values of the isotropic and anisotropic parts of hyperfine tensor of the ¹³C nuclei contributing to the spectra of SQ_A and SQ_B can be obtained from the analysis of the contour lineshape of the cross-peaks. The contour lineshape of cross-peaks in the form of narrow ridges extending along the antidiagonal suggests an axial anisotropic hyperfine tensor for all contributing ¹³C. The analysis providing isotropic and anisotropic hyperfine couplings is straightforward in this case. The ideal cross-peak shape is an arc-type ridge between the points (ν_α⊥, ν_β⊥) and (ν_α∥, ν_β∥) located on the |ν_α ± ν_β| = 2ν_C lines. The shape of the ridge is described by the general equation ν_α=(Qν_β² + G)^1/2, where Q and G are coefficients that are functions of a, T and ν_C.^39,40 This lineshape transforms to a straight line segment in the coordinates (ν_α)² vs. (ν_β)². It should be noted, however, that HYSCORE intensity at points (ν_α⊥,ν_β⊥) and (ν_α∥,ν_β∥), corresponding to orientations of the magnetic field along the A_⊥ and A_∥ principal directions of the hyperfine tensor, is equal to zero and is significantly suppressed in the orientations around the principal directions.⁴¹ Therefore, in HYSCORE spectra, only the central part of the cross-ridge, which corresponds to orientations of the magnetic field substantially different from the principal directions, will possess observable intensity.⁴¹ This means that in real spectra the cross-peak borders do not cross the |ν₁ ± ν₂| = 2ν_C line(s). However, the crossing points (ν_α⊥, ν_β⊥) and (ν_α∥, ν_β∥) can be determined through the linear fitting of the observable parts in (ν₁)² vs. (ν₂)² coordinates.^40,41

Linear regression of the cross-peaks in the spectra of SQ_A and SQ_B plotted in the (ν₁)² vs. (ν₂)² coordinates is shown in Figure 2. Linear regression gives intersection points with the |ν₁ ± ν₂| = 2ν_C curve for each cross-peak. These points define two principal values of the hyperfine tensor. There are two possible assignments to (ν_α⊥, ν_β⊥) or (ν_α∥, ν_β∥) for each crossing point and, consequently, two solutions, one for each assignment. The tensors obtained from the analysis of the SQ_A and SQ_B spectra are summarized in Table 1. Uncertainty in assignment of ν₁ to ν_α or ν_β and, respectively, ν₂ to ν_β or ν_α, allows alternate signs of a and T in both solutions (see footnote to Table 1). Complete information for the linear regression analysis of all cross-peaks and calculation of the tensors is provided in Supporting Information (Table S1 and Figure S2).

Figure 2.

Contour presentation of the HYSCORE spectra of SQ_A (left) and SQ_B (right) in (Figure 1) in ((ν₁)² vs. (ν₂)²) coordinates. The dotted curve is defined by |ν₁ ± ν₂|= 2ν_C. Inserts show the linear regression fits for selected cross-peaks. (Graphs showing inserts for the other fitted ridges are included in the Supplementary Information.)

Table 1.

¹³C hyperfine tensors for SQ_A and SQ_B from linear regression of ridges plotted as (ν₁)² vs. (ν₂)².^a

Site	Set	a (MHz)	T (MHz)
QA	2C	1.3 (± 0.2)	0.4 (± 0.1)
		−1.7 (± 0.2)	0.4 (± 0.1)
	3C	3.2 (± 0.1)	0.4 (± 0.1)
		−3.6 (± 0.1)	0.4(± 0.1)
QB	1C	1.4 (± 0.2)	0.5 (± 0.1)
		−1.9 (± 0.2)	0.5(± 0.1)
	2C	3.6 (± 0.1)	0.4 (± 0.1)
		−4.0 (± 0.1)	0.4(± 0.1)
	3C	4.6 (± 0.1)	0.6 (± 0.1)
		−5.2 (± 0.1)	0.6 (± 0.1)

^ashown signs of a and T are relative, the general form of two solutions: (±a₁, ±T) and (±a₂, −/+T) with equal |2a₁+T| and |2a₂+T|.

In order to choose between the two sets of axial tensors from analysis of ¹³C HYSCORE plotted as (ν₁)² vs. (ν₂)², the ¹³C ridges for Q_A and Q_B were simulated with EasySpin.³⁵ For both Q_A and Q_B, simulated spectra were found to be essentially independent of the relative Euler angles between the ¹³C hyperfine tensors, possibly due to the low level of hyperfine anisotropy. Thus spectra for both solution sets for a and T from Table 1 were simulated and compared without adjusting the Euler angles.

For SQ_A, simulations with one set of axial tensors were in much better agreement with the experimental spectrum than the other. Peak 1_C was simulated, and it was found that its isotropic coupling could not deviate from zero by more than 0.2 MHz for the lineshape to resemble experimental data. For SQ_B, simulations with slightly rhombic tensors showed much better agreement with the experimental spectra for 1_C and 3_C features.

Preferred tensors revealed from the HYSCORE spectra simulations are shown in Table 2. Spectra simulated for two possible tensors for each cross-feature in SQ_A and SQ_B spectra and its comparative analysis are provided in Supporting Information (Table S2 and Figure S3).

Table 2.

¹³C hyperfine tensors determined from the HYSCORE spectra simulations for SQ_A and SQ_B.

Site	Set	a (MHz)	T (MHz)
QA	1C	0	0.4
	2C	1.3	0.5
	3C	−3.6	0.4
QB	1C	1.5	0.4 (δ=0.3)
	2C	−3.9	0.4
	3C	4.7	0.4 (δ=0.1)

DISCUSSION

For the selectively ¹³C labeled samples used here, three ¹³C hyperfine couplings are expected corresponding to the two methoxy groups and the methyl group carbon nuclei. The HYSCORE spectra of Figures 1 and 2 clearly illustrate these three hyperfine couplings for both SQ_A and SQ_B. Analysis of the spectra shows that for SQ_A one of these carbons has a hyperfine coupling close to zero, whereas values of magnitude 1.2 and 3.6 MHz can be estimated for the other two. For SQ_B, values of magnitude 1.9, 4.0 and 4.9 MHz can be estimated. From the spectral lineshapes, low anisotropy is expected, with the estimated values dominated by the isotropic contribution. This is confirmed by a detailed lineshape analysis and spectral simulations, which separate out the isotropic and anisotropic components, as shown in Tables 1 and 2. For the assignment of these measured values to the ¹³C nuclei of the methyl or methoxy groups, it is necessary to compare with previous experimental results in vitro and with values calculated using DFT.

¹³C hyperfine couplings for the methyl group in ubisemiquinone and the related durosemiquinone (DQ) radicals have been previously reported in hydrogen bonding solution (Table 3). For DQ an isotropic coupling of magnitude 3.8 MHz has been reported using ENDOR.^42,43 Similar coupling around 4.0 MHz have been reported for the semiquinone anion radical of UQ-10 in alcohol solution.⁴⁴ Calculated DFT (B3LYP/EPR-II) hyperfine couplings for a hydrogen bonded DQ have been reported and the calculated negative isotropic constant of −3.5 MHz is in good agreement with the experimental magnitude.⁴⁵ These calculations also indicated that a and T have opposite signs in the ¹³C methyl hyperfine tensor. In the current study we have also used SQ_M1 (6-methyl-UQ with 4 H₂O) to simulate the hydrogen bonded ubisemiquinone in solution. After geometry optimization the calculated C5'-methyl group ¹³C isotropic coupling of −3.8 MHz is again in good agreement with the magnitude of the experimental finding, 4.0 MHz. Experimental and calculated values for the ubisemiquinone in the SQ_H site cytochrome bo₃ ubiquinol oxidase and its D75H mutant are also shown. The data in Table 3 indicate an impressive ability of DFT (B3LYP/EPR-II) to reproduce the experimental value of methyl coupling, though all calculated magnitudes are slightly less than the experimental determination.

Table 3.

Comparison of experimental and calculated ¹³C methyl (5’) isotropic constants in SQs.

Semiquinone	aexp(13C), MHz	acalc(13C), MHz
UQ-10	|4.0|a	−3.8
DQ	|3.8|b	−3.5c
SQH	−6.1d	−5.3e
SQH (D75H)	−4.7d	−4.4e
SQA	−3.6	−2.9
SQB	−3.9	−3.5

^afrom ref. 44

^bfrom refs. 42,43

^cfrom ref. 45

^dfrom ref. 27

^efrom ref. 46.

Table 3 also shows the DFT calculated ¹³C isotropic couplings for the C5'-methyl group in the SQ_A and SQ_B site models. For SQ_A, therefore, we can confidently assign the pair a=−3.6 MHz and T=+0.4 MHz (Table 2, set 3_C) to the 5’ methyl group. The isotropic constants from the other two sets (~0 and −1.7 MHz) are too small to be attributed to this group. For SQ_B the calculated value of −3.5 MHz strongly suggests that the pair a=−3.9 MHz and T=+0.4 MHz (Table 2, set 2_C) corresponds to the methyl ¹³C. The other isotropic couplings of value |4.7| MHz and |1.5| MHz are too large and too small respectively.

Based on the above analysis, the two remaining hyperfine tensors must correspond to the ¹³C interaction for the two methoxy groups. In this case analysis is complicated by the expected dependency of the tensor on the orientation of the methoxy group relative to the ring plane. To aid assignment it is first instructive to examine the Mülliken spin populations calculated for each semiquinone and these are given in Table 4.

Table 4.

Mülliken spin populations in selected SQs.

Semiquinone	C2	C3
SQA	0.11	−0.01
SQB	0.09	0.02
SQM2	0.11	0.00
SQM1	0.07	0.07

For SQ_A, values of +0.11 and −0.01 are calculated for positions at C2 and C3, respectively. This arises from the polarization of the spin density distribution caused by the stronger hydrogen bonding to the O1 oxygen atom, as discussed previously.³⁸

The ¹³C isotropic hyperfine coupling of the methoxy group will be proportional to the π(p) spin population of the corresponding ring carbon and the dihedral angle of the methoxy CO bond with respect to the SQ ring plane, see below. Based on the near zero calculated spin population for the C3 position, the diagonal peak 1_C with isotropic coupling close to zero is expected for the 3-methoxy group carbon. The remaining cross-peaks 2_C and corresponding ¹³C coupling with magnitude 1.3 MHz therefore can only arise from the 2-methoxy carbon.

For SQ_B, Table 4 shows that, as for SQ_A, the calculated spin population at position C2 (0.09) is again substantially larger than C3 (0.02). Based on this we expect that the 2-methoxy ¹³C isotropic hyperfine coupling is significantly larger than the 3-methoxy ¹³C value. Thus, from Figure 2 and Table 2, we can assign the larger magnitude 4.7 MHz coupling to the C2 methoxy carbon and the smaller one of magnitude 1.5 MHz to the C3 methoxy. Therefore, the optimized simulation data in Table 2 and DFT calculations (Tables 3 and 4) allows us to make fully consistent assignments of ¹³C hyperfine tensors determined for SQ_A and SQ_B to particular substituents.

The assignments above suggest that the 2-methoxy ¹³C isotropic constant is significantly larger for SQ_B compared with SQ_A. This cannot be explained on the basis of the C2 spin population value, which is larger (0.11 versus 0.09) for the SQ_A. The unpaired spin density giving rise to the ¹³C isotropic hyperfine coupling for the methoxy group carbon atom arises from a combination of spin polarization and hyperconjugation. When the methoxy group is held in the ring plane, hyperconjugation is expected to be zero and small negative hyperfine couplings are expected for in or near to in-plane orientations, due to spin polarization by the methoxy oxygen spin density. As the methoxy orientation is moved progressively out of plane a positive contribution arising from hyperconjugation with the ring carbon π(p) spin density arises. For the SQ_M2 model (6-methyl-UQ with 1 H₂O), which approximates the asymmetric hydrogen bonding of SQ_A and SQ_B, Figure 3 shows that the isotropic coupling exhibits this trend with negative values at in-plane orientations and large positive values for out of plane orientations. The SQ_M2 model spin populations at C2 and C3 are very similar to the SQ_A values, Table 4. The ¹³C data obtained for SQ_B compared with SQ_A, implies that the 2-methoxy group is oriented further out of the ring plane thereby giving rise to the larger isotropic coupling.

Figure 3.

Effect of rotation of the 2-methoxy group on its ¹³C isotropic hyperfine constant for model SQ_M2.

Figure 3 allows us to estimate the difference in 2-methoxy conformation in SQ_A and SQ_B. An isotropic coupling of 1.3 MHz corresponds to a 2-methoxy conformation of ~50° or ~155°. In contrast, the simulated isotropic coupling for 2-methoxy in SQ_B, adjusted to the C2 spin population 0.11, as in the SQ_M2 model, is 5.6 MHz, i.e. multiplied by 1.22, and defines the conformation as ~75° or ~135°. These estimates indicate a difference on the order of ~20-25° between the 2-methoxy conformations in SQ_A and SQ_B.

DFT calculations on the hydrogen bonded ubisemiquinone model SQ_M2 also show that the electron affinity of the quinone increases as the methoxy group is rotated out of the ring plane (Figure 4). The more out of plane orientation exhibited by the 2-methoxy group in SQ_B suggests a greater electron affinity value for Q_B which should lead to a higher redox potential compared with Q_A. Our estimated difference of the 2-methoxy conformation of ~20-25° corresponds to ΔE_A~0.05 eV giving a predicted 50 mV difference in redox potential between Q_A and Q_B. This is a significant fraction of the experimental difference, ΔE_m = 60-75 mV, indicating that the orientation of the 2-methoxy group between Q_A and Q_B may be a crucial factor in determining the functional E_m drop between Q_A and Q_B. Further studies aimed at providing a more quantitative measure of this factor in governing the redox potential of Q_A and Q_B are in progress.

Figure 4.

Effect of rotation of the 2-methoxy group on the electron affinity of model SQ_M2.

CONCLUSIONS

HYSCORE studies on bacterial reaction centers, with Q_A and Q_B site ubiquinones ¹³C-labeled at the headgroup methyl and methoxy groups, have been performed. ¹³C isotropic and anisotropic hyperfine couplings have been obtained and are compared to those calculated by density functional theory. This has allowed firm assignment of measured hyperfine tensors to specific methyl and methoxy carbon nuclei. Further computational analysis indicates a difference in the conformation of the 2-methoxy group between the Q_A and Q_B site semiquinone forms. This difference is attributed to a more out of plane orientation of the 2-methoxy group adopted by the Q_B semiquinone. An out of plane orientation is also shown to be associated with increased electron affinity, indicating that it can contribute significantly to the higher redox potential of the Q_B site ubiquinone in bacterial type II reaction centers. This work shows that ¹³C HYSCORE, coupled with high quality DFT calculations, provide a very sensitive method for characterizing methoxy group orientations in ubisemiquinone, as manifested in the changes of the ¹³C isotropic hyperfine couplings. Preliminary results obtained for other systems indicate that this approach, with appropriate development of the theoretical foundations, will allow quantitative measures for the redox potentials of ubiquinone in other quinone processing sites in proteins.

Abbreviations

2D

two-dimensional

CW

continuous-wave

DFT

density functional theory

B3LYP

Becke3 Lee–Yang–Parr

SQ

semiquinone

RC

reaction center

EPR

electron paramagnetic resonance

ENDOR

electron nuclear double-resonance

ESEEM

electron spin echo envelope modulation

HYSCORE

hyperfine sublevel correlation

UQ-10

ubiquinone-10