Protein-Cofactor Interactions in Bacterial Reaction Centers from Rhodobacter sphaeroides R-26: I. Identification of the ENDOR Lines Associated with the Hydrogen Bonds to the Primary Quinone Q_A^⋅−

Abstract

Hydrogen bonds are important in determining the structure and function of biomolecules. Of particular interest are hydrogen bonds to quinones, which play an important role in the bioenergetics of respiration and photosynthesis. In this work we investigated the hydrogen bonds to the two carbonyl oxygens of the semiquinone in the well-characterized reaction center from the photosynthetic bacterium Rhodobacter sphaeroides R-26. We used electron paramagnetic resonance and electron nuclear double resonance techniques at 35 GHz at a temperature of 80 K. The goal of this study was to identify and assign sets of ¹H-ENDOR lines to protons hydrogen bonded to each of the two oxygens. This was accomplished by preferentially exchanging the hydrogen bond on one of the oxygens with deuterium while concomitantly monitoring the changes in the amplitudes of the ¹H-ENDOR lines. The preferential deuteration of one of the oxygens was made possible by the different ¹H → ²H exchange times of the protons bonded to the two oxygens. The assignment of the ¹H-ENDOR lines sets the stage for the determination of the geometries of the H-bonds by a detailed field selection ENDOR study to be presented in a future article.

INTRODUCTION

Quinones play an important role as cofactors in many proteins (1,2). Their biological function is related to their electrochemical properties, which are influenced by their local protein environment. Quinones can be reduced in two successive reversible one-electron steps. The first step leads to an intermediate, a semiquinone (quinone anion radical); the second step, followed by protonation, leads to the fully reduced quinol (3,4). These properties make quinones ideally suited as electron and proton carriers in the bioenergetic processes of respiration and photosynthesis (5).

Two ubiquinones, Q_A and Q_B, are present in the reaction centers (RCs) of photosynthetic purple bacteria. Q_A accepts one-electron and is not protonated, whereas Q_B accepts, sequentially, two electrons. The second electron transfer is coupled with protonation leading to the formation of ubiquinol, which is weakly bound and diffuses out of the binding site to participate in the formation of the proton gradient required for ATP synthesis (5). Q_A and Q_B form hydrogen bonds to the RC protein. The bonds contribute to the binding and to the chemical properties and function of the quinones. It is, therefore, important to characterize the hydrogen bonds in detail.

Electron paramagnetic resonance (EPR) and electron nuclear double resonance (ENDOR) spectroscopies are ideally suited techniques to study the transient paramagnetic species involved in the primary events of photosynthesis, e.g., the ubiquinone anion radicals, and These studies provide information about the electronic and spatial structure of the transient radicals. In particular, in the photosynthetic RC of Rhodobacter sphaeroides has been investigated extensively using EPR and ENDOR spectroscopies (6–12). These investigations showed that the ENDOR lines of are due to three classes of protons: 1), protons associated with the protein in the vicinity of the binding site (matrix lines);2), nonexchangeable protons on the quinone, e.g., methyl, methoxy, or methylene protons; and 3), exchangeable protons, the most important of which are the protons forming hydrogen bonds with the two carbonyl oxygens of the quinone. The ENDOR lines corresponding to the exchangeable protons were identified by comparing ENDOR spectra of RCs in H₂O and D₂O buffer (6).

The initial ENDOR results suggested that is bound to the RC protein by two nonequivalent hydrogen bonds to the carbonyl oxygens of the quinone (7). Confirming evidence of two H-bonds was obtained by electron spin echo envelope modulation (ESEEM) (13–15). Additional, independent evidence of two nonequivalent hydrogen bonds was obtained by Paddock et al. (16), who showed that there are two different proton-deuterium exchange times associated with two distinct hydrogen bonded protons.

Based on the RC structure (17) the two hydrogen bonds were assigned to the imidazole nitrogen of His (M219) (N_δ-H⋯O₄) and the NH group of Ala (M260) (N-H⋯O₁) (see Fig. 1). However, there remained two problems: 1), in none of the previous experiments (6,7,16) could the ENDOR lines be assigned to protons associated with specific oxygens (i.e., O₁ or O₄, Fig. 1); and 2), there remained the puzzle why only three pairs of ¹H-ENDOR lines were observed instead of the expected four pairs (in randomly oriented molecules each proton gives rise to two pairs of lines corresponding approximately to the parallel and perpendicular components of the axially symmetric hyperfine tensors). In this work we addressed both of these problems. We first determined the different ¹H → ²H exchange times of the two hydrogen bonds by incubating protonated (or deuterated) RCs for different times in D₂O and monitoring the amplitudes of the ¹H-ENDOR lines. By using the results of these experiments we prepared RCs whose hydrogen bonds were preferentially deuterated (or protonated) at either O₁ or O₄. This enabled us to determine which set of ENDOR lines belong to a given oxygen. This solved the problem of the assignment of the observed three pairs of lines. Using the known structure of the RC (17) we postulated an assignment of each set of lines to a particular oxygen. A detailed simulation of the ENDOR spectra and the determination of the geometry of the H-bonds will be presented in a future article.

FIGURE 1.

Structure in the vicinity of Q_A in the RC of Rb. sphaeroides (Brookhaven Protein Data Bank entry 1AIG (17)). The iron is ligated to four histidines (only two are shown) and one glutamate (bidentate ligand). The isoprenoid chain of the quinone is truncated for simplicity. The two carbonyl oxygens of the quinone, labeled 1 and 4, form H-bonds shown by dotted lines.

Hamiltonian of the spin system

In this study, we focus on the interaction between the magnetic moment of the unpaired electron of and the magnetic moments of protons or deuterons that form hydrogen bonds to the quinone oxygens O₁ and O₄ (see Fig. 1). The observed ENDOR spectra were interpreted using a spin Hamiltonian, ℋ, containing the electron and nuclear Zeeman interactions with the applied magnetic field B_o, the hyperfine coupling (hfc) and the nuclear quadrupole coupling (nqc) terms (e.g., Weil et al. (18)):

(1)

where S is the electron spin operator, I is the nuclear spin operator of protons or deuterons in H-bonds to the carbonyl oxygens; A and P are the hfc and nqc tensors in frequency units, g is the electronic g-tensor, h is Planck's constant, g_N is the g-factor of the corresponding magnetic nucleus (proton or deuteron), and β_e and β_N are the electron and nuclear magnetons, respectively.

The first term in Eq. 1 represents the electronic Zeeman term that gives rise to the observed EPR spectrum. The other three terms represent nuclear interactions that give rise to the ENDOR spectra. These can be calculated to first order from Eq. 1 using an expansion in powers of the ratio hA/(gβ_eB_o) (18,19). At the microwave frequency of 35 GHz used in this work, second order terms are small and can be neglected. Furthermore, at 35 GHz the ENDOR free frequency of protons (or deuterons) is much larger than its respective hfc interaction, i.e., g_Nβ_NB_o/h ≫ A/2. Thus, the resonance frequency of the ENDOR transition M_γ ↔ M_γ − 1 is given to first order for a given direction of the magnetic field by (18,19):

(2)

where M_S is the magnetic quantum number of the electron (±½) and M_γ the magnetic quantum number corresponding to the lower energy levels of the nucleus, i.e., +½ for protons and +1, 0 for deuterons.

For protons M_γ = +½ making the term containing the nuclear quadrupole interaction P zero. Thus, for a given magnetic field direction with respect to the molecular axes, one expects two ENDOR lines separated by the hfc A centered around the proton Larmor frequency g_Nβ_NB_o/h. For deuterons M_γ = +1, 0 and each line is additionally split by 3P. In a sample having randomly oriented molecules, the anisotropies of A and P smear out the spectrum and sharp lines are observed only at the extrema (e.g., approximately at A_// and A_⊥).

At 35 GHz (Q-band) the anisotropies of the electronic g-value gives rise to a spectrum shown in Fig. 2. At the positions labeled g_x, g_y, and g_z, the majority of the molecules contributing to the spectrum have their molecular axes along the principal directions x, y, and z. However, it should be noted that at each position there is a significant admixture of molecules having different directions. Thus, only a partial magnetic field selection of orientations is obtained.

FIGURE 2.

EPR powder spectrum of in H₂O. Insets show the quinone orientations with respect to the magnetic field corresponding to g_x, g_y, and g_z. Experimental conditions: T = 80 K, MW frequency = 35.03 GHz, MW power = 1 × 10⁻⁷ W, field modulation = 0.15 mT peak-to-peak at 270 Hz, average of 9 scans, 20 s per scan.

At Q-band the magnetic field B_o is ∼1.25 T and the ENDOR lines for protons and deuterons are observed in different frequency ranges, ∼53 MHz for ¹H and ∼8 MHz for ²H. This enabled us to separate the signals due to protons in H-bonds from those in the quinone ring by using samples of deuterated RCs in H₂O buffer. To observe deuterons in H-bonds we used samples of protonated RCs in D₂O buffer.

MATERIALS AND METHODS

Preparation of reaction centers

RCs from Rb. sphaeroides R-26 were isolated and purified following the procedures of Isaacson et al. (9). The paramagnetic nonheme Fe²⁺ was chemically removed and replaced with diamagnetic Zn²⁺ to reduce the EPR line width of the semiquinone, following the procedure of Utschig et al. (20). The ratio of Zn/RC was determined by atomic absorption spectroscopy and Q-band EPR spectroscopy to be ≥0.90.

Deuteration of reaction centers

Fully deuterated RCs were obtained by growing Rb. sphaeroides R-26 bacteria in a D₂O-based modified Hutner's medium using perdeuterated succinic acid (98% D, Aldrich, Milwaukee, WI) as the sole carbon source as described previously by van der Est et al. (21).

Proton-deuterium exchange

We used two sets of samples in which protons were exchanged with deuterons—one for the determination of the exchange times and the other for obtaining detailed ENDOR spectra, which required a higher signal/noise ratio, i.e., a higher RC concentration.

For the determination of the exchange times, fully protonated and fully deuterated RCs were purified and concentrated to an optical absorbance of The RC samples were diluted into D₂O buffer for different lengths of time, at the end of which was generated photochemically (see next section) and frozen by immersing the sample in liquid nitrogen. We used a fivefold dilution for deuterated RCs and a fourfold dilution for protonated RCs. The final optical absorbance was ∼40 and ∼50, respectively.

To obtain ENDOR spectra in which one of the H-bonds was preferentially deuterated, we prepared the following high concentration () samples:

1. Deuterated RCs in H₂O buffer incubated for 27 min in D₂O buffer (10-fold dilution).

2. Protonated RCs in H₂O buffer incubated for 50 min in D₂O buffer (fivefold dilution).

3. Protonated RCs in D₂O buffer incubated for 190 min in H₂O buffer (300-fold dilution).

The incubation times (see section on ENDOR results) and dilutions were chosen to maximize the preferential deuteration of one of the H-bonds to Q_A. After dilution the samples were concentrated by centrifuging in a Millipore (5–50k) filter to an optical absorbance At the end of the incubation was generated chemically (see next section) and the sample frozen by immersing it in liquid nitrogen.

Generation of semiquinone radical anion

We used two methods to generate : 1), for low concentration samples () we illuminated the RCs with a single, saturating laser flash (Lumen-X DL2100C, λ = 590 nm) in the presence of excess cyt c₂ to reduce the donor, D⁺, and stigmatellin to inhibit electron transfer to Q_B (photochemical method). After the laser flash the sample was plunged into liquid nitrogen. 2), For high concentration samples (), we used Na₂S₂O₄ to reduce Q_A before freezing (chemical method) (6). The quartz sample tubes were type 705PQ (Wilmad, Buena, NJ) (OD 3 mm, ID 2 mm).

Nomenclature

Since we used several combinations of protonated and deuterated RCs and buffer, we define the following notation: a), () for in fully protonated RCs and () for in fully deuterated RCs; b), (H₂O) for fully protonated buffer and (D₂O) for fully deuterated buffer; and c), () for RCs prepared in H₂O and incubated later in D₂O for a specific time (in minutes) and () for the reverse.

EPR instrumentation

EPR and ENDOR measurements were performed at 35 GHz and 80 K. The spectrometer is a home-built superheterodyne-type instrument with a Varian klystron, a cylindrical TE₀₁₁ brass cavity, and an immersion dewar system for temperature control. The cavity and coupler is similar to one described by Jaworski et al. (22). A Li:LiF sample was used as a primary g-value standard (g = 2.00229) (23), and P-doped Si as a secondary standard (g = 1.99891 at 80 K) (24,25). The P-Si marker was permanently attached to the bottom wall of the cavity. ENDOR experiments were performed with the EPR spectra 50% saturated. ENDOR spectra were recorded at the magnetic field positions as indicated in the corresponding traces, using frequency modulation (FM) of ±140 kHz for protons and ±30 kHz for deuterons, at FM rates as indicated in the figure captions. The output of the RF amplifier (ENI 3100L) feeding the ENDOR coils was 50 W for protons and 25 W for deuterons. To improve the signal/noise ratio of the ENDOR signal, many traces were averaged for up to 3 h for protons and 42 h for deuterons.

EXPERIMENTAL RESULTS AND DISCUSSION

EPR results

The Q-band EPR spectrum from in protonated buffer (H₂O), at T = 80 K, is shown in Fig. 2. The broadening of the spectrum is mainly due to protons in solution (exchangeable protons), including the H-bonds to the carbonyl oxygens. Exchanging H₂O with D₂O, decreases the broadening (data not shown).

ENDOR results

¹H ENDOR spectrum corresponding to the H-bonds

Fig. 3 shows the Q-band ¹H ENDOR spectrum at T = 80 K of in H₂O (chemically reduced), arising from ¹H-bonds and from protons in the solvent. The spectrum was recorded at the magnetic field corresponding to g_y (see Fig. 2). The splittings of the ¹H-ENDOR peaks associated with the H-bonds are A₁ (4.82 MHz), A₂ (6.43 MHz), and A₃ (9.07 MHz). Similar splittings were previously observed at X-band by comparing ENDOR spectra of in H₂O and D₂O (6,12).

FIGURE 3.

¹H ENDOR spectrum of in H₂O. The spectrum was taken at the field position corresponding to g_y = 2.0053 (see Fig. 2). The intensities of the transitions are denoted by L₁, L₂, L₃, and the hyperfine coupling constants by A₁, A₂, A₃. Experimental conditions: T = 80 K, MW frequency = 35.03 GHz, MW power = 3 × 10⁻⁶ W; FM = ±140 kHz at a rate of 947 Hz. Number of scans: 500. Scan time: 4 s.

As described in the Spin Hamiltonian section, the single-crystal ¹H ENDOR spectrum corresponding to a proton contains two lines (see Eq. 2). However, at the magnetic field position corresponding to g_y, the spectrum corresponds to a “two-dimensional” powder-type spectrum, with weighted contributions mainly from all in-plane directions. Therefore, a powder spectrum is expected for each H-bond tensor with two sharp features, corresponding approximately to the parallel (A_//) and the perpendicular (A_⊥) components of an axially symmetric hfc tensor (7,26–28), i.e., each proton should give rise to two pairs of lines. For benzoquinone anion radicals in solution, the ENDOR experiments showed that the H-bond hfc tensor is approximately axially symmetric. This is typical for a proton in a normal H-bond. Deviations are expected only for very short H-bonds as described in (29). Since in reaction centers there are two nonequivalent H-bonds to (12–16), in principle four pairs of ¹H ENDOR lines should be observed in the spectrum. However, the experiment shows only three pairs, despite a thorough search for a fourth set of lines. This problem is discussed in a subsequent section.

There are two problems to be solved in the identification of the ENDOR lines associated with the hydrogen-bonded protons: 1), which pair of lines belong to the same proton; and 2), with which oxygen (O₁ or O₄) is a particular set of ENDOR lines associated.

To answer the first question we make use of the selective deuteration of the oxygens as discussed in the next section.

Determination of the ¹H → ²H exchange times of the two protons H-bonded to Q_A

To preferentially deuterate one of the H-bonds to Q_A, it is necessary to determine their respective ¹H → ²H exchange times. We used ENDOR to measure the ¹H → ²H exchange times of the two protons H-bonded to Q_A, in fully deuterated RCs and in fully protonated RCs (R-26). In the fully deuterated RCs in H₂O we measured the amplitudes of the ¹H-ENDOR lines L₁ and L₂ corresponding to the two protons (Fig. 3) at different times of incubation in D₂O. The incubation time started with a fivefold dilution of the H₂O sample in D₂O. The semiquinone was produced photochemically (see Materials and Methods) at the end of the incubation period.

In view of the fivefold dilution of the sample, 20% of the bonds remain protonated and only 80% are subject to exchange with deuterium. Thus, the normalized ratio of amplitudes of L₁ to L₂ is given by:

(3)

where and are the respective initial intensities and τ₁ and τ₂ are the respective ¹H → ²H exchange times for the two protons. For protonated RCs, the H₂O samples were diluted fourfold in D₂O.

Fig. 4 shows the observed time dependence of the normalized intensity ratio of the L₁ and L₂ lines for fully deuterated and fully protonated RCs. The ¹H → ²H exchange times were determined by fitting the experimental data with Eq. 3, with τ₁ and τ₂ as adjustable parameters. The resultant values of τ₁ and τ₂ are shown in the inserts of Fig. 4.

FIGURE 4.

The normalized intensities ratios () for the low frequency ¹H-ENDOR lines (see Fig. 3), as a function of the incubation time in D₂O for deuterated RCs (a) and for protonated RCs (b). The ¹H → ²H exchange times were obtained by fitting the experimental data (dots) with Eq. 3 (solid lines). The values of the exchange times τ₁ and τ₂ are shown in the inserts. Experimental conditions: Microwave frequency 35 GHz, T = 80 K, B_o at position corresponding to g_y (see Fig. 2). Optical absorbance of sample (a) and 50 (b).

The different exchange times observed for protonated and deuterated RCs is believed to reflect the structural changes in the RCs when the bacteria are grown in a fully deuterated medium. The shorter exchange times in deuterated RCs suggest a less compact structure, in which the quinone oxygens are more accessible to the solvent. Paddock et al. (16), who measured the exchange times in protonated RCs of the HC (M266) mutant, found an order of magnitude different times than in the wild-type, suggesting again a structural change. Thus, the exchange times are very sensitive to the details of the structure and may provide a useful probe in investigating structural changes.

Assignment of pairs of lines belonging to the same proton

To solve this problem we used the differential exchange of one proton with respect to the other as discussed in the previous section. The lines associated with a particular proton should show the same reduction in amplitude when the proton is exchanged with deuterons. Fig. 5 shows the low frequency part of the ¹H-ENDOR spectrum of in H₂O (top) and after 27 min of exchanges in D₂O (bottom), both samples were produced chemically. Deuterated RCs were used to eliminate ENDOR signals from protons of the protein. Using the exchange times for the two protons of 10 min and 89 min, respectively (see Fig. 4 a), we predict that one proton should be ∼80% exchanged and the other ∼20%. The reduction in amplitudes of L₁ and L₂ correspond roughly to these values. Thus, the peaks L₁ and L₂ are associated with two different protons, as had been postulated previously (7,16).

FIGURE 5.

Low-frequency part of the ¹H ENDOR spectra of in H₂O (upper spectrum) and in (lower spectrum), at 35 GHz, taken at a magnetic field of 1248.3 mT (see arrow in Fig. 2). The H-bond ENDOR lines are labeled L₁, L₂, L₃, and L₄ in the upper spectrum. Numbers next to each line refer to their amplitudes (in arbitrary units) (see Table 1). Experimental conditions: T = 80 K, MW power = ∼3 × 10⁻⁶ W, FM = ±140 kHz at a rate of 950 Hz. Number of scans: 500 (upper spectrum) and 2860 (lower spectrum). Scan time: 4 s.

We now turn to the assignment of the line corresponding to the largest hyperfine coupling (see left of Fig. 5). Does it represent the partner to L₁ or L₂? We shall show that this line is an overlap of two lines, one being a partner to L₁ and the other to L₂. The proof is as follows: The exchange reduced the amplitude of L₁ 8.1-fold (130/16; see Table 1). If the peak were only the partner of L₁ we would expect it to decrease to 1.1 units. The observed peak is 4 units, i.e., significantly larger, showing that a considerable part of it is associated with another line, L₄, the partner to L₂, thus confirming the overlap. The effect is best observed for a magnetic field position slightly off g_y (at B_o = 1248.3 mT; see Fig. 2).

TABLE 1.

Amplitudes of the ¹H ENDOR lines (in arbitrary units) corresponding to the protons H-bonded to in H₂O (before exchange) and in (after exchange) (see Fig. 5)

	L3 + L4	L2	L1
Before exchange	9	55	130
After exchange	4	34	16

Additional evidence for the overlap is provided by the ²H-ENDOR spectra of protonated RCs in D₂O shown in Fig. 6 (high concentration samples; see Materials and Methods). Each ²H-ENDOR line is split into a doublet by the nuclear quadrupole interaction, 3P (Eq. 2). The center of each doublet (solid arrows in Fig. 6) corresponds to the hfc (18). The top trace shows the low frequency ²H-ENDOR spectrum after incubating the RC for 50 min in D₂O. In these experiments protonated RCs were used to eliminate signals from the protein. With the exchange times of 50 min and 1200 min (Fig. 4 b), we expect one oxygen (O_a) to be ∼50% deuterated, whereas the other (O_b) only ∼3% (see Eq. 3 and inset in Fig. 6, top). Thus, the spectrum is essentially due to deuterons on O_a. In the lower trace the deuterons were exchanged with protons giving rise to an ²H-ENDOR spectrum associated with deuterons on O_b (see inset in Fig. 6, bottom). Comparing the two spectra we see that the low frequency line (at −0.8 MHz) of P₃ is shifted by ∼30 kHz from that of P₄ showing that the two lines belong to different D-bonds. The shift between the two lines is smaller than the line width which explains why the two lines are not resolved when both oxygens are protonated (or deuterated). These results clearly show that the line at −0.8 MHz is associated with deuterons on both O_a and O_b.

FIGURE 6.

Low-frequency part of the ²H ENDOR spectra of in (upper spectrum) and in (lower spectrum), at 35 GHz, taken along the magnetic field position corresponding to g_y (see Fig. 2). The position of the hfc tensor components (A₁, A₂, A₃, and A₄) are indicated by solid arrows, whereas the nqc splittings (3P₁, 3P₂, 3P₃, and 3P₄) are indicated by dotted arrows. Inserts show the preferential deuteration of the carbonyl oxygens. Experimental conditions: T = 80 K, MW power = ∼5 × 10⁻⁶ W, FM = ±30 kHz at a rate of 985 Hz. Number of scans: 8000 (upper spectrum) and 37,000 (lower spectrum). Scan time: 4 s.

Assignment of the ENDOR lines of the hydrogen bonds to a particular (O₁ or O₄) oxygen

In the previous section we refer to O_a and O_b as being associated with a given set of ENDOR lines. Our goal is to relate these results to the specific oxygen, O₁ and O₄ of the quinones (see Fig. 1). This, unfortunately, cannot be accomplished from the ENDOR experiments discussed above. We know, however, from the x-ray structure (17) that Ala (M260), i.e., O₁, is closer to the surface of the RC and is, therefore, expected to exchange its proton faster than the buried His (M219), i.e., O₄. Furthermore, EPR (8,10,30) and FTIR (31,32) experiments on ¹³C substituted quinones showed that the hydrogen bond to O₁ is weaker than to O₄ (reviewed in Lubitz and Feher (12). We, therefore, postulate that the ENDOR line L₁ and L₃ (Fig. 5) are associated with O₁ and L₂ and L₄ with O₄. This assignment will be further validated in a subsequent publication.

CONCLUSION

We have identified the ¹H and ²H-ENDOR lines with the hfcs of the specific protons hydrogen bonded to the two carbonyl oxygens on in reaction centers of Rb. sphaeroides. In a forthcoming study, these assignments will be used to determine the exact geometries of the hydrogen bonds.