Interactions between Cytochrome c₂ and the Photosynthetic Reaction Center from Rhodobacter sphaeroides : The Cation-pi Interaction

Abstract

The cation-pi interaction between positively charged and aromatic groups is a common feature of many proteins and protein complexes. The structure of the complex between cytochrome c₂ (cyt c₂) and photosynthetic reaction center (RC) from Rhodobacter sphaeroides exhibits a cation-pi complex formed between Arg-C32 on cyt c₂ and Tyr-M295 on the RC (Axelrod et. al (2002) J. Mol. Biol. 319, 501–515). The importance of the cation-pi interaction for binding and electron transfer was studied by mutating Tyr-M295 and Arg-C32. The first and second order rates for electron transfer were not affected by mutating Tyr-M295 to Ala indicating that the cation-pi complex does not greatly affect the association process or structure of the state active in electron transfer. The dissociation constant K_D showed a greater increase when Try-M295 was replaced by non-aromatic Ala (3-fold) than by aromatic Phe (1.2-fold) characteristic of a cation-pi interaction. Replacement of Arg-C32 by Ala increased K_D (80-fold) largely due to removal of electrostatic interactions with negatively charged residues on the RC. Replacement by Lys, increased K_D (6-fold) indicating that Lys does not form a cation-pi complex. This specificity for Arg may be due to a solvation effect. Double mutant analysis indicates interaction energy between Tyr-M295 and Arg-C32 of about −24 meV (−0.6 kcal/mole). This energy is surprisingly small considering the widespread occurrence of cation-pi complexes and may be due to the trade-off between the favorable cation-pi binding energy and the unfavorable desolvation energy needed to bury Arg-C32 in the short-range contact region between the two proteins.

Non-covalent molecular interactions play an important role in molecular association processes in biological systems. (1–3) Among these interactions are hydrogen bonds, hydrophobic interactions, and ion-paired salt bridges. (2,3). In addition to these, the cation-pi interaction, the short range electrostatic interaction between a positively charged cation and pi electrons in an aromatic group, (4) has been found to play a role in biological systems. The cation-pi interaction is important in ligand binding (5), in protein structures (6) and in protein-protein complexes (7). This work concerns a cation-pi interaction found in the interface between the cytochrome c₂ (cyt c₂) and reaction center (RC) of Rhodobacter sphaeroides that is formed between Arg-C32 on the cyt c₂ and Tyr-M295 on the RC (8). The importance of this cation –pi interaction on the binding and electron transfer reactions between the two proteins was studied by site directed mutation of both the Arg-C32 on the cyt c₂ and Tyr-M295 on the RC.

The RC (9,10) is a membrane bound pigment-protein complex in photosynthetic bacteria that performs the initial light-induced electron transfer reactions to convert sunlight into chemical energy (11). Light absorbed by the RC induces electron transfer from a special bacteriochlorophyll dimer, D, the primary donor through series of bound electron acceptors to a bound ubiquinone (Q_B). The photo-oxidized donor (D⁺) is reduced by electron transfer from a water soluble cyt c₂ allowing flow of electrons through a membrane associated electron transfer chain in a cycle that is coupled to proton pumping that drives ATP synthesis. Operation of the photocycle relies on efficient reactions between the mobile cyt c₂ and the membrane-bound RC. For efficient operation, cyt c₂ must associate with the RC, transfer electrons and dissociate within the time scale of electron turnover in the cycle (~ 10⁻³ s) (12).

The binding and electron transfer rates of isolated cyt c₂ and RC have been extensively studied using laser pulse kinetic measurements (13–18). The reduction of the oxidized donor D⁺ by reduced cyt c₂ shows two kinetic phases following a single laser flash – a fast (μs) first order phase (independent of cyt c₂ concentration) due to electron transfer from bound cyt c₂ to the photo-oxidized donor of the RC and a slower (ms) second order phase (dependent on cyt c₂ concentration) due to the binding and subsequent electron transfer of free cyt c₂. The observed biphasic kinetics can be explained by the following scheme (17).

$$\begin{array}{c}\text{cyt}{c_2}^{2+}+{\text{DQ}}_{\text{B}}\stackrel{K_{\text{D}}}{\leftrightarrow }\text{cyt}{c_2}^{2+}:{\text{DQ}}_{\text{B}}\\ \downarrow hv\\ \text{cyt}{c_2}^{2+}+{\text{D}}^{+}{{\text{Q}}_{\text{B}}}^{-}\underset{k_{\mathit{off}}}{\overset{k_{on}}{\rightleftarrows }}\text{cyt}{c_2}^{2+}:{\text{D}}^{+}{{\text{Q}}_{\text{B}}}^{-}\stackrel{k_{\text{e}}}{\to }\text{cyt}{c_2}^{3+}:{{\text{DQ}}_{\text{B}}}^{-}\end{array}$$ (1)

where K_D is the dissociation constant, k_on is the association rate constant, k_off is the dissociation rate constant and k_e is the electron transfer rate constant in the bound state. The equilibrium between bound and free cyt c₂ is achieved in the dark. Following a laser flash the re-reduction of D⁺ by cyt c₂²⁺ is biphasic. RCs with a bound cyt c₂ undergo rapid electron transfer with a rate constant k_e ( 10⁶ s⁻¹) (14). RCs without a bound cyt c₂ undergo slower diffusion limited electron transfer with an observed second order rate constant k₂ (~10⁹ s⁻¹M⁻¹) (18). Since k_e ≫ k_off (19), the observed second order rate constant is the association rate, k₂ ~ k_on. (20) The fraction of RCs with a bound cyt c₂ can be determined by the ratio of the fast and slow phases. The issociation constant K_D can be determined from a plot of the fraction of RCs with bound cyt c₂ versus the free cyt c₂ concentration. The importance of electrostatic interactions on binding and electron transfers between cyt c₂ and the RC has been established by the ionic strength dependence of k₂ (13), the effect of site directed mutation of charged residues (17,21), chemical cross-linking (22–24) and by electrostatic modeling (15,25,26). Hydrophobic interactions, particularly those of Tyr L162 on the RC have been shown to be important for binding and electron transfer. (18, 27).

The structure of the cyt c₂:RC complex was obtained by co-crystallizing the two proteins in a photochemically active complex and determining its x-ray crystal structure. (8) The structure of the complex shows the binding site to be between the cyt c₂ and the periplasmic surface of the RC with a central short range contact region in which the exposed heme edge is in contact with Tyr-L162, directly above the BChl dimer (Figure 1). The close contact between the two proteins provides an efficient tunneling pathway for fast electron transfer in the bound state. In the central region the residues from the two proteins make contact through van der Waals, and hydrogen bonding interactions as well as the cation-pi interaction (Figure 1). The closest distance between the positively charged guanidinium group of Arg-C32 and the phenolic group of Tyr-M295 is less than 4 Å, characteristic of a cation-pi complex (6). Surrounding the central region of close contact is a solvent separated region having positively charged residues on the cyt c₂ positioned opposite negatively charged residues on the RC interface between the two proteins. The charged residues do not form salt bridges but are separated by solvent.

Figure 1.

The cation-pi interaction in the RC: cyt c₂ complex (PDB accession code 1L9B, ref. 8). The RC subunits L, M, and H are shown in yellow, blue, and green, respectively. The primary donor (bacteriochlorophyll dimer) of the RC is shown in red. The cyt c₂ shown in lavender binds with its heme cofactor (red) just above the primary RC donor. Tyr L162, in contact with both cofactors is shown in green. Arg-C32 (blue) of cyt c₂ and Tyr-M295 (purple) of the RC form a cation-π interaction in the docked complex. This interaction is enlarged in the box to the right; the view was slightly changed to better emphasize the interaction.

In this work, the functional role of the cation-pi interaction in the binding, association and electron transfer between cyt c₂²⁺ and the RC is addressed in order to answer the following questions. (¹) What is the magnitude of this interaction? (²) Is this interaction important for the association of cyt c₂ and the RC? (³) Is it important for inter-protein electron transfer? To answer these questions site directed mutants of the RC at Tyr-M295 and cyt c₂ at Arg-C32 were constructed. Mutants were created that replaced the aromatic residue Tyr-M295 on the RC with either aromatic or non-aromatic residues, and replaced the cationic residue Arg-C32 on cyt c₂ with either the cationic Lys residue or neutral residues. The effect of these mutations on binding and electron transfer rates were measured by transient absorption spectroscopy using flash photolysis and compared to native values.

MATERIALS AND METHODS

Site Directed Mutagenesis

The site directed mutations on the RC were constructed as previously described using the QuickChange Mutagenesis kit (Stratagene) and a Perkin Elmer PCR System (28). The site directed mutation on the cyt c₂ gene were constructed as follows. First, cycA, the gene coding for the cyt c₂ protein, was transferred on a 2.7 kb Pst I fragment from pC2P404.1 (18) to pBCSK(+) (Stratagene, La Jolla, CA) creating pBCcycA. Mutations were introduced into the plasmid via the QuickChange mutagenesis kit as discussed above. The mutation was confirmed by DNA sequence (UCSD Molecular Pathology Shared Resource in the UCSD Cancer Center). The Pst I fragment carrying cyc A was transferred into pRK404 creating pRKcycA*. pRKcycA* was transformed into S17-1 and conjugated into CYCA1, the Rb. sphaeroides host lacking cyc A (²⁹), for transcription and translation of the modified gene sequence.

Protein Isolation and Purification

The bacteria harboring the modified RC or cyt c₂ gene was grown in the dark as described (28). RCs from Rb. sphaeroides carotenoidless strain R26 (native) and mutant strains were isolated in 15mM Tris-HCl pH 8, 0.025% lauryl dimethylamine-N-oxide (LDAO), 0.1mM EDTA following published procedures (28).. The final ratio of absorbance, A²⁸⁰/A⁸⁰⁰, was ≤ 1.5. The RC samples were then dialyzed for two days against HM (10mM Hepes pH 7.5, 0.04% dodecyl-β-D-maltoside (Anatrace)).

Native and mutant cyt c₂ proteins were isolated and purified as previously reported (30). Samples were purified to an absorbance ration A²⁸⁰/A⁴¹² ≤ 0.3. Mass spectroscopy was performed (UCSD Dept. of Chemistry, mass spectrometry facility) to verify the Arg-C32 →Gln and Arg-C32 →Lys substitutions in the mutant cyt c₂’s.

Quinone/Quinol Preparations

Quinone (Q₀) (2,3-dimethoxy-5-methylbenzoquinone) was obtained from Aldrich with ≥99% purity. Quinol (Q₀H₂) was synthesized by reducing quinone with hydrogen gas in the presence of platinum black also obtained from Aldrich.

Electron Transfer Measurements

Electron transfer kinetics between the RC and cyt c₂ were measured by flash absorption spectroscopy as described (31). Absorbance changes were measured at 865 nm or 595 nm following a laser flash from a Nd:YAG laser with an tunable optical parametric oscillator (Opotek, λ=800nm, τ=10ns). All measurements were conducted at 23° C in a buffer of 10 mM Hepes and 0.04% β-maltoside at pH 7.5. Samples included 50 μM Q₀ and Q₀H₂ to ensure that the cyt c₂ was reduced prior to each laser flash.

The fraction of RC with and without a bound cyt c₂²⁺ were determined from the relative amplitudes of the fast and slow phase of the D⁺ reduction by the cyt c₂²⁺. The concentration of free cyt c₂²⁺ was obtained by subtracting the concentration of bound cyt c₂²⁺ from the total added. The value of the dissociation constant, K_D was determined from the dependence of the fraction bound on the concentration of free cyt c₂²⁺. The change in free energy due to mutation of residue i was obtained from the relation as described in (18)

$$\mathrm{\Delta }\mathrm{\Delta }{G}_i=\text{RTln}\frac{K_{D,i}}{K_{D,0}}$$ (2)

where, K_D,i and K_D,0 are the dissociation constants for mutant and native complexes respectively. The second order rate constant, k₂, was determined from the concentration dependence of the measured rate constant for the slow phase of D⁺ re-reduction. When the concentration of cyt c₂ is large compared to the RC, the reaction becomes pseudo-first order in cyt c₂. To maximize the concentration range over which this condition holds, the RC concentration was kept low (0.2 μM). The first order electron transfer rate (k_e) constants was measured at 595 nm with a higher RC concentration (5 μM RC, 30 μM cyt c₂). With no cyt c₂ present some mutant RCs produced a small transient signal at this wavelength, which was subtracted from the observed kinetic trace.

Determination of the cation-pi interaction energy by double mutant analysis

The specific cation-pi interaction energy can be estimated by double mutant analysis of changes due to mutation of Arg-C32 and Tyr-M295. The relation of the changes in free energy due to mutation and the cation-pi interaction energy is illustrated by the double mutant cycle. (31–33). The interaction energy between two residues can be estimated by the interaction energy ΔΔG_int where

$$\mathrm{\Delta }\mathrm{\Delta }{G}_{\text{int}}=\mathrm{\Delta }\mathrm{\Delta }{G}_{YR}-\mathrm{\Delta }\mathrm{\Delta }{G}_Y-\mathrm{\Delta }\mathrm{\Delta }{G}_R=-\mathit{RT}ln\frac{K_{D,Y}{K}_{D,R}}{K_{D,0}{K}_{D,YR}}$$ (3)

ΔΔG_Y, ΔΔG_R, and are the changes in free energy obtained for the binding reaction of RCs modified at residue Tyr-M295 to native cyt c₂ and cyt c₂ modified at residue Arg-C32 with native RCs and ΔΔG_YR is the change in free energy for the binding between the modified RC and modified cyt c₂. K_D,Y, K_D,R and K_D,YR are the corresponding changes in dissociation constant. K_D,0 is the dissociation constant of native cyt c₂ and RC.

The change in free energy due to mutation of Tyr-M295 is assumed to be due to contributions from the cation pi interaction of-Tyr-M295 with Arg-C32, ΔΔG_cat-pi, and non-specific interactions with other residues ΔΔG_YX.

$$\mathrm{\Delta }\mathrm{\Delta }{G}_Y=\mathrm{\Delta }\mathrm{\Delta }{G}_{\mathit{cat}-pi}+\mathrm{\Delta }\mathrm{\Delta }{G}_{YX}$$ (4)

The major contributions to ΔΔG_cat-pi are the increase in free energy due to the loss of the cation-pi interaction and the decrease due to the loss of the solvation penalty; it will also include changes in other specific interactions between the two residues such as those due to van der Waals contacts. These cannot be separated in this analysis.

The change in binding free energy due to mutation of Arg-C32, ΔΔG_R, is assumed to have contributions from ΔΔG_cat-pi, between Arg-C32 and Tyr-M295 and interactions between Arg-C32 with other residues on the RC, ΔΔG_RX. In addition we specifically include an additional term ΔΔG_elect, which is due to electrostatic interaction of positively charged Arg-C32 with the negative potential of the RC surface that is lost due to mutation to a neutral residue X.

$$\mathrm{\Delta }\mathrm{\Delta }{G}_R=\mathrm{\Delta }\mathrm{\Delta }{G}_{\mathit{cat}-pi}+\mathrm{\Delta }\mathrm{\Delta }{G}_{\mathit{elect}}+\mathrm{\Delta }\mathrm{\Delta }{G}_{RX}$$ (5)

The change in free energy due to the mutation of the two interacting residues, ΔΔG_RY can be expressed as

$$\mathrm{\Delta }\mathrm{\Delta }{G}_{RY}=\mathrm{\Delta }\mathrm{\Delta }{G}_{\mathit{cat}-pi}+\mathrm{\Delta }\mathrm{\Delta }{G}_{\mathit{elect}}+\mathrm{\Delta }\mathrm{\Delta }{G}_{RX}+\mathrm{\Delta }\mathrm{\Delta }{G}_{YX}$$ (6)

where it is assumed that the interactions changed by single mutations are the same in the double mutation, i.e. the structure of the complex containing single mutations is the same (apart from the local changes to the mutated residue) as that containing both mutations. From Eqns. 3–6 the interaction free energy between the mutated residues is the negative of the free energy increase due to loss of the cation-pi interaction.

$$\mathrm{\Delta }\mathrm{\Delta }{G}_{\mathit{int}}=\mathrm{\Delta }\mathrm{\Delta }{G}_{\mathit{cat}-pi}$$ (7)

Thus, the cation-pi interaction energy can be obtained from double mutant analysis.

RESULTS

Dissociation Constants, K_D. K_D and ΔΔG values were obtained from the dependence of the fraction of RC with a bound cyt c₂, on the free concentration of cyt c₂. The binding curves are presented in Figures 2 and 3 and the results are summarized in Table 1.

Figure 2.

Binding data for the native cyt c₂ with the native and mutant RCs. Shown is a plot of f, the fraction of RC with a bound cyt c₂, versus the concentration of free cyt c₂²⁺. Binding of cyt c₂ to RCs with an aromatic residue at M295 are shown as solid symbols (native ■, Tyr →Phe ●, Tyr →Trp ▲). Binding of cyt c₂ to RCs without an aromatic residue are shown as open symbols (Tyr →Leu □, Tyr →Asn ○, Tyr →Ala △). The solid lines are the binding curves (Eqn. 3) using the averaged value of K_D for the aromatic (0.34 μM) and non-aromatic (0.86 μM) mutant RCs. Note that K_D for all of the aromatic residues at M295 are within error indistiguishable. Similarly, K_D for all of the non-aromatic substitutions are within error indistinguishable, but differ by ~3-fold from the native. (Conditions: 10 mM Hepes pH 7.5, 0.04% β-maltoside, 50 μM Q₀, 50 μM Q₀H₂).

Figure 3.

Binding data for mutant cyt c₂ that lack the cationic Arg-C32 with the native and mutant RC lacking the pi system at M295. Diamonds represent mutations of Arg-C32 to Lys, circles represent mutation of Arg-C32 to Gln. Solid symbols indicate the binding mutant cyt c₂ to the native RC. Open symbols indicate the binding data of the mutant cyt c₂ to the Tyr-M295 →Leu mutant RC. The binding curve for the native system is shown for reference. Note that the binding of the mutant cyt c₂ is essentially the same to either the native or mutant RC indicating that the mutations introduce little if any secondary changes that affect the binding. (Conditions as in Figure 2).

Table 1.

Binding and Electron Transfer Parameters for single and double mutant reactions of RCs modified at Tyr-M295 and cyt c₂ modified at Arg-C32 (pH = 7.5)^a

RC Strain	cyt c2	ke(× 106s−1)	k2 (× 108M−1s−1)	KD(μM)	ΔΔG (meV)
Native	Native	1.0	13.1	0.30	0
Tyr-M295 →Phe	Native	1.1	13.4	0.38	6
Tyr-M295 →Trp	Native	0.9	12.5	0.44	10
Tyr-M295 →Asn	Native	1.0	13.2	0.78	24
Tyr-M295 →Leu	Native	0.9	13.1	0.89	28
Tyr-M295 →Ala	Native	1.0	12.1	0.91	28
Native	Arg-C32→Lys	0.9	4.6	1.7	44
Native	Arg-C32→Gln	0.6	7.8	15	100
Native	Arg-C32→Ala	0.8	6.1	24	112
Tyr-M295 →Leu	Arg-C32→Lys	0.9	3.4	1.8	46
Tyr-M295 →Leu	Arg-C32→Gln	0.6	13.1	17	103
Tyr-M295 →Ala	Arg-C32→Ala	0.7	6.7	31	118

^ak_e is the rate constant for the fast, first order electron transfer. k_on is the second order rate constant. K_D is the dissociation constant for the cyt c₂:RC complex. ΔΔG = RTln(K_D,i/K_D,0) is the difference in binding free energy between the mutant and native RCs. The experimental precision for the measured values of the rates and K_D is ± 15%. The uncertainty in the free energies is ± 5meV

Reaction Center Mutations

Mutation of Tyr-M295 on the RC to aromatic residues Phe and Trp had small effects on K_D, (increases by factors of 1.3 and 1.5, respectively) indicating that these aromatic residues can also participate in a cation-pi interaction with Arg-C32. Mutation of Tyr-M295 to Asn, Leu and Ala produced larger changes in K_D (up to a factor of 3). The changes in free energy due to these mutations increased up to a maximum ΔΔG = 28 meV for mutations Tyr-M295 to Leu and Ala. The free energy changes can be attributed largely to the loss of the cation-pi interaction.

Cytochrome Mutations

Mutation of Arg-C32 to Lys, Gln and Ala produced increases in K_D by factors of 5.6, 50 and 80. These changes were larger than those produced by mutation of Tyr- M295. The larger changes free energy (100 and 112 meV) due to the mutation of the cationic Arg to neutral Gln and Ala result from the additional loss of stabilizing electrostatic interaction between Arg and the negative potential due to acidic residues on the RC surface (17).

Second Order Association Rate Constants, k₂

Second order rate constants were obtained from plots of the observed kinetic rate against the concentration of free cyt c₂ (17,31). Mutant RCs had the same second order rate constant within experimental error (k_on=12–13×10⁸M⁻¹s⁻¹) as the native RC for the native cyt c₂. The mutations to cyt c₂ (Arg-C32 →Lys, Arg-C32 →Glu and Arg-C32 →Ala) resulted in smaller values of k_on (5 ×10⁸M⁻¹s⁻¹, 8×10⁸M⁻¹s⁻¹ and 6×10⁸M⁻¹s⁻¹, respectively) that accompanied the increase in dissociation constant K_D.

First Order Electron Transfer Rate Constants, k_e

The first order electron transfer rate constant for the native RC: cyt c₂ system was found to be k_e=1 × 10⁶ s⁻¹, as previously reported (14–18). All mutant RCs bound with native cyt c₂ had the same first order rate constant (k_e = 0.9–1.1 × 10⁶ s⁻¹). Similarly, k_e = 0.9 × 10⁶ s⁻¹ for electron transfer between the bound Arg-C32 →Lys mutant cyt c₂ and the native RC. Small changes were observed in k_e for electron transfer between the native RC and the Arg-C32 →Gln cyt c₂ (k_e = 0.6 × 10⁶ s⁻¹) of the Arg-C32 →Ala cyt c₂ (k_e = 0.8 × 10⁶ s⁻¹).

DISCUSSION

Role of the Cation-pi Interaction in Binding

The mutation of either the cationic or the aromatic partner in the cation pi complex resulted in changes in the binding of the cyt c₂ to the RC. Replacement of Tyr-M295 with residues Ala, Asn or Leu lacking pi electrons had the effect of increasing K_D (by up to 3-fold for Ala) while replacement of Tyr-M295 with the other aromatic residues Phe and Trp had relatively small effect. These results are characteristic of the formation of a cation-pi complex at the cyt c₂:RC interface and indicate that an aromatic residue at position M295 assists the binding of the cyt c₂ to the RC. The removal of the cationic Arg-C32 residue and replacement with neutral residues Ala and Gln resulted in a larger increase in K_D (by up to 80-fold) corresponding to a change in binding free energy of ~110 meV. The larger effect upon mutation of the cationic residue can be attributed to the loss of the electrostatic interactions between the Arg-C32 and residues other than Tyr-M295. The largest interaction is likely to be the electrostatic interaction between Arg-C32 and Asp-L155 at a distance of 7Å away (6) The Coulomb interaction between the two charged residues would be ~100 meV, for a dielectric constant of 20.

The double mutant cycle provides a means to assess the magnitude of the cation-pi interaction energy between the two mutated residues. From the three double mutant cycles involving mutations in this study, (Arg-C32 → Lys/Tyr-M295 → Leu), (Arg-C32 → Gln/Tyr-M295 → Leu), and (Arg-C32 → Ala/Tyr-M295 → Ala), values for ΔΔGint = −26, −25 and −22 meV are obtained from Eqn. 3 giving an average of −24+2 meV. The values obtained for all three mutant cycles are in good agreement with each other. The cation pi interaction energy is close to the magnitude of the free energy change for mutation of Tyr-M295 to Ala indicating that the interactions of Tyr-M295 with residues besides Arg-C32 are small (ΔΔG_YX ~ 4 meV, Eqn.. 4)

The specific site-site interaction between Arg-C32 and Tyr-M295 has a relatively small net contribution to the binding free energy of the complex, ΔΔG_int = −24mev ( −0.6 kcal/mole ~ k_BT). The cation-pi interaction between Arg-C32 and Tyr-M295 in the cyt c₂:RC complex was identified by the CAPTURE program (6) (http://capture.caltech.edu/) that has been successfully used to identify cation pi interactions in many other protein structures. Gas phase estimates of the electrostatic and van der Waals contributions were computed to be ~ −4 kcal/mole (−170 meV) and ~ −2 kcal/mole (−85 meV), respectively (average of 3 structures in 2 crystal forms (pdb id.1L9B and 1L9J). Note that these values are much larger than the smaller observed net effect which includes the competing energy cost required to desolvate the cation. Since the guanidinium group is positively charged, it interacts with water dipoles and forms hydrogen bonds. In forming the cyt c₂:RC complex Arg-C32 becomes largely inaccessible to solvent. A solvent accessibility for Arg-C32 side chain changes from 55% to 3% in forming the cyt c₂:RC complex (determined using the CNS program (34).) Thus, a significant energy is required to desolvate the cationic guanidinium group to bring it into close contact with the Tyr.

Although Lys is capable of participating in a cation pi interaction, the change in the interaction energy between Lys-C32 and Tyr-M295 (ΔΔG_int =−26 meV from the double mutant analysis) suggests that Lys-C32 does not form a cation-pi complex. The larger change in free energy ΔΔG_R of 44 meV for mutation of the charged Arg-C32 to Lys than those found for the mutations the aromatic Tyr-M295 to either Leu or Ala could be due to changes in the electrostatic interaction due to the Arg → Lys mutation, i.e. ΔΔG_elect in Eqn. 5 is not zero. This change could result either form a difference in the Coulomb interaction, due to a change in the position of the charge on the potential surface or due to a differences in the solvation of Lys and Arg.

The inability of Lys to form a cation-pi complex may be because the desolvation energy cost is larger due to the higher charge density for the ammonium cation than the guanidinium cation that has charge delocalized over 3 N atoms. The delocalizalized charge lowers the desolvation energy compared to Lys making the formation of the cation-pi complex with Arg more favorable (34). In addition, steric effects may also contribute. The aromatic ring of Tyr-M295 makes the closest contact with the ε-nitrogen atom of Arg-C32. The ζ-nitrogen atom of Lys-C32 is one bond length longer and may not be able to form a strong cation-pi bond. Crowley and Golovin (7) have recently examined a structural data base for protein structures having cation-pi interactions. They found that the most common cation-pi interaction in protein complexes was between Arg and Tyr with interactions between Lys and Tyr much less frequent. It is interesting to note that Arg-C32 is the only Arg residue in the interface region, compared to 5 Lys residues. This suggests that Arg-C32 is selected for its ability to form a cation-pi complex in close association with the RC surface.

The Role of the Cation-pi interaction in electron transfer

The first order electron transfer rate constants k_e for all mutated RCs in this study were very similar to that of native RCs (k_e = 1.0 × 10⁶ s⁻¹) (Table 1). Since the first order rate is due to electron transfer from cyt c₂ in its bound state on the RC surface, we conclude that the positioning of the cyt c₂ on the RC surface is the same with or without the cation-pi interaction. Small changes in rate were observed for reaction of the mutant cyt c₂ (k_e = 0.6–0.9 × 10⁶ s⁻¹), which could be due to small changes in the redox potential of the mutant cyt c₂, the reorganization energy for electron transfer or in the distance (<1Å) between the cofactors in the bound state (36–38).

These results can be rationalized with results from previous experiments using a model in which the electron transfer rate k_e is determined by the close juxtaposition of the exposed heme edge with the short-range contact region on the RC centered around Tyr-L162 which is in close contact with the BChl₂ (D) (8). Mutation of hydrophobic residues (particularly Tyr-L162) changed the electron transfer rate k_e presumably due to changes in the distance between cofactors in the structure of the modified cyt c₂:RC complex (18). In contrast, mutations of charged residues on the RC, located outside of the short-range contact region, did not change the electron transfer rate k_e, although they modified the binding affinity (17). The modification of cation-pi contacts appears to have effects on k_e similar to the charged mutations even though the cation-pi complex is part of the short range interaction region. This may be because the cation-pi complex is at the boundary of the short-range contact region. Thus, changes in the cation-pi bonding do not affect the structure in the central region around Tyr-L162 which largely determines the rate of electron transfer (39,40).

The Role of the Cation-pi interaction in cyt c₂ association/dissociation

The measured value for the second order association rate constant was the same for all of the mutated RCs used in this study (k₂ =k_on = 12–13 × 10⁸ M⁻¹ s⁻¹). Thus, the cation-pi interaction does not influence the rate of association. This is reasonable since the cation-pi interaction is a short-range interaction between a charged residue and the quadrupole moment of a neutral residue. Thus, the formation of the cation-pi complex is not expected to be important for the formation of the transition state for the association process which is located ~ 8–10 Å above the RC surface (17,25,26). This is consistent with the finding that mutations of hydrophobic residues on the RC do not appreciably affect k_on (18).

Although the cation-pi interaction does not affect the association rate, it does affect the dissociation rate, k_off. Since the loss of the cation– pi results in a 2– 3– fold increase in K_D (Table 1) but no change in k_on, the loss of the cation-pi interaction results in a 2– 3– fold increase in k_off. (K_D = k_off/k_on ). These changes reflect the slightly tighter binding of cyt c₂ to the RC and an increase in the barrier for dissociation in the presence of the interaction. Thus, the cation-pi interaction increases the residence time by ~ 3– fold of the cyt c₂ on the RC surface. If the binding were made substantially tighter, k_off would be correspondingly smaller which would substantially reduce the turnover rate of photosynthetic electron transfer. For example, an increase of 60meV in the cation-pi interaction would result in a 10-fold decrease in k_off and an ~10-fold decrease in the overall turnover rate of electron transfer (τ=25 ms). Thus, a strongly binding cation-pi interaction would have deleterious effect on the photosynthetic efficiency.

Mutations to the charged residue Arg-C32 resulted in decreased association rates (k₂ = 5–8 × 10⁸ M⁻¹s⁻¹). These changes can be attributed to changes in electrostatic interactions rather than cation-pi interactions. The changes in k₂ are consistent with the changes of k₂ due to mutation of charged residues in the interface (17,21) It is interesting to note that the Arg-C32 → Lys mutation had the largest change on k₂ suggesting that the configuration of the two residues is significantly different.

Importance of the Arg-C32 - Tyr-M295 interaction

The magnitude for the cation-pi interaction on increasing the binding in the cyt c₂:RC complex appears to be surprisingly small considering that such complexes have been found in many other systems (7) including the complex between cyt c and cyt bc1 from yeast (41). Thus, the cation-pi interaction in the cyt c₂:RC complex is just one of many interactions (electrostatic, hydrophobic, hydrogen bonding) that optimize the protein-protein interface, rather than a dominant component of the binding energy. The widespread occurrence of cation-pi complexes may be due to the importance of electrostatic interactions (and thus positively charged residues) in the protein association process (7). The cationic Arg-C32 is one of several positively charged residues on the cyt c₂ surface that serve to attract and orient the cyt c₂ on the negatively charged binding surface of the RC form a loosely bound encounter complex. Further short range docking of the cyt c₂ leads through a transition state to the bound state in which electron transfer occurs (26). In the bound state the Arg-C32 is buried in the central short range contact region of the interface where solvent is excluded. In this solvent excluded region the two interfaces are in close contact and optimized for electron transfer. In contrast, the other charged residues in the interface are separated by solvent. (8) The formation of a cation-pi complex provides a means to offset the cost of desolvating the charge (42), thereby providing a means to bury a positively charged group within the solvent excluded short-range interaction domain. Although the overall cation-pi interaction does not greatly increase the binding affinity, the binding is well optimized by other interactions so that a further increase in affinity is not necessary and might slow down the dissociation of the oxidized cyt c₂ and thus reduce the overall turnover of the reaction.

Abbreviations

cyt c₂

cytochrome c₂

RC

reaction center

bc₁

cytochrome bc1

Q₁₀

ubiquinol-10

D

primary electron donor (bacteriochlorophyll dimer)

D⁺

photo-oxidized primary electron donor

Q_B

bound acceptor quinone

k_e

first order electron transfer rate

k₂

observed second order rate constant

K_D

dissociation constant

Q₀

2,3-dimethoxy-5-methyl-1,4-benzoquinone