Origin of Circular Dichroism of Xanthorhodopsin. A Study with Artificial

Dec 12, 2014 - Phone +972 8 9344320. ... The CD spectrum of xR is very unique with a “conservative” character, containing negative and positive lo...
1 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/JPCB

Origin of Circular Dichroism of Xanthorhodopsin. A Study with Artificial Pigments Elena Smolensky Koganov,† Vlad Brumfeld,‡ Noga Friedman,† and Mordechai Sheves*,† †

Department of Organic Chemistry, Weizmann Institute of Science, Rehovot 76100, Israel Department of Chemical Research Support, Weizmann Institute of Science, Rehovot 76100, Israel

‡

S Supporting Information *

ABSTRACT: Xanthorhodopsin (xR) is a retinal protein that contains, in addition to the retinal moiety, a salinixanthin chromophore absorbing at 456, 486, and 520 nm [Balashov, S. P.; et al. Science 2005, 309, 2061]. The CD spectrum of xR is very unique with a “conservative” character, containing negative and positive lobes and resembling the first derivative of the absorption spectrum [Balashov, S. P.; et al. Biochemistry 2006, 45, 10998]. It was suggested that the CD spectrum is likely to be composed of several components and that the salinixanthin interacts closely with the retinal chromophore [Balashov, S. P.; et al. Biochemistry 2006, 45, 10998; Imasheva, E. S.; et al. Photochem. Photobiol. 2008, 84, 977; Lanyi, J. K.; et al. Acta Bioenerg. 2008, 1777, 684; Smolensky, E.; et al. Biochemistry 2009, 48, 8179; Smolensky Koganov, E.; et al. Biochemistry 2013, 52, 1290]. In this work, we aim to further explore the nature and origin of the unique CD spectrum of xR. We follow the absorption and CD spectra at different pHs of wild-type (wt) xR and of artificial xR pigments, characterized by a shifted absorption maximum of the retinal chromophore, as well as their corresponding reduced retinal protonated Schiff base pigments. Our results revealed a protein residue (other than the protonated Schiff base counterion), for which protonation affects the CD spectrum by decreasing the negative lobe at ∼530 nm and the positive lobes at 478 and 455 nm, which might be due to elimination of excitonic coupling between the salinixanthin chromophores, although other possibilities cannot be completely excluded. This spectrum change occurs by the pH decreasing, even in artificial pigment where the absorption of the retinal pigment is significantly shifted from 570 to about 450 nm. The possible excitonic coupling between the salinixanthin chromophores and its contribution to the CD spectrum of xR were supported by a good fitting of the CD spectrum to conservative (excitonic) bands [Zsila, F.; et al. Tetrahedron: Asymmetry 2001, 12, 3125; Zsila, F.; et al. Tetrahedron: Asymmetry 2002, 13, 273]. We propose that the CD spectrum of xR consists of contributions from an excitonic coupling interaction between the salinixanthins chromophores located in different subunits of the 3D structure of xR, the chiral conformation of the salinixanthin within its binding site, and the contribution of the retinal chromophore to the negative lobe at around 550 nm.

■

INTRODUCTION Xanthorhodopsin (xR)1 is a retinal-based proton pump in the cell membranes of the extremely halophilic eubacterium Salinibacter ruber.7 In addition to the covalently bound alltrans retinal, through a protonated Schiff base (PSB) to Lys240, it contains a salinixanthin chromophore. This chromophore is a C40 carotenoid characterized by a glycoside residue and an acyl tail3 as well as a conjugated chain that consists of 11 double bonds and is attached to a ring bearing one double bond and a keto group in the C4 position. The circular dichroism (CD) spectrum of the xR pigment is remarkably different from its apo-xR, which has a very weak CD spectrum similar to free salinixanthin in EtOH solution.1−3 However, following binding of the retinal chromophore to apoxR, it exhibits sharp unique CD lobes at 530 (−), 513 (+), 478 (+), and 455 nm (+) with a bilobe “conservative” character that resembles the first derivative of the absorption spectrum.2 It © 2014 American Chemical Society

was proposed that the structure of the CD spectrum originates from the bands of the bound salinixanthin, from its asymmetric conformation, and from its steric and electronic interaction with the retinal.2,8,9 When the first derivative of the absorption spectrum is subtracted from the CD spectrum, the remaining spectrum has an overall bilobe character, and it was suggested that it may have originated from the dipole interaction between the salinixanthin at 450−500 nm (positive lobe) and from the retinal at 550 nm (negative lobe).2,9 Previous spectroscopic results10 as well the crystallographic studies11 indicated that the rings of these two chromophores are aligned in close proximity to each other, and salinixanthin lies transverse against the outer surface of helix F at a 54° angle Received: October 20, 2014 Revised: December 12, 2014 Published: December 12, 2014 456

DOI: 10.1021/jp510534s J. Phys. Chem. B 2015, 119, 456−464

Article

The Journal of Physical Chemistry B

hydroxylamine by a dialysis procedure versus DDW and were stored at 4 °C to avoid reconstitution with retinal originating from retinal oxime. The retinal analogues were synthesized as described elsewhere.8,18,19 Artificial pigments of xR were prepared by incubating the apo-protein (0.2−0.4 OD according to the absorption at 486 nm) with 2 equiv of the synthetic retinal analogue, 50 mM Tris buffer at pH = 8.3, 30% sucrose, and 0.5 M NaCl at 20 °C overnight. The predicted amount of pigment was estimated based on a bleach process. UV−Vis Absorption Measurements. All of the UV− visible measurements were conducted in the Agilent 4583 diode array spectrophotometer (Agilent Technologies, Palo Alto, CA) equipped with an Agilent 89090A thermostated cuvette holder (Agilent Technologies, Palo Alto, CA). Absorption spectra were corrected for light scattering. CD Measurements. CD spectra were recorded on Chirascan CD spectrometer of Applied Photophysics. The CD spectra are given in ellipticity θ, which is proportional to the difference in absorbance of left and right circularly polarized light [θ = 3300° (AL − AR)20]. A quartz 1 × 1 cm path length cuvette was used. The CD spectra were recorded with a 1 nm bandwidth resolution and in 1 nm steps at 20 °C. The CD spectra were corrected for the baseline distortion by subtracting reference spectra of the corresponding buffer. Reduction of the PSB Bond with NaBH4. Sodium borohydride was added to native xR and to the artificial pigments (in 30% sucrose, 50 mM Tris buffer at pH = 8.3, and 0.5 M NaCl) to a 0.12 M final concentration. The reduction process was carried out under illumination with a halogen lamp (12 V, 150 W) for 1.5 h, and the light was filtered through a long-pass cutoff filter of λmax > 550 nm (Schott, Mainz, Germany). After the reduction process was completed, the pH of the sample was 9.8. To remove all of the traces of sodium borohydride, the sample was dialyzed against 30% sucrose, 50 mM Tris buffer pH = 8.3, and 0.5 M NaCl. Titration Experiments. Titrations were performed in daylight by adding small amounts of HCl (typically 2 μL of 0.25 to 3 M) to the pigments as well as to the reduced pigments (0.25−0.3 OD at 486 nm, 1 mL, in 30% sucrose, 50 mM Tris buffer pH = 8.3, and 0.5 M NaCl), followed by measuring CD and absorbance spectra, after the pH was adjusted. The titration process was confirmed by reversibility of the process following a rapid pH increase. The values in the difference spectra of the CD at 530 nm and of the absorbance at 600 nm were plotted against pH, and the data were fitted to a modified Henderson−Hasselbalch equation

to the membrane, buried at the protein−lipid boundary, and its 4-keto ring is fixed out of the conjugated π-system plane by residues of helices E and F and by the β-ionone ring of the retinal,11 as deduced from a slight blue-shifted absorption maximum upon formation of the xR complex.1,2,11 This asymmetric conformation probably contributes to the chirality and the optical activity detected for the fixed salinixanthin,2 as for other carotenoids in photosynthetic membranes.12 It was later shown that the 4-keto group plays a key role in the salinixanthin binding.9 Binding experiments of synthetic retinal analogues with shifted absorption maxima to the apomembrane of xR8 showed that the CD spectrum of xR can be attributed mainly to the salinixanthin chromophore and that the absorption as well as the CD spectra are formed prior to formation of the PSB linkage.4,8 Moreover, we have shown that modifications in the region of the β-ionone ring significantly affected the rate of the CD spectrum formation.10 Imasheva et al.13 showed that the sharp intense bands in the CD spectrum of the native xR are eliminated following oxidation of salinixanthin with ammonium persulfate. It was further suggested that the remaining low-intensity broad band in the visible range detected in the bleached sample originates from the retinal chromophore due to its asymmetric conformation and/or an asymmetric environment in its binding site.13 Previous titration experiments of xR2,14 showed that lowering the pH from 8.5 to 4.5 causes a minor red shift in the absorption spectrum of xR due to the protonation of the retinal PSB counterion (having an apparent pKa of 6).14 In contrast, lowering the pH induces significant changes in the CD spectrum (having an apparent pKa of 6).2 Therefore, it was proposed that the counterion protonation state is a major factor affecting the CD spectrum and determines the interaction of the retinal chromophore with the salinixanthin antenna.2 In this work, we aimed at shedding further light on retinal− salinixanthin interactions and on the origin of the xR CD spectrum.2,8 It is proposed that the salinixanthins chromophores in the protein subunits interact with each other via exciton coupling interaction and affect the CD spectrum. Furthermore, despite the fact that the retinal and the salinixanthin are situated in close proximity11 and exhibit energy transfer with a quantum efficiency of ∼40% from salinixanthin to the retinal,1,15,16 we were not able clearly detect exciton coupling between the two chromophores.

■

EXPERIMENTAL SECTION Sample Preparation. Growth of Salinibacter ruber was carried out using slightly modified published methods.1,7 The 0.1% sucrose (per liter) was added to the growing medium according to ref 17. The xR membranes samples were prepared using published methods,1 without using 0.15% dodecylmaltoside. The membranes were washed with 0.1 M NaCl, followed by three times washing with DDW (double distilled water). This treatment partially removed unbound salinixanthin. The process yielded a sample in which the ratio between the absorption at 280 and 568 nm was ∼3. The apo-protein was prepared by incubating 0.2 OD of the pigment (at 570 nm) with 0.2 M (0.05−0.1 M for artificial pigments) freshly prepared hydroxylamine at pH = 7.2 and irradiated for 1.5 h with a Schott 250W cold light source (Carl Zeiss Microscopy, Jena, Germany) equipped with a heatabsorbing filter and an optic fiber (level 4B). The light was filtered through a long-pass cutoff filter of λ > 550 nm (Schott, Mainz, Germany). The samples were thoroughly washed from

F (x ) =

1 1 + 10n(pKa − x)

where n is the number of protons participating in the transition, x is the pH, and pKa is the midpoint of the observed transition. Fitting the CD Spectrum to Conservative (Excitonic) Bands. Conservative (excitonic) CD bands were built as a sum of two opposite signed Gaussian peaks. We have initially positioned the bands in a way that the intercept (Δε = 0) of the excitonic band should be at the same wavelength as the absorption peak in the UV−vis spectrum. Then, we fitted the spectrum using a nonlinear least-squares algorithm (Peakfit software, Jandel Ltd.). After the optimization, the experimental spectrum was reproduced with r2 of 0.92 at least. 457

DOI: 10.1021/jp510534s J. Phys. Chem. B 2015, 119, 456−464

Article

The Journal of Physical Chemistry B

Figure 1. Spectra of pH-induced absorbance and CD spectroscopy changes of membrane suspension of xR during the titration in 0.05 M phosphate. (A) CD changes during the titration. The pH range is 9.5−7. (B) Difference of the CD spectra during the titration. The arrows indicate the direction of the titration. The spectrum at the lowest pH was subtracted from all of the spectra in the titration range. (C) Difference absorption spectrum. The spectrum at pH = 7.4 was subtracted from the spectrum at pH = 4.9. (D) pH dependence of the band at 526 nm in the CD spectra and at 598 nm in the absorption spectra. Absolute normalized maxima of the |ΔCD| and |ΔOD| bands are plotted as a function of pH. The pKa values were extracted using the equation F(x) = 1/(1 + 10(n*(pKa−x))). The solid line represents the best fit for the CD spectra with a pKa = 8 (red squares) and pKa = 6.1 (blue circles) for the absorption spectra.

■

RESULTS We have performed titration experiments of wild-type (wt) xR as well as of two artificial pigments, ph-xR and 3,4-deH-xR, in which the absorption maxima of the retinal pigments are shifted relative to wt xR.8 The titrations were monitored using CD and absorbance spectroscopies. Further, titration experiments were carried out on pigments in which their PSB bond was reduced by NaBH4.21,22 Titrations of wt xR. As previously shown,2,14 titration of xR (0.1 M NaCl) monitored by CD and absorption spectroscopies gave similar pKa values in both methods. However, once the titration was carried out in 0.05 M phosphate buffer (Figure 1), the pKa curves of the titrations monitored by the CD and absorbance (Figure 1D) revealed two different pKa values, pKa = 8 by CD and pKa = 6.1 by absorbance spectroscopy (Figure 1D and Table 1). The last CD spectrum of the titration (lowest pH) was subtracted from spectra at various pHs. The obtained difference spectra (Figure 1B), with intersection (Δε = 0) at 486 nm matching the absorption maximum of the salinixanthin

band, may reflect an exciton coupling interaction between salinixanthin chromophores, which is diminished at low pH. We note that other possibilities to account for the CD change cannot be excluded, as will be discussed in the Discussion section. Similar pKa values were obtained with xR in 0.05 M Tris buffer (Figure S1 (Supporting Information) and Table 1). These results indicate that the changes monitored by the CD and absorbance spectroscopies are associated with titrations of two different protein residues. The counterion complex (Asp96 and His-62)11 whose titration was observed by absorbance spectroscopy, as suggested previously, while the protonation state of another protein residue affects the CD spectrum. Titrations of xR in the presence of high salt concentration (50 mM Tris buffer pH = 8.3, 30% sucrose, and 0.5 M NaCl, the medium in which the protein is most stable) shifted the pKas to lower values probably due to alterations of surface potential, 23−25 and both CD (Figure S2, Supporting Information) and absorption (Figure S3, Supporting Information) spectroscopies indicated similar pKa values of 5.3 and 5, respectively (Figure S2C and Table S1, Supporting Information). The larger shift monitored by the CD spectroscopy (2.7 pKa units) versus that monitored by absorption (1.1 pKa units) indicates that the two protein residues are affected differently by the environment and salt concentration. At high salt concentration, the measured pKa reflects the intrinsic pKa of the groups and indicates similar pKa for the two groups. However, the pKa values of the two protein residues are well-separated at

Table 1. Summary of the Obtained pKa Values for xR in 0.05 M Tris and Phosphate pigment

pKa CD

pKa absorption

wt xR in 0.05 M phosphate wt xR in 0.05 M Tris

8 8.1

6.1 7 458

DOI: 10.1021/jp510534s J. Phys. Chem. B 2015, 119, 456−464

Article

The Journal of Physical Chemistry B

A decrease in the salinixanthin absorption band by ∼30% was detected during the reduction process (Figure 2A), which may account for the intensity decrease by ∼40% in the CD spectrum. Partial denaturation of the protein during the reduction of the PSB was excluded because there was no change in the CD spectrum at 220 nm during the reduction (data not shown). Thus, it is plausible that the decrease in the salinixanthin absorption may be due to protein−salinixanthin interaction alterations that lead to a decrease in the salinixanthin absorption intensity. Titration of reduced xR was followed by the CD spectroscopy (Figure 3), and the obtained pKa value was 4.7 (Figure 3C). Following the titration, the negative lobe lost most of its intensity, became much wider, and was red shifted toward 540 nm. The positive lobe at low pH resembles the absorption spectrum of xR (Figure 3A). The difference spectra in Figure 3B may reflect an excitonic coupling, with intersection (Δε = 0) at 486 nm, as was detected in the titration of wt xR. Thus, we suggest that the titration process is associated with a protein residue that may alter the protein 3D structure organization. Consequently, it is possible that excitonic coupling between salinixanthins located in different subunits of the 3D structure is diminished; therefore, at low pH, the residual CD spectrum originates from the chiral conformation of salinixanthin. Phenyl-xR (ph-xR) Titration. The ph-xR, was produced by binding phenyl retinal to the apo-membrane of xR.8 The formed pigment is characterized by a blue-shifted absorption (at around 450 nm),8,10 overlapping the absorption of the salinixanthin and therefore preventing its accurate detection. In addition, there is no contribution of the retinal chromophore analogue in the 530 nm region of the CD spectrum. Titration of the ph-xR was monitored by CD spectra (Figure 4) and gave a pKa value of 5.5 (Figure 4C). The corresponding difference spectra (Figure 4B) indicated that the changes during the titration resembled the changes detected during the titration of wt xR, although the retinyl analogue has a different absorption maximum. This observation indicates that the change in the CD spectrum because of pH decrease is not due to retinal−salinixanthin interaction but rather may originate from cancellation of excitonic coupling between salinixanthins chromophores. Reduction of ph-xR (Figure S4, Supporting Information) induced changes in the CD spectrum similar to those detected after the reduction of wt xR. As the retinal pigment does not contribute to the negative band at 530 nm, in contrast to wt xR reduction, there was no change in the bandwidth or maximum of the negative lobe (Figures 2 and S4 (Supporting Information)). This observation further supports the assumption that loss of the amplitude does not originate from abolishing retinal−salinixanthin interaction and may be due to a salinixanthin extinction coefficient decrease. The reduced ph-xR species was titrated (Figure S5A, Supporting Information) and yielded a pKa of 4.7 similar to reduced wt xR (Table S1, Supporting Information). The difference spectra of the CD titration (Figure S5B, Supporting Information) with a crossing point (Δε = 0) at 486 nm may suggest elimination of excitonic coupling between salinixanthins. At low pH, a small negative lobe remains at around 530 nm (Figure S5A, Supporting Information), similarly to the reduced wt xR. In both cases, as there is no contribution of the retinal chromophore to this band, we attribute the remaining lobe to the salinixanthin transition detected at 545 nm in the absorption difference spectrum.8

lower salt concentration in which the surface potential plays a major role. This observation indicates that the two groups are affected differently by the surface potential. Clarification of this effect needs further studies. Significant alterations in the CD spectra were detected at acidic pH, namely, much lower amplitude of the negative lobe and its red shift toward 536 nm (a shift of ca. 8 nm), which may suggest that the negative lobe originated from two contributions, and by decreasing the pH, one of the contributions disappeared. In order to investigate the contribution of the retinal chromophore to the CD spectrum, the PSB bond in xR was reduced with NaBH4 (Figure 2), leading to a blue shift of the

Figure 2. Reduction of xR retinal PSB with NaBH4, in 30% sucrose, 0.5 M NaCl, and 50 mM Tris buffer at pH = 8.3. (A) UV−vis absorption alterations caused during the reduction of xR. 1, Initial absorption spectrum of xR, before adding NaBH4; 2, after completion of the reduction reaction with NaBH4 and dialysis. (B) CD spectra changes caused following the reduction of xR. Red, initial CD spectrum of xR, before adding NaBH4; blue, after completion of the reduction reaction with NaBH4 and dialysis, pH = 8.3; Green, difference spectrum between spectra 1 and 2.

retinal chromophore absorption from 568 to 360 nm (as evident by disappearance of the absorption shoulder at ∼570 nm; Figure 2A), therefore minimizing possible retinal− salinixanthin interactions.15,22 The reduction process was followed by the CD spectroscopy, indicating a decrease in the amplitude of the positive lobes, a disappearance of a shoulder at around 540 nm (from the negative lobe), as well as formation of a new positive lobe at around 405 nm, suggesting that the reduced retinal polyene exhibits chirality. 459

DOI: 10.1021/jp510534s J. Phys. Chem. B 2015, 119, 456−464

Article

The Journal of Physical Chemistry B

Figure 3. Spectra of pH-induced CD changes of the membrane suspension of reduced xR following the titration in 30% sucrose, 0.5 M NaCl, and 50 mM Tris buffer at pH = 8.3. (A) CD spectra changes following the titration. The pHs range from 8.3 to 3.8. (B) Difference of the CD spectra following the titration. The spectrum at the lowest pH was subtracted from all of the spectra in the titration range. (C) pH dependence of the band at 530 nm in the CD spectra. Absolute normalized maxima of the |ΔCD| bands are plotted as a function of pH. Extraction of the pKa value in the range of pHs of 3.8−8.3, using the equation F(x) = 1/(1 + 10(n*(pKa−x))). The solid line represents the best fit with a pKa = 4.7.

Figure 4. Spectra of pH-induced CD changes of the membrane suspension of ph-xR during the titration in 30% sucrose, 0.5 M NaCl, and 50 mM Tris buffer at pH = 8.3. (A) Spectra of CD changes following the titration. The pHs range from 6.7 to 3.6. (B) Difference of the CD spectra during the titration. The spectrum at the lowest pH was subtracted from all of the spectra in the titration range. (C) pH dependence of the band at 530 nm in the CD spectra. Absolute normalized maxima of the |ΔCD| bands are plotted as a function of pH. Extraction of the pKa value in the range of pHs of 3.6−6.7, using the equation F(x) = 1/(1 + 10(n*(pKa−x))). The solid line represents the best fit with a pKa = 5.5.

Titration of 3,4-deH-xR. The absorption maximum of 3,4deH-xR is red shifted by 20 nm to 590 nm.8 Titration experiments of the pigment monitored by absorption and CD spectroscopies indicated similar pKa values of 6.7 and 6.6, respectively (Figures 5 and S6 and Table S1 (Supporting Information)). These pKa values are much higher compared to those of wt xR and ph-xR, suggesting a change in the protein conformation that might affect the pKa values of the protein

residues due to the change of the local environment. The additional ring double bond affects the ring conformation, which may alter its interaction with the 4-keto ring of salinixanthin, thereby altering the protein conformation. The 460

DOI: 10.1021/jp510534s J. Phys. Chem. B 2015, 119, 456−464

Article

The Journal of Physical Chemistry B

The PSB bond of 3,4-deH-xR was reduced with NaBH4, and the process was monitored by CD spectroscopy. The amplitude of the negative lobe at 535 nm was decreased by approximately 50%, and its width was reduced as well (Figure S7, Supporting Information), suggesting that prior to the reduction reaction, the negative lobe in 3,4-deH-xR is composed of two bands at about 530 and 560 nm. Our results show that there is a correlation between the width of the negative lobe of the CD spectrum and the absorption of the pigment, namely, ph-xR has the most narrow lobe (pigment at around λmax = 450 nm), wt xR has a wider one (λmax = 568 nm), and 3,4-deH xR has the widest (λmax = 590 nm). It supports the assumption that the retinal chromophore contributes to the negative lobe, affecting mostly its width. The contribution of the retinal is about ∼30% of the band intensity.2,8 Titration of reduced 3,4-deH-xR (Figure S8A, Supporting Information) was followed by CD spectroscopy and revealed a pKa value of 6.2 (Table S1, Supporting Information), higher than the pKa of reduced wt xR. The CD difference spectra (Figure S8B, Supporting Information) of the titration show weakening of both negative and positive lobes with a pKa of 6.6 (Figure S8C, Supporting Information), and Δε = 0 at 486 nm. The final CD spectrum, at pH = 4.6, lacks completely the negative band and exhibits high similarity to the absorbance spectrum (Figure S8A, Supporting Information), unlike the CD spectra of reduced wt xR and reduced ph-xR at acidic pH, which still presents a weak negative band. Fitting of the Four Excitonic Bands of the CD Spectrum. The vibrational fine structure of the CD bands indicates that the vibronic components split individually.5,6 We have tried to fit the CD spectrum to conservative (excitonic) bands and examine the possibility that the spectrum of wt xR is partially due to excitonic coupling. For this purpose, we have built conservative (excitonic) CD bands that are the sum of two opposite signed Gaussian peaks on the measured CD spectra of the titration at pH = 7.6 and 4 and their difference spectrum. We have initially positioned the bands in a way that the intercept (Δε = 0) of the excitonic band will be at the same wavelength as the λmax of the absorption peak. Then, we fitted the spectrum using a nonlinear least-squares algorithm. The final width, position, and amplitude of the conservative and Gaussian bands are summarized in Tables S2 and S3 (Supporting Information). The result shown in Figure 6A suggests that the CD spectrum of wt xR at pH = 7.6 may involve excitonic coupling as we found a good fitting to two conservative bands (Figure 6A and Table S2, Supporting Information) and two gaussians. Next, we applied the same analysis to the CD spectrum of xR at pH = 4, the last spectrum of the titration. However, including the four conservative spectra in the deconvolution of the spectra taken at a pH of 4 resulted in a lack of convergence. It was best fitted to seven gaussians (five positive and two negative, R2 = 0.99; Figure 6B) and not to conservative bands. We suggest that the positive lobes originate from the chiral conformation of the fixed salinixanthin, and the negative lobe is composed of two components, chiral salinixanthin (there is a vibrational transition of the salinixanthin at around 545 nm8) and the chiral conformation of the retinal. The same analysis was performed on the difference CD spectrum of xR titration, which could be fitted to four conservative bands (Figure 6C and Table S3 (Supporting Information)), implying that an excitonic coupling is weakened during the titration. These

Figure 5. Spectra of pH-induced CD changes of the membrane suspension of 3,4-deH-xR during the titration in 30% sucrose, 0.5 M NaCl, and 50 mM Tris buffer at pH = 8.3. (A) Spectra of CD changes during the titration. The pHs range from 8.4 to 4.8. (B) Difference of the CD spectra during the titration. The spectrum at the lowest pH was subtracted from all of the spectra in the titration range. (C) pH dependence of the band at 530 nm in the CD spectra. Absolute normalized maxima of the |ΔCD| bands are plotted as a function of pH. Extraction of the pKa value in the range of pHs of 4.8−8.4, using the equation F(x) = 1/(1 + 10(n*(pKa−x))). The solid line represents the best fit with a pKa = 6.6.

intensity of the CD spectrum was decreased during the titration (Figure 5A), similar to wt xR and ph-xR, and a negative lobe at around 535 nm was still detected. The difference spectra (Figure 5B) of the titration monitored by CD spectra with a crossing point at 486 nm may suggest that excitonic coupling between salinixanthins chromophores was abolished at low pH. 461

DOI: 10.1021/jp510534s J. Phys. Chem. B 2015, 119, 456−464

Article

The Journal of Physical Chemistry B

originated from both components, excitonic coupling and chirality, we were not able to fit the wt xR spectrum at pH = 7.6 to four conservative bands. We were able to fit the difference spectrum, where the excitonic coupling is decreased, to four excitonic bands because the fraction originating from the chirality of the salinixanthin did not contribute to the spectrum. The analysis may provide explanation for the unusual shape of the CD spectrum because it originates from summation of four excitonic bands. In acidic pH, the remaining fraction of the CD spectrum originates mainly from the chiral salinixanthin and some contribution of the chiral retinal pigment at around 550 nm, as explained before. These analyses support our hypothesis that the CD spectrum of wt xR is formed mainly due to the excitonic coupling (presumably between salinixanthins); however, it contains also a fraction that is formed due to the chirality of the fixed salinixanthin and the retinal chromophore, as was already shown previously.8

■

DISCUSSION We have studied wt xR and its artificial pigments characterized by a shifted absorption maxima. Following the titration of wt xR, artificial pigments, and their reduced pigments, its CD spectra at low pHs show high similarity to the absorption spectrum of xR with three positive lobes at 450, 486, 520 nm as well as a small negative lobe at around 545 nm, which may indicate the disappearance of excitonic coupling due to a pH decrease. Titrations of wt xR in 0.05 M Tris or phosphate buffers revealed different pKa values extracted from absorption and CD spectra, demonstrating two different processes detected by these spectroscopies. The process monitored by CD has a high pKa value of around 8, while the titration detected by absorption has a pKa value of around 6. Therefore, we suggest that each process is controlled by a different protein residue. The process that has a pKa value of 6 is ascribed to the PSB counterion complex (Asp-96−His-62) pKa.14 Thus, the protonation of the counterion red shifts the absorption maximum of the pigment but does not affect significantly the CD spectrum. In contrast, the protonation of the protein residue that has a pKa value of 8 (at low salt concentration) does not affect the absorption spectrum but affects the CD spectrum, a process that we propose leads to disappearance of an excitonic coupling. At high salt concentration, the two protein residues have a similar pKa value, suggesting that the two groups are affected differently by salt. At low pH, the remaining CD spectrum resembles the absorption spectrum of the salinixanthin and probably originates from the chiral conformation of the salinixanthin within its binding site. The diminishing of the CD spectrum at low pH can be explained by several possibilities. A plausible explanation is based on excitonic coupling between salinixanthin chromophores located at different subunits of the 3D structure of xR. The unusual shape of the CD difference spectra (Figure 1B) can be derived from the vibronic structure of the salinixanthin absorption. The possibility of excitonic coupling between the salinixanthins is supported by the fact that the crossing point (Δε = 0) of the positive and negative lobes is at 486 nm, equal to the absorption maximum of the salinixanthin in xR, as expected for the excitonic coupling between salinixanthins.26 Futhermore, fitting the CD spectra to excitonic bands fits the difference spectra with a high value of R2 but deviates from the spectrum at high pH or low pH. The CD spectrum of bacteriorhodopsin was also explained by excitonic coupling between the retinal chromophores in the trimeric structure.27 A

Figure 6. Fitting of CD spectra of xR at different pHs to excitonic bands. In the insets, the black curve in each figure represents the observed spectrum. The red represents the generated spectrum for the fitting. In each figure, different colors represent different fitting to the lobes of the CD spectrum. (A) Analysis of the xR spectrum at pH = 7.6. The spectrum was fitted to two conservative bands and three gaussians. (B) Analysis of the xR spectrum at pH = 4. The spectrum was fitted to seven Gaussian spectra. (C) Analysis of the difference spectrum pHs of 7.6−4. The spectrum was fitted to four conservative bands.

results may imply that the CD spectrum of wt xR is composed of excitonic coupling and a contribution from the chiral conformation of the salinixanthin. As the CD spectrum 462

DOI: 10.1021/jp510534s J. Phys. Chem. B 2015, 119, 456−464

Article

The Journal of Physical Chemistry B

event leads to the addition of a retinal band in the 530 nm region but does not affect significantly the positive carotenoid band.8 Therefore, this observation indicates that there is no strong excitonic coupling between the retinal and the salinixanthin. We propose that the CD spectrum of xR consists of contribution from three factors: (1) excitonic coupling between the salinixanthins chromophores located in different subunits of the 3D structure of xR, (2) chiral conformation of the salinixanthin within its binding site probably due the fixation of the 4-keto ring in a specific twisted conformation, and (3) contribution of the retinal chromophore to the negative lobe at around 550 nm, as was proposed before.2,8 Identification of the protein residue that affects the CD spectrum and its protonation eliminating the excitonic coupling, needs further work. However, it is possible that this residue is located in the vicinity of the protein surface. Protonation of this residue perturbs the 3D structure of the protein, thereby weakening the interactions between the salinixanthin chromophores located in the different subunits of the organized 3D structure and eliminating their excitonic coupling. Although we have observed that the pKa of this residue is affected by modification in the retinal ring structure, this residue can be located far away from the retinal binding site because retinal modification can affect its interaction with the salinixanthin chromophore, thereby affecting the protein conformation. Another possible explanation to account for the effect on the CD spectrum of xR involves protonation of His-62 which is part of the retinal PSB counterion, as was found for the gR protein.31 Recent SEC analysis (size exclusion chromatography)31 with the H87M mutant of the gR protein that exhibits high similarity to xR (as the corresponding residue of bR is methionine) and where Asp-121 is considered to form a cluster with His-87, indicated that the pH dependence of the trimer/monomer transition almost disappears and that the monomer is dominant. It was also found that when Asp-121 is protonated, the monomer is dominant. It was concluded that His-87 is a key residue for trimer formation, and in the absence of His-87, the deprotonation of carboxyls (mainly Asp-121) does not lead to trimer formation.31

different explanation attributed the biphasic band shape to the retinal−protein heterogeneity, leading to two or more types of bR in which their retinals have an opposite sense of intramolecular rotational distortion along their retinal long axis.27−30 Similar possibility can be suggested for xR. Accordingly, the negative band at 530 nm is part of the negative salinixanthin bands that are cancelled by the positive salinixanthin bands because of their summation in the apparent CD spectrum. Another possibility is that the negative lobe at 530 nm in the CD spectrum of xR originates from the salinixanthin band and contribution of the retinal, which is diminished at low pH due to an altered protein environment, affecting also the positive lobes. Although these possibilities cannot be completely excluded, we propose that a decrease in the intensity of the CD spectrum due to a protein residue protonation2 is because of excitonic coupling loss between the salinixanthin chromophores, and it is the most plausible explanation to account for the CD spectrum difference obtained following a pH decrease. The question arises whether the excitonic coupling involves the retinal and the salinixanthin chromophores that are located in close proximity or if it originated mainly from interaction between salinixanthins in different units of an organized xR structure. Several observations suggest that the contribution of the retinal chromophore to the excitonic coupling is minor. Reduction of xR PSB blue shifts the retinal absorption considerably by almost 200 nm, and it is expected that its possible involvement in excitonic coupling with the salinixanthin will be dramatically reduced. However, loss of excitonic coupling is detected by lowering the pH of the reduced xR, similar to the wt xR. In addition, the absorption maximum of the artificial pigment derived from aromatic retinal is blue shifted relative to xR by about 100 nm,10 but still, the excitonic coupling is detected by lowering the pH, similar to that detected in native xR. The presence of the negative lobe at 530 nm was detected as well, despite the fact that the retinal chromophore does not contribute to this lobe in the artificial aromatic pigment anymore as it was significantly blue shifted. Therefore, it can be proposed that the excitonic coupling originated from interaction of salinixanthins located in different subunits of the xR structured organization. The question remains why the close proximity and the angle of 46° between the long axis of the two chromophores11 do not lead to a strong excitonic coupling between salinixanthin and the retinal. As the dihedral angle between the electric transition moments of the two chromophores is a crucial factor in determining the excitonic coupling, no exciton coupling can occur if the dihedral angle between the chromophores planes is 0°.26 Therefore, a dihedral angle between the planes of retinal and salinixanthin, which is close to 0° according to the crystallographic structure, will prevent an efficient excitonic coupling between the chromophores. Moreover, the poor overlapping between the absorption maxima of salinixanthin and retinal (486 versus 570 nm, respectively) becomes an essential cause of excitonic coupling not being detected in our measurements, in contrast to the complete overlap between two molecules of salinixanthins. In addition, this suggestion is strongly supported by our previous results where we showed that the formation of the CD spectrum of xR is triggered by the occupation of the binding site by the retinal chromophore, leading to the salinixanthin fixation.8 Moreover, the CD spectrum (including the negative band at about 530 nm) is formed before the formation of the retinal−protein covalent bond. The latter

■

ASSOCIATED CONTENT

S Supporting Information *

Detailed information about the titrations’ excitonic coupling in xR as well as the table with specific pKa values of the obtained pKa values of the pigments. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail [email protected]. Phone +972 8 9344320. Fax +972 8 9343026. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors would like to thank Professor Aharon Oren (the Hebrew University of Jerusalem, Israel) for providing a strain of Salinibacter ruber M31. We would like also to express our gratitude to Professor Sandy Ruhman (the Hebrew University of Jerusalem, Israel) for the fruitful discussions and ideas. This work was supported by grants from the Kimmelman center for 463

DOI: 10.1021/jp510534s J. Phys. Chem. B 2015, 119, 456−464

Article

The Journal of Physical Chemistry B

of Halobacterium−Halobium  Contribution of Bacteriorhodopsin and Halorhodopsin to Photosensory Activity. Photochem. Photobiol. 1983, 38, 417−423. (19) Matsumoto, H.; Asato, A. E.; Denny, M.; Baretz, B.; Yen, Y. P.; Tong, D.; Liu, R. S. H. Aromatic Retinal Analogs and Their Interaction with Cattle Opsin. Biochemistry 1980, 19, 4589−4594. (20) Buchecker, R.; Noack, K. Carotenoids; Birkhäuser Verlag: Basel, Boston, Berlin, 1995; Vol. 1B. (21) Balashov, S. P.; Imasheva, E. S.; Wang, J. M.; Lanyi, J. K. Excitation Energy-Transfer and the Relative Orientation of Retinal and Carotenoid in Xanthorhodopsin. Biophys. J. 2008, 95, 2402−2414. (22) Gdor, I.; Zhu, J. Y.; Loevsky, B.; Smolensky, E.; Friedman, N.; Sheves, M.; Ruhman, S. Investigating Excited State Dynamics of Salinixanthin and Xanthorhodopsin in the Near-Infrared. Phys. Chem. Chem. Phys. 2011, 13, 3782−3787. (23) Eliash, T.; Weiner, L.; Ottolenghi, M.; Sheves, M. Specific Binding Sites for Cations in Bacteriorhodopsin. Biophys. J. 2001, 81, 1155−1162. (24) Szundi, I.; Steckenius, W. Effect of Lipid Surface-Charges on the Purple-to-Blue Transition of Bacteriorhodopsin. Proc. Natl. Acad. Sci. U.S.A. 1987, 84, 3681−3684. (25) Szundi, I.; Stoeckenius, W. Purple-to-Blue Transition of Bacteriorhodopsin in a Neutral Lipid Environment. Biophys. J. 1988, 54, 227−232. (26) Harada, N.; Nakanishi, K. Circular Dichroic Spectroscopy-Exciton Coupling in Organic Stereochemistry; University Science Books: Mill Valley, CA, 1983. (27) Cassim, J. Y. Unique Biphasic Band Shape of the Visible Circular-Dichroism of Bacteriorhodopsin in Purple Membrane  Excitons, Multiple Transitions or Protein Heterogeneity. Biophys. J. 1992, 63, 1432−1442. (28) Friedman, N.; Ottolenghi, M.; Sheves, M. Heterogeneity Effects in the Binding of All-Trans Retinal to Bacterio-Opsin. Biochemistry 2003, 42, 11281−11288. (29) El-Sayed, M. A.; Lin, C. T.; Mason, W. R. Is There an Excitonic Interaction or Antenna System in Bacteriorhodopsin. Proc. Natl. Acad. Sci. U.S.A. 1989, 86, 5376−5379. (30) Wu, S.; El-Sayed, M. A. CD Spectrum of Bacteriorhodopsin. Best Evidence against Exciton Model. Biophys. J. 1991, 60, 190−197. (31) Tsukamoto, T.; Kikukawa, T.; Kurata, T.; Jung, K. H.; Kamo, N.; Demura, M. Salt Bridge in the Conserved His-Asp Cluster in Gloeobacter Rhodopsin Contributes to Trimer Formation. FEBS Lett. 2013, 587, 322−327.

Biomolecular Structure and Assembly. M.S. holds the KatzirMakineni professorial chair in chemistry.

■

ABBREVIATIONS bR, bacteriorhodopsin; PSB, protonated Schiff base; xR, xanthorhodopsin; CD, circular dichroism; wt, wild-type

■

REFERENCES

(1) Balashov, S. P.; Imasheva, E. S.; Boichenko, V. A.; Anton, J.; Wang, J. M.; Lanyi, J. K. Xanthorhodopsin: A Proton Pump with a Light-Harvesting Carotenoid Antenna. Science 2005, 309, 2061−2064. (2) Balashov, S. P.; Imasheva, E. S.; Lanyi, J. K. Induced Chirality of the Light-Harvesting Carotenoid Salinixanthin and Its Interaction with the Retinal of Xanthorhodopsin. Biochemistry 2006, 45, 10998−11004. (3) Lutnaes, B. F.; Oren, A.; Liaaen-Jensen, S. New C-40-Carotenoid Acyl Glycoside as Principal Carotenoid in Salinibacter Ruber, an Extremely Halophilic Eubacterium. J. Nat. Prod. 2002, 65, 1340−1343. (4) Imasheva, E. S.; Balashov, S. P.; Wang, J. M.; Smolensky, E.; Sheves, M.; Lanyi, J. K. Chromophore Interaction in Xanthorhodopsin−Retinal Dependence of Salinixanthin Binding. Photochem. Photobiol. 2008, 84, 977−984. (5) Zsila, F.; Bikadi, Z.; Simonyi, M. Induced Chirality Upon Crocetin Binding to Human Serum Albumin: Origin and Nature. Tetrahedron: Asymmetry 2001, 12, 3125−3137. (6) Zsila, F.; Bikadi, Z.; Simonyi, M. Further Insight into the Molecular Basis of Carotenoid−Albumin Interactions: Circular Dichroism and Electronic Absorption Study on Different Crocetin− Albumin Complexes. Tetrahedron: Asymmetry 2002, 13, 273−283. (7) Anton, J.; Oren, A.; Benlloch, S.; Rodriguez-Valera, F.; Amann, R.; Rossello-Mora, R. Salinibacter Ruber Gen. Nov., Sp Nov., a Novel, Extremely Halophilic Member of the Bacteria from Saltern Crystallizer Ponds. Int. J. Syst. Evol. Microbiol. 2002, 52, 485−491. (8) Smolensky, E.; Sheves, M. Retinal−Salinixanthin Interactions in Xanthorodopsin: A Circular Dichroism (CD) Spectroscopy Study with Artificial Pigments. Biochemistry 2009, 48, 8179−8188. (9) Balashov, S. P.; Imasheva, E. S.; Choi, A. R.; Jung, K. H.; LiaaenJensen, S.; Lanyi, J. K. Reconstitution of Gloeobacter Rhodopsin with Echinenone: Role of the 4-Keto Group. Biochemistry 2010, 49, 9792− 9799. (10) Smolensky Koganov, E.; Hirshfeld, A.; Sheves, M. Retinal β− Ionone Ring−Salinixanthin Interactions in Xanthorhodopsin: A Study Using Artificial Pigments. Biochemistry 2013, 52, 1290−1301. (11) Luecke, H.; Schobert, B.; Stagno, J.; Imasheva, E. S.; Wang, J. M.; Balashov, S. P.; Lanyi, J. K. Crystallographic Structure of Xanthorhodopsin, the Light-Driven Proton Pump with a Dual Chromophore. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 16561−16565. (12) Christen, Rl; Kohler, B. E. Low Resolution Optical Spectroscopy of Retinyl Polyenes  Low-Lying Electronic Levels and Spectral Broadness. Photochem. Photobiol. 1973, 18, 293−301. (13) Imasheva, E. S.; Balashov, S. P.; Wang, J. M.; Lanyi, J. K. Removal and Reconstitution of the Carotenoid Antenna of Xanthorhodopsin. J. Membr. Biol. 2011, 239, 95−104. (14) Imasheva, E. S.; Balashov, S. P.; Wang, J. M.; Lanyi, J. K. pHDependent Transitions in Xanthorhodopsin. Photochem. Photobiol. 2006, 82, 1406−1413. (15) Polivka, T.; Balashov, S. P.; Chabera, P.; Imasheva, E. S.; Yartsev, A.; Sundstrom, V.; Lanyi, J. K. Femtosecond Carotenoid to Retinal Energy Transfer in Xanthorhodopsin. Biophys. J. 2009, 96, 2268−2277. (16) Zhu, J. Y.; Gdor, I.; Smolensky, E.; Friedman, N.; Sheves, M.; Ruhman, S. Photoselective Ultrafast Investigation of Xanthorhodopsin and Its Carotenoid Antenna Salinixanthin. J. Phys. Chem. B 2010, 114, 3038−3045. (17) Oren, A.; Mana, L. Sugar Metabolism in the Extremely Halophilic Bacterium Salinibacter Ruber. FEMS Microbiol. Lett. 2003, 223, 83−87. (18) Schimz, A.; Sperling, W.; Ermann, P.; Bestmann, H. J.; Hildebrand, E. Substitution of Retinal by Analogs in Retinal Pigments 464

DOI: 10.1021/jp510534s J. Phys. Chem. B 2015, 119, 456−464