HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Jegerschöld, C. Articles by Bassi, R. Search for Related Content PubMed PubMed Citation Articles by Jegerschöld, C. Articles by Bassi, R. Social Bookmarking                      What's this?

J Biol Chem, Vol. 275, Issue 17, 12781-12788, April 28, 2000

Calcium Binding to the Photosystem II Subunit CP29^*

Caroline Jegerschöld^§, A. William Rutherford, and Tony A. Mattioli^¶

From the  Section de Bioénergétique, Département de Biologie Cellulaire et Moléculaire, CEA/Saclay and CNRS URA 2096, 91191 Gif-sur-Yvette cedex, France and the ^¶ Section de Biophysique des Protéines et des Membranes, Département de Biologie Cellulaire et Moléculaire, CEA/Saclay and CNRS URA 2096, 91191 Gif-sur-Yvette cedex, France

Massimo Crimi^**, and Roberto Bassi^**

From the ^** Facoltà di Scienze MMFFNN, Biotecnologie Vegetali, Università di Verona, Strada Le Grazie, 37134 Verona, Italy

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

We have identified a Ca²⁺-binding site of the 29-kDa chlorophyll a/b-binding protein CP29, a light harvesting protein of photosystem II most likely involved in photoregulation. ⁴⁵Ca²⁺ binding studies and dot blot analyses of CP29 demonstrate that CP29 is a Ca²⁺-binding protein. The primary sequence of CP29 does not exhibit an obvious Ca²⁺-binding site therefore we have used Yb³⁺ replacement to analyze this site. Near-infrared Yb³⁺ vibronic side band fluorescence spectroscopy (Roselli, C., Boussac, A., and Mattioli, T. A. (1994) Proc. Natl. Acad. Sci. U. S. A. 91, 12897-12901) of Yb³⁺-reconstituted CP29 indicated a single population of Yb³⁺-binding sites rich in carboxylic acids, characteristic of Ca²⁺-binding sites. A structural model of CP29 presents two purported extra-membranar loops which are relatively rich in carboxylic acids, one on the stromae side and one on the lumenal side. The loop on the lumenal side is adjacent to glutamic acid 166 in helix C of CP29, which is known to be the binding site for dicyclohexylcarbodiimide (Pesaresi, P., Sandonà, D., Giuffra, E., and Bassi, R. (1997) FEBS Lett. 402, 151-156). Dicyclohexylcarbodiimide binding prevented Ca²⁺ binding, therefore we propose that the Ca²⁺ in CP29 is bound in the domain including the lumenal loop between helices B and C.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Photosystem II (PSII)¹ of cyanobacteria and higher plants is a multisubunit, membrane-spanning enzymatic complex responsible for oxygen evolution resulting from water oxidation. The oxidation of water occurs at the lumenal surface of the thylakoid membrane through the stepwise oxidation of a manganese cluster, (Mn)₄, which forms the catalytic core of the oxygen evolving enzyme (see reviews, see Refs. 1-3). PSII water oxidation requires both calcium and chloride ions, and although the calcium requirement has been intensively studied, insight regarding the number of Ca²⁺ sites and the precise role of Ca²⁺ in the photosynthetic oxidation of water remains murky (for review, see Refs. 1-4). It is known that the absence of Ca²⁺ in PSII inhibits proper photochemical function preventing oxygen evolution, and it is now becoming clear that one Ca²⁺-binding site is associated with these calcium-depleted effects (1-3, 5-8). The marked effects which the removal of Ca²⁺ has on the functioning of the oxygen evolving enzyme strongly suggest that this Ca²⁺ site is located in the reaction center core of PSII (5). Site-directed mutagenesis studies have allowed suggestions concerning the location of the Ca²⁺-binding site in the PSII reaction center core (9, 10).

Photosystem II possesses a second Ca²⁺-binding site not directly related to oxygen evolution (5, 11, 12). This second Ca²⁺ ion is very tightly bound and appears to be associated with a light-harvesting protein (5, 12-14). There are conflicting reports as to which of the light-harvesting proteins the strongly bound Ca²⁺ is associated. Irrgang and co-workers (12) have characterized the strong Ca²⁺-binding properties of a 30-kDa Chl a/b protein and reported its implication as being a Ca²⁺-binding protein. This 30-kDa protein is most likely CP29 (see Ref. 15).

CP29 is the largest of the so-called "minor" light harvesting complex (LHC) proteins of PSII and consists of approximately 257 amino acids. Various data indicate that the protein binds 6 Chl a, 2 Chl b, and 2 xanthophyll molecules (16-18). CP29 is situated between the major LHCII complex and the PSII core and is thought to play a role in the regulation of the concentration of the Chl a excited states (16, 19, 20). CP29 also undergoes reversible phosphorylation under photoinhibitory conditions (21). It seems then that CP29 probably plays important roles in photoregulation in PSII.

In this contribution, we identify a specific Ca²⁺-binding site in the CP29 polypeptide of maize PSII. CP29 was reconstituted with the lanthanide Yb³⁺ and the metal-binding site in CP29 was characterized using near-infrared Yb³⁺ vibronic side band (VSB) fluorescence spectroscopy (22, 23). In this technique, we exploit the well known phenomenon of the specific replacement of Ca²⁺ in proteins with trivalent lanthanide metal ions (24, 25). After Ca²⁺ is replaced selectively with Yb³⁺, then the fluorescence spectrum of Yb³⁺ in the Ca²⁺ site is recorded and analyzed. When coordinated, the Yb³⁺ fluorescence spectrum exhibits two characteristic types of features: 1) a sharp 4f-4f electronic feature called the zero phonon line (ZPL) that corresponds to the Yb^{3+ 2}F_5/2  ²F_7/2 electronic transition, and, 2) weak VSBs shifted to longer wavelengths with respect to the ZPL. These VSBs are the result of short-range dipole-dipole interactions that couple the ZPL electronic transition dipole of Yb³⁺ with the dipoles of the vibrational modes of its ligands. These VSBs reflect the vibrational modes of the Yb³⁺ ligands: the spectral energy difference between the ZPL and the various VSBs correspond to vibrational energies (or frequencies) of the corresponding ligand vibrational modes. VSBs, in fact, provide a specific and selective vibrational spectrum of the ligands and only the ligands defining the binding site of Yb³⁺ in the protein (see Refs. 22 and 23 and references therein). As with Raman or infrared vibrational spectra, Yb³⁺ VSB "fingerprint" spectra can be used to identify the chemical nature of the metal binding ligands. Since inspection of the primary sequence of CP29 (26) does not reveal an obvious Ca²⁺-binding site such as a classic EF-type motif, the vibrational information obtained from the Yb³⁺ VSB spectra provides valuable information concerning the chemical nature of the binding site of Yb³⁺ reconstituted into the Ca²⁺-binding site in CP29, and a plausible location may be proposed.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Native CP29 and LHCII were purified from maize PSII membranes as described previously (27, 28). Proteins were transferred to a nitrocellulose membrane (Bio-Rad) with a dot blot apparatus (Bio-Rad) (CP29 = 1.6 µg of Chl; LHCII = 1.7 µg of Chl; calsequestrin = 3.8 µg of protein).

Detection of ⁴⁵Ca²⁺ binding was performed essentially as in Ref. 29 in the presence of 10 µCi of ⁴⁵CaCl₂ and 2 µM CaCl₂ for 20 min at pH 7.5; nitrocellulose was then washed 3 times for 2 min with 10 ml of 30% ethanol. Radioactivity was detected with a Instant Imager, Packard.

The proteins were probed for the binding of ⁴⁵Ca²⁺ either in the presence or absence of DCCD (dissolved in ethanol). Control proteins were treated with the same volume of ethanol utilized for DCCD treatment (< 1.5% of the final volume) and incubated 10 min at room temperature. Calsequestrin from rabbit muscle was used as positive control.

All further procedures used plastic tubes to avoid contamination from other cations, and media made from deionized water (Elgastat Maxima, Elga Ltd., Bucks, United Kingdom; resistivity 18 M), and passed through a Chelex 100 (Bio-Rad) column. Fresh stock solutions of Yb³⁺ were prepared just before use from chloride crystals (YbCl₃·6H₂O, Aldrich) kept in a dessicator. The buffers used, i.e. HEPES, MES, and acetate, are known to be poor or weak lanthanide coordinators (25).

The reconstitution of Yb³⁺ into CP29 was performed on initial solutions of 1.85 mM CP29 in "Solution A" (0.03% dodecyl--maltoside (-DM), 10 mM HEPES, pH 7.6, 0.1 M sucrose); all steps were performed at 4 °C in dim light. The CP29 protein solution was first diluted to 0.075 mM in 0.03% -DM, 20 mM MES, 0.1 M sucrose, 1 mM EDTA all at pH 4.4. Similar pH conditions were found for the deionization procedure for the removal of Ca²⁺ and Mg²⁺ ions in bacteriorhodopsin (23). The resulting solution was then centrifuged in a Centricon 30 microconcentrator (Amicon) to remove the aqueous buffer. UV-visible absorption spectroscopy (Shimadzu UV-160) was used after every centrifugation step to verify no chlorophyll or carotenoid molecules were present in the wash. The CP29 protein was then re-diluted in a solution containing 0.03% -DM, 0.1 M sucrose, 20 mM MES, pH 6.6, and 0.5 mM Yb³⁺, representing a Yb³⁺:CP29 ratio of approximately 10:1, and this solution was allowed to incubate overnight on ice in the dark. This Yb³⁺-containing solution was then exchanged for the original Solution A by washing several times with Solution A using the Centricon 30 microconcentrators; again, UV-visible absorption spectrophotometry was used to verify no loss of protein or spectral changes in the native and Yb³⁺-treated CP29. The Yb³⁺ content of the resulting CP29 protein, and after several consecutive washings in Solution A using the microconcentrators, was determined by graphite furnace atomic absorption spectrophotometry (Perkin-Elmer 2280) using the carefully determined linear relation observed for Yb³⁺ standard solutions.

Low temperature (15 K) near infrared Yb³⁺ vibronic side band spectra of Yb³⁺-reconstituted CP29 were obtained as described in Roselli et al. (22, 23). Near-infrared light used for exciting the Yb³⁺ fluorescence was provided by a Ti⁴⁺:sapphire laser (Spectra Physics 3900 S) pumped by a continuous wave Ar⁺ laser (Coherent Innova 100) operating in the "all-lines" mode. Laser power at the sample was about 100 mW. Yb³⁺ near-infrared fluorescence was detected using a modified Bruker IFS 66 Fourier transform interferometer equipped with a Bruker FRA 100 Raman module as described by Roselli et al. (22, 23). Samples were held in an optical cold helium gas-flow cryostat (TBT-SMC, Grenoble, France). CP29 protein samples in D₂O were prepared by repeated cycles of solution exchanges with D₂O, 10 mM HEPES buffer, pD 8.0, containing 0.03% -DM and 0.1 M sucrose using Centricon 30 concentrators; D₂O solutions are used as a precaution in Yb³⁺ VSB spectroscopy to avoid lanthanide fluorescence quenching often observed in H₂O (25). Spectral baseline correction, Fourier deconvolution, and second-derivative analyses were performed using the GRAMS 32 software package (Galactic Industries, Salem, NH).

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

⁴⁵Ca²⁺ Binding in CP29 and LHCII-- We have studied the binding of calcium to the minor antenna protein CP29 and to the major LHCII polypeptide. The dot blot analysis of purified CP29 and LHCII proteins (Fig. 1) demonstrates that CP29 is a calcium-binding polypeptide whereas LHCII did not show any significant binding (position B2 and B1, respectively, in Fig. 1). It should be pointed out that a weak signal is also present in the control LHCII protein, this signal was attributed to small contamination of minor antenna proteins (CP26 and CP29) in this preparation as confirmed by SDS-polyacrylamide gel electrophoresis analysis (not shown).

View larger version (21K): [in this window] [in a new window]   Fig. 1.   Effects of DCCD on calcium-binding properties of isolated LHCII and CP29. Samples (after DCCD binding or treatment with the same volume of ethanol) were transferred on a nitrocellulose membrane with a dot blot apparatus. A1 and A2, LHCII and CP29, respectively, incubated 10 min with 200 µM DCCD at room temperature; B1 and B2, control LHCII and CP29 respectively; A3, blank; B3, calsequestrin.

Fig. 1 also shows that DCCD prevented ⁴⁵Ca²⁺ binding to CP29 (position A2). This is in agreement with previous observations that radioactive calcium was specifically bound by unidentified PSII antenna proteins rather than by PSII core proteins and was displaced by DCCD (14).

One Persistent Yb³⁺ Ion Binding in CP29-- Yb³⁺ can be reconstituted into CP29 by first lowering the pH to 4.4, washing, and then incubating CP29 in a 10-fold excess of Yb³⁺ in a buffered solution at pH 6.6. Excess Yb³⁺ ions, and those that were weakly bound, were removed by simple washing with Solution A with no chelator present. Fig. 2 shows the number of Yb³⁺ atoms per CP29, determined using graphite furnace atomic absorption spectrophotometry, as a function of the number of washing steps after Yb³⁺ reconstitution. This figure clearly shows that after only two washing steps, essentially all of the excess, and nonspecifically bound, Yb³⁺ is removed from CP29, resulting in the stoichiometric (1:1) binding of Yb³⁺ to the protein. These conditions for Yb³⁺ reconstitution procedure were determined after preliminary studies at higher pH.

View larger version (10K): [in this window] [in a new window]   Fig. 2.   The amount of Yb3+ in CP29, determined by graphite furnace atomic absorption spectroscopy, as a function of the number of washing steps.

Fluorescence of Yb³⁺ Reconstituted into CP29-- Fig. 3 shows the Yb³⁺ fluorescence spectrum of the Yb³⁺-reconstituted CP29 protein (Yb³⁺-CP29) in the zero phonon line (or low frequency) region as compared with those of Yb³⁺ in the buffer/-dodecyl maltoside D₂O solution (Yb³⁺--DM) and Yb³⁺ in buffer/D₂O solution (Yb³⁺-D₂O) (see legend in Fig. 3 for details). The fluorescence spectra in Fig. 3 are plotted in terms of absolute energy expressed in wavenumbers (cm¹); the zero phonon lines are denoted "ZPL" and their absolute energies are indicated (10,236, 10,270, and 10,277 cm¹ for Yb³⁺-CP29, Yb³⁺--DM, and Yb³⁺-D₂O, respectively). The vibronic side bands observed at lower energies are indicated as the shift (in cm¹) relative to the ZPLs (0 cm¹). Fig. 3 clearly shows that the Yb³⁺ VSB spectra of these different Yb³⁺ complexes (i.e. Yb³⁺-CP29, Yb³⁺--DM, and Yb³⁺-buffer/D₂O) are all distinctly different. These different VSB spectra permit us to arrive at specific conclusions concerning the Yb³⁺-CP29 complex. The VSB spectrum of Yb³⁺ complexed with D₂O is relatively strong and exhibits distinct characteristic VSB bands at 41, 193, 289, 369, and 454 cm¹ in the low frequency region. In comparing the low frequency VSB spectra of the Yb³⁺-D₂O and Yb³⁺-CP29 complex shown in Fig. 3 it is clear that the Yb³⁺ ion in CP29 is in a quite different coordination environment with few or no water molecules as ligands. In addition, comparison of the Yb³⁺-CP29 and Yb³⁺--DM low frequency VSB spectra indicates that we have no interference from possible Yb³⁺-detergent complex artifacts in the Yb³⁺-CP29 VSB spectrum (this point will be discussed further below).

View larger version (20K): [in this window] [in a new window]   Fig. 3.   Low temperature (15 K) Yb3+ fluorescence spectra plotted in terms of absolute energy expressed in wave numbers (cm1) for Yb3+-reconstituted CP29 protein in D2O buffer (Yb-CP29, excited at 887.78 nm), Yb3+ complexed with 10-fold excess -dodecylmaltoside in D2O buffer (Yb-DM, excited at 891.17 nm), and Yb3+ complexed with D2O in D2O buffer (Yb-D2O, excited at 822.22 nm). The energy of the ZPL are indictated in the figure (shown horizontally) and are expressed in absolute energy (cm1). The vibronic side band frequency values shown in the figure (shown vertically) are calculated from the shift in energy (cm1) from the ZPL. Spectra were shifted on the vertical scale for pictorial clarity.

The low frequency VSB spectrum of Yb³⁺-reconstituted CP29 shown in Fig. 3 allows us to determine the ZPL of Yb³⁺ in CP29 along with some low-frequency ligand vibrational modes of the Yb³⁺ ligands, however, these low frequency modes are not as distinctive as higher frequency vibrational modes which serve as "vibrational fingerprints" of the ligands coordinated with Yb³⁺ in CP29. Fig. 4 shows the Yb³⁺ VSB spectra of Yb³⁺-CP29, in the high frequency region, recorded using several excitation wavelengths. The VSBs are readily identified by their invariance in energy as the excitation wavelength is changed while, in contrast, Raman bands follow the change in excitation wavelength. Observed Raman bands in Fig. 4 are designated by "R"; with respect to the excitation wavelength, these clusters of Raman bands correspond to vibrational frequencies at about 2028-2414-2474 cm¹ (in Raman shift) and can be attributed to the D₂O ice in the samples. In contrast, there are three distinct bands in Fig. 4 at about 8900-8650 cm¹ (on the absolute energy scale) that do not shift as the excitation wavelength is changed and can be thus confidently assigned to vibronic side bands of the Yb³⁺ in CP29. These three distinct VSB bands in Fig. 4 are seen at 1,314, 1,397, and 1,597 cm¹, as calculated from the shift of the ZPL of the Yb³⁺-CP29 complex observed at 10,236 cm¹ (Fig. 3). We also note that these VSBs are unique to the Yb³⁺-CP29 complex and are not observed for the Yb³⁺·-dodecylmaltoside complex (Fig. 4) nor for the Yb³⁺·D₂O complex (data not shown). Furthermore, at these wavelengths of excitation we could not detect any interference from any Raman contributions from either the protein or its chlorophyll/carotenoid pigments in the VSB spectral region studied.

View larger version (17K): [in this window] [in a new window]   Fig. 4.   Low temperature (15 K) fluorescence spectra of Yb3+ in Yb3+-CP29 excited at 943.75 nm (a), 957.85 nm (b), 964.79 nm (c), and fluorescence (d) of Yb3+ in Yb3+--dodecylmaltoside complex excited at 957.67 nm. The bands marked with "R" to designate Raman bands of D2O ice are seen to shift on the absolute energy scale (cm1) as the excitation wavelength is changed. The Yb3+ VSB do not shift as the excitation wavelength changes, and prominent VSBs are observed in Yb3+-CP29 at 1,314, 1,433, and 1,597 cm1 (indicated vertically in the figure) as calculated from the shift with respect to the ZPL at 10,236 cm1 (see Fig. 2). These latter VSBs are unique for Yb3+ in CP29 and are not observed in the Yb3+--dodecylmaltoside complex (d). Spectra were shifted on the vertical scale for pictorial clarity.

As with any fluorescence phenomena, the intensity of the fluorescence band is excitation wavelength dependent. We have analyzed the behavior of the VSB bands as a function of excitation energy. Fig. 5 shows the variation in intensity of the ZPL, two low frequency VSBs, as well as the relatively intense 1397 and 1597 cm¹ VSBs, in the high frequency region, of Yb³⁺ in CP29, as a function of excitation wavelength. The intense, sharp 2028 cm¹ Raman band of the D₂O ice (denoted R in Fig. 4) was used as an internal intensity standard. As expected, the intensities of the two VSBs vary as the excitation wavelength is changed but more importantly, the trends of the VSBs are all the same. In addition, in the low frequency region, the 399 cm¹ VSB and the two unresolved 146 and 200 cm¹ VSBs follow the same trend as the ZPL; for these two latter bands, the intensity was monitored near the center frequency of 170 cm¹. In the high frequency region,² the 1397 and 1597 cm¹ VSBs exhibit similar trends as the wavelength of excitation is changed and indicate the same maximum at about 960 nm. If these two VSB bands were originating from two or more spectroscopically different Yb³⁺ sites, then it is expected that their excitation wavelength-dependent behavior would be different and would manifest this difference by variation in the trends of the VSB bands. The Yb³⁺ VSB technique is very sensitive to this type of behavior, as we have previously reported (23, 30). The lack of differences in the observed VSB intensities as a function of excitation wavelength as seen in Fig. 5 indicates that the population of Yb³⁺ ions detected in CP29 protein is spectroscopically homogenous. We note that Yb³⁺ VSB spectroscopy is a very sensitive technique and can distinguish energy differences in Yb³⁺-binding sites as small as 15 cm¹ (30).

View larger version (10K): [in this window] [in a new window]   Fig. 5.   Variation of the intensities of: (A) the ZPL () and the low frequency 170 () and 399 () cm1, and (B) the high frequency 1397 () and 1597 () cm1 vibronic side bands of Yb3+ in CP29 as a function of excitation wavelength. The intensities were normalized to the 2028 cm1 Raman band of D2O which was used as an internal standard. The data for the 170 cm1 () VSB is taken as the central frequency of the unresolved VSBs at 146 and 200 cm1 (see Fig. 3).

Fig. 6 shows the unsmoothed average of eight high frequency VSB spectra of Yb³⁺ in CP29. For clarity of presentation, a baseline correction has been performed. In addition to the prominent VSBs observed at 1314, 1397, and 1597 cm¹, we have performed Fourier deconvolution and second-derivative analyses (not shown) to accurately determine the frequencies of unresolved shoulders seen at 1433, 1533, and 1623 cm¹ in the VSB spectrum. We note that the comparison of the Fourier deconvolution and second-derivative spectra were both in agreement in identifying spectral components at 1433, 1533, and 1623 cm¹. The Yb³⁺ VSB spectrum shown in Fig. 6 is similar to that of Yb³⁺ reconstituted into the Ca²⁺-binding site of rabbit muscle parvalbumin (22) and of other complexes of Yb³⁺ predominantly bound by carboxylic groups (22, 23, 30). The VSB spectrum in Fig. 6 unambiguously indicates that the majority of the ligands of Yb³⁺ bound in CP29 are carboxylic acids, as would be expected for a Ca²⁺-binding site. Details of the spectrum and band assignments will be discussed below.

View larger version (16K): [in this window] [in a new window]   Fig. 6.   Low temperature (15 K) high frequency vibronic side band spectrum of Yb3+ reconstituted into CP29. This spectrum represents an average of 8 individual VSB spectra recorded at various wavelengths. For clarity, a baseline has been subtracted. Spectral components designated at 1433, 1533, and 1623 cm1 were obtained by Fourier transform deconvolution and second-derivative analyses.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Lanthanide (Yb³⁺) Reconstitution into Protein Ca²⁺-binding Sites-- It is well established that lanthanides readily substitute and strongly bind in protein Ca²⁺-binding sites (22-25, 30, 31). Trivalent lanthanides have ionic radii (32) which are comparable to that of Ca²⁺ (about 1 Å). Like Ca²⁺, lanthanide bonding is essentially ionic (25). Crystallographic, spectroscopic, and chemical data indicate that the most common ligand atoms of Ca²⁺-binding sites in biological systems are oxygen atoms (33-35). Similarly, lanthanides exhibit a strong preference for ligands providing oxygen donor atoms (25). Lanthanides have been used extensively to probe Ca²⁺-binding sites in proteins (reviewed in Refs. 25, 31, and 36) and x-ray crystallographic protein structures provide strong evidence for the isomorphous replacement of Ca²⁺ by trivalent lanthanides (25, 36). Several structures suggest that it is quite generally possible to replace Ca²⁺ by Ln³⁺ with minimal disruption to the binding site or the overall structure of the protein. It should be expected, however, that the Ca²⁺ to Ln³⁺ substitution could increase the co-ordination number by one (36), as has been observed in one Ca²⁺ site of carp parvalbumin which was substituted with Yb³⁺ (37).

Concerning the activity of the Ln³⁺-substituted proteins, there are examples cited in the literature where retention of full biological activity, diminished activity, and inhibitory behavior are reported. The first case will likely result in situations where the Ln³⁺ like the Ca³⁺ play a purely structural role. If Ca²⁺ is present at a catalytic site, then some change in activity is to be expected (36).

Although the comparable ionic radii and co-ordination chemistries of Ca²⁺ and the Ln³⁺ ions render near-perfect structural replacements, the decreased lability toward ligand exchange and stronger binding of the lanthanides must be considered when using these ions as probes. Although they are not perfect analogues of Ca²⁺ in all functional respects, their spectroscopic properties render them extremely valuable in probing spectroscopically silent Ca²⁺-binding sites. (36).

The Ca²⁺- and Yb³⁺-binding Sites in CP29-- In this work we have identified and confirmed CP29, the 29-kDa chlorophyll a/b-binding protein associated with Photosystem II, as a Ca²⁺-binding protein (Fig. 1). Treatment of CP29 at pH 4.4 and reconstitution with Yb³⁺ results in the 1:1 stoichiometric binding of Yb³⁺ ion in CP29 as determined by graphite furnace atomic absorption spectrophotometry (Fig. 2). Lanthanides can be replaced for Ca²⁺ in Ca²⁺-binding proteins either by placing the native or the apoprotein in contact with the lanthanide solution (25). The method we have used (see "Experimental Procedures") was developed for the reconstitution of Yb³⁺ into a fragile, temperature- and light-sensitive, detergent-solubilized protein of limited quantity. The treatment of CP29 at pH 4.4 before significant incorporation of Yb³⁺ into the protein metal-binding site was observed is consistent with the weakened binding or release of Ca²⁺ from carboxyl group ligands (25) to facilitate metal exchange with Yb³⁺. In addition, the lack of observed changes in relative intensities of the Yb³⁺ fluorescence vibronic side bands as excitation wavelength was changed indicates that the Yb³⁺ site being probed is originating from one site or one species. All these data indicate that the Yb³⁺ ion which has been reconstituted into CP29 is strongly and specifically bound in a well defined metal-binding site. The above behavior is consistent with Yb³⁺ replacing Ca²⁺ in the metal-binding site of CP29.

It should be mentioned that in many studies on Ca²⁺ removal from the intact PSII complex, only the Ca²⁺ involved in oxygen evolution, and thus the Ca²⁺ most closely associated with the PSII reaction center core, was reported to be removed (reviewed in Ref. 1; see also Ref. 6) whereas the other Ca²⁺ ion associated with the extrinsic polypeptides remained. Some treatments, such as that described by Ono and Inoue (38), descend in pH values lower (i.e. pH 3) than that used in the Ca²⁺ deionization step for CP29 in our work. However, in those studies the treatments were performed on intact PSII membrane fragments and not on the isolated single polypeptide as is the case in our work here where the replacement of Ca²⁺ in CP29 by Yb³⁺ may have been greatly facilitated due to the fact that the protein is in an isolated detergent-solubilized form rendering the Ca²⁺-binding site more susceptible to ion exchange.

The ligands of Yb³⁺/Ca²⁺ in CP29-- Based on its similarity with the Yb³⁺ vibronic side band spectra of Yb³⁺-carboxylic ligand complexes and with Yb³⁺-reconstituted Ca²⁺-binding proteins with carboxylic ligands (see Fig. 2 in Ref. 22), the VSB spectrum of Yb³⁺ bound in CP29 clearly indicates a majority of carboxylic acid ligands. This is what would be expected for a Ca²⁺-binding site. Indeed lanthanides and calcium are known to have a strong affinity for oxygen ligands (25). Previously, we have systematically studied the infrared absorption and the fluorescence VSB spectra of many various Yb³⁺ model complexes, several of which contain carboxylate groups (COO), carbonyl groups (C=O), as well as alcohol and phenolic C-O groups (22, 23, 30). VSB assignments of these groups could easily be made based on the well known characteristic vibrational frequencies (or fingerprints) of these groups. This approach was successfully applied to the case of well known Ca²⁺-binding proteins, such a rabbit muscle parvalbumin whose Ca²⁺-binding sites are rich in COO ligands, and include protein backbone C=O groups as well as C-O groups from alcoholic side chains such as serine (22).

Like calcium, lanthanides can bind carboxylic acids in a monodentate or bidentate manner. The symmetric, _sym, and antisymmetric, _antisym, vibrational stretching modes of COO are expected at about 1400-1450 cm¹ and about 1550-1600 cm¹, respectively (22, 23, 39). For monodentate coordination, the difference, , in these two vibrational modes is about 200 cm¹ while for bidentate coordination is about 100 cm¹ (39). In the VSB spectrum of Yb³⁺, the most intense bands are observed at 1397 and 1597 cm¹; these bands are assignable as the symmetric _s and antisymmetric _antisym vibrational modes of the COO ligands of Yb³⁺. The difference in vibrational frequency, , of these two modes is 200 cm¹ indicating COO ligands of monodentate coordination. This indicates that most of the COO ligands are coordinated to Yb³⁺ in a monodentate manner. However, we cannot conclude that all of the COO ligands are monodentate. The 1397 and 1597 cm¹ bands are broad (about 100 cm¹ full width at half-maximum), asymmetric, and clearly indicate the presence of unresolved shoulders. Two shoulders, whose vibrational frequencies have been determined by Fourier deconvolution and second-derivative analyses to be 1433 and 1533 cm¹ can also be assigned to _sym and _antisym modes, respectively. For this case, the difference in frequency, , is 100 cm¹ and thus indicates bidentate binding of the COO group(s) giving rise to the 1433 and 1533 cm¹ VSBs. It is difficult to accurately quantify the number of monodentate and bidentate COO ligands to Yb³⁺ in CP29, but considering the intensities of the 1397/1597 cm¹ VSBs relative to the intensities of the weaker 1433/1533 cm¹ VSBs, the bidentate COO ligands are a minority, perhaps representing only one bidentate COO ligand to three or four monodentate COO ligands.

The 1623 cm¹ VSB component we have identified using Fourier deconvolution and second-derivative analyses on the VSB spectrum in Fig. 6 is assignable to the vibrational stretching frequency of a carbonyl C=O group coordinated to Yb³⁺ (Table I in Ref. 22, and references therein). Carbonyl (C=O) groups are also good Ca²⁺ and Yb³⁺ ligands. These C=O groups are expected to vibrate at about 1620 cm¹, based on the infrared and Yb³⁺ VSB spectra of Yb³⁺ complexes with ligands possessing such groups (22). Because of their similar vibrational frequencies, the C=O bands (about 1620 cm¹) and the monodentate COO _antisym modes (about 1600 cm¹) are often difficult to resolve. The VSB spectrum of Yb³⁺ in CP29 indicates that at least one C=O group, most likely from the protein backbone, is a ligand.

The 1314 cm¹ VSB in Fig. 6 is assignable to a C-O group (22). This value is slightly too high to arise from a phenolic C-O group which would be expected at about 1270-1280 cm¹ such as that from a tyrosine residue (23), but is more consistent with a C-O group from an alcoholic side chain, such as serine.

Based on the above analysis we can deduce the following ligands for Yb³⁺ reconstituted in CP29: a majority of monodentate COO groups and at least one bidentate COO group from carboxylic acid residue side chains, at least one C=O group from the protein backbone, and at least one C-O group from an alcoholic side chain. In the absence of an intense VSB at about 2700 cm¹, we find no evidence for the presence of D₂O as a ligand of Yb³⁺ reconstituted into CP29.

Location of the Yb³⁺-/Ca²⁺-binding Site in CP29-- CP29 is homologous to the major light harvesting complex of Photosystem II (LHCII) whose structure has been recently determined (40). Accordingly, three transmembrane hydrophobic helices (A-C) and a lumenal surface amphiphilic helix (D) can be identified connected by two helix-helix loops on either side of the membrane, a stromal exposed N-terminal extension and a lumen exposed C terminus. According to the availability of clustered acidic residues, two domains are eligible for Yb³⁺-/Ca²⁺-binding: the N terminus, protruding into the stromal space, and the lumen exposed domain including part of helix C and the loop extending to helix B. A structural model of the CP29 protein has been recently proposed (41).

In order to probe the two possible locations for the Ca²⁺-binding site, we have treated the protein with DCCD, a protein modifying agent reacting with acidic residues in hydrophobic environments, such as those within -helices. In the case of CP29, DCCD has been shown to bind to a glutamic acid residue (Glu-166) in helix C (26) (see Fig. 7) which is part of the lumen-exposed putative Ca²⁺ binding sequence. In the case Glu-166 is involved in binding Ca²⁺ or is near the Ca²⁺-binding site, we would expect that DCCD treatment would prevent Ca²⁺ binding.

View larger version (44K): [in this window] [in a new window]   Fig. 7.   Model for the location of the calcium-binding site in CP29. The residue involved in the binding of DCCD (Glu-166) is located in an hydrophobic region of helix C. Three acidic residues are present in the lumenal loop and they could be directly involved in calcium binding.

Fig. 1 showed that DCCD prevented ⁴⁵Ca²⁺ binding to CP29 (position A2) thus confirming that Glu-166 is part of, or near to, the Ca²⁺-binding site. Within our structural model, Glu-166 is found in helix C and is the only acidic residue in CP29 that is in a hydrophobic environment and is not charge compensated by neighboring arginine residues. On the basis of competition with DCCD for the Glu-166 residue (26), the region involved in Ca²⁺ binding can be located in the domain including the lumenal loop between helices B and C, and, perhaps, part of the hydrophobic helix C. In this region, three acidic residues are present (2 Asp and 1 Glu) that could be part of the binding site. A structural model for CP29 (41) and its putative Ca²⁺-binding site is shown in Fig. 7.

The helix-loop-helix EF-hand is by far the most common motif recognized for intracellular Ca²⁺-binding proteins (35). In the so-called EF-hand motif, the Ca²⁺ ion is bound in a loop consisting of 10-14 amino acids forming a tight turn between two helices. The proposed Ca²⁺-binding domain in the CP29 structural model of Fig. 7 is the loop between helices B and C. This loop is not recognized as typical of a regular EF-hand because it is too long (i.e. more than 12 amino acid residues). Interestingly, we have noted significant homology in the partial sequence alignment of the purported Ca²⁺-binding domain of the polypeptide CP29 and that of the Ca²⁺-binding protein(s) -lactalbumin (see Fig. 8). As in an EF-motif, the Ca²⁺-binding site in -lactalbumin is also formed by two helices and a loop joining them but compared with the EF-hand, the loop is shorter by about 2 amino acid residues and the arrangement of the structural units are different (42). However, this helix-loop-helix "elbow" motif in -lactalbumin would not be consistent with the structural model of CP29 in Fig. 7 which shows a relatively large loop domain between helices B and C. Thus, the Ca²⁺-binding site domain of CP29 identified in this work may be regarded as having a structural motif atypical or uncommon as compared with the known binding motifs of most Ca²⁺-binding proteins. The unusual character of the Ca²⁺-binding site in CP29 may be related to the Ca²⁺ ion's proximity to chlorophyll and carotenoid molecules and/or to a specific function or role of the Ca²⁺ ion.

View larger version (12K): [in this window] [in a new window]   Fig. 8.   Partial amino acid sequences of CP29 in the lumenal loop region between the purported transmembrane -helices B and C aligned with that of LHCII (Ref. 26), and compared with the Ca2+-binding site of -lactalbumin, -LA (Ref. 42). The residues which coordinate the Ca2+ in -lactalbumin are indicated in boldface (three aspartic acid residues binding via their carboxyl groups) and underlined (residues binding via their backbone carbonyl groups). The CP29 residues shown in boldface indicate possible Ca2+-binding residues, possessing COO or OH groups; the underlined residues in italics WQDAG are part of the carotenoid binding pocket.

It is surprising that CP29 is the only Ca²⁺-binding antenna protein since it belongs to a protein family which includes at least 10 members, six of which, Lhca1-4, Lhcb4-5, carry a glutamate residue in positions homologous to Glu-166 in CP29. However, sequence comparison in this region shows that CP29, with respect to other Lhc gene products, is unique in carrying not only the glutamic residue in position 166, but also the three acidic residues in the loop. Site-directed mutagenesis work has recently shown that Glu-166 is also involved in the binding of chlorophyll as well as in carotenoid (violaxanthin) binding (41).

It is interesting to observe that the glutamic acid residues involved in chlorophyll coordination in LHCII (40) were found to be charge-compensated by arginine residues. However, in the case of CP29, no basic residue can be found in positions compatible with Glu-166 charge compensation. We can therefore hypothesize that the Ca²⁺ ion may perform the role of the arginine in compensating the charge on the glutamate, thus allowing it to act as a ligand to the chlorophyll. It was proposed (43) that photoprotective dissipation of excess energy in the PSII antenna, elicited by low lumenal pH, is triggered by protonation of an acidic residue in a hydrophobic environment thus leading to a structural change involving a closer association between the chlorophyll and a carotenoid molecule. Within this model, Ca²⁺ coordination to the acidic residues making up the coordination sphere of the Ca²⁺ could be involved in such a pH triggered structural change.

	ACKNOWLEDGEMENTS

We thank Dr. Alessandra Nori (University of Padua) for advice in Ca²⁺ overlay procedures and for the kind gift of purified calsequestrin. Roberta Croce is thanked for help in the purification of CP29 and helpful discussion. Dr. Alain Boussac (CEA/Saclay) is thanked for helpful discussions.

	FOOTNOTES

^* This work was supported by a long-term FEBS fellowship (to C. J.) and a grant from "Progetto Finalizzato Biotecnologie" of CNR (to R. B.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

^§ Current address: Dept. of Biochemistry, Imperial College of Science, Technology and Medicine, London, SW7 2AY, England.

To whom correspondence should be addressed: Tel.: 33-1-6908-4166; Fax: 33-1-6908-8717; E-mail: mattioli@dsvidf.cea.fr.

² For technical reasons related to the spectral characteristics of the laser-rejection filters used in our apparatus, we are unable to observe the ZPL line when measuring the high frequency VSBs. Consequently, the VSB data are depicted separately in the low frequency and the high frequency VSB spectral ranges.

	ABBREVIATIONS

The abbreviations used are: PSII, Photosystem II; -DM, dodecyl--maltoside; Chl, chlorophyll; CP, chlorophyll protein; DCCD, dicyclohexylcarbodiimide; MES, 2-[morpholino]ethanesulfonic acid; LHCII, light-harvesting complex II; VSB, vibronic side band; ZPL, zero phonon line.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

				A. Boussac, F. Rappaport, P. Carrier, J.-M. Verbavatz, R. Gobin, D. Kirilovsky, A. W. Rutherford, and M. Sugiura Biosynthetic Ca2+/Sr2+ Exchange in the Photosystem II Oxygen-evolving Enzyme of Thermosynechococcus elongatus J. Biol. Chem., May 28, 2004; 279(22): 22809 - 22819. [Abstract] [Full Text] [PDF]

				L. Bao, C. Kaldany, E. C. Holmstrand, and D. H. Cox Mapping the BKCa Channel's "Ca2+ Bowl": Side-chains Essential for Ca2+ Sensing J. Gen. Physiol., April 26, 2004; 123(5): 475 - 489. [Abstract] [Full Text] [PDF]

				M. Gastaldelli, G. Canino, R. Croce, and R. Bassi Xanthophyll Binding Sites of the CP29 (Lhcb4) Subunit of Higher Plant Photosystem II Investigated by Domain Swapping and Mutation Analysis J. Biol. Chem., May 23, 2003; 278(21): 19190 - 19198. [Abstract] [Full Text] [PDF]

				P. Dominici, S. Caffarri, F. Armenante, S. Ceoldo, M. Crimi, and R. Bassi Biochemical Properties of the PsbS Subunit of Photosystem II Either Purified from Chloroplast or Recombinant J. Biol. Chem., June 14, 2002; 277(25): 22750 - 22758. [Abstract] [Full Text] [PDF]

				H. Teramoto, T.-a. Ono, and J. Minagawa Identification of Lhcb Gene Family Encoding the Light-harvesting Chlorophyll-a/b Proteins of Photosystem II in Chlamydomonas reinhardtii Plant Cell Physiol., August 1, 2001; 42(8): 849 - 856. [Abstract] [Full Text] [PDF]