Role of the N- and C-terminal regions of the PufX protein in the structural organization of the photosynthetic core complex of Rhodobacter sphaeroides Francesco Francia 1,2 , Jun Wang 1, *, Hans Zischka 1, †, Giovanni Venturoli 2 and Dieter Oesterhelt 1 1 Department of Membrane Biochemistry Max-Planck-Institute for Biochemistry, Martinsried, Germany; 2 Department of Biology, Laboratory of Biochemistry and Biophysics, University of Bologna, Italy The core complex of Rhod obacter sphaeroides is formed by the a ssociation o f the light-harvesting antenna 1 (LH1) and the reaction center (RC). The P ufX protein is essential for photosynthetic gr owth; i t i s located within the core in a 1 : 1 stoichiometry with the RC. PufX is required for a fast ubiquinol exchange between the Q B site of the RC and the Qo site of the cytochrome bc 1 complex. In vivo the LH1– PufX–RC complex is assembled in a dimeric form, where PufX is involved as a structural organizer. We have modified the PufX protein at the N and the C-terminus with pro- gressive deletions. T he nine mutants obtained have b een characterized for th eir ability f or photosynthetic growth, the insertion of PufX in the core LH1–RC complex, the stability of the dimers and the kinetics of flash-induced reduction of cytochrome b 561 of the cytochrome bc 1 complex. Deletion of 18 residues a t t he N-terminus destabilizes the dimer in vitro without preventing photo synthetic growth. The dimer (or a stable dimer) does not seem to be a necessary requisite for the photosynthetic phenotype. Partial C-terminal d eletions impede the insertion of PufX, while the complete absence of the C -terminus leads to t he insertion of a PufX protein composed of only its first 53 residues and does not affect the photosynthetic growth of the bacterium. Overall, the results point to a complex role of the N and C domains in the structural organization of the core complex; the N-terminus is suggested to be responsible mainly for d imerization, while the C-terminus is thought to be involved mainly in PufX assembly. Keywords: LH1-RC; photosynthesis; PufX; Rhodobacter sphaeroides. The purple b acterium Rhodobacter (Rb.) sphaeroides can grow photosynthetically or heterotrophically via aerobic or anaerobic respiration. When growing photosynthetically, it uses light energy as a driving force to form ATP via a cyclic electron transfer. Photons are captured from the light- harvesting (LH) complex(es) and the excitation energy funnelled towards a bacteriochlorophyll (BChl) special pair (P), located in the reaction centre (RC). The excited P delivers an electron via an accessory BChl and a bacterio- pheophytin molec ule to a primary ubiquinone acceptor (Q A ). In a m uch slower reaction the electron is transferred to a second ubiquinone acceptor (Q B ). The full reduction of the quinone m olecule at Q B to quinol requires a second photoexcitation of the RC and is coupled to the uptake of two protons from the cytoplasmic space. The formed ubiquinol dissociates from the RC and is released into the membrane lipid phase [1]. Ubiquinol molecules are oxidized at the Qo site of the cytochrome bc 1 complex (cyt bc 1 ). Here the electron pathway branches into a high and a low potential chain. The first electron reduces in series an iron cluster centre and a cytochrome c 1 in the high potential chain, while the second electron reduces the low poten tial chain, composed of cytochrome b 566 , cytochrome b 561 and a ubiquinone molecu le located at the Qi site. A second ubiquinol oxidized at the Qo site brings the electron to fully reduce the ubisemiquinone to ubiquinol on Qi. From the cytochrome c 1 , the electron is transferred to a soluble cytochrome c 2 that is the physiological electron donor to the oxidized P. From the Qo site, protons are released into the periplasmic space of the cell. This cyclic mechanism of redox reactions acts as a proton pumping system, moving protons from the cytoplasmic to the periplasmic space. The formed H + gradient is the driving force for synth esis of A TP that is used to power the metabolic reactions in the cell [2]. In Rb. sphaeroides the ability of the RC to capture light energy is largely increased by the presence of two LH complexes: LH1 and LH2. The LH1 complex is intimately associated with th e RC in a fixed stoichiometry to form the core complexes (LH1–RC), while the LH2 is arranged peripherally with respect to the core. Both LHs are organized in circular supramolecular complexes, resulting from the repetition of a minimal building block commonly Correspondence to F. Francia, Department of Biology, Laboratory of Biochemistry and Bi ophysics, Un iversity of Bologna, Via Irnerio n.42, 40126 Bologna, Italy. Fax: + 39 051 242576, Tel.: + 39 051 2091300, E-mail: francia@alma.unibo.it Abbreviations: BChl, bacteriochlorophyll; cyt bc 1 , cytochrome bc 1 complex; ICM, intracytoplasmic membranes; LH, light-harvesting complex; PMC, photosynthetic membrane complex; Q A ,Q B , primary/ secondary electron acceptor; Qi, quinone reductase site of the cyt bc 1 complex; Qo, quinol oxidase site of the cyt bc 1 complex; RC, reaction center. *Present address: Department of Plant Biology, The Ohio State Uni- versity, Columbus, OH, USA. Present address: GSF Forschungszentrum, Institut fu ¨ r Humangene- tik, Oberschleißheim, Germany. (Received 6 November 2002, revised 12 February 2002, accepted 13 February 2002) Eur. J. Biochem. 269, 1877–1885 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.02834.x referredtoasthea,b heterodimer. The a and b polypeptides span the membrane with a single hydrophobic a helix. This circular protein scaffold binds the pigments that are maintained in a spatial orientation that maximizes the efficiency of the energy transfer reactions. Structures of the LH2 [3,4] and RC [5], as well as of cyt bc 1 [6] are known at atomic r esolution but came from different organisms; on the contrary, high resolu- tion structural data of the core complex are not yet available. In Rb. sphaeroides and Rb. capsulatus photosynthetic growth requires the presence of the PufX protein [7,8]. When an intact LH1–RC core complex is present, PufX is essential to p romote an efficient ubiquinone/ubiquinol exchange between the RC and cyt bc 1 [9], but is not necessary when the LH1 system is absent or reduced in size [10–12]. T his evidence points to a complex structu ral relationship between the components of the photosynthetic system, in which PufX plays a central role [13]. Recently, several works have indicated that PufX is involved directly in the supramolecular organization of the photosystem: (a) the core complexes of Rb. sphaeroides are organized in a dimeric form [14], in which the p resence of PufX induces a specific orientation of the RC inside the LH1 complex as well as the formation of a long range regular array of LH1– RC in the photosynthetic membrane [15]; (b) biochemical studies have shown that PufX is present in the LH1–RC complex in a 1 : 1 stoichiometry with the RC, and that the dimeric form of the core complex c ould only be isolated in the presence of PufX [16]; (c) the PufX protein has a strong tendency to i nteract with the LH1 a polypeptide, while no interaction was detected with the LH1 b polypeptide [17]; (d) the deletion of PufX increases t he number o f LH1- associated Bchls per RC, suggesting an increased number of a,b heterodimers in the LH1 [18]. Moreover in the presence of PufX, electron density maps of the dimeric LH1–RC show unequivocally interruptions in the LH1 ring encircling theRC.ThetopviewoftheLH1–RCcorecomplex presents two rings of LH1 in close contact forming a pattern which resembles the shape of the letter S ; each interrupted ring contains an electron dense nucleus attributed to the RC [14]. All of these experimental results are consistent with the idea that PufX is responsible for these interruptions, allowing a faster lateral diffusion of ubiquinone/ubiquinol molecules toward/from the RC Q B site. A previous work on Rb. sphaeroides demonstrated the role of the C-terminal amino-acid residues of the LH1 a polypeptide in the organization of the LH1–RC complex [12]. In the present work, we have investigated the possible involvement of the N-terminus and of the C-terminus of PufX in protein–protein interactions stabil- izing the LH1–RC complex. To this aim two sets of mutant strains have been constructed. The N-terminal domain has been progressively shortened by deletions extending from the second residue of the primary sequence, while the C-terminal portion has been progres- sively shortened by introducing stop codons by site- directed mutagenesis. We have obtained information on the involvement of the N- and C-terminal portions of PufX in its insertion in the membrane and dimerization of the core complexes. EXPERIMENTAL PROCEDURES Bacterial strains, plasmid, gene transfer, growth conditions, membrane preparations Bacterial strains and the plasmid used in this w ork were as described previously [16]. Growth conditions for Escherichia coli and Rb. sphaeroides have also been described [16]. All o f the Rb. sphaeroides strains were grown semi-aerobically. Photoheterotrophic growth tests in liquid culture were monitored with a Klett–Summerson colorimeter as described by Farchaus et al. [7]; kanamycin and tetracycline were added at 25 lgÆmL )1 and 2 lgÆmL )1 , respectively. Cultures were illuminated by two 120 W incandescent light bulbs; excessive warming was prevented by placing a 40-cm water bath between the lamps an d the cultures. Intra- cytoplasmic membranes (ICM) were prepared as described previously [19]. PufX mutagenesis The PufX N-terminal deletion and C-termin al stop codon series were constructed using the pRKX plasmid [16] as DNA template a nd introducing the desired mutation/ deletion with the method given by Ausbel et al. [20]. The external primers used for the cited sequential PCR muta- genesis anneal, respectively, 330 bp upstream and 441 bp downstream the pufX gene on the plasmid. This fragment contains the Hin dIII and ClaI (8 bp upstream and 171 bp downstream the pufX gene, respectively) unique sites and allows, after digestion with HindIII and ClaI, the ligation of the final PCR product containing the mutated pufX gene with the pRKX vector digested with the same restriction enzymes. After transformations of E. coli S17-1 cells, single colonies of the putative t ransformants were grown over- night in 5 mL Luria–Bertani media with 10 l M tetracycline. The p RKX derived harbouring mutated pufX gene co n- structs f rom t he E. coli S17-1 cells were introduced into Rb. sphaeroides DQ x/g cells by conjugation [16]. Muta- tions were confirmed by sequencing the plasmids isolated from transformed Rb. s phaeroides cells with the Qiagen Minikit. Isolation of core complexes and SDS/PAGE Core complexes were extracted from (ICM) according to the method described previously [16,21] except that the NaBr washing step was performed a t 0.6 mgÆmL )1 total protein. The concentration of LH1–RC complex in the isolated bands was estimated on the basis of the total photooxid- izable RC measured by flash kinetic spectrophotometry as described before [16]. Aliquots containing the same number of LH1–RC moles were treated with 10 vol. cold acetone/ methanol (7 : 2, v/v), vortexed for 2 min and centrifuged. The organic phase was discarded, and the protein pellet was dried at 40 °C for 30 min. The pellets were redissolved in SDS/PAGE loading buffer to final concentration of 2 l M LH1–RC in all of the samples. SDS/PAGE was carried out accordingly to Scha ¨ gger & Von Jagow [22], with a separating gel o f 19.5% (w/v) acrylamide, 0.5% (w/v) bis-acrylamide. 1878 F. Francia et al. (Eur. J. Biochem. 269) Ó FEBS 2002 MS and sequencing of PufX54* protein Proteins were separated by SDS/PAGE as described above, and stained with Coomassie G250. Gels were washed extensively with H 2 Otoremove residual acid from the destaining process. The b and of interest was cut out with a razor blade and transferred to reaction tubes. Proteins were then subjected to a limited protease treatment overnight (0.5–1 lgÆband )1 endopro- teinase LysC, Roche Molecular Biochemicals) [23]. Peptides were extracted from gel slices by altered incubation with 10% formic acid and acetonitrile. Pooled fractions were dried i n a speed vac concentrato r. D ried peptides were re-dissolved in 10 lL 10% acetonitrile/0.1% trifluoroacetic acid. Between 0.5 and 1 lLwereusedfor MALDI-TOF a nalysis (adapted from [2 4]). The r esidual sample was applied to a reversed-phase HPLC system to separate peptides. Purified peptides were subjected to automated Edman degradation (with kind support of J. Kellermann, Max Planck Institute for Biochemistry, Martinsried). Time resolved spectroscopy on ICM The kinetics o f cytochrome b 561 reduction induced by a single actinic flash were measured under the following conditions: ICM were resuspended in a buffer composed of 50 m M Mops, 100 m M KCl, pH 7.0; valinomycin and nigericin were added at 10 l M to collapse the transmem- brane proton gradient and to avoid spectral interference du e to BChl and carotenoid electrochromic effects; 5 l M antimycin A was used to inhibit the Qi site of the cyt bc 1 . Measurements were performed in a nitrogen atmosphere under controlled redox conditions as described by Venturoli et al. [25]. One micromolar each of phenazine methosulfate and phenazine ethosulfate; 2 l M of 2,3,5,6-tetra-methyl-p- phenylenediamine; 10 l M each of p-benzoquinone, duro- quinone, 1,2-naphthoquinone, 1 ,4-naphthoquinone were used as redox mediators. The experimental apparatus i s as described in Francia et al. [16]. Traces of cytochrome b 561 reduction (Fig. 3) were ana- lysed numerically in terms o f pseudo fir st-order kinetics following an initial lag period, as described by Barz et al. [9]. In order to determine the best fitting parameters, the lag period following the time of the flash was varied stepwise: for each lag period the amplitude and r ate constant of the exponential function were optimized using a nonlinear v 2 minimization routine [26] an d a plot of the minimized v 2 vs. the lag period was constructed for each kinetic trace. For all traces this procedure yielded a minimum reduced v 2 (v 2 min ) between 0.8 and 1.2. The confidence interval in the determined value of the lag was obtained by using an F-statistic to determine the probability p of a particular fractional increase in v 2 according to: v 2 =v 2 min ¼ 1 þ½m=ðn À mÞ Fðm; n À m; 1 À pÞ where m is the number of parameters, n is the number of data points, and F is the upper (1–p) quantile for Fisher’s F distributions with m and (n–m) d egrees of free dom [27]. Confidence intervals within 1 SD (P ¼ 0.68) calculated by this procedure are given in Table 2. These intervals are generally asymmetrical, due to the nonlinear nature of deconvolution. RESULTS Construction of the PufX N-terminally deleted series and C-terminally truncated series To obtain the strains with the mutated PufX protein reported in Table 1, two sets of plasmids were constructed. The first series consists of a progressive deletion at the N-terminus, extending from the second residue of the primary sequence; the second series consists of a progressive truncation of the C-terminal domain of PufX (Fig. 1) obtained by the introduction of stop codons in the gene sequence of pufX.Inallthecases,theRb. sphaeroides host strain was DQ x/g [10]. The pseudo wild-type strain used in this work was obtained reintroducing the complete puf operon via the plasmid pRKX (in trans) into the host Rb. s phaeroides DQ x/g, deprived of the chromosomal copy of the puf operon. Table 1. Bacterial strains and plasmids. The plasmid host strain in all the cases was Rb. sphaeroides DQx/g. Strain Plasmid name Wild-type pRKX N-Terminus series PufXD2–4 pRKXD2–4 PufXD2–7 pRKXD2–7 PufXD2–19 pRKXD2–19 PufXD2–26 pRKXD2–26 C-Terminus series PufX54* pRKX54* PufX68* pRKX68* PufX72* pRKX72* PufX76* pRKX76* PufX81* pRKX81* PufDX pRKDX Fig. 1. Nature of the deletions and truncations on PufX. The helix transmembrane r egion of the Pu fX p rotein , pre dicted w ith the program PHDHTM [32], is indicated at the top of the figure and represented as an empty recta ngle in the primary sequ ences of PufX showed below. The related Rb. sphaeroides strains are give n on th e lef t. Ó FEBS 2002 Core complex organization in Rb. sphaeroides (Eur. J. Biochem. 269) 1879 Photosynthetic growth curves Aliquots of bacteria corresponding t o 1 absorbance unit at 700 n m from precultures grown semi-aerobically in dark- ness were transferred to 13 m L final volume of fresh media in 15 mL glass tubes. Air was eliminated from the tubes by using a vacuum water pump; tubes were then exposed to the light in a 30 °C chamber. The results of this photosynthetic assay a re shown in Fig. 2. Mutants of the N -terminally deleted series (Fig. 2A) exhibit photosynthetic growth with the exception of the P ufXD 2)26 strain. The curves in Fig. 2A show a lag phase varying between 15 and 50 h. B y comparing several independent growth curves for each mutant (data not shown), it appeared that a similar, large variability could be observed in any strain including wild- type (compare Fig. 2A,B). Therefore the observed lag phase did not show any correlation with the phenotype. Clearly photosynthetic-negative phenotypes are e videnced by the PufDX (as already reported previously [28]) and PufXD 2)26 curves. A lso t he C-terminus mutants P ufX76*, PufX72* and PufX68* exhibit a nonphotosynthetic phenotype, whereas the mutant with the shortest truncation (PufX81*) is photosynthetically competent (Fig. 2B). Surprisingly, the most extended truncation (mutant PufX54*) does not affect the ability of photosynthetic growth. Kinetics of cytochrome b 561 reduction induced by a single-turnover flash on ICM The rate of electron transfer through the Qo site of cyt bc 1 can be measured in ICM by monitoring the reduction of the cytochrome b 561 induced by a short actinic light fl ash in the presence of the inhibitor a ntimycin A [ 29]. Reduction of the cytochrome b 561 typically shows a lag period prior to the onset of the reaction at its maximal rate. In wild-type ICM, the initial rate of this reaction, as well as the lag phase, depends on the redox state of the ubiquinone pool and on the ubiquinone/RC stoichiometry [2,25]. For a normal size of ubiquinone pool (% 25 ubiquinone molecules/RC )1 ), upon decreasing the ambient redox potential (Eh) from 250 to 100 mV at pH 7.0, the initial rate of cytochrome b 561 reduction increases progressively, while the lag becomes shorter. This behaviour has been attributed to the increased availability of prereduc ed ubiquinone molecule s in t he pool reactingattheQositeofcytbc 1 . K eeping t he Eh high enough, the only ubiquinol molecule which can react at Qo and reduce t he cytochrome b 561 is the one released by the RC following photoexcitation, as the ubiquinone pool is completely preoxidized [30]. Under this condition the lag period is maximal, typically 1 m s in wild-type ICM. A drastic increase of the lag phase, paralleled by a decrease in the initial reduction rate is observed in pufX-deleted strains as compared with wild-type [9]. Both of these effects are maximal at Eh > 180 mV (i.e. when the ubiquinone pool is fully oxidized) and reflect a dramatic impairment in t he redox interaction betw een the Q B site of the RC and the Qo site of the cyt bc 1 in the pufX-deleted strain. We have measured the k inetics of cytochrome b 561 reduction of all the N- and C-terminal PufX mutants (at Eh 180–220 mV) on ICM prepared from cultures grown semi-aerobically the d ark. It must be pointed out that under these growth conditions there is no photosynthetic selective pressure that could induce t he suppression phenomenon reported in [10]. Kinetic traces recorded from the mutants PufXD 2)26 and PufX54* are shown in F ig. 3. The continuous curves are best-fits to an exponential function; the lag dura- tion was d etermined numerically as outlined in Experimental procedures. T he properties of the complete N-terminally deleted and C-terminally truncated series are liste d in Table 2 . While the lag period of PufX54* is comparable to that measured in a typical wild-type, the PufX68*, PufX72*, PufX76* as well as the PufXD 2)26 exhibit an increased lag period usually observed in the pufX-deleted strain. The Pu fXD 2)4 ,PufXD 2)7 and PufX81* show a lag period like that of wild-type, indicating that a short deletion a t the N- and C-terminus does not affect this parameter. Measure- ments carried out on PufX D 2)19 revealed an intermediate lag duration (see Table 2), making ambiguous the attribution of the PufXD 2)19 strain to the wild-type or the PufDXcluster. Isolation of the photosynthetic complexes from the mutant strains As described previously [16], the photosynthetic complexes (PMCs) could be extracted by detergent solubilization from Fig. 2. Growth curves of the control and mutated PufX strains under photosynthetic conditions. The growth of the coltures was monitered by a Klett–Summerson colorimeter. (A) Wild-type (j), PufXD2–4 ( s), PufXD2–7 (n), PufXD2–19 (,), PufDX(h); the growth curve of PufXD2–26 (not shown in the figure for visual clarity) coincides with that of PufDX. (B) Wild-type (j), PufX54*(s), PufX68* (n), PufX72* (,), PufX81* ( e); the growth curve of Pu fX 76* (not shown) coincides with those of PufX68* and PufX72*, i.e. reveals inability of photosynthetic growth. 1880 F. Francia et al. (Eur. J. Biochem. 269) Ó FEBS 2002 the membranes and purified by centrifugation in continuous sucrose density gradients. Briefly, the final wild-type pattern consists of four bands, named PMC1, PMC2, PMC3 a nd PMC4 from top to bottom of the tubes (Fig. 4, tube 1). PMC1, PMC2, PMC3 and PMC4 r epresent LH2, LH1 Ôempty ringsÕ, LH1–RC monomers and LH1–RC dimers, respectively. A consequence o f the deletion of the pufX gene is the lack of LH1–RC dimer bands in the gradients (see Fig. 4A, tube 6). Fig. 4 s hows the patter ns of the N -terminally deleted (Fig. 4 A) and C-terminally truncated (Fig. 4B) series. PufXD 2)4 and the PufX81* (the shortest deletion and truncation, respectively) mutants exhibit the wild-type like pattern, with all four bands present. In the mutant PufXD 2)7 , a very faint band in the correct position of PMC4 could be seen in the original photograph, whereas in the other N-terminally deleted strains, PufX D 2)19 and PufXD 2)26 PMC4 is undetectable. Also in the mutants PufX68*, PufX72* and PufX76* PMC4 is not seen; interestingly a fourth band not clearly separated from PMC3, with a position intermediate between those of PMC3 and PMC4 (between the LH1–RC monomer and the dimer), is present in the PufX54* gradient profile (Fig. 4B, tube 2). SDS/PAGE of the isolated LH1–RC complex The RC : PufX stoichiometry in i solated LH1–RC com- plexes is 1 : 1 in t he wild-type strain. The same unitary stoichiometry was determined in the monomeric (PMC3) and dimeric (PMC4) core complex bands [16]. The presence of the mutated PufX can therefore be assessed in PMC3 isolated from the mutants. Fig. 5A shows an S DS/PAGE of isolated PMC3 from the N-terminally deleted s eries. Only the small molecular weight r egion is s hown. In Fig. 5A, lanes 2 and 3, corresponding to PMC3 from PufXD 2)4 and PufXD 2)7 , respectively, above the dominant LH1 a and b bands, a band attributable to the mutated PufX is clearly visible. In lane 4 a faint band, attributable to PufXD 2)19 is indicated by an arrow, while no band, except those of LH1 a and b, can be seen in the mutant P ufXD 2)26 and in the PufDXstrain. Table 2. Summary of experimental data. Strain Light growth Cytochrome b 561 reduction lag (ms) a Sucrose gradient- isolated PMC Detection of PufX in PMC3 b Wild-type Yes 0.7 (0.2–0.8) PMC 3 ,PMC 4 Yes PufXD2-4 Yes 1.1 (0.9–1.3) PMC 3 ,PMC 4 Yes PufXD2-7 Yes 0.5 (0.0–1.0) PMC 3 (PMC 4 ) c Yes PufXD2-19 Yes 3.8 (3.5–4.2) PMC 3 Yes PufXD2-26 No 5.3 (4.3–5.8) PMC 3 No PufX54* Yes 1.2 (0.5–1.4) PMC 3 ,PMC 3/4 Yes PufX68* No 10.7 (9.5–11.1) PMC 3 No PufX72* No 6.3 (6.2–6.8) PMC 3 No PufX76* No 4.7 (4.0–5.8) PMC 3 No PufX81* Yes 0.7 (0.5–1.0) PMC 3 ,PMC 4 Yes PufDX No 8.8 (8.8–9.1) PMC 3 No a The confidence interval within 1 SD is given in parentheses. b Detected by SDS/PAGE on sucrose gradient-isolated PMC 3. c Detectable as a weak band in the original photograph. Fig. 3. Cytochrome b 561 reduction kinetics induced by single flash pho- toexcitation in ICM. The continuous vertical line indicates the instant when the actinic flash pulse was fired (time ¼ 0), the dotted vertical line marks the beginning of t he cytochrome b 561 reduction at its maximal rat e. Th e t ime i nter val b etween the continuous a nd the dotted vertical lines corresponds to th e l ag p hase of the reduction kinetics. Lag duration was evaluated by a numerical procedure as outlined under experimental procedures. The experimental t race is represented b y a continuous lin e connec ting the points sampled by the rec ording apparatus; the best-fitting mono-exponen tial functio n is indicated by a continuous cu rve. (A) ICM from strain PufXD2-26; (B) ICM from strain PufX54*. Ó FEBS 2002 Core complex organization in Rb. sphaeroides (Eur. J. Biochem. 269) 1881 In Fig. 5B, data for the C-terminally truncated series are shown. The presence of PufX81* is evident in lane 2, while no PufX band was observed in the mutants PufX76*, PufX72* and PufX68* (lanes 3, 4, 5, respectively). AthinbandveryclosetotheLH1a band, indicated by the arrow, is apparent in the PufX54* mutant (lane 6 ). This band was excised from the gel after SDS/PAGE and the protein was identified by using the method described in the Experimental procedures. Briefly, the band was digested wi th 0.5 lg of the proteolytic enzyme e ndopro- teinase Lys-C and parts of the resulting peptide mixture were analysed by MS (MALDI-TOF). The peptide mass fingerprint obtained corresponded to fragments 17–29 (TNLRLWVAFQMMK) and 5–16 (TIFNDHLNTNPK) of the PufX p rotein. The identity of PufX in the band isolated was examined further by subjecting part of t he peptide mixture to separation by reversed-phase HPLC. Purified p eptides w ere subjected to sequence analysis b y Edman degradation. The sequence KTIFNDHLNTN, corresponding to the 4–14 fragment of PufX, was identi- fied. DISCUSSION Effects of N-terminal and C-terminal PufX deletion on LH1–RC dimerization In the LH1–PufX–RC core complex o f Rh. sphaeroides,the RC : PufX stoichiometry is 1 : 1 [16]. The role o f PufX as a structural organizer of the core complex has been discussed recently in several works (rewiewed in [13]). Data on the sequence of assembly of the LH1–PufX–RC complex in vivo [31] and on protein–protein interactions between the single polypeptides o f the complex [17] are consistent with the hypothesis t hat PufX interrupts the continuity of the LH1 ring and switches the structure of the complex from a Ôclosed Õ monomeric form to an ÔopenÕ dimeric form. Moreover, linear dicroism studies have demonstrated the role of PufX in the orientation of the RC inside the LH1 Fig. 4. Isolation of the PMCs on a sucrose gradient. The final detergent extracts from the ICM were loaded on the top of a 10–40% sucrose gradient a nd centrifuged for 1 9 h at 23 0 · 10 3 g. T he gradient was buffered with 50 m M Na-glycylglycine to pH 7.8, the detergents octyl- glucoside and Na-cholate were added to the gradien t at a final con- centration of 0.6% and 0.2% (w/v), respectively. (A) T ube 1, wild-type; tube 2, PufXD2–4;tube3,PufXD2–7; tube 4, PufXD2–19; tube 5, PufXD2–26; tube 6, PufDX. (B) Tube 1, wild-type; tube 2, PufX54*; tube 3, PufX68*; tube 4, PufX72*; tube 5, PufX76*; tube 6, PufX81*. Fig. 5. SDS/PAGE on sucrose g radient-isolated core complexes. The proteins o f the PMC3 bands isolated from the sucrose gradients (see Fig. 4) were sub jected to SDS/PAGE accordin g to Scha ¨ gger & Von Jagow [22]. The concentrations of acrylamide and bis-acrylamide were 19.5% and 0.5% (w/v), respectively, in the separating gel and 3.9% and 0.1% in the stacking gel. For each lane, 24 pmol PMC3, corre- sponding to the monomeric form of the core complex, was loaded. Only the region of low molecular mass proteins is shown i n the figure. (A)Lane1,wild-type;lane2,PufXD2–4; lane 3, PufXD2–7; lane 4, PufXD2–19; lane 5, PufXD2–26; lane 6, PufDX. (B) Lane 1, wild-type; lane 2, PufX81*; lane 3, PufX76*; lane 4, PufX72*; lane 5, PufX68*; lane 6, PufX54*; la ne 7, Puf DX. The position of t he faint band attributed to the PufXD2–19 protein is indicated by an arrow in lane 4 A, th e position of t he detec ted PufX54* is ind icated by a n arrow in lane 6 B. 1882 F. Francia et al. (Eur. J. Biochem. 269) Ó FEBS 2002 [15]. These results indicate that the PufX protein is in contact with the LH1 and the RC subunits inside the core complexes. When secon dary structure prediction was p erformed on PufX [32] the final output revealed a strong tendency to build ahelices at both t he N- and C-termini [33] and a transmembrane a helix in the central region (Fig. 1 ). On this basis, and in view of the finding that the C-terminal part of the LH1 a polypeptide plays an important role in the structure of the core complex [12], we decided to investigate the possible structural role of the N-terminus and the C-terminus of PufX. To this aim, nine strains of Rb. sphaeroides with mutated PufX were constructed. The dimeric form of the core complex purified from ICM [16] has been confirmed by electron microscopy [14]. We consider, therefore, the pr esence of the d imeric form (PMC4) upon isolation as an indication for dimerization in v ivo. The shortest deletion in the PufXD 2)4 and the shortest truncation in PufX81* do not impair the ability of PufX to facilitate dimerization, as a clear PMC4 b and can be detected in the gradient (Fig. 4). Interestingly in the gradient of the N -terminus mutant PufX D 2)7 averyfaintPMC4 band is visible in the original gradient photograph (unde- tectable in Fig. 4). Apparently this deletion strongly desta- bilizes the dimer to th e extent that it cannot withstand fully the m embrane d etergent extraction. The p resence o f t he dimer in vivo in the mutant PufXD 2)7 and presumably in PufXD 2)19 is therefore not excluded. We have shown previously that in vitro an irreversible dissociation of the dimeric to the monomeric form of the complex from the wild-type exists: the dimer dissociates gradually into the monomer when the octyl-glucoside concentration is increased from 0.6 to 1.2% [16]. This result suggested that hydrophobic interactions are involved in maintaining the dimeric form. The data obtained on t he PufXD 2)7 and PufXD 2)19 strains indicate that important protein–protein hydrophobic interactions are made by the Pu fX N-terminus. In the case of the longest N-terminal deletion (strain PufXD 2)26 ), the PufXD 2)26 protein i s not detectable in the core complex (see below and Fig. 5A, lane 5). Correspond- ingly only the monomeric form of the complex can be seen in the gradient (Fig. 4A, tube 5). Two main points of interest a rise from the results obtained from the C-terminal truncation series. First, three mutants (characterized by a nonphotosynthetic phenotype), PufX76*, PufX72* and PufX68* show no dimers of the isolated core complex, whereas from the P ufX54* s train a fourth band, with different sedimentation characteristics on sucrose gradients, has been isolated. In the following we refer to this band, located in an intermediate position between t he monomer (PMC3) and the dimer (PMC4), as PMC3/4. We propose three alternative interpretations: (a) PMC3/4 represents a dimeric form in which the LH1 rings assume a d ifferent curvature, leading t o a different sedi- mentation coefficient; (b) PMC3/4 is formed by two LH1 rings that lost one or two r eaction centers; (c) when t he C-terminal part of PufX is deleted the equilibrium between the monomer and the dimer is not attained during sedimentation. The second interesting point is that PufX76*, PufX72* and PufX68* mutants are photosynthetically incompetent, whereas the PufX54* mutant grows photosynthetically, demonstrating that a complete removal of the C-terminus is tolerated by the cell, while a partial truncation is photosyn- thetically lethal. The absence of PufX in PufX76*, PufX72* and PufX68* (Fig. 5B) could in principle either reflect an impairment in the insertion into the membrane of the shortened protein and/or in the assembly of PufX in the LH1–RC, or resides at transcriptional/post-translational level. The PufX54* protein possesses only the N-terminus and the hydrophobic transmembrane h elix, whereas the other mutants have in addition part of the C-terminus. We suggest that the presence of a partial C-terminus leads to a misfolding that impedes the insertion/assembly of PufX in the membrane complex. Parkes-Loach et al. [34] have recently reported that mature forms of PufX extracted from cells of Rb. sphaer- oides and Rb. capsulatus contains 12 and nine fewer amino acids, respectively, at the C-terminal end of the protein than are encoded by their pufX genes. These data are inconsistent with our previous repo rt [16], where a PufX with a C-terminal six-histidine tail has been used to determine the RC : PufX stoichiometry by Western blot analysis with anti-His6 antibodies. However the genetic background of the strains used is different: in our studies (present paper and [16]) both the LH2 and the LH1 antenna systems a re present, while in the work of P arkes-Loach et al. an LH2 – , LH1 – strain and an LH2 – strain from Rb. sphaeroides and from Rb. capsulatus, respectively, have been used to extract PufX. We can suppose that the discrepancy is related to the presence of the LH2 which could influence the shortening processes of the assembled PufX protein. The exchange of ubiquinone between the RC and the cyt bc 1 in the presence of mutated PufX protein The role of the PufX protein in facilitating the ubiquinone/ ubiquinol exchange between the Q B site of the RC and the ubiquinone pool has been demonstrated in Rb. sphaeroides wild-type strains [7,8]. It has been proposed that PufX facilitates ubiquinone exchange by determining the struc- tural supramolecular organization o f the LH1–PufX –RC complex [12]. In this work, PufX has been detected by SDS/PAGE in core complexes (Fig. 5) isolated from the N -terminus mutants PufXD 2)4 ,PufXD 2)7 ,PufXD 2)19 and from the C-terminus mutants PufX54*, PufX81*. The evidence that these a re the only mutants which are ph otosynthetically competent ( see Table 2) is in accordance with previous results on the requirement of the PufX protein for photosynthetic growth and suggests that the assembly of the wild-type or mutated PufX protein in the core complex is necessary for efficient light energy transduction. In the other mutants examined, PufXD 2)26 , PufX68*, PufX72* and PufX76* no PufX protein could b e d etected on SDS/ PAGE after isolation of the complex. Assaying on ICM the reduction kinetics of the cyto- chrome b 561 induced by a single actinic flash in the mutants PufXD 2)4 ,PufXD 2)7 , PufX54* and PufX81*, we found a lag time betwee n the fl ash excitation a nd the onset of cytochrome b 561 reduction close to that observed in wild- type ICM. This is indicative of a fast ubiquinone exchange between the reaction center Q B site and the cyt bc 1 Qo site. In the case of the shortest N-terminal deletion and C-terminal truncation (PufXD 2-4 and PufX81*, respectively) Ó FEBS 2002 Core complex organization in Rb. sphaeroides (Eur. J. Biochem. 269) 1883 this result was expected; in these two mutants a d imeric form of the core complex could be i solated. On the contrary, we obtained evidence of a less stable dimer in mutan t PufXD 2)7 and observed a band intermediate between that of the monomeric and the dimeric form ( see above) in mutant PufX54*. As in these last two mutants a short lag was observed (see Table 2), apparently the presence of a stable dimer is not a n ecessary requisite for a fast RC/bc 1 redox interaction, which is associated with a photosynthetic phenotype. As an alternative e xplanation the monomeric a nd the dimeric form of t he LH1–PufX–RC could both be presen t in vivo; i n the presenc e of an intact PufX the dimeric form would prevail, while altered equilibria arising from muta- tions on PufX could affect the stationary concentration o f the dimer in the membranes. In the PufXD 2)19 strain the dimeric form is even more destabilized, as no PMC4 can be isolated. Measurements of the lag in cytochrome b 561 reduction in ICM from this mutant yielded values intermediate between those usually obtained in t he wild-type and in the P ufDX s train, with some variability between preparations from different cul- tures. Considering t hat the same amount of LH1–RC has been loaded in all lanes of the SDS/PAGE gel in Fig. 5A, the weaker intensity of the PufXD 2)19 band (lane4) suggests that the amount of PufX per LH1–RC complex is lower in this mutant. T here fore it is likely that a mixture of monomeric LH1–RC with and without PufXD 2)19 is isolated on the sucrose gradient. The occurrence of a mixed population of LH1–RC core complex in the ICM would explain the variability i n the duration of the lag of cytochrome b 561 reduction kinetics. The presence of the PufXD 2)26 within the isolated core complex cannot be excluded in the SDS/PAGE shown in Fig. 5A, due to a possible overlapping with the a subunit of the LH1 complex. However a significant efficiency of PufXD 2)26 insertion in the core complex seems unlikely, due to the nonphotosynthetic phenotype of this strain and to the pronunced lag in the cytochrome b 561 reduction kinetics (see Table 2 ), systematically found in chromatophores from the pufX-deleted strain. Organization of the Q-cycle complexes The dimeric organization of the LH1–PufX–RC has been demonstrated directly in the membranes of Rb. sphaeroides by electron microscopy [14]. In this paper, the authors tentatively attribute a positive electrondense region in the two-dimensional projection o f the dimer to cyt bc 1 and interpret the S-shaped structure of the projection map as a supercomplex formed by the LH1–RC and cyt bc 1 in a 2 : 1 stoichiometr y. Some considerations on this point can be made in the light of our results. The presence in vitro of a less stable dimer in the mutant PufXD 2)7 neither affects the photo- synthetic capability of the bacteria nor the efficiency of exchange of the ubiq uinol molecules between th e RC and the bc 1 , a s judged from the reduction kinetics of cyto- chrome b 561 measured in ICM. Also the mutant PufD 2)19 , in which the dimeric form cannot be detected in the isolated core complex, exh ibits a p hotosynthetic phenotype. In these two mutants, t he photosynthetic phenotype suggests the presence in viv o of an o pen monomeric complex (or a prevalence of it with respect to the wild-type situation), consisting of an incomplete single LH1 ring containing one RC. The photosynthetic ability in P ufX54* is consistent with the fast RC/bc 1 ubiquinol exchange observed in ICM; on the other hand, the structural organization of the core complexes and/or the possible monomer–dimer equilibrium seem to be appreciably perturbed a lso in this m utant as judged from the d ifferent position of t he PMC3/4 band (Fig. 4 B) after isolation of the photosynthetic complexes on linear sucrose gradien t. It is possible that the dimer form is not required as long as a r eorganized core complex can efficiently shuttle quinones between the RC a nd the bc 1 complex. The PMC3/4 isolated complex stimulates our interest, and further studies are in progress to understand the nature of the PufX54* mutant. In conclusion, our data indicate that both the N- and C-terminal portions of the PufX protein play a complex role in organizing the structure of the LH1–RC complex; the N-terminal region would be responsible mainly for the formation o f a stable dimer, whereas the C-terminal portion would be involved mainly in PufX insertion/assembly. The transmembrane helix region of PufX appears to be sufficient to allow a fast quinone exchange between the core and the cytochrome b 561 complex. Interestingly this c onclusion fits well with the recent work of Parkes-Loach et al. [34], showing that the interaction between the hydrophobic PufX region and the LH1-a polypeptide has an inhibitory effect on the formation of the LH1 complex. This result suggests that the central core of the PufX protein is responsible of the break in the continuity of the LH1 ring in vivo [14], allowing a faster diffusion of the quinone molecules from/toward the RC Q B site. ACKNOWLEDGEMENTS We thank B . A. Melandri and P. Turina for fruitful discussions, C. Wey rauch, U. Schimanko and N. Mele for technical assistance. F. F. and J . W. were recipients of M .P.I. postdoctoral fe llowships. This work was supported b y grant PRIN/99, Bioenergetica e t rasporto di membrana from the Italian MURST and from The Fonds der Chemischen Industrie. REFERENCES 1. Okamura, M.Y., Paddock, M.L., Graige, M.S. & Feher, G. (2000) Proton and electron transfer in bacterial reaction centers. Biochim. Biophys. Acta 1458, 148–163. 2. Crofts, A.R. & Wraight, C.A. (1983) The electrochemical domain of photosynthesis. Biochim. Biophys. Acta 72 6, 149–186. 3. McDermott, G., Prince, S.M., Freer, A.A., Hawthornthwaite- Lawless, A.M., Papiz, M.Z., Cogdell, R.J. & Isaacs, N.W. (1995) Crystal structure of an integral membrane light-harvesting com- plex from photosynthetic bacteria. Nature 374, 517–521. 4. Koepke,J.,Hu,X.,Muenke,C.,Schulten,K.&Michel,H.(1996) The crystal struc ture o f t he light-ha rvesting c omplex II (B800–850) from Rhodospirillum molischianum. Structure 4, 581–597. 5. Feher, G., Allen, J .P., Okamura, M.Y. & R ees, D.C. (1989) Structure and function of bacterial photosynthetic reaction cen- ters. Nature 339, 111–116. 6. Xia,D.,Yu,C.A.,Kim,H.,Xia,J.Z.,Kachurin,A.M.,Zhang,L., Yu, L. & Deisenhofer, J. (1997) Crystal structure of the cyto- chrome bc 1 complex from bovine h eart mitochondr ia. Science 277, 60–66. 1884 F. Francia et al. (Eur. J. Biochem. 269) Ó FEBS 2002 7. Farchaus, J.W., Barz, W.P., Gru ¨ nberg, H. & Oesterhelt, D. (1992) Studies on the expression of the pufX polypeptide and its requirement for ph otoheterotrophic growth in Rhodobacter sphaeroides. EMBO J. 11, 2779–2788. 8. Lilburn, T.G., Haith, C.E., Prince, R.C. & Beatty, J.T. (1992) Pleiotropic effects of pufX gene deletion on the structure and function of the pho tosynthetic apparatus of Rhodobacter capsu- latus. Biochim. Biophys. Acta 1100, 160–170. 9. Barz, W.P., Vermeglio, A., Francia, F., Venturoli, G., Melandri, B.A. & O esterhelt, D. (1995) Role of the PufX protein in photo- synthetic growth of Rhodobacter sphaeroides.2.PufXisrequired for efficient ubiquinone/ubiquinol exchange between the reaction center Q B site and the cytochrome bc 1 complex . Biochemistry 34, 15248–15258. 10. B arz, W.P. & Oesterhelt, D. (1994) Photosynthetic deficiency of a pufX deletion mutant of Rhodobacter sphaeroides is suppre ss ed by point mutations in the light-harvesting complex genes pufB or pufA. Biochemistry 33, 9741–9752. 11. Lilburn, T.G., Prince, R.C. & Beatty, J.T. (1995) Mutation of the Ser2 codon of the light-harvesting B870 alpha polypeptide of Rhodobacter capsulatus partially suppresses the pufX phenotype. J. Bacteriol. 177, 4593–4600. 12. McGlynn, P., Westerhuis, W.H.J., Jones, M.R. & Hunter, C.N. (1996) Consequences for the organization of reaction center-light harvesting antenna 1 (LH1) c ore complexes of Rhodobacter sphaeroides arising from deletion of amino acid residues from the C-terminus of the LH1 alpha polypeptide. J. Biol. Chem. 271, 3285–3292. 13. Loach, P.A. (2000) Supramolecular complexes in photosynthetic bacteria. Proc. Natl Acad. Sci. USA 97, 5016–5018. 14. Jungas, C., Ranck, J L., Rigaud, J L., Joliot, P. & Vermeglio, A. (1999) Supramolecular organization of the photosynthetic appa- ratus of Rhodobacter sphaeroides. EMBO J. 18, 534–542. 15. Frese, R .N., Olsen, J.D., Branvall, R., Westerhuis, W.H.J., Hunter, C.N. & van Grondelle, R. (2000) The long-range supraorganization o f the bacterial photosynthetic unit: a key role for PufX. Proc. Natl Acad. Sci. USA 97, 5197–5202. 16. Francia, F., Wang, J ., Venturoli, G., Melandri, B.A., Barz, W.P. & Oesterhelt, D. (1999) The reaction center-LH1 antenna complex of Rhodobacter sphaeroides contains one PufX molecule which is invol ved in dimerization of this complex. Bi ochemistry 38, 6834–6845. 17. Recchia, P.A., Davis, C.M., Lilburn, T.J., Beatty, J.T., Parkes- Loach, P.S., Hunter, C.N. & Loach, P.A. (1998) Isolation of the PufX protein from Rhodobacter capsulatus and Rhodobacter sphaeroides: evidence for its i nteraction with the alpha-polypeptide of the core light-harvesting complex. Biochemistry 37, 1 1055– 11063. 18. Mc Glynn, P., Hunter, C.N. & Jones, M.R. (1994) The Rhodo- bacter sp haeroides P ufX p rotein is not required for photosynthetic competence in the a bsen ce of a ligh t h arv esting system. FEBS Lett. 349, 349–353. 19. Bowyer, J.R., Tierney, G.V. & Crofts, A.R. (1979) Secondary electron transfer in chromatoph ores of Rhodopseudomonas cap- sulata A1a pho. Binary out -of-p hase osc illations i n ubisemiqui- none formation and cytochrome b 50 reduction with consective light flashes. FEBS Lett. 101, 201–206. 20. Ausbel, M.F., Brent, R., Kingston, R.E., Moore, D.D., Seidman, J.G., Smith, J.A. & Struhl , K. ( 2000) Current Protocols in Molecular Biology. John Wiley & Sons, Inc, US A 21. Gabellini, N., Gao, Z., Oesterhelt, D., Venturoli, G. & Melandri, B.A. (1989) Reconstitution of cyclic electron transport and pho- tophosphorylation by incorporation of the reactio n center, cyto- chrome bc 1 complex and ATP synthase from Rhodobacter capsulatus into ubiq uinone-10/p hospho lipid v esc icles. Bioc him. Biophys. Acta 974, 202–210. 22. Scha ¨ gger, H. & Von Jagow, G. (1987) Tricine-sodium dodecyl sulfate-polyacrylamide gel electrophoresis for the separation of proteins in the ran ge from 1 to 1 00 kDa. Anal. Biochem. 166, 368–379. 23. Houthaeve, T., Gausepohl, H., Mann, M. & Ashman, K. (1995) Automation of micro-preparation and enzymatic cleavage of gel electrophoretically separated proteins. FEBS Lett. 376, 91–94. 24. Shevchenko,A.,Wilm,M.,Vorm,O.&Mann,M.(1996)Mass spectrometric sequencing of proteins silver-stained polyacryla mide gels. Anal. Chem. 68, 850–858. 25. Venturoli, G., Ferna ´ ndez-Velasco, J.G., Crofts, A.R. & Melandri, B.A. (1986) Demonstration of a collisional interaction of ubiqui- nol with the ubiquinol–cytochrome c 2 oxidoreductase complex in chromatophores from Rhodobacter sphaeroides. Biochim. Biophys. Acta 851, 340–352. 26. Bevington, P.R. (1969) Data Reduction and Error Analysis in the Physical Sciences.McGraw-Hill,NewYork. 27. Beechem., J.M. (1992) Global analysis of b iochemical and bio- physical data. Methods Enzymol. 210, 37–54. 28. Farchaus, J.W., Gru ¨ nberg, H. & Oesterhelt, D . (1990) Com- plementation of a reaction center-deficient Rhodobacter sphaer- oides pufLMX deletion strain in trans with pufBALM doe s not restore the photosynthesis-positive phenotype. J. Ba ct eriol. 172, 977–985. 29. Crofts, A.R., Meinhardt, S.W., Jones, K.R. & Snozzi, M. (1983) The role o f the quinone pool in the cyclic electron-transfer chain of Rhodopseudomonas sphaeroides. Biochim. Biophys. Acta 723, 202–218. 30. T akamiya, K.I. & D utton, P.L. (1979) Ubiq uinone in Rho- dopseudomonas sphaeroides. Some thermodynamic properties. Biochim. Biophys. Acta 546, 1–16. 31. Pugh,R.J.,McGlynn,P.,Jones,M.R.&Hunter,C.N.(1998)The LH1–RC core complex of Rhodobacter sphaeroides: interaction between components, time-dependent assembly, and topology of the PufX protein. Biochim. Biophys. Acta 1366, 301–316. 32. Rost, B., Casadio, R ., Fariselli, P. & Sander, C. (1995 ) Transmembra ne helices predicte d at 95% accura cy. Protein S ci. 4, 521–533. 33. Francia, F ., Turina, P., Melandri, B .A. & Venturoli, G. (1998) The molecular role of the PufX protein in bacterial photosynthetic electron transfer. In Biophysics of Electron Transfer and Molecular Bioelectronics (Nicolini, C., ed.), pp. 103–116. Plenum Press, New York. 34. Parke s-Loach, P.S., L aw, C .J., Recc hia, P.A., Keh oe, J., Ne hrlich, S., Chen, J. & Loach, P.A. (2001) Role of the core region of the PufX protein in inhibition of reconstitution of the core light-har- vesting complexes of Rhodobacter sphaeroides and Rhodobacter capsulatus. Biochem ist ry 40, 5593–5601. Ó FEBS 2002 Core complex organization in Rb. sphaeroides (Eur. J. Biochem. 269) 1885 . between the RC and the cyt bc 1 in the presence of mutated PufX protein The role of the PufX protein in facilitating the ubiquinone/ ubiquinol exchange between the Q B site of the RC and the ubiquinone. Role of the N- and C-terminal regions of the PufX protein in the structural organization of the photosynthetic core complex of Rhodobacter sphaeroides Francesco Francia 1,2 ,. important role in the structure of the core complex [12], we decided to investigate the possible structural role of the N-terminus and the C-terminus of PufX. To this aim, nine strains of Rb. sphaeroides