Heterologous Production of the Photosynthetic Reaction Center and Light Harvesting 1 Complexes of the Thermophile Thermochromatium tepidum in the Mesophile Rhodobacter sphaeroides and Thermal Stability of a Hybrid Core Complex

ABSTRACT

The photosynthetic complexes of the thermophile Thermochromatium tepidum are of considerable interest in biohybrid solar cell applications because of the ability of thermophilic proteins to tolerate elevated temperatures. Synthetic operons encoding reaction center (RC) and light harvesting 1 (LH1) pigment-protein complexes of T. tepidum were expressed in the mesophile Rhodobacter sphaeroides. The T. tepidum RC (TRC) was assembled and was found to be functional with the addition of menadione to populate the Q_A pocket. The production of T. tepidum LH1 (TLH1) was increased by selection of a phototrophy-capable mutant after UV irradiation mutagenesis, which yielded a hybrid RC-TLH1 core complex consisting of the R. sphaeroides RC and T. tepidum TLH1, confirmed by the absorbance peak of TLH1 at 915 nm. Affinity chromatography partial purification and subsequent sucrose gradient analysis of the hybrid RC-TLH1 core complex indicated that this core complex assembled as a monomer. Furthermore, the RC-TLH1 hybrid core complex was more tolerant of a temperature of 70°C than the R. sphaeroides RC-LH1 core complexes in both the dimeric and monomeric forms; after 1 h, the hybrid complex retained 58% of the initial starting value, compared to values of 11% and 53% for the R. sphaeroides RC-LH1 dimer and monomer forms, respectively.

IMPORTANCE This work is important because it is a new approach to bioengineering of photosynthesis proteins for potential use in biophotovoltaic solar energy capture. The work establishes a proof of principle for future biohybrid solar cell applications.

INTRODUCTION

One of the challenges of using biological proteins as light-harvesting material in photovoltaic applications is their susceptibility to degradation due to thermal energy. The effects of heat and the degradation kinetics have been documented for the photosynthetic reaction center (RC) from the purple nonsulfur Rhodobacter sphaeroides alphaproteobacterium (1). Several possible solutions have been investigated to mitigate irreversible denaturation. For example, the use of the detergent N-dodecyl β-d-maltoside (DDM) in place of lauryldimethylamine-N-oxide (LDAO) or styrene maleic acid copolymer has yielded greater stability than the use of RCs in a cellular membrane or LDAO detergent (2); additionally, hydrogen bonds have been added to further increase the thermal stability of the RC (3).

Thermochromatium tepidum is a gammaproteobacterium isolated from a hot spring in Yellowstone National Park, USA, which grows optimally at 50°C, in contrast to the mesophile R. sphaeroides, which grows optimally at 30°C (4). There have been studies elucidating the properties that confer heat tolerance to the T. tepidum photosynthetic machinery (5), and such properties could be exploited in applications such as biohybrid photovoltaic cells operating at elevated temperatures.

Like R. sphaeroides, T. tepidum produces an RC (here called the TRC), but the TRC is more similar to that from Blastochloris viridis (6). As in the case of the R. sphaeroides RC, the TRC is comprised of three polypeptides (H, L, and M) and 10 cofactors, including bacteriochlorophylls (BChls), bacteriopheophytins (BPhes), quinones, a carotenoid, and a nonheme iron. However, in contrast to the R. sphaeroides RC but like the B. viridis RC, the TRC is associated with a tightly bound, tetraheme cytochrome (cyt) c on the periplasmic side. Additionally, like B. viridis, T. tepidum produces two different quinones, such that ubiquinone is found in the Q_B pocket whereas menaquinone binds in the Q_A pocket. Lastly, a different type of carotenoid, spirilloxanthin/2,2′-diketo-spirilloxanthin, is produced and assembled in the TRC, in lieu of the spheroidene/spheroidenone present in R. sphaeroides (7).

R. sphaeroides produces light harvesting complex 1 (LH1) and LH2, as does T. tepidum (here called TLH1 and TLH2). T. tepidum TLH1 is comprised of α and β polypeptide subunits and forms a monomeric core complex with the TRC similar to that of Rhodopseudomonas palustris (8). Unlike the dimeric core complex of R. sphaeroides, there is no PufX homologue; therefore, the T. tepidum monomeric core complex is a single TRC surrounded by 16 TLH1 αβ subunit heterodimers. It is thought that the role of PufX in RC-LH1 core complexes of Rhodobacter species is to permit quinone diffusion through the LH1 (9). In the absence of a PufX-like protein (or of a ω-like protein, as in R. palustris) to enable quinones to diffuse through the LH1 ring, a different mechanism appears to be used by T. tepidum. The crystal structure of the T. tepidum core complex was recently solved and shows that the TLH1 ring contains channels between the partners in each αβ protein pair through which quinones are thought to diffuse (7).

It is difficult to perform genetic work on T. tepidum, in contrast to R. sphaeroides, because there are no plasmids described in the literature that T. tepidum will maintain; furthermore, T. tepidum is an obligate anaerobe and requires high temperature and sulfide for optimal growth, making it technically difficult to cultivate. Hence, making site-directed mutations to study structural and functional aspects of the protein or even adding a His tag to expedite purification would be difficult, if not impossible. To overcome these difficulties in the genetic study of the T. tepidum photosystem, we developed methods to express T. tepidum photosynthesis protein genes in the genetically tractable organism R. sphaeroides (10). Here, we describe these methods and their use to create a thermally stable hybrid core complex.

RESULTS

Recombinant expression of the T. tepidum TRC and formation of a charge-separated state.

We initially investigated whether the TRC could be assembled in an R. sphaeroides strain (RCx^R) that lacks all of the genes encoding RC or LH proteins. Similarly to plasmid pIND4-RC (10), where the R. sphaeroides puhA, pufL, and pufM genes are transcribed as a synthetic operon driven by an IPTG (isopropyl-β-d-thiogalactopyranoside)-inducible promoter, plasmid pIND4-TRC was created with T. tepidum puhA, pufL, and pufM genes (here called TtpuhA, TtpufL, and TtpufM, respectively), encoding a 6-His tag on the C terminus of the RC H subunit. Cell cultures of R. sphaeroides strain RCx^R(pIND4-RC) and RCx^R(pIND4-TRC) grown under aerobic, chemotrophic conditions were induced in the early exponential phase with 1 mM IPTG and reached a cell density at 700 nm of 1 to 2 absorbance units (AU). Expression of T. tepidum genes in R. sphaeroides resulted in impaired growth (data not shown), although induction in the late exponential phase minimized this effect. Figure 1 shows an absorbance spectrum comparison of the levels of the RC and TRC after cells were disrupted by sonication. There was a large amount of chromatophores (pigmented membrane vesicles) released from cells upon lysis, indicative of cytoplasmic membrane invaginations in vivo. The TRC assembled, as indicated by the 800-nm absorbance peak corresponding to the accessory BChls, although at levels lower than the RC. Improperly assembled or unstable RCs would yield an increase in the 760-nm peak relative to the 800-nm peak, indicating free BChls (1), which was not observed.

FIG 1.

Assembly and in vivo levels of the RC and TRC. Plasmids pIND4-RC and pIND4-TRC were expressed in R. sphaeroides RCx^R. (A) Absorbance spectra of broken cells expressing RCs (blue) and TRCs (red). The spectra were normalized to an A₆₅₀ value of 0.2. (B) Relative levels of RC production compared to that seen with the wild-type R. sphaeroides RC (pIND4-RC). Levels were calculated by subtracting the RCx^R baseline from the peak amplitude and normalizing the difference to the wild-type RC value. Error bars represent standard deviations of results from biological replicates (n = 3).

The absorbance spectrum of the purified TRC was similar to that of the RC (Fig. 2) and closely matched native TRC spectra previously reported in the literature (11). As previously described for the native TRC (11), there were a few differences compared to the RC, such as a blue shift of the accessory BChl from 804 to 800 nm and a red shift of the BChl special pair from 865 to 870 nm; the amplitudes and wavelengths of the BPhe peak at 760 nm were similar in the RC and TRC. These spectra showed that a native-like TRC was assembled in R. sphaeroides. As expected, because the only carotenoid biosynthesis genes present were those of the host cell genome, the carotenoid present in the TRC was spheroidenone as determined on the basis of the characteristic absorbance at around 515 nm (not shown).

FIG 2.

Absorbance spectra of the purified RC (blue) and TRC (red). The spectra were normalized to an A₆₅₀ value of 0.2.

The electron-transfer functionality of the TRC was determined by measuring the formation of a charge-separated state in vitro. Upon illumination, charge separation occurs by the rapid transfer of an electron from the special pair to the primary and secondary quinones. One of the defining characteristics of the charge-separated state is the bleaching of the 870-nm P-band (12). A charge-separated state in the TRC was measured, as indicated by a decrease in the amplitude of the P-band, but only partial bleaching was obtained (Fig. 3A), perhaps indicating that the Q_A pocket was not fully populated with menadione because supplementing the RLB medium (LB with the addition of 810 μM MgCl₂ and 510 μM CaCl₂) with 10 μM menadione during growth of cultures was insufficient. In vitro addition of menadione at a 10-fold molar excess resulted in a greatly bleached 870-nm P-band, indicating a complete separation of charge (Fig. 3B), consistent with the idea that the Q_A pocket was not fully populated with menadione during the growth of cultures.

FIG 3.

Absorbance spectra of the Q_y transition region of the purified TRC. The dark-adapted (blue) and light-adapted (red) states were measured using an 870-nm peak-emission LED for illumination without additional menadione (A) and with a 10-fold molar excess of menadione (B). The dark-adapted spectra were normalized to an A₆₅₀ value of 0.2, and the light-adapted spectra were normalized using the same factor.

Recombinant expression of T. tepidum TLH1 in an RC-TLH1 hybrid core complex.

Previous studies showed that the thermal stability of the TRC is due to the presence of the T. tepidum TLH1 and Ca²⁺ ions in the TLH1 structure, rather than to that of the TRC itself (5, 11). Therefore, given that TLH1 is believed to confer stability to the core complex, the TLH1 was expressed in R. sphaeroides as part of a hybrid core complex with the RC to see whether TLH1 could increase the thermal stability of the RC, as it does for the TRC.

Because the pIND4-RC1 plasmid produced RC and LH1 at the native ratio in an operon with all the coding and noncoding genetic elements recognized by R. sphaeroides (10), it was thought that simply replacing the R. sphaeroides coding regions with T. tepidum sequences and maintaining the noncoding genetic material might yield the native stoichiometry of TLH1. The pIND4-RCT1 plasmid was created by replacing the R. sphaeroides pufB and pufA coding regions with the corresponding T. tepidum TtpufB and TtpufA coding sequences; this construct, illustrated in Fig. 4, was designed to create a hybrid core complex containing the R. sphaeroides RC and T. tepidum TLH1.

FIG 4.

Schematic of the T. tepidum and R. sphaeroides gene combinations for protein expression constructs. In the pIND4-RC1 construct, the synthetic operon is comprised of the puhA gene encoding the H subunit of the RC followed by the pufQBALMX genes encoding the LH1 αβ polypeptides and the L and M subunits of the RC. The intercistronic RNA stem-loop located in the intergenic region between pufA and pufL is represented. In pIND4-RCT1, the R. sphaeroides pufB and pufA coding regions were replaced with the analogous sequences from T. tepidum. The R. sphaeroides genetic coding and noncoding elements are outlined in black and the T. tepidum components in gray.

TLH1 levels were measured at the peak absorbance at 915 nm, due to the pair of BChls in each αβ subunit, and it was assumed that the peak amplitude would indicate the degree of TLH1 complex assembly. The TLH1 levels were much lower than those of LH1 (Fig. 5), indicating that despite simply replacing the coding regions of the R. sphaeroides pufQBALMX operon with the corresponding TLH1 pufBA gene sequences, the TLH1 protein levels were much lower than those of the RC-LH1 control. Because the intercistronic mRNA stem-loop that stabilizes pufBA messages (13) was present between TtpufA and TtpufL and because the translation of all these genes was initiated with the R. sphaeroides native Shine-Dalgarno sequences used successfully in pIND4-RC1, it was hypothesized that the issue of low expression originated at the posttranslational level, perhaps during protein complex assembly.

FIG 5.

Assembly and in vitro levels of the RC-TLH1 hybrid core complexes capable of supporting phototrophic growth. The construct pIND4-RCT1 was expressed in R. sphaeroides RCx^R prior to and in RCx^R-TcPS⁺ after UV mutagenesis. (A) Absorbance spectra of broken cells expressing the RC-LH1 dimer core complex (blue), the RC-TLH1 hybrid core complex (red), and the RC-TLH1 hybrid core complex, phototrophy-capable mutant (green). The spectra were normalized to an A₆₅₀ value of 0.2. (B) Relative levels of TLH1 production compared to the wild-type LH1 (pIND4-RC1) level. Levels were calculated by subtracting the RCx^R baseline from the peak amplitude and normalizing the difference with respect to wild-type LH1. These comparisons do not take into consideration the extinction coefficients and thus are rough approximations. Error bars represent standard deviations of results from biological replicates (n = 3).

Increasing the levels of TLH1 by UV mutagenesis and selection for improved phototrophic growth.

Because the nature of the block in TLH1 assembly was unclear, we used random mutagenesis and genetic selection for phototrophic growth, rather than attempting a rational design to genetically manipulate R. sphaeroides to produce more TLH1. Cells of the RCx^R strain containing plasmid pIND4-RCT1 were irradiated with UV for random mutagenesis, and survivors of UV mutagenesis were cultivated under the selective pressure of anaerobic, phototrophic conditions to select for mutants exhibiting improved growth.

After mutagenesis and selection for mutants capable of improved phototrophic growth, a phototrophic mutant was obtained. In this mutant, called RCx^R-TcPS⁺, containing plasmid pIND4-RCT1, there was an increase in the T. tepidum TLH1 levels as shown in Fig. 5A, where the absorbance spectrum of strain RCx^R(pIND4-RCT1) seen before mutagenesis is overlaid on the spectrum of a mutant obtained after selection, RCx^R-TcPS⁺(pIND4-RCT1). However, the levels of LH1 seen with the mutant did not match the wild-type R. sphaeroides levels (Fig. 5B), even though the mutants were able to grow phototrophically (Fig. S1). Clearly, despite the low TLH1 levels, there were enough light harvesting complexes to support growth under the conditions employed. The elevated levels of the RC, relative to TLH1, would have been unlikely to contribute much to phototrophic growth, because R. sphaeroides strains lacking both LH1 and LH2 (i.e., possessing RC only) have impaired growth (14, 15). DNA sequencing of the plasmid did not show any mutations in the photosynthesis genes expressed in trans on pIND4. Therefore, the mutation(s) induced by UV mutagenesis appears to be located in the genome of strain RCx^R-TcPS⁺, and this issue was not investigated further.

The low levels of TLH1 relative to the RC in strain RCx^R-TcPS⁺ could conceivably be explained by the presence of empty or partial TLH1 rings, although there have been no reports of TLH1 rings forming independently of the TRC in T. tepidum. Alternatively, there could be two populations, an RC-only pool and an RC-TLH1 pool, coexisting in the membrane. There is strong evidence that, once LH1 units start assembling around an RC in R. sphaeroides, the circularization process continues until the ring is completed; there has been no evidence to date to support the idea of the formation of a partially encircled RC core complex (16).

Partial purification of RC-TLH1 hybrid core complex.

The detergent DDM did not fully solubilize the core complex (only the RC was solubilized), whereas the detergent Deriphat 160C solubilized well, and so Deriphat 160C was used for the initial solubilization and partial purification. Comparison of the purified RC-TLH1 to the RC-LH1 core complex showed the TLH1 characteristic absorbance peak at 915 nm, whereas the LH1 absorbed maximally at 875 nm (Fig. 6), indicating near-native stoichiometric assembly of an RC-TLH1 hybrid core complex in strain RCx^R-TcPS⁺.

FIG 6.

Absorbance spectra of the purified RC-LH1 dimer (blue) and RC-TLH1 hybrid (red) core complexes. The R. sphaeroides LH1 absorbs maximally at 875 nm, whereas the T. tepidum TLH1 absorbs at 915 nm. The spectra were normalized to an A₆₅₀ value of 0.2.

Role of the PufX protein in the RC-TLH1 hybrid core complex.

The RC-LH1 core complex from R. sphaeroides assembles in vivo as a dimer, due to the presence of protein PufX. Without PufX, the monomeric form of the core complex predominates. Furthermore, PufX appears to create a break in the ring of LH1 surrounding the RC to permit the diffusion of quinone molecules in and out of the core (9). It was shown that Rhodobacter pufX null mutants are impaired in photosynthetic growth capability (17, 18). In T. tepidum, there is no PufX equivalent in the TRC-TLH1 core complex, resulting in the monomeric form. Interestingly, despite having a completely enclosed ring of TLH1, T. tepidum is able to grow photosynthetically, apparently because of a gap between the αβ dimers of the TRC-TLH1 core complex permitting the diffusion of quinones (7).

In expressing TLH1 in R. sphaeroides as part of a synthetic operon containing the pufX gene (Fig. 4), there was no indication from the absorbance spectra of whether the PufX-replete hybrid core complexes were in the dimer form (Fig. 5 and 6). Therefore, as in previous work (19, 20), sucrose gradients were run to determine whether the RC-TLH1 was in a monomeric state or a dimeric state (Fig. 7). The RC-LH1 dimer from a crtD mutant (which yields a greater amount of the dimer) was run as a control to show the sedimentation of the dimer core complex (9), and the RC-LH1 monomer from a strain lacking pufX was run to show the sedimentation of the monomer core complex. The TLH-containing core complex from the PufX⁺ strain did not sediment to a position similar to that of the R. sphaeroides RC-LH1 dimer but instead sedimented to a position more similar to that of the RC-LH1 monomer, as seen in Fig. 7. This result indicates that the RC-TLH1 hybrid is likely in a monomeric form and that the PufX protein does not interact with TLH1.

FIG 7.

Sucrose gradients of the RC-LH1 dimer, RC-LH1 monomer, and RC-TLH1 hybrid core complexes. The RC-LH1 dimer was isolated from a strain lacking crtD, resulting in a green-colored complex containing methoxyneurosporene as the primary carotenoid. The strain from which the RC-LH1 monomer was purified was grown aerobically, under which conditions spheroidenone is produced, creating red coloring. The strain from which the RC-TLH1 hybrid was purified was grown anaerobically, under which conditions spheroidene is synthesized, resulting in green coloring similar to that of the RC-LH1 dimer.

Absorbance spectrum analysis of the heat stability of core complexes.

The purpose of expressing T. tepidum TLH1 with an RC was to explore the feasibility of creating an RC-TLH1 hybrid core complex in R. sphaeroides that might be highly tolerant of the damaging and denaturing effects of heat. The enhanced thermal stability of the TRC in the core complex has been attributed to Ca²⁺ bound to TLH1 protein; hence, if it were possible to increase the stability of the RC with the simple addition of a TLH1 ring structure, it could be highly advantageous in biohybrid solar cell downstream refinements and applications. A secondary benefit of using TLH1 in these applications would be the increased absorbance spectrum range, especially in the infrared (IR) range, which could be used to absorb photons and generate photocurrents (21, 22).

To study thermal stability, all the protein complexes were purified and obtained in a final buffer containing DDM detergent (see Materials and Methods). The protein complexes were subjected to temperatures of 30, 50, 70, and 90°C, and the stability of the complex was measured by absorbance spectroscopy; a decrease in the LH1 875-nm or TLH1 915-nm peak amplitude indicated a loss of BChl from the complex, along with an increase in absorbance of free BChl in solution at 775 nm as shown by data from an experiment performed with R. sphaeroides RC-LH1 (Fig. 8). The decay kinetics of several native and hybrid complexes are given in Fig. 9, and parameters of the two-phase exponential decay fits of the samples at 70°C are listed in Table 1.

FIG 8.

Changes in the absorbance profile of the RC-LH1 dimer due to thermal denaturation at 70°C. Characteristic changes in the absorbance profile due to denaturation are indicated by a decrease of the LH1 875-nm peak (downward-pointing arrow) and an increase in the 775-nm peak corresponding to protein-free BChl and BPhe (upward-pointing arrow). The denaturation progress was measured at the indicated times after the temperature was increased. Samples were normalized to the same starting concentration (A₈₇₅ = 0.2); data represent averages of results from three biological replicates.

FIG 9.

Decay kinetics of the LH1 875-nm peak and the TLH1 915-nm peak due to exposure to heat. The protein samples were incubated at temperatures of 30, 50, 70, and 90°C for up to 60 min. (A) RC-LH1 dimer. (B) RC-LH1 monomer. (C) RC-TLH1 hybrid. (D) Decay profiles of the RC-LH1 dimer (green), the RC-LH1 monomer (red), and the RC-TLH1 hybrid (blue) incubated at 70°C are shown. Lines represent two-phase exponential-decay fits. Error bars represent standard deviations of results from biological replicates (n = 3).

TABLE 1.

Two-phase exponential decay fit parameters at 70°C of the RC-LH1 dimer, RC-LH1 monomer, and RC-TLH1 hybrid core complexes shown in Fig. 9

Mutant	y0a	Afastb	t1/2fast (min)c	Aslowd	t1/2slow (min)e	R2f
RC-LH1 dimer	9 ± 2	53 ± 4	1.7 ± 0.2	38 ± 3	15 ± 3	0.99773
RC-LH1 monomer	53 ± 1	31 ± 2	1.5 ± 0.2	16 ± 2	17 ± 6	0.99544
RC-TLH1 hybrid	43 ± 12	18 ± 2	1.2 ± 0.2	39 ± 10	44 ± 23	0.98881

^ay₀, final minimum value reached at the bottom plateau of the curves.

^bA^fast, maximum value in the fast phase of the curves.

^ct_1/2^fast (min), half-life of the fast phase.

^dA^slow, maximum value in the slow phase.

^et_1/2^slow (min), half-life of the slow phase.

^fR², residual value of the fit.

The RC-TLH1 hybrid core complex consisting of an RC core and TLH1 showed the greatest tolerance to 70°C heat during the 60-min incubation time, because the RC-TLH1 hybrid core complex decayed to only 58% of the starting TLH1 absorbance peak amplitude, compared to 11% and 53% for the RC-LH1 dimer and monomer, respectively. Furthermore, the exponential decay is better fitted by an initial fast decay followed by a slow decay (Fig. S2) than by a single exponential. The RC-TLH1 hybrid core complex shows the greatest slow-phase half-time value, indicating greater stability. The RC-LH1 monomer showed higher stability than the RC-LH1 dimer, because the latter resulted in nearly complete degradation by the end of the heat incubation, indicating that the smaller, compact monomer is more stable than the dimer complex. As seen in the sucrose gradients, the RC-TLH1 hybrid appears to be a monomer and this form may contribute much to overall stability. The further increase in stability over the RC-LH1 monomer is attributed to contributions from TLH1 and the binding of Ca²⁺, as published by others (5, 11). These results indicate that although the monomeric form of the RC-LH1 core complex is more tolerant of high temperature than the dimer, TLH1 produced in concert with the RC in R. sphaeroides yields greater thermal stability than either form of the R. sphaeroides native complex. The pH of the Tris-HCl buffer used in these experiments, and used in other work on the thermal stability of R. sphaeroides complexes (1, 2), decreases about 0.025 units for each 1°C increase in temperature in the range of 25 to 37°C. If the same relationship had been linear up to 70°C, then the pH at that temperature would have been about 6.9, and the RC is stable over this range of pH (23).

DISCUSSION

Characterization of the TRC and in vitro charge separation.

The TRC was produced in and purified from R. sphaeroides; such heterologous expression of purple sulfur bacterial genes has not been reported in the literature previously as far as we know. Although T. tepidum is a purple sulfur bacterium of the gammaproteobacteria, the synthetic mRNA encoding the TRC subunits was translated, and proteins were assembled into complexes in the alphaproteobacterium R. sphaeroides. Figure 2 shows the Q_y transition region of the purified TRC expressed in R. sphaeroides with the better-studied RC for comparison, and the differences in these absorbance spectra are consistent with the previously reported absorbance spectrum of the T. tepidum native RC (11).

The charge-separated state of the TRC, which indicates electron transfer through the A-branch to either the primary or secondary quinone, indicated that the Q_A pocket was partially occupied. This interpretation is supported by the finding that the bleaching increased along with an increased amount of menadione, consistent with the idea that the menadione bound to the Q_A pocket (24).

The successful expression, functional assembly, and purification of the TRC in R. sphaeroides now provide a platform for studying this pigment-protein complex without resorting to the more complicated growth conditions of T. tepidum, sidestepping the need for a mechanism for mobilization of cloned genes from Escherichia coli into T. tepidum (to, for example, express His-tagged or site-directed mutant genes) and obviating the need to cultivate mutant strains of T. tepidum that might be incapable of phototrophic growth.

Reduced levels of TLH1 and selection of phototrophy-capable mutants.

The greatest contributing factor in T. tepidum core complex thermal stability is TLH1 (5). Therefore, the coding regions of the genes encoding TLH1 were used to create plasmid pIND4-RCT1 by replacing the R. sphaeroides pufB and pufA coding regions with T. tepidum sequences and maintaining the R. sphaeroides noncoding genetic elements in a new synthetic puf operon. This strategy was used because the analogous pIND4-RC1 plasmid produces high levels of the RC and LH1 complexes in R. sphaeroides strains lacking puf genes. However, the amount of TLH1 produced was much less than the amount of LH1 produced by the pIND4-RC1 control, based on a crude comparison of the respective absorbance peak amplitudes.

UV irradiation mutagenesis followed by selective pressure for improved growth under anaerobic, phototrophic conditions was used to select for mutants that produced more TLH1. Despite the isolation of a phototrophy-capable mutant after UV irradiation, the TLH1/RC ratio remained lower than that seen with the R. sphaeroides complexes. This presented an interesting case in which there were not enough TLH1 units to encircle each RC, based on the structures of the R. sphaeroides and T. tepidum core complexes and the absorbance spectra (7, 9). Recent data have shown that once LH1 units begin to assemble around a core RC, the assembly goes to completion (16). It was assumed, therefore, that there were two populations of complexes in the RCx^R-TcPS⁺(pIND4-RCT1) strain, consisting of RC-only pools and RC-TLH1 pools. Sucrose gradients confirmed this assumption because a dominant, discrete band corresponding to the monomeric form of the core complex was detected in the case of the RC-TLH1 hybrid core complex; this is in contrast to a smear or broad band in the gradient, which would have been expected if there had been a distribution of RCs that were partially encircled to different extents.

In R. sphaeroides, maximal production of LH1 and LH2 complexes requires the proteins LhaA and PucC, respectively; in the absence of either protein, phototrophic growth is impaired, but not abolished, suggesting that LhaA and PucC are not absolutely required for the formation of core complex (16). No lhaA homologue was found by BLASTp analysis of the T. tepidum genome using a Rhodobacter LhaA protein sequence as the query, although a pucC homologue was found (W. Swingley, personal communication). Whether the T. tepidum PucC is specific for TLH2 or also assembles TLH1 units is an issue that was not explored; however, although the exact functions of LhaA and PucC are not known, both proteins appear to permit or assist in the assembly of the noncognate light harvesting complex in R. sphaeroides (16). It is possible that in R. sphaeroides, LhaA and/or PucC did not recognize the T. tepidum αβ polypeptides, resulting in low-efficiency assembly of TLH1 around the RC. Therefore, in general, heterologous expression of photosynthetic pigment-protein complex genes may be complicated by the absence of cognate assembly factors in the heterologous host cell. As we showed in the case of TLH1 in R. sphaeroides, there is a viable strategy of increasing levels by using mutagenesis to select for mutants exhibiting improved phototrophic growth. Expression in this case may also be improved by inclusion of the tetraheme cytochrome gene TtpufC and/or by the use of a genetically tractable species such as Rubrivivax gelatinosus, which assembles an RC-tetraheme cyt c complex, as a platform (25).

The PufX protein may not be involved in the assembly of RC-TLH1 hybrid core complexes.

In R. sphaeroides, the PufX protein facilitates formation of the dimeric form of the RC-LH1 core complex and produces a gap in the encircling LH1 to allow quinones to diffuse into and out of the complex (9). A monomeric RC-LH1 is produced in a pufX gene knockout and does not support phototrophic growth (17). The T. tepidum core complex, however, is found as a monomer naturally. Despite lacking a PufX homologue, the closed LH1-ring monomeric form supports phototrophic growth, and crystallographic data show gaps between TLH1 αβ dimers that are thought to allow the passage of quinone through the TLH1 ring (7).

The RC-TLH1 variant was capable of phototrophic growth after UV irradiation and selection under anaerobic, phototrophic conditions. Because PufX is required for phototrophic growth of R. sphaeroides, we addressed the issue of whether PufX played a role in the assembly of the hybrid core complex; i.e., was the RC-TLH1 present as a monomer or as a dimer? Sucrose gradients of the control RC-LH1 dimer and RC-LH1 monomer showed the relative differences in sedimentation (Fig. 7). As expected, the RC-LH1 dimer form, which is approximately twice the molecular weight of the monomer, sedimented to a lower position than the monomer. In the case of the RC-TLH1 hybrid core complex, the predominant band migrated to a position similar to that of the RC-LH1 monomer. The other faint bands most likely represent contamination given that solubilized chromatophores of the hybrid core complexes were run in the gradient rather than purified protein; these regions contained pigments but did not appear to contain core complexes. Absorbance spectroscopy of the RC-TLH1 dominant band showed the characteristic RC and TLH1 peaks and ratio. Therefore, it appears that PufX does not associate well with TLH1 and is not necessary for assembly of the core complex.

The monomer form of the hybrid core complex supported phototrophic growth, which indicates that TLH1 in the hybrid core complex assembles and functions as in T. tepidum, with the channels between the TLH1 αβ dimers permitting the diffusion of quinone. Similar results were obtained with R. sphaeroides and Rhodobacter capsulatus pufX mutants in which suppressor mutations that restored phototrophic growth were found in the LH1 α and β polypeptides which were thought to allow quinone diffusion through the LH1 ring in the absence of PufX (26, 27). In the case of the RC-TLH1 hybrid core complex, the DNA sequence of the noncoding regulatory regions and puf genes located between the NcoI and HindIII restriction sites on plasmid pIND4-RCT1 from phototrophy-capable strain RCx^R-TcPS⁺ did not reveal any mutations, indicating that no mutations of plasmid-borne genes were involved.

Thermal stability of hybrid core complexes.

The purpose of creating a hybrid RC-TLH1 was to determine if the hybrid would be more tolerant of heat than the native RC-LH1 dimer. It was determined previously that the TRC has a tolerance of heat similar to that of the RC (28), and the addition of TLH1 and Ca²⁺ ions was needed to increase the thermal stability of the TRC within the core complex (5). Therefore, it was thought that the conceptually simple step of adding a TLH1 ring structure around a core RC might increase thermal stability, thereby establishing a proof of principle for future biohybrid solar cell applications.

The RC-LH1 dimeric and monomeric forms were used as controls. As shown in Fig. 9, about half of the RC-LH1 dimer denatured within 4 min at 70°C and essentially full denaturation was seen after 1 h. As generally expected with proteins from mesophiles, the RC-LH1 dimer did not show great tolerance of heat above 30°C. In contrast, the RC-LH1 monomeric form performed much better than its dimeric counterpart and did not sustain losses of more than 50%, even after 1 h at 70°C. This result may be due to the formation of a rigid core complex surrounded by a LH1 ring and decreased flexibility possibly due to the presence of a single, planar core rather than the larger, bent conformation of the dimer with a gap in the ring structure due to PufX (9), which might be more susceptible to increasing thermal energies.

The controls showed greater stability than that represented by the published data of the TRC-TLH1 core complex at 70°C. It would be expected that TRC-TLH1 from the native organism would be the most tolerant of heat, but TRC-TLH1 showed complete denaturation within 20 min (5), whereas the RC-LH1 dimer denaturation took about an hour, and the denaturation seen with the RC-LH1 monomer did not go below 50% during the same period. One possible reason for the increased thermal enhancement of the mesophilic complexes may have been the use of the nonionic DDM detergent in buffer solutions, whereas the zwitterionic detergent N-decylphosphocholine was used in the previous work on the native TRC-TLH1 (5). A previous report suggested that DDM has a stabilizing effect on the RC compared to zwitterionic detergents, such as LDAO (2).

The RC-TLH1 hybrid core complex showed even greater heat tolerance than the RC-LH1 monomer at 70°C throughout the duration of the 1-h experiment (Fig. 9D). The greater heat tolerance over the RC-LH1 monomer is attributed to the TLH1 that substituted for the LH1. The fast-phase exponential-decay component may be due to the presence of a small population of core complexes that did not assemble properly and that were thus the most susceptible to the denaturing effects of heat. The slow-phase half-time constant of the RC-TLH1 hybrid form indicates that it is more stable than the RC-LH1 dimer. This may be due in part to a synergistic, stabilizing effect of the combination of the monomeric, planar form of the hybrid core complex and the addition of the thermally stable TLH1.

The results of these experiments show that, in principle, the strategy of adding heat-stable TLH1 to an RC core can be used to improve the heat tolerance of the core complex. Even though an RC-LH1 monomer has almost the same tolerance of heat as the RC-TLH1 hybrid core complex, the fact that the monomer does not provide phototrophic support (i.e., passage of quinones through the ring of LH1) limits such a system for biohybrid solar cell applications using an electrolyte containing charge mediators, whereas the RC-TLH1 hybrid core complex allows for a greater range of configurations.

These initial results open the way for investigating other changes to improve heat stability. A TRC-TLH1 that contains the bound tetraheme cyt c may yield even better tolerance of heat, given that the protein-protein interactions between the TRC and TLH1 found naturally could be more thermally stable than the interactions in the hybrid complex of RC and TLH1. Furthermore, other groups have investigated the use of detergents, such as DDM, or nondetergent molecules such as styrene maleic acid copolymer to stabilize photosynthetic pigment-protein complexes in the presence of light or heat (2). We found that other strategies, such as using nanodisc scaffold proteins, also increased the heat stability of the RC (data not shown). Another approach would be to use homologous complexes from other thermophilic species, such as Chloroflexus aurantiacus. Both the R. sphaeroides RC and T. tepidum TRC are not stable for long times at temperatures over 40°C, whereas the RC from C. aurantiacus is able to withstand a temperature of 65°C for at least 15 min in vitro (29). A combination of these approaches may result in a greatly heat-stable complex suitable for applications in solar biophotovoltaic technologies, where degradation due to heat is an issue.

Conclusion.

There are several advantages of producing T. tepidum pigment-protein complexes in R. sphaeroides. First, there are genetic tools, such as controlled expression plasmids, that can be used to generate synthetic operons with different T. tepidum gene combinations for expression in R. sphaeroides. Second, it is logistically simpler and faster to work with and cultivate R. sphaeroides in large culture volumes, because the organism grows aerobically or semiaerobically in the nonilluminated facilities available in most microbiology laboratories, whereas T. tepidum is an obligate anaerobe and requires 50°C and illumination for optimal growth. The successful expression of T. tepidum TRC and TLH1 genes in R. sphaeroides has yielded a tractable platform that may be used to easily make modifications in T. tepidum genes for future functional and structural studies. Finally, work on the expression of the hybrid RC-TLH1 core complex in R. sphaeroides has shown that, as a proof of concept, TLH1 can be added to an RC to yield a hybrid core complex that is more tolerant of the denaturing effects of thermal energy than the native RC-LH1 core complex of R. sphaeroides.

MATERIALS AND METHODS

Bacterial strains and plasmids.

A list of strains is given in Table 2 and a list of plasmids in Table 3. E. coli DH5α was used for cloning and site-directed mutagenesis. E. coli was grown at 37°C in lysogeny broth (LB), defined as 1% tryptone, 0.5% yeast extract, and 1% NaCl. The media were supplemented as appropriate with 150 μg/ml ampicillin for the pTZ18U, pTZ19U, and pTZN cloning vectors or with 50 μg/ml kanamycin for pIND4 derivatives. The R. sphaeroides strain has been subjected to genome editing to delete all genes encoding RC and light harvesting (LH) proteins and the oxygen-responsive photosynthesis gene regulator PpsR (10). R. sphaeroides cultures were grown at 30°C in LB or in a modified medium, RLB, which is LB with the addition of 810 μM MgCl₂ and 510 μM CaCl₂ (10). For anaerobic, photoheterotrophic growth, cultures were inoculated into 16.5-ml screw-cap tubes or plates in anaerobic jars and grown in temperature-controlled aquariums at 30°C and 75 μE M⁻² s⁻¹ light intensity with illumination provided by tungsten filament lamps. Large-scale aerobic cultures for protein purification were grown in Erlenmeyer flasks filled to 25% of the nominal capacity with RLB medium and shaken at 200 rpm. Media were supplemented as appropriate with 25 μg/ml kanamycin for pIND4 derivatives, 1 mM IPTG for gene expression, or 10 μM menadione for expression of the genes encoding the TRC.

TABLE 2.

Bacterial strains

Strain	Description of phenotypea	Reference or source
E. coli
DH5α	Chemically competent	30
DH5α λpir	Chemically competent, pir	Biomedal (Spain)
R. sphaeroides
RCxR	RC− LH1− LH2− RshI− PpsR− Rifr	10
RCxR-TcPS+	RC− LH1− LH2− RshI− PpsR− Rifr PS+	This study
ΔcrtD	Lacks methoxyneurosporene dehydrogenase, accumulates methoxyneurosporene	This study (spontaneous mutation; isolated as previously described [9])

^aRif^r, rifampin resistant; PS⁺, phototrophy capable.

TABLE 3.

Plasmids^a

Plasmid	Description	Relevant genes	Source or reference
pTZ18U/pTZ19U	Cloning vector, Apr		Pharmacia (USA)
pTZN	pTZ18U with the SphI site changed to NcoI		10
pTZN::TtpuhA	pTZN containing the T. tepidum TtpuhA gene (0.8 kb) as a NcoI-BamHI fragment with a 6-His tag on the C terminus		This study
pTZ18U::TtpufL	pTZ18U containing the T. tepidum TtpufL gene (0.9 kb) as a BamHI-XbaI fragment		This study
pTZ18U::TtpufM	pTZ18U containing the T. tepidum TtpufM gene (1.0 kb) as a XbaI-HindIII fragment		This study
pTZ18U::B−TtpufM	pTZ18U::TtpufM with an internal BamHI site removed		This study
pTZ19U::TtpufLM	pTZ19U containing the BamHI-XbaI fragment from pTZ18U::TtpufL (0.9 kb) and the XbaI-HindIII fragment from pTZ18U::B−TtpufM (1.0 kb)		This study
pTZ19U::pufQBALMX	pTZ19U containing the R. sphaeroides pufQBALMX operon (3.2 kb) as a BamHI-HindIII fragment		This study
pTZ19U::Rs TtpufBA pufLMX	pTZ19U containing the R. sphaeroides pufQBALMX operon (3.2 kb) as a BamHI-HindIII fragment and pufBA replaced with T. tepidum TtpufBA		This study
pIND4	R. sphaeroides expression vector, Knr		31
pIND4-RC	pIND4 containing the R. sphaeroides puhA gene (0.8 kb) as a NcoI-BamHI fragment with a 6-His tag on the C terminus and the pufLM operon (1.8 kb) as a BamHI-HindIII fragment	puhA, pufL, pufM	10
pIND4-RC1	pIND4 containing the R. sphaeroides puhA gene (0.8 kb) as a NcoI-BamHI fragment with a 6-His tag on the C terminus and the pufQBALMX operon (3.2 kb) as a BamHI-HindIII fragment	puhA, pufL, pufM, pufB, pufA	10
pIND4-TRC	pIND4 containing the T. tepidum puhA gene (0.8 kb) as a NcoI-BamHI fragment with a 6-His tag on the C terminus and the TtpufLM operon (1.9 kb) as a BamHI-HindIII fragment	TtpuhA, TtpufL, TtpufM	This study
pIND4-RCT1	pIND4 containing the R. sphaeroides puhA gene (0.8 kb) as a NcoI-BamHI fragment with a 6-His tag on the C terminus and the pufQBALMX operon (3.2 kb) as a BamHI-HindIII fragment and pufBA replaced with T. tepidum TtpufBA	puhA, pufL, pufM, TtpufB, TtpufA	This study

^aAp^r, ampicillin resistant; Kn^r, kanamycin resistant.

A list of primers used to clone the T. tepidum genes is shown in Table 4. The TtpuhA gene was PCR amplified from T. tepidum chromosomal DNA using Phusion DNA polymerase (NEB) and the TpuhA NcoI-BamHI primer set and was cloned into pTZN as an NcoI-BamHI fragment. A 6-His tag was introduced at the C terminus of the TtpuhA gene. TtpufL (TpufL BamHI-XbaI primer set) and TtpufM (TpufM XbaI-HindIII primer set) were similarly amplified and cloned into pTZ18U, resulting in pTZ18U::TtpufL and pTZ18U::TtpufM, respectively. Artificial ribosomal binding site sequences were introduced for both TtpufL and TtpufM. The BamHI site in pTZ18U::TtpufM was silently removed by site-directed mutagenesis using the Vent polymerase (NEB) and the TpufM antiBam primer set, resulting in pTZ18U::B⁻TtpufM. Plasmid pTZ19U::TtpufLM was created by directionally cloning the TtpufL BamHI-XbaI fragment (from pTZ18U::TtpufL) and the TtpufM XbaI-HindIII fragment (from pTZ18U::B⁻TtpufM) into pTZ19U.

TABLE 4.

Primers used to clone the T. tepidum genes

Primer name	Sequence (5′ → 3′)
TpuhA NcoI-BamHI F	TCTAACCATGGCTGCTGGCATCACT
TpuhA NcoI-BamHI R	GAATAGGATCCTCAGTGGTGGTGGTGGTGGTGCAGCAGGGGACCG
TpufL BamHI-XbaI F	TCTAAGGATCCGAGGAGAAATTAACCATGGCCATGCTCAG
TpufL BamHI-XbaI R	GAATATCTAGATTACCACAGCGGGA
TpufM XbaI-HindIII F	TCTAATCTAGAGAGGAGAAATTAACCATGCCAGAATATCAAAAT
TpufM XbaI-HindIII R	GAATAAAGCTTTCATTGCATCACCTCC
TpufM antiBam F	ACGGCGGGGATCGGAATACTTCTGAGCGGGAC
TpufM antiBam R	GTCCCGCTCAGAAGTATTCCGATCCCCGCCGT
Rs TpufB FC F	GATCCGGAGGATAGCATGGCTGAACAGAAG
Rs TpufB FC R	AACATGATCTTGTTCTCCTTACAGCCAAGGACG
Rs TpufA FC F	TGTAAGGAGAACAAGATCATGTTCACGATGAAT
Rs TpufA FC R	GGGCCGGCCCGATTATTTCTTGCCCAGCGC
Rs pufL:lacZ FC F	GCGCTGGGCAAGAAATAATCGGGCCGGCCCTCCGTCGCG
Rs pufL:lacZ FC R	CTTCTGTTCAGCCATGCTATCCTCCGGATCGTAAGACTG

Using the FastCloning method (32), plasmid pTZ19U::Rs TtpufBA pufLMX was created by inverse PCR amplification of pTZ19U::pufQBALMX with the Rs pufL/lacZ primer set. TtpufB and TtpufA were PCR amplified from chromosomal DNA using Rs TpufB and Rs TpufA primer sets, respectively. The PCR products were subjected to DpnI digestion and annealed together prior to transformation into E. coli DH5α.

Measurement of pigment-protein complexes in disrupted cell samples.

Levels of RC, LH1, and TLH1 were measured as described previously with some modifications (10). A volume of 1 ml of cells was disrupted by sonication with a Microson ultrasonic cell disruptor for 30 s at a power output setting of 5 W. Following centrifugation at 14,000 × g for 1 min, the supernatant liquid was retained for absorbance measurements using a Hitachi U-3010 spectrophotometer. Spectra were normalized to 0.2 AU at 650 nm. For photobleaching, a StellarNet BLACK-Comet-SR diode-array spectrophotometer was used to measure absorbance spectra at 600 to 1,000 nm with an integration time of 70 ms and a pixel boxcar smoothing level of 4. Actinic light was from an 870-nm light-emitting-diode (LED) lamp (DigiKey catalogue no. 751-1207-ND).

Purification of the TRC and RC-TLH1.

Purification of the photosystems followed a modified published protocol of using a His tag and Ni²⁺-nitrilotriacetic acid (Ni²⁺-NTA) affinity chromatography to purify proteins (10, 33). Cells were inoculated with 50 ml of overnight starter culture into 2-liter Erlenmeyer flasks with 500 ml RLB medium, 25 μg/ml kanamycin, and 1 mM IPTG (and 10 μM menadione when expressing the TRC). Cultures were grown at 30°C, shaken at 200 rpm, grown to the late logarithmic-growth phase (approximately 18 to 21 h), and subsequently centrifuged at 7,000 × g.

The TRC was purified by suspending the cell pellet in 10 mM Tris (pH 8.0)–150 mM NaCl; for every gram of cell paste, 2 ml of buffer was added, as well as 1 μl of 1 M MgCl₂ and a few crystals of DNase I. Cells were broken in a French press at 20,000 to 25,000 lb/in² and centrifuged at 10,000 × g for 20 min to pellet cell debris and unbroken cells. DDM and imidazole were added to reach final concentrations of 1% and 5 mM, respectively, in the supernatant for 15 to 20 min at room temperature in the dark. The mixture was centrifuged at 40,000 rpm for 20 min at 4°C in a Beckman-Coulter type 70 Ti rotor to pellet insoluble materials. The supernatant was subsequently bound to a Ni²⁺-NTA chromatography column equilibrated with 10 mM Tris (pH 8.0), 150 mM NaCl, and 0.04% DDM at 4°C. The column was washed with this buffer until the 280-nm absorbance value was less than 0.02 AU. The TRCs bound to the column were eluted with 10 mM Tris (pH 8.0), 150 mM NaCl, 0.04% DDM, and 100 mM imidazole. The eluted product was exchanged into a buffer containing 10 mM Tris (pH 8.0) and 0.04% DDM using a Centricon (Millipore) centrifugal filter unit with a 50-kDa cutoff.

The RC-TLH1 hybrid core complex was purified by suspending the cell pellet in 10 mM Tris (pH 8.0) and was broken as described above. The chromatophore membranes were solubilized in 1% Deriphat 160C and centrifuged to pellet insoluble materials. A DEAE anion chromatography column (Toyopearl) was equilibrated with 1 mM Tris (pH 8.0)–0.04% DDM at room temperature. The column was loaded with supernatant equal in volume to the DEAE resin, and RC-TLH1 core complexes were eluted with the same volume of 10 mM Tris (pH 8.0)–0.04% DDM; the RC-TLH1 did not readily bind to the column, whereas the RC did.

UV mutagenesis of R. sphaeroides.

Semiaerobically grown R. sphaeroides cultures were centrifuged, the supernatant was removed, and the cell pellet was resuspended in ice-chilled 0.1 M sterile MgSO₄, followed by an additional centrifugation step. A volume of cold 0.1 M MgSO₄ was added to the cell pellet such that the final cell concentration was about 10⁹ CFU per ml (using a Klett-Summerson photoelectric colorimeter to measure turbidity). Cells were chilled for 15 min prior to placing 8-ml cultures in a Bio-Rad GS Gene Linker UV chamber for UV irradiation for periods of 0, 10, 20, and 30 s. Samples of 2 ml were taken from the irradiated cultures and inoculated into 2 ml of RLB medium supplemented with 25 μg/ml kanamycin in aluminum foil-wrapped tubes. The cultures were incubated at 30°C and agitated at 150 rpm for 4 h. Afterward, an additional 4 ml of RLB medium was added and the cultures were incubated overnight at 30°C with agitation of 150 rpm. The overnight cultures were diluted to 50 Klett units (KU), induced with 1 mM IPTG, and incubated overnight again. Four-milliliter volumes from the 0-, 10-, 20-, and 30-s cultures were transferred into 16.5-ml screw-cap tubes containing RLB medium supplemented with 25 μg/ml kanamycin and 1 mM IPTG. The cultures were grown in temperature-controlled aquariums at 30°C and 75 μE M⁻² s⁻¹ light intensity to select for phototrophy-capable mutants. Mutants were isolated by streaking on RLB agar plates containing 25 μg/ml kanamycin and 1 mM IPTG and grown in anaerobic jars to isolate individual colonies. The mutant capable of phototrophic growth while containing plasmid pIND4-RCT1 (expressing genes encoding the R. sphaeroides RC and T. tepidum TLH1) was called RCx^R-TcPS⁺.

Heat stability measurements of the core complex.

Core complexes in a buffer consisting of 10 mM Tris (pH 8), 0.04% DDM, and 1 mM CaCl₂ were normalized to a concentration such that the absorbance at 875 nm (LH1) or 915 nm (TLH1) was approximately 0.2 AU. Samples were subjected to temperatures of 30, 50, 70, and 90°C for 2, 4, 8, 15, 30, and 60 min in water baths in triplicate. A StellarNet BLACK-Comet-SR diode-array spectrophotometer was used to measure absorbance spectra from 600 to 1,000 nm with an integration time of 70 ms and a pixel boxcar smoothing level of 4.