Introduction

Chloracidobacterium (Cab.) thermophilum was discovered in 2007 in microbial mats of hot springs in Yellowstone National Park, WY, USA (Bryant et al. 2007). It is the first chlorophototrophic species in the phylum of Acidobacteria, and at the time of its discovery, it extended the number of phyla containing chlorophototrophic bacteria to six. Recently, a phototrophic species named Gemmatimonas phototrophica strain AP64T from a seventh bacterial phylum (Gemmatimonadetes) has been described (Zeng et al. 2014). Cab. thermophilum and “Ca. Thermochlorobacter aerophilum” are the only known aerobic chlorophototrophs that have type-1 homodimeric reaction centers (RC) (Bryant and Frigaard 2006; Liu et al. 2012; Tsukatani et al. 2012; Tank and Bryant 2015a, b; Tank et al. 2017). RCs are able to convert light energy into chemical energy and are, therefore, the key to the photochemical process of chlorophyll-based phototrophy (i.e., chlorophototrophy) (Tsukatani et al. 2012). There are two groups of photosynthetic RCs: type-1 RCs that utilize iron–sulfur clusters as terminal electron acceptor, and type-2 RCs that contain quinone molecules for this purpose (Golbeck 2003). Photosystem I of cyanobacteria and higher plants is a heterodimeric, type-1 RC. Homodimeric type-1 RCs are also found in the anaerobic heliobacteria and green sulfur bacteria. Previously, this was explained by the sensitivity of the Fe4–S4 clusters towards oxygen (Overmann 2001; Madigan 2001). However, it was later found that, unless it is illuminated, Chlorobaculum (Cba.) tepidum is tolerant to oxygen, suggesting that reactive oxygen species causes the damage (Li et al. 2009). The RCs of purple bacteria, including aerobic anoxygenic phototrophic bacteria and G. phototrophica AP64T, and the filamentous anoxygenic phototrophic bacteria contain type-2 RCs. Photosystem II of plants and cyanobacteria are also type-2 RCs (Golbeck 2003; Oh-oka 2007). All type-2 RCs have heterodimeric cores and utilize (bacterio)-pheophytin molecules as the primary electron acceptor.

(Bacterio-)chlorophylls [(B-)Chls] act in various processes in chlorophototrophic bacteria. In addition to the function of light harvesting and the associated transfer of excitation energy, they act as primary electron donor and primary electron acceptor in photosynthetic RCs. Based on a porphyrin structure, Chls and BChls comprise a central atom (usually a Mg2+ ion but sometimes Zn2+) and different substituents at the outer macrocycle carbons. Pheophytins, which do not coordinate to a central atom, act as primary electron acceptors in type-2 RCs. In the RC complex of Cab. thermophilum, reverse-phase high-performance liquid chromatography (RP-HPLC) revealed the presence of three different Chls: BChl a, Chl a, and Zn-BChl a′ in a ratio of about 12.8:8.0:2.0 (Tsukatani et al. 2012). Zn-BChl a′ was only found in the RC and was not present in chlorosomes, the light-harvesting antenna complexes of this organism, nor in the FMO antenna protein that only contained BChl a (Wen et al. 2011; Tsukatani et al. 2012). Zn-BChl a′ was suggested to function as the primary electron donor (Tsukatani et al. 2012). In the literature, 1H and 13C chemical shifts of Zn-BChl a have been reported (Kobayashi et al. 1998, 2016). In this work, 15N liquid-state NMR data from this cofactor are presented to provide reference values for the 15N photo-CIDNP MAS NMR spectra.

The solid-state photo-CIDNP effect (Jeschke and Matysik 2003; Daviso et al. 2008a; Bode et al. 2013; Sosnovsky et al. 2016) provides selective access to information on the electron donor and the primary electron acceptor that form a spin-correlated radical pair. It allows the detection of electron transfer chain cofactors of isolated RCs, and it can be used to detect these cofactors at nanomolar concentrations in intact cells (Prakash et al. 2006; Janssen et al. 2010). The electronic ground-state structures of the electron donor and acceptor forming a spin-correlated radical pair is reflected in the observed chemical shifts. Electron spin densities can be deduced from the photo-CIDNP MAS NMR intensities. There are up to three mechanisms that occur in parallel and that contribute to the solid-state, photo-CIDNP effect: (1) three-spin mixing (TSM), generally yielding emissive signals; (2) differential decay (DD); and (3), if the lifetime of the donor triplet state is sufficiently long, differential relaxation (DR) giving enhanced absorptive signals (Jeschke 1998; McDermott et al. 1998; Polenova and McDermott 1999). Recently, the spin dynamical system has been re-interpreted in terms of level crossings and anti-crossings within an electron–electron–nuclear spin scheme (Sosnovsky et al. 2016).

The solid-state photo-CIDNP effect has already been observed in the RCs of nine species, including plants (Alia et al. 2004; Diller et al. 2005; Roy et al. 2007b; Janssen et al. 2010, 2012), bacteria (Zysmilich and McDermott 1994; Schulten et al. 2002; Prakash et al. 2005, 2006; Diller et al. 2008; Roy et al. 2007a), and Chromista (Zill 2017a). An overview of these species is presented in Fig. 1. In addition, the effect has been observed for the blue-light photoreceptor phototropin LOV1-C57S (Thamarath et al. 2010) from the green alga Chlamydomonas reinhardtii. There are seven bacterial phyla with chlorophototrophic members (Tank et al. 2017; Thiel et al. 2017). The effect has not yet been observed in the phyla Acidobacteria, Chloroflexi, and Gemmatimonadetes. In the present work, the solid-state photo-CIDNP effect was studied in the membranes of Cab. thermophilum from the phylum Acidobacteria. The fact that the solid-state photo-CIDNP effect has been observed in all taxonomic kingdoms and in a large variety of phyla implies that the effect is an inherent property of natural photosystems (Matysik et al. 2009) and that it probably reflects conserved features of light-induced charge separation in reaction centers with important functional relevance (Céspedes-Camacho and Matysik 2014).

Fig. 1
figure 1

Taxonomic overview referring to Cavalier-Smith (1998) modified. Domains and kingdoms which contain photochemically active species are labeled in green. Until now, the solid-state photo-CIDNP effect has been observed in ten species. For reference see Alia et al. (2004), Diller et al. (2005, 2008), Bryant et al. (2007), Roy et al. (2007a, b, 2008), Janssen et al. (2010, 2012), Thamarath et al. (2010), Zill (2017a), and Zill et al. (2017b). In this work, the effect is reported for the first time from the phylum Acidobacteria

Full size image

Materials and methods

Medium preparation and growth of 15N-labeled Cab. thermophilum cells

Chloracidobacterium thermophilum was grown in CTM medium (1 L) according to the protocol described in Tank and Bryant (2015a, b). Slight changes in the medium were made to achieve the growth of Cab. thermophilum on 15N-labeled substrates. Bacto peptone and the 20 amino acid mix of the CTM medium were replaced by uniformly 15N-labeled algal amino acid mix NLM-2161 (Cambridge Laboratories, Tewksbury, MA, USA) as the sole nitrogen source. The 15N-labeled amino acid mix (100 mg) was added to the medium prior autoclaving. Cells were incubated at 51–52 °C and were illuminated with a tungsten lamp at ≤ 50 mol photons m−2 s−1 for 7–10 days.

Isolation of chlorosome-depleted total membranes containing reaction centers

15N-labeled cells of Cab. thermophilum were harvested by centrifugation at 7500×g for 20 min. Cell pellets were resuspended in phosphate-buffered saline solution (100 mL; PBS = 10 mM potassium phosphate, 150 mM sodium chloride, pH 7.2). Cell suspensions were aliquoted equally (25 mL each) into 50-mL screw-capped plastic tubes. Lysozyme (3 mg mL−1) was added, and the suspensions were incubated at room temperature in the dark for 25 min. The cell suspensions were kept on ice while the treated cells were disrupted by four passages through a French pressure cell (138 MPa). Cellular components, including membranes, were pelleted by ultracentrifugation at 220,000×g for 2 h. This and subsequent ultracentrifugation steps were performed at 4 °C. Each of the four pellets was resuspended in ~ 3 mL of PBS, and the resuspended membranes were subsequently loaded on continuous PBS–sucrose density gradients (10–50%, w/v) and centrifuged at 220,000×g for 18 h to separate the RC-containing membranes from other cell components. The Cab. thermophilum membrane fraction appeared as an orange-colored band underneath a dark green band that contained the remaining BChl c-containing chlorosomes in the PBS–sucrose density gradients (Fig. S1A). The chlorosome band was carefully removed by pipetting prior to collection of the membrane fraction to avoid cross-contamination. The membrane fractions were recovered by pipetting and were pooled in an ultracentrifugation tube. The membrane fraction was diluted with five volumes of PBS to reduce the sucrose concentration of the sample in two subsequent washing steps. The pooled membranes were centrifuged at 250,000×g for 2 h. The supernatants were discarded and the membrane pellets were resuspended in ~ 25 mL PBS and centrifuged again at 220,000×g for 2 h. In a final step the membrane pellet was resuspended in a minimal volume of ~ 1.5 mL PBS isolation buffer. The absorption spectrum of these chlorosome-depleted membranes showed a strong enrichment of the characteristic absorption features of FMO and RCs of Cab. thermophilum, which have a maximum at 809 nm and a shoulder at 823 nm (Fig. S1B). Washed membranes were stored at 4 °C in the dark until they were required for analysis.

For photo-CIDNP MAS NMR experiments, the volume was reduced using Centricon Plus-20 centrifugal filter devices (Amicon®, Millipore Corporation, Bedford, USA, 30 kDa molecular weight cutoff) performing centrifugation for 12 min at 4000×g. Membranes containing RCs were reduced by adding 1 M sodium dithionite solution up to a final concentration of 100 mM in an oxygen-free atmosphere.

MAS NMR measurements

Solid-state NMR measurements were performed with an Avance-III 400-MHz NMR spectrometer equipped with a 4-mm MAS probe. The sample was loaded in a 4-mm sapphire rotor and inserted into the MAS probe. Freezing slowly helped to obtain homogeneous sample distribution against the rotor wall (Fischer et al. 1992). The light and dark spectra were collected with a Hahn-echo pulse sequence and two-pulse phase modulation (TPPM) proton decoupling (Bennett et al. 1995). 15N-MAS NMR spectra were obtained at a temperature of 250 K under continuous illumination with white light (1000 W xenon arc lamp) (Matysik et al. 2000). The rotational frequency for MAS was 8 kHz and the cycle delay was 0.5, 4, or 12 s. Artificial line broadening of 40 Hz was applied prior to Fourier transformation. All the 15N-NMR spectra were referenced using the amino NH2 response of histidine ⋅HCl at 49.09 ppm.

Preparation of 15N-labeled Zn-BChl a

Cultures of Rhodobacter (Rba.) sphaeroides were grown phototrophically using [15N]-ammonium sulfate (95%) as the sole nitrogen source in anoxic M22+ medium. Zn-BChl a was prepared from Mg-BChl a by slight modification of a previously described method (Chen et al. 2016). Briefly, one volume of extracted 15N-labeled Mg-BChl a in methanol was mixed with excess zinc acetate, 50 mM sodium ascorbate, and six volumes of glacial acetic acid. The mixture was incubated at 90 °C for 1 h. After centrifugation at 10,000×g for 5 min, the supernatant was transferred to a clean tube and dried in a vacuum concentrator at 30 °C. The dried pigments were dissolved in methanol and purified by reversed-phase HPLC on an Agilent 1100 HPLC system (Agilent Technologies) with a semi-preparative Discovery C18 column (10 mm × 25 cm) (Supelco, Sigma-Aldrich) using a modified method (Van Heukelem et al. 1994). Solvents A and B were 80:20 (v/v) methanol–500 mM ammonium acetate and 80:20 (v/v) methanol–acetone, respectively. Pigments were eluted at 3.5 mL min−1 with a linear gradient of 92–93% solvent B over 10 min, increasing to 100% to wash the column. Elution of Zn-BChl a was monitored at 770 nm. Approximately 9 mg of 15N-labeled Mg-BChl a was extracted from the cells of a 200-mL culture of Rba. sphaeroides, and the final yield of 15N-labeled Zn-BChl a was approximately 1 mg. The identity of the compound was established by its absorption spectrum (Fig. 2), which had maxima at 355 and 763 nm (Wakao et al. 1996). Mass spectrometry showed that the purified 15N–Zn-BChl a had m/z of 955.5 (M+) (Fig. S2) compared to Mg-BChl a (calculated mass = 910.6) and 14N–Zn-BChl a (calculated mass = 950.5) (Tsukatani et al. 2012; Kobayashi et al. 2006). As expected, the mass confirmed the replacement of Mg by Zn (+ 39.9) and the substitution of the four 14N pyrrole nitrogens with 15N (+ 4). Furthermore, the purified 15N–Zn–BChl a exhibited the characteristic isotopic abundance pattern for compounds containing Zn (Tsukatani et al. 2012). After the NMR experiments, infrared (Fig. S3) and Raman spectra (Fig. S4) were also obtained.

Fig. 2
figure 2

The absorption spectrum of Zn-BChl a showing the characteristic maxima at 355 and 763 nm

Full size image

Solution-state NMR measurements

For the solution-state NMR measurements, 15N–Zn-BChl a (~ 3 mg) was dissolved in d8-tetrahydrofuran (THF-d8). The 1H-NMR spectrum and the 15N–1H-HMQC were measured at 600 MHz using a cryoProbe. The 15N-NMR spectrum and the 13C-APT spectrum were acquired in the same solvent at 500 MHz using a Prodigy-BBO probehead. 1H and 13C spectra were referenced to TMS. 15N spectra were referenced to liquid ammonia using a solution of 90% formamide in DMSO as external standard.

Quantum chemical calculations of 15N-chemical shifts

The Zn-BChl a geometry was optimized at the B3LYP/6-31G* level of theory (Becke 1993; Rassolov et al. 2001), including the Grimme’s dispersion correction (Grimme et al. 2011) and the SMD implicit solvent model (Marenich et al. 2009) with THF as the solvent. In order to reduce complexity, the phytyl tail was replaced with a methyl group. In addition, we have added an explicit THF molecule to saturate the axial coordination of zinc. There are two options for THF ligation, called “α” and “β” as shown in Fig. 3. In the case of α ligation, the THF molecule and the C-171 atom are on opposite faces of the tetrapyrrole macrocycle (Fig. 3 top), while for β ligation, the THF molecule is on the same side of the macrocycle as the C-171 atom (Fig. 3 bottom).

Fig. 3
figure 3

Zn-BChl a with one explicit THF molecule in α and β positions. Top THF molecule at the α position. The THF and atom C171 are in opposite faces of the tetrapyrrole macrocycle. Bottom THF molecule at the β position. The THF and atom C171 are at the same faces of the tetrapyrrole macrocycle

Full size image

The isotropic shielding tensors were calculated using B3LYP/6-31G* level of theory and the gauge-invariant atomic orbitals (GIAO) method (Ruud et al. 1993), and chemical shifts were taken with respect to the isotropic shielding of ammonia 254.97 ppm, which was calculated at the same level of theory. All calculations were performed using the Gaussian 09 software package (Frisch et al. 2009).

Results and discussion

15N photo-CIDNP MAS NMR spectra from Cab. Thermophilum

Figure 4 shows the 15N-MAS NMR spectra of uniformly 15N-labeled membranes of Cab. thermophilum. All spectra (A–F) were recorded for about 2 days with different cycle delays. Spectra A, C, and E were collected in the dark, and no signal is observed. Due to the relatively low concentration of the sample, no signal from the protein is observed. Spectra B, D, and F were measured under continuous illumination with white light. Spectra A and B were collected with a cycle delay of 12 s (16,300 scans). Spectra C and D were measured with a cycle delay of 4 s (45,000 scans), and a cycle delay of 0.5 s (345,000 scans) was used for spectra E and F. The strongest light-induced signals are observed in spectra B and D. In spectrum F only a signal at 216.0 ppm can be observed. This is due to a too short cycle delay of 0.5 s, which does not allow for sufficient relaxation. Hence, the optimal cycle delay is around 4 s. All signals are light-induced and emissive (negative). In the 15N-photo-CIDNP MAS NMR spectrum, no splitting of the signals is observed as in the special pair of Rba. sphaeroides (Daviso et al. 2008b). Hence, we can conclude that the observed set of signals occur either from a perfect dimer or a monomeric donor (i.e., in a strongly and functionally asymmetric dimeric donor). To obtain a maximal light-induced signal, a detailed view of the sum of spectra B and D of Fig. 4 is provided in Fig. S5. Upon illumination, three signals emerge in the aromatic region at 250.1, 216.0, and 191.4 ppm. The observation of these three signals implies that the signals occur from only one cofactor, which is either the electron donor or the primary electron acceptor. It would be possible that both donor and acceptor would be of the same type of cofactor; however, one would not expect that they would show perfectly overlapping chemical shifts.

Fig. 4
figure 4

15N-photo-CIDNP effect observed on Cab. thermophilum: a dark, cycle delay = 12 s, 16,300 scans; b light, cycle delay = 12 s, 16,300 scans; c dark, cycle delay = 4 s, 45,000 scans; d light, cycle delay = 4 s, 45,000 scans; e dark, cycle delay = 0.5 s, 345,000 scans; f light, cycle delay = 0.5 s, 345,000 scans, *spectrometer signal

Full size image

In the RCs of Cab. thermophilum, three types of cofactors have been identified (Chl a, BChl a, and Zn-BChl a′) (Tsukatani et al. 2012). Careful repetition of reversed-phase HPLC and reversed-phase HPLC–MS analysis again revealed only the same three pigments, and no 81-OH Chl a was detected. 15N-chemical shifts of Chl a and Zn-BChl a was prepared to obtain 15N-chemical shifts under solution-state conditions.

Solution-state NMR spectra from Zn-BChl a

Figure S6 shows the 1H15N-HMQC spectrum of Zn-BChl a. Analysis of this spectrum confirmed the assignment of the three protons H-5, H-10, and H-20 of the porphyrin ring (for numbering, see Fig. S7). Each of these protons is coupled via three bonds to an sp3- and an sp2-hybridized N atom. It is a characteristic feature of BChl a and, hence, also Zn-BChl a′, that each pyrrole nitrogen is coupled to a different number of protons. The resonance at 180.1 ppm is coupled to one proton only and must, therefore, be induced from the nitrogen of ring-III. Nitrogen N-I arises at 178.1 ppm which is confirmed by the 3J-coupling to two protons at 9.11 ppm (H-5) and 8.52 ppm (H-20). The signal at 261.5 ppm is coupled to three protons. Hence, these resonances can be assigned to N-IV. In addition, a weak fourth crosspeak occurs due to the 4J-coupling to H-132 at 6.05 ppm. This nitrogen N-II resonates at 259.4 ppm. Table 1 summarizes the experimental 15N assignments. The macrocycle of Zn-BChl a with 55 C-atoms and 74 protons shows overlapping of a number of signals and the assignment of the resonances of the phytyl side chain remains incomplete. The structure is presented in Fig. S7.

Table 1 Tentative assignment of the chemical shifts obtained by 15N photo-CIDNP MAS NMR in comparison to other photo-CIDNP MAS NMR data and calculations
Full size table

An overview on the chemical shifts assigned to the 13C and 1H positions in comparison to literature data is provided in Tables 2 and 3, respectively. The assignment in this work is almost identical to the chemical shifts observed in the literature (Kobayashi et al. 1998, 2016). In addition to the known signals, four 13C-resonances and two 1H-resonances have been assigned for the first time. Hence, the signals of Zn-BChl a are assigned completely except for those associated with the phytyl side chain.

Table 2 Comparison of 13C chemical shifts (ppm) of Zn-BChl a in different solvents
Full size table
Table 3 Comparison of 1H chemical shifts (ppm) of Zn-BChl a in different solvents
Full size table

Calculation of the 15N-chemical shifts from Zn-BChl a

Figure S8 shows the optimized geometries of Zn-BChl a and Zn-BChl a′ with one THF molecule as the fifth ligand at either the α or β positions. It was found that in the case of Zn-BChl a, the α-ligand is energetically preferable by ~ 4.21 kJ mol−1 (~ 0.0436 eV). However, in the case of Zn-BChl a′, the β-ligand is more stable by ~ 3.15 kJ mol−1 (~ 0.0327 eV). In addition, the two geometries of 81-OH Chl a were calculated (Fig. S9), showing that the S configuration is ~ 3.52 kJ mol−1 (~ 0.0365 eV) more stable than the R configuration.

Table 1 also compiles the theoretically calculated values. Clearly, the trends in the experimental data are well reproduced and Chl a and Zn-BChl a can be readily distinguished. The differences between the values of the four calculated forms of Zn-BChl a are too small to conclude the state of that molecule in THF solution. The calculations of the 15N-chemical shifts of Zn-BChl a and Zn-BChl a′ show that the chemical shift differences between the epimers are nearly identical and differ by only a few ppm. Furthermore, solely on the basis on 15N NMR, it would also be difficult to distinguish Chl a and 81-OH Chl a.

Assignment of the 15N-photo-CIDNP MAS NMR signals from Cab. thermophilum

Having obtained the 15N-chemical shifts from Zn-BChl a, the sets of 15N-chemical shifts are now available for all cofactors associated with the RCs of Cab. thermophilum. Table 1 provides an overview on the photo-CIDNP MAS NMR signals from Cab. thermophilum in comparison to solution-state NMR data for the pure cofactors. Furthermore, 15N-photo-CIDNP MAS NMR data are presented for other RCs of plants (Diller et al. 2007a), bacteria (Thamarath et al. 2012), and Chromista (Zill 2017a), together with the 15N-chemical shift calculations from Zn-BChl a and Zn-BChl a′ as well as from Chl a and 81-OH Chl a determined in this study. On the basis of the chemical shift values obtained by experiments and calculations (summarized in Table 1), we can identify the (B)Chl cofactor that best matches the three 15N photo-CIDNP MAS NMR signals observed in Cab. thermophilum (Fig. S5).

In the 15N-photo-CIDNP MAS NMR spectrum of Cab. thermophilum, the strongest signal is observed at 216.0 ppm. In comparison to the three (B)Chls found in Cab. thermophilum, this very characteristic signal matches best to the N-II of Chl a at 206.0 ppm. The resonance at 250.1 ppm can conveniently be assigned to N-IV and the signal at 191.4 ppm corresponds to N-III. Hence, we conclude that the three 15N photo-CIDNP MAS NMR signals that are observed in membranes of Cab. thermophilum originate from Chl a, although the signals appear slightly disturbed, especially at ring II. A comparison to the signals observed in the other three RCs, however, reveals that the N-II resonance differs by only a few ppm. Hence, the chemical shifts of the three 15N signals reported here match well to those of Chl a incorporated in other RCs.

The question now arises whether Chl a acts as the donor or the acceptor within the spin-correlated radical pair. The RC complex of Cab. thermophilum contains three pigments in the ratio 1.6 Zn-BChl a′, 10.3 Chl a, and 6.4 BChl a molecules per P840 donor. The light-induced optical difference spectrum (P+ minus P) exhibits maximal photobleaching near 840 nm, which indicates that Chl a is certainly not the primary donor for these RCs. There are probably at least two Zn-BChl a′ molecules in the RC, implying that these two molecules could form a dimeric donor in the electron transfer cofactor chain. All known type-1 RCs have at least one epimer form of (B)Chl as the electron donor. For example, heliobacterial RCs contain two BChl g’ molecules as the symmetric dimeric donor, and these RCs further contain two BChl g’ molecules within the core antenna portion of the RC (Gisriel et al. 2017). These observations strongly suggest that Zn-BChl a′ acts as the electron donor in Cab. thermophilum. This implies that the observed Chl a portion of the spin-correlated radical pair acts as the electron acceptor. The spin-correlated radical pair would therefore be formed by Zn-BChl a′ as the donor and Chl a as the acceptor. The occurrence of a single set of signals from the acceptor can be interpreted to indicate that either only one branch of cofactors is active, or that both branches are active and contain identical acceptor cofactors. Because the RC is a homodimer of PscA subunits (Tsukatani et al. 2012), the results are consistent with the latter option of a symmetric dimer. The absence of signals from the donor in this photo-CIDNP MAS NMR experiment might be an experimental accident because of the field strength chosen. It might be that the positive (enhanced absorptive) contributions by the DD and DR mechanisms cancel the negative (emissive) contributions by the TSM mechanism.

A comparison of the 15N-photo-CIDNP MAS NMR spectrum of Chl a to those of other species reveals a high similarity to PS I from plants (Diller et al. 2007a) and diatoms (Zill 2017a). Cba. tepidum from the phylum Chlorobi also utilizes Chl a as electron acceptor (Roy et al. 2007a). Furthermore, in the RCs of heliobacteria, 81-OH Chl a acts as primary electron acceptor (Thamarath et al. 2012; Gisriel et al. 2017). The type-1 RCs from H. mobilis only showed signals induced by the donor for the anaerobic Braunstoff preparation and predominantly resonances from the acceptor in case of the aerobic Grünstoff (Thamarath et al. 2012). Comparison of 15N-photo-CIDNP MAS NMR spectra from Cab. thermophilum and the aerobic RCs from H. mobilis also reveals good matching. As we have seen above, 15N NMR is unable to distinguish Chl a and 81-OH Chl a, in the present case; however, only because Chl a is present and 81-OH Chl a can be excluded biochemically, we conclude that the primary acceptor must be Chl a.

All 15N photo-CIDNP MAS NMR signals (Figs. 4 and S5) appear with emissive sign, which can be explained by a predominance of the TSM over the DD mechanism. The DR mechanism does not contribute to signals from the acceptor. The lineshape of the observed signals is Lorentzian, indicating that the natural lineshape remains undisturbed. The strongest signal at 216.0 ppm is slightly broader (full width half maximum = FWHM of 70 Hz) than the signal at 250.1 ppm (FWHM of 48 Hz). The narrow linewidth is typical for photosynthetic RCs and is explained by a rigid inner part of the RC (Fischer et al. 1992; Diller et al. 2007a).

Figure 5 shows the photo-CIDNP intensity pattern of Chl a, which acts most likely as the primary electron acceptor in Cab. thermophilum. The depicted signal intensities correlate to the local electron spin densities at this position (Diller et al. 2007b). As in plant PS I (Diller et al. 2007a) and Grünstoff of heliobacteria (Thamarath et al. 2012), N-II shows the highest intensity.

Fig. 5
figure 5

15N photo-CIDNP MAS NMR intensity pattern of Cab. thermophilum

Full size image

Conclusions

The solid-state photo-CIDNP effect is observed in the membranes of Cab. thermophilum by 15N-MAS NMR. This is the first time that the effect has been found in RCs of an organism from the phylum Acidobacteria. The high number of species as well as the large variety of taxonomic ranks indicates that the effect is an intrinsic property of all photosynthetic species and their RCs. Hence, photosynthetic functionalities appear to be exceptionally constant on very long evolutionary timescales. The results suggest that details concerning the mechanism of primary charge separation are conserved in all types of RCs, which may indicate that primary charge separation in RCs occurs by a single, physically similar process in all RCs. This finding is in general agreement with the widely accepted idea that photochemical reaction centers evolved once and subsequently diversified by processes including gene duplication, divergence, and horizontal gene transfer (e.g., see recent reviews by Cardona 2015; Fischer et al. 2016; Martin et al. 2018).

A set of three 15N NMR signals is observed in the photo-CIDNP MAS NMR spectrum of the type-1 homodimeric RCs of Cab. thermophilum. To identify the cofactor observed with photo-CIDNP MAS NMR, 15N-chemical shifts of Zn-BChl a were determined from solution-state NMR experiments as well as from calculations. The chemical shifts of Zn-BChl a as well as values of Mg-BChl a known from literature do not match the resonances observed with 15N-photo-CIDNP MAS NMR. The photo-CIDNP signals, however, match well to the chemical shifts of Chl a. We conclude that Chl a acts as primary electron acceptor, implying that Zn-BChl a′ is the electron donor in these RCs. The absence of signal splitting is consistent with the observed acceptor cofactor being a part of a symmetrical electron transport chain in this homodimeric RC.