The thylakoid NAD(P)H dehydrogenase-like (NDH) complex is a large protein complex that reduces plastoquinone and pumps protons into the lumen generating protonmotive force. In plants, the complex consists of both nuclear and chloroplast-encoded subunits. Despite its perceived importance for stress tolerance and ATP generation, chloroplast-encoded NDH subunits have been lost numerous times during evolution in species occupying seemingly unrelated environmental niches. We have generated a phylogenetic tree that reveals independent losses in multiple phylogenetic lineages, and we use this tree as a reference to discuss possible evolutionary contexts that may have relaxed selective pressure for retention of ndh genes. While we are still yet unable to pinpoint a singular specific lifestyle that negates the need for NDH, we are able to rule out several long-standing explanations. In light of this, we discuss the biochemical changes that would be required for the chloroplast to dispense with NDH functionality with regards to known and proposed NDH-related reactions.
In the chloroplast electron transport chain  (Figure 1), light energy is harnessed to drive the transfer of electrons from water to soluble electron carriers in the stroma in the form of NADPH and ferredoxin (Fd) to feed downstream metabolic processes. At several points in the chain, electron transfer is coupled to proton deposition into the lumen. Protons enter the lumen at photosystem II (PSII), where light-induced water oxidation supplies electrons for the reduction of plastoquinone (PQ) to plastoquinol (PQH2) and protons are released. Protons are also released into the thylakoid lumen when PQH2 is oxidized via the cytochrome b6f (bf) complex. These points of proton-coupled electron transport generate an electrochemical potential across the thylakoid membrane, or protonmotive force (pmf) for ATP synthesis via the ATP synthase. The pmf consists of two thermodynamically equivalent components, the membrane potential (Δψ) and the proton gradient (ΔpH). While both components are known to have important regulatory functions, the latter is necessary for the photoprotection of PSII and photosynthetic control of electron transfer through the bf complex [2–4].
Proton and electron transport across the thylakoid membrane.
This minimal conceptualization of the electron transport chain theoretically generates a fixed ratio of NADPH to ATP to supply downstream metabolic processes, including the Calvin–Benson–Bassham (CBB) cycle. To avoid down-regulation of electron transfer or the accumulation of potentially harmful reactive intermediates, the chloroplast must match the production of ATP and NADPH with downstream consumptive metabolic processes. The chloroplast has evolved multiple alternative electron transfer pathways to adjust ATP/ADP + Pi and/or NADPH/NADP+ pools [2,5]. Particularly, when an ATP deficit occurs, the chloroplast may activate cyclic electron flow (CEF) around photosystem I (PSI) , a process by which pmf is generated by shunting electrons from downstream metabolism back into the PQ/PQH2 pool. This allows for additional proton deposition into the lumen via PQH2 oxidation at the bf complex, thereby increasing ATP production without the net production of NADPH. CEF has been proposed to be catalyzed by at least three distinct pathways (reviewed in ). In seed plants, the three main routes of CEF that have been proposed are (i) the antimycin A-sensitive Fd-dependent quinone reductase (FQR) , (ii) PQ reduction at the Qi site via heme ci in the bf complex  and (iii) a pathway involving the NAD(P)H dehydrogenase-like (NDH) complex (, Figure 1). While these routes are often cited as redundant, they may instead have distinct physiological roles, as evidenced by their differential activation [9–11] and energetics (i.e. NDH is protonmotive, while the FQR and heme ci pathways are not [12,13]).
The NDH complex, an analog of respiratory complex I (CI) present in bacteria and the inner mitochondrial membrane, is of particular interest due to its implications for thylakoid bioenergetics [12,14], and its relatively high degree of evolutionary conservation in seed plants . Given this strong conservation of NDH in higher plants, those species lacking this complex present an interesting evolutionary conundrum. Mutants lacking NDH are not strongly affected by this loss under controlled growth conditions, but are significantly impacted when subjected to environmental stress [16,17]. Despite the stress sensitivity of the NDH lacking mutants, there have been many NDH-deficient species described in nature, some of which are capable of surviving harsh environments [15,18,19]. This provokes the question: What is altered in the physiology of these species that allows for loss of this photosynthetic complex?
In this review, we discuss NDH in an evolutionary context, with a primary focus on what lineages have lost ndh genes from the chloroplast genome. We further address the proposed functions of NDH in photosynthesis and what conditions might lead to a relaxation of selective pressure to maintain this complex.
CI and NDH evolution
The respiratory CI can be divided into three modules based on their function: (i) the electron donor N-module, (ii) the quinone reductase Q-module and (iii) the proton-translocating P-module. While the evolutionary specifics are still debated, simplistically, the N-module arose from a hydrogenase, the Q-module arose from a quinone reductase and the P-module arose from a complex of MRP (multiple resistance and pH locus) Na+/H+ antiporters [20–22]. Within the context of CI evolution, NDH lies within a class of 11-subunit complexes lacking the N-module, and comprising the most minimal CI found in nature [21,23–25]. This complex is present in the photosynthetic membranes of cyanobacteria and green plants, with a few notable exceptions . In plants, the NDH complex has a relative molecular mass of ∼550 kDa  and consists of ∼30 subunits. While the majority of the ndh genes are encoded in the nucleus, 11 are encoded in the chloroplast genome [26–28]. These ‘core’ subunits are the basis of the aforementioned homology with CI, and are responsible for most of the proton and electron movement within the complex (reviewed in [12,29]).
To understand the structure-function relationships in NDH, it is useful to draw comparisons to what is known about CI, as the respiratory analog has been studied in much finer detail. CI contains an integral membrane hydrophobic arm containing three proton channels, in which proton pumping is partially attributed to polar residues conserved in the integral membrane subunits of NDH [12,30]. The hydrophilic domain of CI contains eight iron-sulfur clusters, seven of which facilitate electron transfer from NADH via a flavin mononucleotide (FMN) . In contrast, the NDH lacks the N-module and, therefore, is missing five iron-sulfur clusters and FMN of the CI electron transport chain ([24,25,27]; discussed in ). This large structural difference between CI and NDH can be seen by electron microscopy of cyanobacterial NDH-1 , and the recent atomic resolution cryo-EM structures of NDH-1 from Thermosynechococcus elongatus [24,25]. In these, NDH-1 presents with a shorter hydrophilic arm than CI due to the lack of the N-module. In the minimal 11-subunit CI and NDH, the absence of the N-module has led to questions concerning the electron donor [21,23,29,33,34]. For NDH, the current consensus is that the electron donor is Fd, and not an NAD species [12,34], and formation of a PSI-NDH supercomplex has been proposed to facilitate substrate channeling of reduced Fd to the NDH complex when CEF is active [35–37].
In both plants and cyanobacteria, NDH consists of not only the minimal core of 11 subunits, but also oxygenic photosynthesis-specific subunits (NdhL-P, S and V) associated with the hydrophilic arm. Of these photosynthesis-specific subunits, NdhS is thought to be the Fd-binding subunit for electron donation , an assignment which has recently been challenged, and NdhO instead proposed as the Fd-binding site . However, conflicting to this new interpretation is the observation that knock-out mutants of ndhO in Synechocystis appear to have increased NDH-CEF , an impossibility if this subunit was the point of electron injection into the chain. The situation in plants is likely different, as the deletion of NDHO leads to loss of NDH activity . Due to this apparent discrepancy between cyanobacteria and higher plants, it has been proposed that NDHO function has drastically diverged during evolution. Further differences between plants and cyanobacteria include the cyanobacteria-specific NDHQ, which in the recent cryo-EM structures appears to stabilize the transverse helix of NDHF [24,25]. In addition to these differences in the photosynthesis-specific subunits, cyanobacteria also possess alternative subunits for the core subunits NDHF and NDHD. These appear in non-classical configurations of the NDH complex that have been implicated in non-CEF-mediating complexes that function in the carbon-concentrating mechanism or are thought to associate with different membranes (reviewed in [38,41]). Therefore, the identification of novel plant-specific NDH-associated proteins, and deviations in their perceived role within the complex as exemplified by NDHO, suggests that the plastid NDH has diverged in structure and regulation from CI and cyanobacterial NDH-1 .
NDH losses in plant evolution
The plastid-encoded ndh genes maintain a relatively high degree of evolutionary conservation across the green lineage. This conservation is surprising due to the lack of an obvious growth phenotype in ndh mutants of tobacco and Arabidopsis thaliana under unstressed environmental conditions [17,42]. There are several exceptions to the conservation of the NDH complex in the green lineage that have been discussed in earlier works [15,19,43–45]. In a broad review of NDH losses throughout evolution, Martín and Sabater  pointed out ndh gene loss in gymnosperms and algae, and hypothesized a role for RNA editing to restore the function of genes that had experienced periods of dispensability. In a typical seed plant, more than half of the plastid RNA editing sites reside in ndh genes . Since this review, the number of publicly available plastid genome sequences has vastly increased, giving us the opportunity to assess the evolutionary history of the NDH complex throughout the green lineage in greater detail.
To obtain a broad picture of the evolutionary dynamics in the plant kingdom, we selected a total of 60 species spanning across different orders from the gymnosperm and angiosperm clades, and also included representative photosynthetic organisms outside the seed plants. To show evolutionary relatedness of these species, we constructed a phylogenetic tree using maturase K (MatK), a highly conserved plastid genome-encoded protein (Figure 2A,B). To determine the pattern of NDH retention, we performed a pBLAST search in the NCBI Refseq database (restricted to plastids) in all selected organisms using the amino acid sequences of the different plastid-encoded subunits of the NDH complex from A. thaliana (i.e. NdhA-K; Figure 3) as a query.
NDH complex genes were independently lost multiple times across the green lineage.
Presence/absence of plastid
ndh genes across the green lineage.
The MatK amino acid sequences from all organisms were aligned by MUSCLE and used to create a phylogenetic tree using PhyML (500 bootstraps). Based on this analysis, and assuming that the loss of most of the subunits in the plastid genome likely leads to the complete loss of NDH activity, we color-coded the tree branches in green or blue, respectively, to illustrate the presence or absence of the complete suite of NDH complex genes. As had been previously highlighted, many independent loss events occurred in both gymnosperms and angiosperms. It is important to note that the absence of any or even all NDH subunits encoded in the plastid genome leaves open the possibility of these genes having been transferred to the nucleus. To examine this possibility, we took advantage of the draft genomes available for some of the species lacking plastid-encoded NDH subunits, including the gymnosperms Picea abies (GCA_900067695.1) and Pinus taeda (GCA_000404065.3), the orchid Phalaenopsis equestris (GCF_001263595.1) and the cactus Carnegiea gigantea (GCA_002740515.1; Figure 2A; red). Using the nucleotide sequences of plastid-encoded ndh subunits from A. thaliana, we performed an nBLAST search limited to these four species and A. thaliana as a reference. Interestingly, A. thaliana possesses two putative pseudogenes of ndhF (or the mitochondrial encoded nad5) on chromosome 2. However, from this search, we obtained no significant matches for any of the other species under study, in agreement with previous studies [47,48]. While most of these genome drafts have low coverage and may need further improvement, these findings provide no evidence for the existence of endosymbiotic gene transfer events (i.e. gene transfer from the plastid to the nuclear genome; ) in these organisms. The absence of endosymbiotic gene transfer has also been inferred from the lack of nucleus-encoded NDH transcripts in species lacking plastid-encoded ndh sequences .
Tracing back the loss of
ndh genes from plastomes during gymnosperm evolution
As previously noted [15,19,43–45], there have been multiple losses of plastome-encoded NDH complex genes within the gymnosperms. To determine the frequency of ndh gene loss, we selected at least two species from each of the four gymnosperm subclasses Cycadidae, Ginkoidae, Gnetidae and Pinidae. Their analysis revealed that the orders Pinales, Welwitschiales, Ephedrales and Gnetales have presumably lost the NDH complex. In fact, all the species under study but one (Pinus koraiensis that has retained a putative ndhK) lost all the components of the NDH complex encoded in the chloroplast genome (Figure 3). In the Cycadaceae family, the genus Cycas is of particular interest due to the observation that two out of the three species have presumably lost the NDH complex, even though their plastomes still retain 3 or 4 subunits of the full set of 11 subunits. All three species from this genus have retained ndhF, one of the transmembrane proton channels that have arisen from an MRP ancestor. While this may indicate the progression of ndh gene loss (assuming endosymbiotic gene transfer to the nucleus has not occurred), it is also possible that there has been selective pressure for the retention of these specific subunits. Cyanobacterial NDH subunits assemble into multiple configurations, with each type having a distinct function [51,52], and the Escherichia coli CI homolog of NdhF, NuoL, has been shown to complement the loss of the MRP protein MrpA in Bacillus subtilis. Therefore, it seems possible these genes have been retained, because they confer some advantage. In summary, our phylogenetic analysis (Figure 2) agrees with the previous suggestion of multiple independent losses of NDH throughout gymnosperm evolution, but does not reveal the genetic and/or environmental factors that allow plants to dispense with NDH.
Looking for the selective pressure underlying NDH evolution in angiosperms
Angiosperms, or flowering plants, are a group of plants that have not only diversified rapidly, but also conquered a wide range of habitats. Similar to gymnosperms, phylogenetic analysis shows that angiosperms have lost the ndh genes independently in multiple lineages throughout evolution. The relationships between the species analyzed by us (Figure 2) do not allow us to pinpoint the reasons for these losses, but a few hypotheses are testable with the currently available dataset. Within the angiosperms, orchids present an interesting situation, as pseudogenization and deletion of ndh genes are widespread throughout the Orchidaceae family [48,54–57]. Because many orchids (order Asparagales) show lack of plastid genes for NDH subunits, a trend also reported for parasitic plant species from multiple clades (e.g. [58–60]), of which there are many example species within the orchids (e.g. [61,62]), we considered the possibility that NDH loss is due to the transition to a non-photosynthetic lifestyle (i.e. the transition to parasitism). In addition to parasitism, loss of plastid genes for NDH has also been shown in carnivorous plant species, and may be associated with the appearance of mixotrophy in these organisms . It is known that transition to a non-photosynthetic lifestyle in plants results in a reduction in the chloroplast genome. Typically, this process starts with the loss of the ndh genes, followed by other photosynthesis-related genes. The former may be due to a relaxed dependency on photosynthesis for carbon assimilation, with the stress response of NDH-mediated CEF being the most dispensable reaction under the metabolic dependency on photosynthesis [60,64]. Higher-resolution trees of the orchid family have been constructed with sequence data, which show that heterotrophic orchids have lost genes for a functional NDH complex within their plastome . While all Orchidaceae species proposed to still retain the complex are autotrophic [57,65], there is no clear distinction between hetero- and autotrophy, as there are examples of many autotrophs within the family that likely have no functional plastid ndh genes. Therefore, it appears likely that the loss of NDH in the heterotrophic orchid species is a result of the transition to the non-photosynthetic lifestyle. However, this hypothesis cannot explain the loss of NDH in multiple autotrophic orchid species, or the losses within other non-Orchidaceae angiosperms such as Erodium (Geraniaceae) and C. gigantea (Cactaceae; Figure 2A; [47,66]).
Alternatively, and because many orchids are epiphytic or lithophytic, we considered a scenario where the NDH loss within the orchid family is related to nutrient uptake and/or plant water relations. A model family to test this hypothesis is the Orchidaceae family, as it contains terrestrial (soil-growing) or non-terrestrial (epiphytic or lithophytic) species. We constructed a higher resolution tree containing epiphytic, lithophytic and soil-growing species (Figure 2B). In green or blue are those taxa likely possessing or lacking NDH, respectively. In addition, epiphytic and lithophytic plants are colored in gray, while terrestrial species are colored in black. This tree shows examples of NDH retention in non-terrestrial (e.g. Masdevallia species) and NDH loss in terrestrial orchids (e.g. Goodyera fumata), leading us to reject the hypothesis that water relations or nutrient uptake are factors relevant to NDH loss within the photoautotrophic orchids.
Crassulacean acid metabolism (CAM) photosynthesis has arisen multiple times and has been shown to correlate with epiphytism in the orchids . Since several other CAM species have been found to lack NDH (e.g. C. gigantea) one could hypothesize that CAM could lead to the dispensability of ndh genes (e.g. due to thermodynamic constraints on PQ reduction by pmf, of which little is known in CAM species). However, within the orchids, there are several examples that challenge this explanation. The genera Cattleya and Sobralia employ CAM, yet retain ndh genes, while Goodyera uses the C3 photosynthetic pathway, but has lost the ndh genes [57,67].
The curious case of ndhB
In orchids, it has been noted in several lineages that the ndhB gene in the plastid genome is conserved when all other subunits are lost . This could be explained by two alternative hypotheses: (i) ndhB is retained due to its presence in the inverted repeat region of the plastid genome (which has a lower mutation rate than the two single-copy regions of the plastid genome; [68,69]), and retention of the gene is an example of spandrel evolution, or (ii) the NdhB protein has adopted a different (NDH complex-independent) function in these species. To assess the conservation of the retained B subunit genes, we performed an alignment of the predicted amino acid sequences of NdhB with high resolution in the orchid genus Dendrobium. This analysis revealed strong sequence similarity, with half of the deviations from the A. thaliana reference sequence being shared with at least one other monocot. Of the remaining deviations, only two are conserved between the cyanobacterial sequence and A. thaliana (L371S and L398S, A. thaliana numbering; Figure 4A), and are not expected to be RNA editing sites. To see how changes in these residues may effect function, we modeled the Dendrobium brymerianum NdhB protein structure using the recent cryo-EM structure of cyanobacterial NDH-1 as a template ([24,70]; Figure 4B). Both of these residues are predicted to lie at the stromal side of the proton channel at the end and start of helices 12 and 13, respectively. It is unclear how these residue changes would impact any ion transport activity possibly remaining in a solitary NdhB subunit. However, only L371 is conserved in CI of some (but not all) lineages. It is possible that substitution of a polar residue within the hydrophobic stretch spanning across L371 disrupts the structure. However, the hydrophobicity of this region is disrupted in other CI lineages without a defect in proton translocation of this channel (see , Supplemental data). Therefore, we cannot rule out that, if expressed, NdhB is functional independently of the complex in these species.
Conservation of the NdhB protein structure in
In the absence of proteomic or transcriptomic data, it is currently unclear whether the retained ndhB genes are expressed and their gene product accumulates at the protein level in any of these lineages. Gene retention may, in fact, be a consequence of increased stability of the inverted repeat region of the plastid genome . However, similar to the case of NdhF, since the CI homolog of NdhB in E. coli, NuoN, has been shown to complement B. subtilis mutants of the MRP protein MrpD , the possibility that these genes have been retained due to a new, or reverted to an old, function independent of the PQ-oxidoreductase complex remains in the absence of contrary experimental data.
The function of NDH in plants: comments on dispensability in evolution
Within our phylogenetic tree, we have been unable to pinpoint a shared trait or environment that all NDH-deficient species share, and we, therefore, are still left with the question: what has relaxed selective pressure to maintain NDH? We may begin to answer this by addressing each proposed function of NDH and what changes would need to occur to render that function redundant. In this section, we address all the current models of NDH function in chloroplast physiology and offer testable hypotheses of changes that may have occurred to allow for each to be lost.
In C3 plants, even under permissive conditions (e.g. unstressed, low photorespiration rates), an ATP deficit is predicted to occur due to the ratio of proton coupling to electron transfer, c-ring stoichiometry of ATP synthase, and the energetic demands of the CBB cycle (discussed in ). When plants are under stress or employ a carbon-concentrating mechanism (i.e. C4 or CAM), the ATP demand is further increased, and, in the absence of other pathways, an increase in electron flux through CEF is required. NDH-mediated CEF (Figure 1) augments ATP production in the chloroplast by shunting electrons from PSI into the PQ/PQH2 pool. Due to the H+/e− stoichiometry of NDH and the Q-cycle , for each PQ reduced by NDH-mediated CEF, 8H+ are transferred from the stroma to the lumen, thus increasing the output of ATP/NADPH from the thylakoid electron transport chain more efficiently than those CEF pathways that only reduce PQ (with an overall H+/e− stoichiometry for NDH-CEF of 4). In C4 plants, NDH has been observed to accumulate in the chloroplasts of cells in which the CO2-concentrating mechanism leads to an increased ATP demand (i.e. bundle sheath cells in NADP-ME type C4 species and mesophyll cells in NAD-ME type C4 species) [72,73]. Additionally, mutants in C4 species lacking NDH have decreased CO2 assimilation efficiencies and increased rates of photorespiration [74,75], supporting the idea that the C4 cycle relies on ATP generated via NDH-mediated CEF. Mutants lacking NDH show delayed growth under certain environmental conditions [17,76], and several mutants have been identified with increased rates of CEF through the NDH complex, likely in response to an increase in the accumulation of reactive oxygen species signaling ATP/NADPH imbalance [9,77–79].
The low content of NDH in C3 thylakoids  and the lack of perceivable growth phenotypes presented by ndh mutants seem to support the redundancy of NDH in the CEF pathway, at least in the C3 lineages. It is possible that increased flux through the other alternative electron transfer pathways present in chloroplasts [2,5,19,50], such as the FQR or a type II (non-protonmotive) NADH dehydrogenase [80–82], alone is sufficient to augment ATP production in lineages that have lost NDH.
In terms of energetics, instances of NDH loss in CAM species such as the cactus C. gigantea, several species within the orchids, and possibly Welwitschia mirabilis (for discussion see [83,84]) are intriguing, as CAM has similar energetic demands as C4 photosynthesis . While NDH was likely lost before the divergence of Welwitschiales, Ephedrales and Gnetales, higher phylogenetic resolution of Cactaceae is needed to see the extent of NDH loss within this clade before a hypothesis can be proposed as to the genetic dispensability of the complex. Currently, there is only one sequenced chloroplast genome from a species within this family . Still, the question remains as to how the energetic demands of CAM photosynthesis are met in these species. NDH is the most efficient of the alternative electron transport pathways in terms of ATP generation, so it is surprising to see instances in which loss of the complex coincides with increased ATP demand. As little is known about the extent of pmf in these species, it is possible that pmf is thermodynamically limiting PQ reduction through NDH, making the less efficient pathways for generating ATP favored and NDH dispensable.
It is also possible that NDH, like its respiratory analog CI, has multiple roles within the thylakoid beyond quinone reduction. The canonical reaction of CI, and therefore NDH, is a 2e− reduction of PQ to PQH2, with 4H+ pumped across the membrane generating pmf . In most conditions, the reduction of PQ is thought to be the thermodynamically favored reaction (i.e. ΔG < 0) and leads to efficient ATP generation. Under conditions of higher pmf, some prokaryotes have been shown to run CI in the reverse direction, consuming pmf to drive the oxidation of quinol to generate reducing power [87,88]. Under some conditions, this reaction has been proposed to occur in the NDH complex as well, as the protonmotive nature of NDH  allows for a reversible complex to use pmf to drive the oxidation of PQH2 and supply reduced Fd/NADPH when there is a surplus of ATP (Figure 5) and high pmf (>∼180 mV). These are conditions which may occur during the induction of photosynthesis . Utilization of pmf to drive electron transfer from PQH2 to stromal reducing equivalents is a distinct feature of a protonmotive NDH, as it is an energetic impossibility for the FQR and the heme ci routes of CEF, as these routes do not directly pump protons across the thylakoid membrane.
The reverse reaction of NDH.
The bioenergetic consequences of the PQ reduction and PQH2 oxidation reactions via NDH are opposite, increasing or decreasing the ATP/NADPH production of the thylakoid, respectively. Within our current understanding of the CEF pathways the chloroplast employs to alleviate an ATP deficit, the NDH is the only route that is expected to be capable of reverse electron transfer (i.e. able to oxidize PQH2 and reduce Fd) outside of extreme redox poising of the PQ/PQH2 and NADP+/NADPH pools . Therefore, assuming the reverse reaction of NDH plays a significant role in the flexibility of photosynthesis, dispensability of this reaction would require the consumption of ATP via another pathway under conditions where there is a deficit of reducing power. This could be achieved, for example, by futile cycles that burn ATP, or, perhaps more productively, by an increase in the flexibility of the chloroplast/cytosol ATP exchange machinery  in lineages lacking NDH.
The cation antiporter legacy of the thylakoid CI
As discussed above, CI has been proposed to have evolved in a modular fashion. The proton-pumping module (P-module) arose from either duplication of a cation antiporter or recruitment of an antiporter complex closely resembling MRP Na+/H+ antiporters [22,91,92]. There is some evidence that, in the respiratory CI, this ancestral antiporter activity is retained [91,93–95]. As NDH shares this modular ancestry with CI, it may also be capable of antiporter activity. An NDH complex that also functions as an antiporter would be capable of altering the composition of pmf, by interconversion of ΔpH and Δψ. Loss of NDH would affect the permeability of the thylakoid to the counterions and may lead to the sustained shift of the partitioning of pmf into ΔpH or Δψ. As the ΔpH component is regulatory, this would have consequences for feedback regulation of both light harvesting and electron transfer (e.g. [2,96,97]). Figure 6 illustrates our expected consequences, if NDH acts as an antiporter. If we assume that in the absence of NDH activity pmf is equally distributed between ΔpH and Δψ (Figure 6A), activation of cation co-transport in NDH coupled to PQ reduction would shift pmf toward ΔpH (Figure 6B). Conversely, PQH2 oxidation via NDH would shift pmf towards Δψ (Figure 6C). In this model, lineages that have lost NDH may have been able to compensate by adjusting the composition or activation of the various other antiporters in the thylakoid membrane (e.g. [98–101]).
Effects of proton/cation antiporter activity coupled to electron transfer reactions of NDH on
For all of the above scenarios, it is assumed that loss or pseudogenization of NDH genes represents loss of NDH activity. It is important to note that this assumption is made in the absence of biochemical characterization of the thylakoid electron transport chain in those species. Often, publications describing loss of plastid ndh genes propose that endosymbiotic gene transfer has occurred, and the NDH complex is still active in these plants. However, there has been no evidence supporting this claim from any nuclear genomic or transcriptomic datasets that have been surveyed so far. Future research is needed to determine if gene transfer to the nucleus has occurred, as well as if import into the chloroplast of the large hydrophobic NDH subunits is possible. Additionally, to fully understand the dispensability of NDH, we need to know (i) if NDH activity is taken up by alternative processes in species lacking the complex, (ii) which functions of NDH are complemented by these alternative pathways and (iii) what aspects of chloroplast physiology are different in NDH-deficient species, and how these have changed the selective pressure on ndh gene maintenance.
Crassulacean acid metabolism
cyclic electron flow
ferredoxin-dependent quinone reductase
multiple resistance and pH locus
Fd-dependent quinone reductase
transmembrane proton gradient
D.D.S. and R.B. conceived of this review article. L.D. constructed the phylogenetic trees. D.D.S., L.D. and R.B. wrote the paper.
Work related to NDH in the authors’ laboratory is funded by the Max Planck Society and a grant from the Deutsche Forschungsgemeinschaft (DFG; SFB-TRR 175) to R.B.
We thank Daniel Karcher and Nicholas Fisher for helpful discussions.
The Authors declare that there are no competing interests associated with the manuscript.