Enhancing (crop) plant photosynthesis by introducing novel genetic diversity

Abstract

Although some elements of the photosynthetic light reactions might appear to be ideal, the overall efficiency of light conversion to biomass has not been optimized during evolution. Because crop plants are depleted of genetic diversity for photosynthesis, efforts to enhance its efficiency with respect to light conversion to yield must generate new variation. In principle, three sources of natural variation are available: (i) rare diversity within extant higher plant species, (ii) photosynthetic variants from algae, and (iii) reconstruction of no longer extant types of plant photosynthesis. Here, we argue for a novel approach that outsources crop photosynthesis to a cyanobacterium that is amenable to adaptive evolution. This system offers numerous advantages, including a short generation time, virtually unlimited population sizes and high mutation rates, together with a versatile toolbox for genetic manipulation. On such a synthetic bacterial platform, 10 000 years of (crop) plant evolution can be recapitulated within weeks. Limitations of this system arise from its unicellular nature, which cannot reproduce all aspects of crop photosynthesis. But successful establishment of such a bacterial host for crop photosynthesis promises not only to enhance the performance of eukaryotic photosynthesis but will also reveal novel facets of the molecular basis of photosynthetic flexibility.

This article is part of the themed issue ‘Enhancing photosynthesis in crop plants: targets for improvement’.

1. Background

Because the human population is currently growing by about 1.2% per year [1], agricultural productivity must be raised by more than 50% in order to cater for the needs of approximately 9 billion people by 2050 [2]. Loss of arable land, soil depletion and rising demands for biofuels and food of animal origin will further aggravate the situation. Optimization of photosynthesis—the basis of photoautotrophic growth—by breeding and genetic engineering is an obvious route to crop improvement ([3,4] and references therein). However, unlike carbon allocation and plant anatomy [5], the efficiency of light conversion to biomass (in the following termed as ‘photosynthetic productivity’) has proven to be refractory to improvement so far [6,7]. Must we therefore accept that the light reactions are as efficient as they can be [7], and concentrate our efforts on traits, such as pathogen resistance, which clearly are susceptible to enhancement? Or might a new paradigm in photosynthesis research enable us to overcome recognized obstacles to improving this fundamental process? In this opinion piece, we argue for this latter option and suggest a new experimental approach.

2. Do photosynthetic light reactions grant maximum crop productivity?

Enhancement of the metabolic reactions associated with carbon fixation has received much attention (see e.g. [8–11]), but the light reactions of photosynthesis have acquired an aura of perfection. In part, this can be attributed to the conspicuous lack of publications reporting the success of attempts to boost photosynthetic productivity over the total growth period. Moreover, several characteristics of photosynthesis indeed give an impression of (near) perfection. The quantum efficiency of charge separation is near unity ([12–16] and references therein) and the efficiency of kinetic control due to appropriate cofactor spacing and energy gaps between components [17] is deemed close to ideal. Furthermore, the mechanisms that redirect or effectively dissipate excess light energy ([18–24] and references therein) exhibit a high level of evolutionary optimization. Nevertheless, overexpression of some components involved in the dissipation of excess light energy has very recently been shown to accelerate recovery from photoprotection and to increase biomass of tobacco grown in the field [25]. Another concept for improving photosynthetic productivity focuses on the enhancement of light harvesting by expanding the photosynthetically active light spectrum [26,27], although the use of low-energy light will inevitably lead to increased production of reactive oxygen species ([17] and references therein). Also the reduction of chlorophyll antenna size (reviewed in e.g. [28,29]) to limit excessive light absorption has been considered to increase crop yield by providing more homogeneous light energy supply throughout crop canopies. All these concepts, however, were either not applied to crops yet (expanding light spectrum) or did not translate into consistently increased crop yields so far (reduced antenna size, discussed in [7]) or were only tested for their potential to increase biomass in tobacco (accelerated recovery from photoprotection). Moreover, in addition to the photosynthetic characteristics targeted by the approaches outlined above for improvement, other aspects of the light reactions also appear to be suboptimal. Inefficiencies arise, for instance, from limitations in exciton energy transfer [30,31], photoinhibition of PSI and PSII [32,33], and the interplay between the two photosystems [34].

While only about 60 different polypeptides make up the core of the photosynthetic apparatus of a flowering plant (reviewed in [35]), many more serve essential functions as so-called auxiliary photosynthetic proteins. For instance, the TAIR gene ontology terms ‘photosystem’ and ‘photosynthesis’ yield 143 and 293 distinct protein-coding gene models with full-length cDNAs, respectively (arabidopsis.org; see electronic supplementary material, table S1). Many of these accessory proteins are involved in the assembly, turnover and functional readjustment of the photosynthetic complexes, while the molecular functions of others remain vague or unknown. The waste of energy resulting from protein (complex) instability, lack of catalytic specificity and excessive light harvesting reveals potential targets for improving yield capacity. The multitude of auxiliary factors, together with their inferred increase in number in the green lineage of photoautotrophs [36–38], suggests that in the course of evolution the main reactions needed to be adjusted and supported in various ways [17,39]. But have such adjustments eliminated the ‘imperfections’ or do they represent makeshift solutions? We believe that the entire process of photosynthesis is fundamentally flawed, because it evolved and was optimized for low-light (i.e. marine) conditions in the absence of oxygen, and the process was subsequently not fundamentally changed but ‘tweaked’ to function reasonably well in terrestrial, oxygen-rich environments (reviewed in: [17,39–41]). But in addition to this fundamental flaw with respect to its origin under conditions very different from today, several lines of evidence suggest that our current crop plants have not been fully optimized for photosynthetic productivity in the course of evolution, as outlined below.

3. Has evolution already maximized photosynthetic productivity?

Considering the dramatic environmental changes that land plants have repeatedly experienced in the course of their evolution ([42–44]; figure 1), photosynthesis has evolved to be ‘robust’ rather than highly efficient. By ‘robust’ we mean that photosynthetic productivity was maintained as high as possible under many adverse conditions rather than maximally efficient under ‘convenient’ conditions as provided in modern farming. Moreover, traits other than photosynthesis have doubtless been exposed to much stronger selective pressures. For instance, low soil temperatures strongly constrain water uptake in many modern plant species (reviewed in [54]) and the recent 100 000-year ice-age cycles during the Quaternary [55] must have exerted such selective pressure on Palaearctic and Nearctic species periodically. On more restricted taxonomic scales, selective pressures imposed by pathogens can be enormous, especially in genetically uniform populations [56]. For instance, during the twentieth century, Fusarium oxysporum f. spec. cubense (which causes Panama disease) almost wiped out the then dominant banana cultivar Gros Michel [57,58]. In such cases pathogen-induced selective pressure can easily exceed that encumbering photosynthetic productivity. We reason that strong positive selection acting temporarily on traits different from photosynthetic productivity might have impaired evolutionary optimization of the latter. Similar effects can be expected to have resulted from the five documented prehistoric mass extinction events that have occurred during the evolution of land plants [59,60]. A multitude of coinciding environmental hazards characterizes these events, such as changes in atmospheric composition, temperature and interception of solar radiation (reviewed in [60]), which probably defined key determinants of plant fitness other than photosynthetic productivity. We expect such alterations of the relative contribution of photosynthesis to plant fitness to have alleviated the selective pressure resting on the corresponding genes and thus to have favoured genetic continuity of suboptimal ‘bioenergetic hardware’. Because we presume that the primary effect of selection during plant evolution is robustness of the photosynthetic process in terms of optimized plant fitness, enhancement of the productivity of the process should in principle be possible.

Figure 1.

Modern crops are depleted of genetic potential for enhancing photosynthetic productivity. As an example, photosynthetic evolution in the cereal lineage and its ancestors is illustrated qualitatively. Multiple severe bottlenecks occurred during the rise of land plants (approx. 475 Mya), seed plants (approx. 324 Mya), angiosperms (approx. 209 Mya) and monocots (approx. 136 Mya) [45], as well as during the evolution of grasses (approx. 78 Mya) [46]. Domestication (approx. 10 kya) [47] and elite-line breeding during the Green Revolution (over the past 60 years) [5,48] leading to the fixation of domestication/high performance traits caused additional losses of genetic variability [49]. Deleterious environmental variations such as the Carboniferous O₂ maximum (COM; 286 Mya) [50], which probably favoured low photorespiration (PR) modes of photosynthesis, the Griesbachian and Smithian temperature maxima (GSTM; approx. 252 and approx. 251 Mya) with average sea-surface temperatures of up to 41°C [51] and the Brunhesian glaciation cycles (BGC; indicated by sea-surface temperature minima approx. 23 kya, approx. 150 kya, approx. 265 kya, approx. 370 kya and at least four more cycles fitting an approx. 100-ky glaciation periodicity) [52,53], will all have exerted strong selective pressures that did not necessarily promote optimization of photosynthetic productivity. Thus large amounts of the adaptive variation once available have presumably been lost during periods of extreme environmental change (diagram area, curved arrows).

4. Crop plants offer a poor basis for enhancing photosynthetic productivity

With few exceptions such as nori (rhodophyte Porphyra spec.) or nut pines (gymnosperm Pinus spec.), the vast majority of current crops are angiosperms. The evolutionary origin of angiosperms has been dated to 194 (210–162) Mya, while the first land plants evolved 475 (480–471) Mya [45]. Today, cereals are our major crops, being cultivated on about 61% of arable land [61] and accounting for more than 50% of human dietary energy supply [62]. All cereals belong to a single plant family (Poaceae) which emerged about 82–74 Mya [46]. Thus, a single, relatively young evolutionary lineage is now responsible for feeding the majority of the world's population.

Conventional breeders largely rely on existing natural genetic variation, but angiosperms display low levels of genetic and protein structural diversity in their photosynthetic apparatus (see e.g. [63–66]). Only marginal photosynthetic differences have been detected even in wild relatives of modern crops (e.g. [67,68]), and with regard to photosystem (PS) II, PSI and plastocyanin (PC) oxidation kinetics, among-group variability markedly surpasses within-group variability in angiosperm and gymnosperm species [69].

We believe that this dearth of natural (genetic) variation for photosynthetic light reactions can in part be explained as follows. The lineage leading to cereal crops experienced its last common bottleneck event (82–72 Mya) at a point when over 80% of the period during which photosynthesis adapted to life on land had already elapsed (figure 1). Moreover, angiosperms have been in existence for about only 40% of this time (since about 210 Mya). Consequently, the plant species that are crucial for global food security have effectively been deprived of most of the variation that made earlier evolutionary inventions possible (figure 1). Furthermore, genes or alleles that were intrinsically favourable for photosynthetic productivity could have been lost during evolutionary bottlenecks if other factors were more essential (see chapter above). In addition, most ground-breaking innovations in land plant evolution—such as vascularity, seed set, flower production and zoophily [70]—are not directly related to photosynthetic productivity. Thus, selective sweeps or genetic hitch-hiking effects [71,72] might have caused severe linkage drags if photosynthetic variants were swept to fixation while being chromosomally linked to functionally unrelated genes. Such effects can decimate the genetic ‘seed capital’ of descendant clades, in particular when they coincide with genetic bottlenecks [73]. Finally, land plants (like most organisms) display low genome-wide mutation rates per generation (in the range of 1–10 mutations per genome and generation) [74,75], and, compared to nuclear and mitochondrial DNA, mutation rates are even further reduced in chloroplasts [76] the genomes of which encode many core subunits of the photosystems. This makes the photosynthetic apparatus likely to stay trapped on inherited local fitness maxima defined by certain allele combinations [77], especially in light of the elaborate mechanisms that buffer the activity of photosynthesis.

The assumption that selection has favoured the robustness, rather than the sheer productivity, of photosynthesis, together with the notion that alleles (or allele combinations) favouring highly efficient photosynthesis might easily have been lost during genetic bottlenecks and are hard to ‘re-invent’ due to the genetic conservatism of higher plants, provides a plausible explanation for the scarcity of genetic variation for photosynthesis in angiosperms. In addition, serial genetic bottlenecks in the course of domestication and elite-line breeding have further depleted modern crops of genetic diversity [78–80] (reviewed in [81]) aggravating their already pronounced, phylogeny-related molecular limitations.

These considerations make it likely that in the course of angiosperm evolution, a combination of periods of weak selection for photosynthetic productivity and repeated loss/rebound of genetic diversity within an already genetically restricted system has led to a situation in which crop plants possess a robust but not maximally efficient photosynthetic machinery with little relevant genetic variation. If such variants exist, they presumably survive at low frequency in wild relatives, at the family level or among all angiosperms. However, screening the 295 000 known species of angiosperms [82] for photosynthetic variations is not really a practical proposition.

5. Beyond angiosperms: photosynthesis can be done differently—and probably better

As outlined above, angiosperms harbour little genetic variation relating to photosynthesis, but recent findings imply that more stress-resilient modes of oxygenic photosynthesis can be found elsewhere. In fact, the green alga Chlorella ohadii, which occurs in desert sand crusts, has a D1 protein that is virtually insensitive to photodamage and can convert amounts of light energy that exceed the maximum level of solar radiation received on Earth into photosynthates [83]. Therefore, C. ohadii provides an example for enhanced ‘specific robustness’ of photosynthesis, because the beneficial effects of this photosynthetic modification are most probably limited to high light conditions but might be detrimental under other adverse environmental conditions. Reversibility of photobleaching—a valuable adaptation to long-term exposure to high light in crops—has been reported for chlorophytes [84,85] and rhodophytes [86]. Moreover, in cyanobacteria, the use of different D1 isoforms, such as the ‘low-light’ versus ‘high-light’ D1 isoforms has been reported, a capacity absent in photosynthetic eukaryotes [87]. Biogeological records imply that plants might once have possessed other traits that would be considered highly valuable today. For instance, plant physiology must have adapted to concentrations of atmospheric O₂ of up to 35% that prevailed in the Late Carboniferous period. Such adaptations probably included anatomical changes [88], carbon concentration mechanisms [89,90] or even more selective RuBisCO variants. However, such adaptations would have become obsolete after the disappearance of the high-O₂/low-CO₂ atmosphere of the Upper Carboniferous and were probably lost again. Nevertheless, signatures of positive selection acting on RuBisCO in red algae have recently been found to correlate well with declines in atmospheric CO₂ levels [90].

Taken together, these findings suggest that evolution is quite capable of generating photosynthetically more efficient eukaryotes. They currently exist among algal lineages and could have been present in the progenitor lineage of land plants. These considerations lead to two conclusions: (i) The potential to enhance crop photosynthetic productivity exists, given that tools are available to identify the required genetic diversity, and (ii) photosynthesis could, in principle, be enhanced by introducing algal/bacterial ‘inventions’, such as carboxysomes [10,91–93], into land plants.

6. Can we enhance the photosynthetic light reactions with our current toolbox?

The relative paucity of extant natural variation nevertheless poses a formidable challenge. Ancient variants cannot be easily resurrected due to the lack of molecular fossils, and such extinct alleles must be inferred phylogenetically [94]—which, in the case of rare alleles (presumably the most interesting) is virtually impossible. Tools for the engineering of protein complexes have been employed for more than 20 years [95,96]. However, to the best of our knowledge, no targeted modification of the photosynthetic light reactions has predictably enhanced their productivity. This can be generally attributed to our limited understanding of how the properties of protein assemblies emerge from their primary structures, and more specifically to the complexity of the regulation, biogenesis and functioning of the photosynthetic complexes.

Evolutionary breeding in crop plants seems unsuitable to provide us with enhanced photosynthetic machineries in a reasonable time frame. Spontaneously new allelic variants arise with (for this endeavour far too) low overall efficiency as implied by nucleotide substitution rates of angiosperm nuclear genomes ranging from 1.5 × 10⁻⁸ to 6.1 × 10⁻⁹ per site and year [97–99]. In addition, breeding populations necessarily have to be small, since they compete with producer populations. Random mutagenesis has been used to quickly generate new genotypes in most cereal species, but its potential is greatly limited by the nature of their genomes and limits on screenable population sizes, so that the approach has not yielded superior photosynthesizers so far [100]. In light of these practical restrictions on the rapid evolution of superior traits in crop plants with small (experimentally accessible) populations and long generation times, we need to turn to a fundamentally different kind of genetic platform.

7. Outsourcing crop photosynthesis to cyanobacteria—anticipating plant adaptation ex planta

In principle, adaptive evolution of crop plants must be outsourced to an organism that is easy to manipulate, can rapidly generate large populations of selectable individuals and is compatible with photoautotrophic metabolism. At first glance, a green algal species like Chlamydomonas reinhardtii seems to be the system of choice because it hosts a eukaryote-type photosynthetic machinery and has a short generation time. However, its inaccessibility to efficient (prokaryote-type) genetic engineering technologies like efficient transformation, homologous recombination and shot-gun complementation [101] makes C. reinhardtii not attractive for complex adaptive evolution approaches that require the introduction and modification of entire batteries of plant genes. In other words, crops need to be reduced to a bacterial scale rather than an algal scale—i.e. bacteria should be used as proxies for crop plants. Concepts involving synthetic biology approaches to the transfer of photosynthetic modules into prokaryotes and their optimization in their new hosts have already been introduced [39,102,103]. Instead of employing in vitro evolution and protein engineering approaches, we suggest that in vivo evolution accelerated by mutagenesis could provide a far more versatile, pragmatic and promising way to enhance photosynthesis.

Cyanobacteria like Synechocystis sp. PCC6803 meet all the platform standards mentioned above [39,101]. Some strains even allow one to bridge fitness valleys of intermediate genotypes by facultative heterotrophic growth. In order to undertake large-scale alterations of plant photosynthesis ex situ by random mutagenesis, these strains must be converted into suitable genetic chassis (i.e. ‘free-living chloroplasts’; figure 2) by replacing the endogenous cyanobacterial components of photoautotrophy with their plant homologues, creating synthetic prokaryotes capable of eukaryote-style photosynthesis. This includes not only the proteins constituting the photosynthetic multi-protein complexes but also enzymes for producing plant-specific pigments, auxiliary proteins for the assembly, maintenance and regulation of plant photosynthesis and regulatory elements to allow expression of plant genes in a cyanobacterium. Therefore, this poses a formidable challenge, and experimental solutions will employ approaches at the interface of genetic engineering and synthetic biology. Such strains could then be mutagenized and selected for photosynthetic machineries that are resilient against current and predicted environmental hazards. In this way, genetic changes of adaptive value could be identified after isolating superior cells based on non-invasive measurement of chlorophyll fluorescence parameters or simple growth advantage under photosynthesis-relevant conditions. The identified genetic changes could then be transferred into plants in far less time than would be required for them to evolve naturally in planta.

Figure 2.

Constructing a synthetic prokaryote capable of eukaryotic photosynthesis. Insets depict the architecture of pro- and eukaryotic protein complex architecture and indicate degrees of subunit conservation: light/medium shading, highly/moderately conserved proteins; dark shading, species-specific subunits. Plant-specific components of the light-harvesting complexes (LHCI + II), cyclic electron flow (PGRL1, PGR5; [104]) and state transitions (TAP38, STN7; [105]) are indicated. In the interests of clarity, ATP synthase complexes are not shown.

A simple calculation illustrates the potential of outsourcing crop evolution to prokaryotes. When grown photoautotrophically, the doubling time of Synechocystis is about 10 h [106], so that 1000 generations can be grown in a little more than 1 year. The Neolithic revolution, which encompassed the advent of agriculture, took place about 10 000 years ago [47]. Assuming the same mutation rate and population size for our synthetic prokaryote as for the donor crop (with a generation time of 1 year), in Synechocystis one could recapitulate natural evolutionary innovations that occurred in a crop species since that time within just over a decade. Moreover, the natural mutation rate of 2 × 10⁻⁹ to 5 × 10⁻⁷ for Synechocystis [107] is up to 30 times higher than that of the fastest-evolving angiosperms (see above), and artificial mutagenesis, as described by Tillich and colleagues [108], can easily be expected to accelerate Synechocystis mutation rates by a factor of 10. Taking all of this together, all gain-of-function mutations that a single crop lineage may have accumulated since its domestication could in theory be generated in a synthetic Synechocystis strain within about 13 days (figure 3).

Figure 3.

Synthetic bacteria have unmatched evolutionary potential for crop improvement. Simulated data for accumulation of allelic variants in photosynthesis-related genes in single genetic lineages are presented. A total DNA sequence length of 349 399 bp for 293 gene models was inferred from TAIR database GO annotations for ‘photosynthesis’ (see above; electronic supplementary material, table S1). We assume mutation rates of 5×10⁻⁷ for Synechocystis (Synnormal) and 1.5×10⁻⁸ for angiosperm crops (Angio), as well as a potential 10-fold increase in Synechocystis mutation rates inducible by artificial mutagenesis (Synhigh_mut). The large panel represents the accumulation of point mutations in a crop plant lineage since its domestication. The inset represents the timespan from the start of a hypothetical Synechocystis adaptive evolution experiment until the inferred date of point mutation parity between Synechocystis and crop lineages. Mutagenized Synechocystis strains are predicted to accumulate an equal amount of point mutations as a crop lineage has since its domestication approximately 10 kya within 12.6 days (i.e. 52.9 SNPs in approx. 350 kb of photosynthetic genes). By increasing the number of individuals by a factor of 283, syntheticSynechocystis could generate within a single year as many allelic variants as have arisen in a single crop plant lineage since the evolution of the Poaceae (82 Mya). Increasing this factor to 726 would allow one to ‘recapitulate’ the level of allelic variation accumulated since the rise of the angiosperms (210 Mya). (Online version in colour.)

Moreover, the number of selected individuals can easily be increased beyond the current total population size of any given crop species. In the USA, 33.6 × 10⁶ ha of harvested area were devoted to maize cultivation in 2014 (faostat3.fao.org). Maize is currently cultivated at densities of about 80 × 10³ plants ha⁻¹ [109], such that about 3 trillion (2.7 × 10¹²) individuals are planted in the USA per year. With about 7 × 10⁷ cells ml⁻¹ at a culture density of OD_{730 nm} = 1 [110], more Synechocystis cells can be grown in a culture volume of just 38 l than the total number of maize plants grown in the USA in a year. This estimate obviously exceeds by several orders of magnitude the size of the maize population used for breeding purposes, underlining the enormous evolutionary potential of a bacterial platform.

Synechocystis offers additional advantages, e.g. it can tolerate a much higher mutation load than crop plants, which tend to develop male or female sterility upon artificial mutagenesis [111] (reviewed in [100,112]). This greatly increases the probability of accumulating the level of multisite alterations presumably needed to bridge fitness valleys between different photosynthetic variants. Because Synechocystis is polyploid, many selectable beneficial mutations are likely to be gain-of-function mutations. This facilitates the transfer of modified versions of existing components and the introduction of new ones into crop plants.

8. The cyanobacterial chassis: limitations and challenges

Every adaptation of the photosynthetic machinery to certain conditions/stresses may be maladaptive under other conditions. Moreover, the unicellular Synechocystis system has clear limitations with respect to inter-organellar (e.g. plastid-to-nucleus signalling) and multicellular (e.g. functional differentiation of tissues, anatomical changes, hormone networks and their regulatory versatility) aspects of photosynthesis, as well as sink-source relationships. Moreover, the carbon-concentrating mechanisms of cyanobacteria function to guarantee inorganic carbon uptake from water that contains low amounts of dissolved inorganic carbon, such that plant CO₂ fixation mechanisms might be difficult to optimize in Synechocystis. Hence, the main potential of the system lies in the improvement of the molecular foundation of photosynthetic productivity, for instance by increasing protein complex stability or reinforcing protective mechanisms. Furthermore, proof for the feasibility of this ‘forced evolution’ approach has been provided recently by generating thermo-tolerant Synechocystis strains [108,113].

The adaptive evolution strategy represents a brute force approach that compensates for the very low incidence of truly superior genetic variants in random mutagenesis screens by using astronomically high numbers of organisms. But considering the limitations of conventional crop breeding (with respect to both resources and understanding), this approach could prove to be the most powerful and practical way to identify novel variation for a trait like photosynthesis, for which the level of currently available genetic diversity is low. Clearly we cannot expect to generate perfect crops by this approach alone, and other strategies/disciplines will no doubt contribute to further improvement of photosynthesis. Still, the approach outlined has great potential for generating new building blocks for the photosynthetic patchwork and for elucidating basic features of its function. It might even constitute the first step towards adaptive-change libraries that provide us with a multitude of specialized and situation-dependently applicable photosynthetic alleles. Furthermore, generating bacterial or crop lineages that excel under highly specific and controlled conditions may turn out to be very useful when it comes to providing roof-top farms in cities, or even cultivation complexes on future space stations and extra-terrestrial outposts with tailored primary producers.

Taken together, the suggested transfer of plant photosynthetic machineries into a prokaryotic chassis poses a formidable challenge, regardless of whether all the relevant genes or ‘just’ the major protein complexes are targeted. Many obstacles will have to be overcome to achieve true metabolic compatibility, but at the same time many new discoveries relating to biogenesis, assembly and maintenance of photosynthetic protein complexes will certainly be made.

Authors' contributions

M.D. and D.L. conceptualized the manuscript. M.D. drafted the manuscript and designed the figures. M.D. and D.L. revised and finalized the manuscript outline and content.

Competing interests

We have no competing interests.

Funding

D.L. received funding for this work from the DFG: Graduate School GRK 2062 (Molecular Principles of Synthetic Biology) and the Transregional Collaborative Research Centre TRR 175 (The Green Hub—Central Coordinator of Acclimation in Plants).