Genetic, Genomics, and Responses to Stresses in Cyanobacteria: Biotechnological Implications

Cyanobacteria are widely-diverse, environmentally crucial photosynthetic prokaryotes of great interests for basic and applied science. Work to date has focused mostly on the three non-nitrogen fixing unicellular species Synechocystis PCC 6803, Synechococcus PCC 7942, and Synechococcus PCC 7002, which have been selected for their genetic and physiological interests summarized in this review. Extensive “omics” data sets have been generated, and genome-scale models (GSM) have been developed for the rational engineering of these cyanobacteria for biotechnological purposes. We presently discuss what should be done to improve our understanding of the genotype-phenotype relationships of these models and generate robust and predictive models of their metabolism. Furthermore, we also emphasize that because Synechocystis PCC 6803, Synechococcus PCC 7942, and Synechococcus PCC 7002 represent only a limited part of the wide biodiversity of cyanobacteria, other species distantly related to these three models, should be studied. Finally, we highlight the need to strengthen the communication between academic researchers, who know well cyanobacteria and can engineer them for biotechnological purposes, but have a limited access to large photobioreactors, and industrial partners who attempt to use natural or engineered cyanobacteria to produce interesting chemicals at reasonable costs, but may lack knowledge on cyanobacterial physiology and metabolism.


Introduction
Cyanobacteria are ancient Gram-negative prokaryotes that perform oxygenic photosynthesis and are phylogenetically close to recently discovered non-photosynthetic bacteria termed Melainabacteria and Sericytochromatia [1,2]. Cyanobacteria are regarded as the producer of the atmospheric oxygen (O 2 ) of Earth [3] and the ancestors of the plant chloroplast [4]. Cyanobacteria capture solar energy at high efficiencies (3-9%) [5] to power up their efficient photoautotrophic metabolism, which fixes huge amounts of inorganic carbon (CO 2 , NaHCO 3 , and Na 2 CO 3 ) and nitrogen (N 2 , NH 4 , NO 2 , NO 3 , or urea) [6,7], into an enormous biomass [8] that supports a large part of the food chain.
By colonizing aquatic ecosystems (fresh, brackish, and marine waters) and soils (including deserts) of our planet, cyanobacteria are inevitably exposed to multiple stresses such as solar ultraviolet radiations and variations in light intensity and quality, inorganicnutrients availabilities, temperatures (high and low), salinity, pH (acidic and basic), drought, and pollutants (herbicides and heavy-metals). In addition, cyanobacteria are involved in numerous interactions with competitors, predators, or symbiotic hosts [9]. Consequently, it is not surprising that cyanobacteria have evolved as a widely diverse organisms, which are of high interest for basic and applied research [10]. They display various metabolisms and morphologies [11,12], and numerous species can differentiate cells, akinetes (spores) and/or heterocysts, which are dedicated to growth or survival under adverse conditions [13,14]. Most of our knowledge on cyanobacteria came from studying the three non-nitrogen fixing, unicellular cyanobacteria Synechocystis PCC 6803, Synechococcus PCC 7942 (formerly named "Anacystis nidulans") and Synechococcus PCC 7002 (formerly named "Agmenellum quadruplicatum PR6") that are (i) straightforward to culture under laboratory conditions, (ii) amenable to genetic manipulation, and (iii) freezable for long-term storage. In this review, we will summarize the genetic and physiological properties of these models, emphasizing on their tolerance to stresses, and the recent progress in their genetics. We will also put forward that in spite of more than three decades of intensive research the genomes of these models still contain a large number of genes and small RNAs (sRNAs) of unknown function. Furthermore, many of the genes annotated "by sequence analogy" with those genes characterized in intensively studied, non-photosynthetic, models such as Escherichia coli might in fact have a different function in cyanobacteria. This situation makes comparative genomics and metabolic modeling difficult. Consequently, we will discuss that to better understand and exploit the wide biodiversity of cyanobacteria, strong efforts should be put in large-scale analysis of genes and sRNAs functions in model cyanobacteria. Finally, we will emphasize that we need to identify and thoroughly study Most of our knowledge on cyanobacteria came from studying the three non-nitrogen fixing, unicellular cyanobacteria Synechocystis PCC 6803, Synechococcus PCC 7942 (formerly named "Anacystis nidulans") and Synechococcus PCC 7002 (formerly named "Agmenellum quadruplicatum PR6") that are (i) straightforward to culture under laboratory conditions, (ii) amenable to genetic manipulation, and (iii) freezable for long-term storage. In this review, we will summarize the genetic and physiological properties of these models, emphasizing on their tolerance to stresses, and the recent progress in their genetics. We will also put forward that in spite of more than three decades of intensive research the genomes of these models still contain a large number of genes and small RNAs (sRNAs) of unknown function. Furthermore, many of the genes annotated "by sequence analogy" with those genes characterized in intensively studied, non-photosynthetic, models such as Escherichia coli might in fact have a different function in cyanobacteria. This situation makes comparative genomics and metabolic modeling difficult. Consequently, we will discuss that to better understand and exploit the wide biodiversity of cyanobacteria, strong efforts should be put in large-scale analysis of genes and sRNAs functions in model cyanobacteria. Finally, we will emphasize that we need to identify and thoroughly study new cyanobacteria endowed with natural properties of interest for basic or applied researches, and test whether the synthetic biology tools developed for model strains can be used to facilitate the engineering of these newly identified cyanobacteria so as to turn their promises into industrial realities.

Interest and Current Limitations of Comparative Genomics
The comparative analysis of cyanobacterial genomes allows to determine which genes are present in any particular genome and which ones are absent. These findings serve to determine the pan-genome, which describes the entire genes set of all cyanobacteria analyzed. It includes (i) a core-genome, which comprises the sets of genes that exist in each of the strains analyzed [71] and (ii) the dispensable-genome, which comprises the genes present in a subset of the studied strains and species-specific genes [72]. For such analyses, strains can be selected on the basis of their natural habitats [73], their physiology and/or morphology [74], and/or their phylogenetic position determined by 16S-rRNA sequence comparison [75]. Such comparative genome analyses revealed habitatspecific (or heterocyst-specific) cyanobacterial proteins [72]. It was also revealed that cyanobacteria occupy a unique position among prokaryotes as a hub between anaerobes and obligate aerobes [76] and that the earliest cyanobacteria were small and unicellular, while filamentous forms appeared shortly afterwards [77]. Comparative genome studies were also used to discuss plastid evolution [78]. When applied to two closely related cyanobacteria, comparative genomics can be used to predict the behavior of one strain relative to the other one. As a rare example of such comparison, Synechocystis PCC 6803 was found to be more resistant to zinc excess than Synechocystis PCC 6714 as it was predicted [79]. Comparative genomic information is also crucial for genome-based reconstruction of an organism's metabolism as done for Synechocystis PCC 6803 to predict which metabolic reaction or pathway should be engineered to increase the production of biotechnologically interesting chemicals [80][81][82][83][84][85].

Danger of Genome Annotation Based Only on Sequence Comparison
Generating robust and accurate genome-scale models (GSMs) is an iterative process dependent on the expansion and updating of draft simulation models with available experimental data obtained after measurement of fluxes and metabolic pool sizes [80,86,87]. However, GSMs are largely based on gene function predicted "by sequence analogy" with genes characterized in other model organisms, which may behave differently in the studied organism [88]. In Synechocystis PCC 6803, the best studied model, only ca. 1050 coding sequences (~30%) have assigned functions [80,89], which may turn out to differ from those determined in well-studied heterotrophic models. For examples, the Synechocystis PCC 6803 LexA protein regulates the expression of genes involved in carbon assimilation [90], not DNA repair as it occurs in E. coli. Furthermore, the NAD(P)H dehydrogenase transcription factor (NdhR, Sll1594), homologous to the proteobacterial regulators of the Calvin−Benson−Bassham metabolic pathway (CbbR, hereafter referred to as Calvin cycle), does not regulate the transcription of Calvin cycle genes [91]. Instead, NdhR regulates numerous genes involved in the transport and assimilation of inorganic carbon (CO 2 and bicarbonate) [6,92,93].
In addition, a single cyanobacterial protein can have more than one function. The fructose-1,6/sedoheptulose-1,7-bisphosphatase (FBP/SBPase) operates in both the Calvin cycle and gluconeogenesis [94]. The CrtLdiox protein is also dual-function enzyme with both lycopene cyclase and dioxygenase activity [95]. The IsiA protein is active in both the light-harvesting ability of PSI and the photoprotection of PSII [96]. The AgrE enzyme catalyzes two sequential reactions in arginine catabolism, in converting arginine to ornithine, and then ornithine into proline [97,98]. The ApalaDH enzyme transforms pyruvate to L-alanine, L-alanine back to pyruvate, and glyoxylate to glycine [99]. The sucrose-synthesis SPS enzyme has sucrose-phosphate synthase and sucrose-phosphate phosphatase activities [100]. The KatG enzymes have both catalase and peroxidase activities, which are also capable of oxidizing chloride, bromide, and iodide compounds [101,102].

Identification of Essential Genes at the Level of a Whole Genome for a Better Understanding of the Genotype-Phenotype Relationships
To use genomic data for reconstruction of an organism's metabolism it is also important to identify the comprehensive set of genes that are essential to cell growth in well-defined conditions. This has been addressed using transposon mutagenesis in Synechocystis PCC 6803 [103,104] and Synechococcus PCC 7942 [105-108]. For the same purpose, the construction of a whole genome library of gene insertion plasmids has been undertaken in Synechocystis PCC 6803 [7]. These works will certainly contribute to decrease the high number of proteins with still unknown function. In Synechocystis PCC 6803, transposon mutagenesis has been useful to study (i) photosynthesis [109-111], (ii) the transport of CO 2 or bicarbonate [112,113], (iii) the production of the poly-3-hydroxybutyrate (PHB) biodegradable bioplastic [114], and (iv) the NADPH:plastoquinone oxidoreductase complex operating in plastoquinone reduction and cyclic electron transfer (CET) around photosystem I [115][116][117][118][119][120][121]. Transposon mutagenesis has been also useful in Synechococcus PCC 7942 to identify a gene in fatty acid production [122].

Importance of Deciphering the Selectivity/Redundancy of Multiple Gene Families
To generate robust and predictive model of the metabolism of cyanobacteria, it is also important to unravel the redundancy/selectivity of multigene families, such as those that code for the stress-responsive redox proteins ferredoxins, glutaredoxins, and glutathione transferase.
Ferredoxins (Fed), are small (acidic) proteins present in most organisms. They use their iron-sulfur cluster (Fe-S) to distribute electrons to various metabolic pathways involved in nutrients assimilation [123]. The best studied cyanobacterium Synechocystis PCC 6803 possess nine Feds. The Fed1-6 proteins possess a (2Fe-2S) center, while Fed7 harbors a (4Fe-4S) cluster. In contrast both Fed8 and Fed9 have two clusters: (3Fe-4S) (4Fe-4S) and (4Fe-4S) (4Fe-4S), respectively. The highly abundant Fed1 protein, essential to photosynthesis [124,125], is encoded by a light-inducible gene [126]. The low-abundant Fed2-Fed9 proteins are encoded by stress-responsive genes (light, carbon, herbicides, or heavy metals), which have differential importance (crucial/dispensable) for the photoautotrophic growth and/or the resistance to stresses [45,123,125,127]. Fed1, Fed7, and Fed9 participate in a ferredoxin-glutaredoxin-thioredoxin crosstalk pathway that operates in the protection against oxidative and metal stresses [45]. Fed7, but not Fed9, interacts with a DnaJ-like protein, an interaction that has been strengthened in photosynthetic eukaryotes in the form of a Fed7-DnaJ fusion protein [123]. Fed7 also has a regulatory role under photooxidative stress [ Like ferredoxins, the evolutionary-conserved enzymes glutaredoxins (Grxs) are widely distributed in cyanobacteria [25,131]. Grxs use electrons provided by glutathione (GSH), or the thioredoxin reductase enzyme [45,132], to reduce the oxidative-stress-generated disulfides occurring between two cysteinyl-residues of either the same or two proteins, or a protein and a molecule of glutathione, which otherwise affect protein activity [33,39]. The dithiol Grxs, which possess a CXXC redox center (C and X stand for cysteine and any other amino acid, respectively), catalyze the reduction in protein disulfides or glutathioneprotein mixed disulfides (the latter activity is named "deglutathionylation"). The monothiol Grxs, which have a CXXS redox center (S stands for serine), operates in iron sensing and trafficking the biogenesis of iron-sulfur clusters of proteins and deglutathionylation [30]. All cyanobacteria possess a monothiol Grx-encoding gene and a variable number of dithiol Grx genes [25]. Synechocystis PCC 6803 possesses three Grxs, which are all dispensable to the standard photoautotrophic growth [25,133] [26]. This feature, which has been conserved in Grx3 orthologs from cyanobacteria to plants and mammals, likely operates in Fe sensing and distribution of (2Fe-2S) cluster [26,30,136].
Another important multiple genes family encodes the evolutionary-conserved glutathione transferase (GT) enzymes, which can conjugate glutathione (GSH) on diverse toxics (oxidants, chemicals, and heavy metals) thereby generating water-soluble complexes that can then be degraded or excreted out of the cell [33, 137,138]. GST also operate in the glutathionylation/deglutathionylation process [42,43]. Glutathione-S-transferases (GST) are commonly divided in three different families: (i) cytosolic GSTs (the largest family), (ii) mitochondrial GSTs, and (iii) microsomal (membranous) GSTs designated as MAPEGs (membrane-associated protein involved in ecosanoïd and glutathione metabolism) [139]. Little is known about GSTs in cyanobacteria though they are regarded as having evolved the oxygen-generating photosynthesis, as well as GSH and GSH-dependent enzymes to protect themselves against the toxic ROS (reactive oxygen species) massively produced by their active photosynthesis [3]. Based on phylogenetic tree analyses, 12 GST classes were identified in cyanobacteria [140,141]. Cyanobacterial GSTs have been studied mostly in Synechocystis PCC 6803, which possesses six GST, namely, Sll0067, Sll1147, Sll1545, Sll1902, Slr0236, and Slr0605 [142][143][144]. While the role of Sll1902 and Slr0605 is still unknown, Sll1545 and Slr0236 were shown to operate in the protection against photo-oxidative stress triggered by high light or H 2 O 2 [32]. Sll1147 plays a prominent role in tolerance to mem-

Properties of the Intensively Studied Unicellular Cyanobacteria Synechocystis PCC 6803, Synechococcus PCC 7942, and Synechococcus PCC 7002
The three non-nitrogen fixing models Synechocystis PCC 6803, Synechococcus PCC 7942, and Synechococcus PCC 7002 have specific physiological properties that can influence their potential to serve as cell factories for biotechnological projects. Synechocystis PCC 6803 in being one of the few cyanobacteria capable to grow under photoautotrophic, mixotrophic, or heterotrophic conditions [147] has become the best studied cyanobacterium. It serves as model to study the photoautotrophic metabolism [6,7,93], the crucial importance of the carbon/nitrogen metabolic balance [6,93], and the responses to stresses [35]. Such investigations are also carried out, to a lesser extent, with Synechococcus PCC 7002 and Synechococcus PCC 7942. The latter model is well used to investigate cell division [158], the carbon/nitrogen metabolic crosstalk [93], the circadian rhythm [159], and the biogenesis of the CO 2 -concentrating carboxysome organelle [160]. Synechocystis PCC 6803, the best studied model, harbors a 3.57 Mbp circular chromosome [48], which has a copy number of about 10-22 per cell [161,162], and seven plasmids of sizes 2.3 [163,164], 2.4 [165], 5.2 [166], 44, 103, 106, and 120 kbp [167]. The ploidy of the three smaller plasmids 2.3-5.2 kbp was shown to be similar and increased from 3 to 8 copies per chromosome in cells reaching the stationary phase of growth. In contrast, the ploidy of the four larger plasmids, ranging from ∼0.3 to 1.2 per chromosome depending on the studied plasmid, varied little with the growth phase [168]. The 5.2 kbp plasmid appeared to be dispensable to the standard photoautotrophic growth of Synechocystis PCC 6803 [169]. In contrast, large plasmids appeared to be essential (cells could not lose these plasmids), likely because they encode several toxin-antitoxin systems mediating plasmid maintenance [170,171] and they operate in beneficial functions such as potassium transport [172], copper resistance [173], and the metabolic tricarboxylic acid cycle [174].
The other model cyanobacteria Synechococcus PCC 7942 and Synechococcus PCC 7002 possess a circular chromosome of, respectively, 2.69 Mbp and 3.0 Mbp, each occurring at two to five copies per cell [175][176][177]. In addition, Synechococcus PCC 7942 has two plasmids of 7.84 [178] and 46.4 kbp [179], while Synechococcus PCC 7002 contains six plasmids [180] from 4.8 [181] to 186 kbp. The 7.84 kbp plasmid of Synechococcus PCC 7942 is not essential to its photoautotrophic growth [182,183]. In comparison, the well-studied filamentous cyanobacterium Anabaena (Nostoc) PCC 7120 has a 7.  [149,183,186]. However, Synechococcus PCC 7002 requires vitamin B12 (cobalamin) to grow [146], unlike Synechocystis PCC 6803 and Synechococcus PCC 7942 [147]. Thus, the cost of vitamin B12 supplementation should be considered when Synechococcus PCC 7002 is to be used for biotechnological purposes requiring large-scale cultures. The vitamin B12 auxotrophy of Synechococcus PCC 7002 is due to the fact that it uses a cobalamin-dependent methionine synthase (MetH) for the synthesis of methionine, though it cannot synthesize cobalamin de novo. Recently, a cobalamin-independent methionine synthase metE gene from Synechococcus PCC 73109 was expressed in Synechococcus PCC 7002 to relieve its cobalamin auxotrophy [187], but this modified Synechococcus PCC 7002 sub-strain has been little employed yet.
All three model cyanobacteria are growing well on ammonium and nitrate, the usual nitrogen sources for cyanobacteria. In addition, both Synechocystis PCC 6803 and Synechococcus PCC 7002 can grow on urea (a frequent pollutant) as the sole nitrogen source [188][189][190], unlike Synechococcus PCC 7942 [190]. Furthermore, Synechococcus PCC 7002 and Synechocystis PCC 6803 are salt resistant (Synechococcus PCC 7002 is a marine strain), unlike the freshwater strain Synechococcus PCC 7942 [191]. Thus, Synechococcus PCC 7942 is not a suitable cell factory for future projects aiming at the photosynthetic production of chemicals in waters polluted by urea and/or salt to save the costs of potable waters.
As iron is an essential enzyme cofactor for oxygenic photosynthesis, cyanobacteria utilize multiple strategies to maintain iron levels within a desired range. One of them is the synthesis, export, and re-import of ferric ion chelators called siderophores. Both Synechococcus PCC 7942 and Synechococcus PCC 7002 can synthesize siderophore, unlike Synechocystis PCC 6803 that can only import the siderophore produced by other organisms [192]. These three model cyanobacteria have other metabolic differences. The RbcX chaperone operating in assembly of the CO 2 -fixing RubisCO enzyme is essential in Synechococcus PCC 7002 [193], whereas it is dispensable in Synechococcus PCC 7942 [194] and Synechocystis PCC 6803 [195]. In addition, Synechocystis PCC 6803 has four flavodiiron proteins (Flv1−Flv4) [196], which function as heterodimers Flv1/3 and Flv2/4. Flv1/3 catalyzes the NAD(P)H-driven reduction in oxygen to water on the acceptor side of PSI [196][197][198], while Flv2/4 operates both in the photoprotection of the photosystem II [199] and in an oxygen-dependent alternative electron flow [200]. Unlike Synechocystis PCC 6803, both Synechococcus PCC 7942 and Synechococcus PCC 7002 possess only Flv1/3, not Flv2/4 [201].
Finally, both Synechocystis PCC 6803 and Synechococcus PCC 7002 secrete extracellular polymeric substances mainly composed of exopolysaccharides, which act in the formation of biofilm and the protection against salt and metals stresses [202][203][204][205], unlike Synechococcus PCC 7942 that do not normally form biofilms [206]. Attesting the importance of the glutathione system for the tolerance to photo-oxidative stress and cell detoxication, most cyanobacteria possess the glutathione system. Its composition varies depending on the species and the environmental challenges they face. Most cyanobacteria have a glutathione reductase (GR) enzyme as Synechococcus PCC 7942, unlike both Synechocystis PCC 6803 and Synechococcus PCC 7002 [25,45]. Furthermore, Synechocystis PCC 6803 and Synechococcus PCC 7002 possess three glutaredoxins (Grxs) operating in tolerance to oxidative and metal stresses [34,45,133-135], whereas Synechococcus PCC 7942 has only two Grxs [25,131]. Another common feature shared by Synechocystis PCC 6803 and Synechococcus PCC 7002 is the fact that they both have an orange carotenoid protein (OCP) operating in photoprotection, whereas Synechococcus PCC 7942 has no OCP [207]. Synechocystis PCC 6803 and Synechococcus PCC 7942 have a MAPEG-type glutathione transferase enzyme operating in the tolerance to temperature stresses [29], which could occur during culture in open ponds, unlike Synechococcus PCC 7002.
Concerning the also important DNA repair system, which has been poorly studied in cyanobacteria, it is worth noting that the three model species have the following common and specific features. Synechocystis PCC 6803 possesses the recD gene, which is absent in Synechococcus PCC 7002 and Synechococcus PCC 7942 [58]. Furthermore, Synechocystis PCC 6803 has recB and recJ, which are duplicated in Synechococcus PCC 7002 and Synechococcus PCC 7942.
Synechocystis PCC 6803 and Synechococcus PCC 7002, which can both use exogenous carbohydrates to accelerate their growth [6], possess lexA, which encodes a transcription regulator involved in carbon assimilation, not DNA repair [90]. In contrast, the obligate photoautotroph Synechococcus PCC 7942 has no lexA [58].
Synechocystis PCC 6803, Synechococcus PCC 7002, and Synechococcus PCC 7942 have also different features in the simplest DNA repair system, which removes only the basemodifying agent in one single step catalyzed by the AlkB demethylase, the Ogt alkyltranferase, and the Phr photolyase. All three model cyanobacteria possess phr, whereas Synechocystis PCC6803 has alkB but not ogt, Synechococcus PCC7942 has ogt (duplicated) but not alkB, and Synechococcus PCC7002 has neither alkB nor ogt [58]. To better characterize the common and different physiological features exhibited by these three model cyanobacteria, it would be very interesting to perform simultaneous analyses and comparisons of their growth and tolerance to stresses in the same laboratories. This task is of high interest for not only basic science but also for biotechnological projects, which use of robust cell factories capable to withstand the environmental challenges imposed by the industrial process and the possible toxicity of the intended products. So far, a very limited number of comparative studies have been carried out. Synechocystis PCC 6803 was shown to be more tolerant than Synechococcus PCC 7942 to UV (and gamma radiations) [90] and to the undecane hydrocarbon [208]. Reciprocally, Synechococcus PCC 7942 appeared to be more tolerant to chromate than Synechocystis PCC 6803, possibly because the sulfate transporters of Synechococcus PCC 7942 have lower affinity to chromate than those of Synechocystis PCC 6803 [209]. Furthermore, chromate generated more ROS (reactive oxygen species) in Synechocystis PCC 6803, as compared to Synechococcus PCC 7942, likely because Synechocystis PCC 6803 has intrinsic levels of superoxide dismutase, catalase and 2-Cys-peroxiredoxin than Synechococcus PCC 7942 [210]. Additionally, interestingly, Synechocystis PCC 6803 and Synechococcus PCC 7002 were reported to be more tolerant to ethanol than Synechococcus PCC 7942 [208,211].
Another important aspect to study is the genetic stability of the recombinant cyanobacteria generated for applied research. It has been reported that many engineered cyanobacteria appeared to be genetically unstable [58]. In one Synechocystis PCC 6803 recombinant strain, the instability was caused by the transposition of an IS5 insertion sequence leading to the inactivation of a regulatory gene [56]. Thus, it is of interest to note that the genomes of Synechococcus PCC 7942 and Synechococcus PCC 7002 are predicted to contain one and approximatively 10 transposase encoding genes, whereas Synechocystis PCC 6803 possesses a single transposase gene.

As Observed in Synechocystis PCC 6803, the Sub-Strains of a Single Cyanobacterium Cultivated in Various Laboratories Can Have Different Behaviors
Another level of complexity is the un-surprising finding that a single cyanobacterium cultivated for some times in different laboratories tend to develop genetic and physiological differences from one laboratory to another. In the case of Synechocystis PCC 6803, originally isolated from a fresh water lake in California [146] and deposited in both the American Type Culture Collection (strain number ATCC 27184) and the Pasteur Culture Collection (strain number PCC 6803), it was reported that several sub-strains cultivated in different laboratories varied in genotypes and/or phenotypes. As compared to the original genome sequence [48], several specific mutations were identified in genes related to photosynthesis, transport, or motility [212][213][214][215][216][217][218]. Furthermore, variability in genome copy number was also observed [162]. Moreover, phenotypic variations were also reported including photosynthesis [217], cell size [219], motility [212,218], capability to grow on glucose as the carbon source [212,220], or resistance to temperature and salt (NaCl) stresses [221].

Genetic Characteristics of the Model Cyanobacteria Synechocystis PCC 6803, Synechococcus PCC 7942, and Synechococcus PCC 7002
Shortly after their selection based on their robustness and simple (unicellular) morphologies [146,147] Synechocystis PCC 6803, Synechococcus PCC 7942, and Synechococcus PCC 7002 were found to be naturally competent for genetic transformation (see below).

Synechocystis PCC 6803, Synechococcus PCC 7942, and Synechococcus PCC 7002 Are Naturally Competent for Genetic Transformation
Natural competence for transformation of prokaryotes refers to their capability to take up DNA from their environment and incorporate it into their own genome. Originally, antibiotic resistant mutants, generated after UV or chemical mutagenesis, were selected, and their genomic DNA was isolated and used to transform wild type cells selecting for antibiotic resistance. Three naturally transformable species have emerged from such studies, namely, Synechococcus PCC 7942 [222], Synechococcus PCC 7002 (Agmenellum quadruplicatum PR6) [223,224], and Synechocystis sp. PCC 6803 [225]. In Synechococcus PCC 7942, the efficiency of transformation was found to decrease at pH lower than 7.0 or temperature ≥40 • C, two conditions being also unfavorable for growth. Furthermore, transformation was best effective with cells reaching the transition from the first to the second exponential phases of growth, supporting the notion that transformation depends upon the physio-logical stage of the culture [183,226]. In addition, transformation appeared to be more efficient when expression of the transferred genes was allowed for 24 h on non-selective solid medium, prior to introducing the selective antibiotic underneath the agar of the plate for its slow diffusion toward the cells for gentle selection of the transformants [183]. Additionally, interestingly, the transformation to Synechocystis PCC6803 was reported to be strongly stimulated after the deletion of the exonuclease recJ gene [227]. Recently, the natural competence for transformation was shown to involve pilus appendages [228] and to be regulated by the circadian clock [229].
In Synechococcus PCC 7002, it was shown that single-stranded DNA cannot transform competent cells. Furthermore, cells in the stationary phase of growth or deprived of nitrogen or light before exposure to donor DNA tend to lose their competence for transformation. In contrast, significant improvement in transformation frequency were achieved by increasing the nitrate content of the culture medium or lowering the temperature from 39 • C (the optimal temperature for growth) to 30 • C before exposure to donor DNA [230].
Later, it was established in Synechococcus PCC 7942 [231], Synechococcus PCC 7002 [232], and Synechocystis PCC 6803 [161,163,233] that the donor DNA is integrated in the chromosome, or an endogenous plasmid (see below) of the recipient cells, through homologous recombination (double crossing-over or gene conversion) occurring on each side of the transferred DNA sequence in the region of homology between both the donor DNA and the recipient DNA. This process, similar to what had been described earlier in Bacillus subtilis [234], was extensively used for deleting cyanobacterial genes (or a part of them) through their targeted replacement by an easy selectable antibiotic resistance marker. Soon after its discovery, natural transformation of cyanobacteria was used in pioneering in vivo analyses of proteins involved in photosynthesis and/or stress resistance, which were performed with Synechocystis PCC 6803 [235][236][237][238][239][240], Synechococcus PCC 7942 [241][242][243][244][245], and Synechococcus PCC 7002 [246,247].
The molecular mechanism of natural transformation has been recently described in Synechocystis PCC 6803 [248]. First, DNA uptake from the environment is mediated by binding to the (type IV) pili appendages, which are also involved in cell adhesion and biofilm formation as well as twitching motility. During pilus retraction, the DNA is pulled into the periplasmic space, where one DNA strand is degraded while the other is translocated further across the cytoplasmic membrane by the Com (competence) proteins. The singlestranded DNA arrived in the cytoplasm is protected from nucleases and incorporated into the genome of recipient cells via homologous recombination.

Interest and Limitation of the Polyploidy of Synechocystis PCC 6803, Synechococcus PCC 7942, and Synechococcus PCC 7002
In cyanobacteria, genomic modification is a time-consuming process because these organisms are polyploid. For example, Synechococcus PCC 7942, Synechococcus PCC 7002 and Synechocystis PCC 6803 harbor about 2-5 and 10-12 chromosome copies per cell [161,[175][176][177]. Thus, to create a homoploid mutant, a segregation procedure must be applied to ensure that all chromosome copies in the transformants carry only the modified DNA. This requires multiple rounds of culture streaking in the presence of the selective antibiotic, which can last several weeks. However, the polyploidy of cyanobacteria is not a merely negative trait. It allows to study genes that are essential to cell life. In such cases, we obtain the corresponding heteroploid mutants, which possess both mutant and WT chromosomes copies (with and without the studied vital gene). Such mutants survive because they retain a limited but sufficient amount of the studied essential protein, and they usually have a phenotype different from the wild-type strain. This difference often serves as a guide to infer a role of the studied crucial protein. Transformation has been used extensively for the introduction of endogenous or heterologous genes into neutral chromosomal or plasmid sites, i.e., loci that can be disrupted with no negative effect on cellular viability. Many neutral sites localized inside a dispensable gene or intergenic region have been identified in the chromosome or in an endogenous plasmid of Synechocystis PCC 6803 [163,[249][250][251][252], Synechococcus PCC 7942 [253][254][255] and Synechococcus PCC 7002 [256][257][258]. These neutral cloning sites were frequently used for cloning and expression of endogenous, or heterologous, genes involved in cell metabolism, stress responses, or the engineering of recombinant strains for the photosynthetic production of chemicals (see Tables 1-3).
In Synechocystis PCC 6803, the frequently used neutral loci are (i) the gene slr0168 or the intergenic region between slr2030 and slr2031, which have no known function; (ii) the cpcB gene involved in the synthesis of the photosynthetic pigment phycocyanin; and (iii) the three psbA genes (Table 1). Although psbA1 is not expressed [235,259], psbA2 and psbA3, encoding the D1 protein subunit of the photosystem II, are expressed and dispensable, but they cannot be inactivated simultaneously [259]. The deletion of psbA2 gene is compensated by an up-regulation of psbA3 [260]. Neutral sites on the endogenous plasmids pCC5.2 and pCA2.4 have also been identified and evaluated for genetic integration and expression [251,261]. One neutral site on pCC5.2 was used for cloning the limonene synthase genes from Mentha spicata or Citrus limon, which directed the production of limonene [262]. Interestingly, the production of fluorescent proteins directed from neutral sites in pCC5.2 or pCA2.4 were, respectively, 14-or 100-fold higher than those observed after chromosomal integration [249,261,263,264]. These data are consistent with the finding that the endogenous small plasmids have a higher copy number than the chromosome [168]. These findings suggest that to be well-expressed genes should be cloned preferentially in a small endogenous plasmid than in the chromosome, but this assumption remain to be verified with other genes. Table 1. Literature on the utilization of neutral chromosomal cloning sites in Synechocystis PCC 6803.

Intergenic region Synpcc7942_0893 and Synpcc7942_0894
Photoproduction of 2,3-butanediol [384] The chromosomal cloning sites, written in bold cases are highlighted in grey color.
In Synechococcus PCC 7002, the few neutral cloning sites employed (Table 3) are mainly the chromosomal genes glpK, which had been thought for some time to harbor a frameshift mutation preventing the production of a functional glycerol kinase enzyme [385], and acsA, which encodes an acetyl-CoA ligase, the inactivation of which conferred resistance to (3-hydroxy)propionate [386]. Table 3. Literature on the utilization of chromosomal neutral cloning sites in Synechococcus PCC 7002.

Neutral Cloning Sites
Objective of the Gene Manipulation and References glpK (SYNPCC7002_A2842) encoding the glycerol kinase [385,386] Removal of carboxysomes for containment of genetically modified strains [387] acsA gene (SYNPCC7002_A1838) encoding an acetyl-CoA ligase and glpK Analysis of an organic acid-based counter selection system [386] acsA and glpK Analysis of promoters and ribosome binding sites [388] Integration between SYNPCC7002_A0935 and SYNPCC7002_A0936 Photoproduction of bisabolene and limonene [388] glpK and desB (SYNPCC7002_ A0159 encoding a ω3 acyl-lipid desaturase [389] and integration between SYNPCC7002_A0935 and SYNPCC7002_A0936 Development of genetic tools [390] glpK and integration between A0935-A0936 Engineering a strain for melamine degradation [391] Intergenic regions between SYNPCC7002_A0932 and SYNPCC7002_A0933, SYNPCC7002_A1202 and SYNPCC7002_A1203, SYNPCC7002_A1778 and SYNPCC7002_A1779 Development of genetic tools [257] SYNPCC7002_A1838, SYNPCC7002_A2542, and SYNPC-C7002_A2842 Photoproduction of L-lysine [392] The inactivation of the glycerol kinase glpK gene (SYNPCC7002_A2842), which was mistakenly annotated as having a frameshift mutation preventing the production of a functional protein [385], has no influence on the physiology of Synechococcus PCC 7002 [386]. Similarly, the inactivation of desB (SYNPCC7002_ A0159, ω3 acyl-lipid desaturase) has no detrimental influence at temperature above 22 • C [389]. The inactivation of the acsA gene (SYNPCC7002_A1838) encoding an acetyl-CoA ligase confers resistance to 3-hydroxypropionate and propionate [386].

Development of Transformable Shuttle Vectors Based on the Endogenous Plasmids of Synechocystis PCC 6803, Synechococcus PCC 7942, and Synechococcus PCC 7002
Because transformation and autonomously replicating plasmids have played a crucial role for gene manipulation in Escherichia coli, several groups tried to introduce an E. coli plasmid (pBR322 and its pUC derivatives) into cyanobacteria by transformation. All attempts were unsuccessful [163,178,231], in spite of a single report [393] that was never confirmed thereafter. These findings indicated that these E. coli plasmids are not able to replicate in cyanobacteria. Consequently, chimeric plasmids capable to replicate both in E. coli and a transformable cyanobacteria were constructed by cloning a small (cryptic) cyanobacterial plasmid (or a part of it) into an E. coli plasmid [394]. This approach was initiated in Synechococcus PCC 7942 using its smaller 7.84 kbp endogenous plasmid [182,395]. Then, the transformation efficiency was increased by plating the transformation mixture (recipient cells plus transforming DNA) on solid medium and incubating the plate in standard condition for 24 h prior to adding the selective antibiotic underneath the agar of the plate [183]. The transformation was more efficient when the recipient cyanobacterial cells were taken at the transition from the first to the second exponential growth phases [183], similarly to what has been observed for the transformation with linear chromosomal DNA [226]. The influence of the growth phase on transformation, which was not confirmed by other workers [396], was also observed in Synechocystis PCC 6803 [163].
Synechococcus PCC 7942 was found to be transformed more efficiently by chimericplasmid DNA isolated directly from this cyanobacterium rather than from E. coli [182]. This was explained by assuming that Synechococcus PCC 7942 and E. coli DNA are differently modified, for instance, by damor dcm-like DNA methylation systems [182]. Similarly, Synechococcus PCC 7002 was more efficiently transformed by its chimeric plasmid when it had been isolated directly from this host rather than from E. coli [397]. Furthermore, the elimination from this Synechococcus PCC 7002 biphasic plasmid of the AvaI restriction site (cleaved by the AvaI-isoschizomer AquI endonuclease of Synechococcus PCC 7002) strongly increased the efficiency of transformation to Synechococcus PCC 7002 [397]. The importance of DNA modification for genetic transformation and cell fitness of cyanobacteria were firmly established in Synechocystis PCC 6803 [398][399][400][401].
The biphasic plasmids autonomously replicating in E. coli and a specific cyanobacterium were improved by several groups through the addition of restriction sites for facile gene cloning and/or several antibiotic-resistance genes for effective selection in Synechococcus PCC 7942 [402][403][404], Synechocystis PCC 6803 [163], and Synechococcus PCC 7002 [405]. These shuttle vectors were used for complementation analyses selecting for both the antibiotic resistance of the vector and the wild-type phenotype [406]. This approach allowed the identification and analysis of genes encoding the key stress-defense proteins RecA [405] and the Mn SOD [407]. One Synechocystis PCC 6803 shuttle vector served for analyzing the activity of several E. coli promoters, such as the tac promoter which appeared to be as effective in Synechocystis PCC 6803 as in E. coli [408]. This promoter-probe vector was also used to show that the lambda phage Ci 857 gene (encoding a temperature sensitive repressor) and associated P R promoter can be employed for strong and tight temperature-controlled gene expression [408]. Latter these gene expression devices were also shown to work well in other cyanobacteria [409,410].
The Synechococcus PCC 7002 and Synechococcus PCC 7942 shuttle vectors were employed for the production of heterologous larvicidal proteins in cyanobacterial cells, which turned out to be toxic when ingested by mosquito larvae [411,412]. Other studies reported the cloning of heterologous genes in Synechococcus PCC 7942 to increase its resistance to cadmium [413] or salt [414,415]. Similarly, the Zymomonas mobilis genes encoding the pyruvate decarboxylase (pdc) and alcohol dehydrogenase II (adh) were cloned into a Synechococcus PCC 7942 shuttle vector to engineer an ethanol producer [416]. The same strategy of cloning heterologous genes into a shuttle vector was used to generate a Synechococcus PCC 7942 recombinant strain for the production of ethylene [417][418][419]. Chimeric shuttle vectors were also used to overexpress endogenous metabolic genes to improve CO 2 -fixation and/or biomass production in Synechococcus PCC 7942 [420].

The Chimeric Shuttle Vectors Based on an Endogenous Cyanobacterial Plasmid Tend to Have a Narrow-Host-Range of Replication
The abovementioned biphasic shuttle plasmids based on a cyanobacterial replicon appeared to replicate in one or a few genetically manipulable cyanobacteria. For examples, the Synechococcus PCC 7942 chimeric plasmids could not transform Synechocystis PCC 6803 [163], and reciprocally, the Synechocystis PCC 6803 shuttle vectors could not transform Synechococcus PCC 7942 [163]. In contrast, a shuttle vector based on an endogenous plasmid from Synechococcus PCC 7942 was shown to replicate in Anabaena PCC 7120, but not in Synechocystis PCC 6803 [421].
Plasmid vectors that can be transferred to, but cannot replicate in, a host are interesting for transporting DNA that must either transpose (e.g., a transposon, useful for random mutagenesis) or by homologous recombination in order to be stably maintained.

The Development of Autonomously Replicating Vectors Derived from the Broad-Host-Range Conjugative Plasmid RSF1010 Has Boosted the Genetics of Cyanobacteria
The interest of chimeric plasmid vectors that uses two narrow-host-range replicons originating from E. coli and a cyanobacterium is limited by the fact that they can shuttle only from E. coli to mostly the corresponding cyanobacterium. Such narrow-host-range vectors are not suitable for rapid tests and comparisons of a gene function in diverse cyanobacterial hosts to select (and engineer) a cyanobacterial chassis for an effective photosynthetic production of an industrially interesting chemical. Thus, several groups turned their attention to the naturally-occurring RSF1010 plasmid [422], a member of the incQ incompatibility group that replicates in a wide range of Gram-negative bacterial genera (cyanobacteria are also Gram-negative), including Acetobacter, Acinetobacter, Agrobacterium, Alcaligene, Azotobacter, Escherichia, Klebsiella, Methylophilus, Providencia, Pseudomonadales, Rhizobiaceae, Rhodopseudomonas, and Salmonella SP-and Serratia. The value of RSF1010 as a shuttle vector is further enhanced by its ability to be transferred by conjugation from an E. coli strain also carrying a self-transmissible (incP group) plasmid, such as RP4. The RSF1010-derived plasmids can be used to carry novel genetic information to those bacteria that are not capable of transformation [423][424][425].
Shortly after, the Synechocystis PCC6803 pioneering report [426], RSF1010 was used for the development of the first conjugative plasmid vector for promoter analysis [409] and regulated protein production in cyanobacteria [437]. The promoter-probe vector used the promoter-less chloramphenicol-acetyl-transferase (cat) gene as the reporter. When expressed by a studied promoter, cat directs the production of the CAT enzyme, the activity of which can be monitored by a spectrophotometric assay and confers the resistance to chloramphenicol [409]. This vector was used for the analysis of constitutive or regulated promoters [126, 438,439] and references therein. The conditional expression vector [437] harbors the lambda-phage gene cI 857 encoding a temperature-sensitive repressor that tightly controls the activity of the otherwise strong p R promoter. Together these elements allow a tight temperature-controlled expression of the studied genes (no production of the corresponding proteins at temperatures below 30 • C, moderate level at 34-36 • C, and high production at 39-40 • C). This vector and its derivative harboring the gfp gene encoding the green-fluorescent reporter protein have been employed for proteins involved in photosynthesis [125], response to stress [34], cell division [440][441][442], and biogenesis of the carboxysome [410].
Other RSF1010-derived plasmids vectors have been used to analyze ribosome binding sites and transcription terminators [264], light-emitting proteins GFP, YFP (yellowfluorescent protein), and luciferase [255,264,432,434]. RSF1010-derived vectors were also employed to analyze the role of carbon stores glycogen and PHB (polyhydroxybutyrate) in the tolerance to stress [443] and systems for the control of gene expression (in Synechococcus PCC 7942) [372], as well as various proteins such as the Synechocystis PCC 6803 photolyase enzyme PhrA [444] and to improve carbon fixation [445].
Interestingly, in Synechocystis PCC 6803, it has been shown that RSF1010 and the pCC5.2 endogenous plasmid could be used for cloning, respectively, two pentose phosphate pathway native genes and the limonene synthase genes (lims) from either Mentha spicata or Citrus limon, which directed the production of limonene [262].

Interest and Limitation of the CRISPR/Cas Genome Editing Technology
Recently, the CRISPR/Cas system (CRISPR stands for clustered regularly interspaced short palindromic repeats and Cas for CRISPR-associated endonuclease) has facilitated the way genomes are edited in cyanobacteria, such as Synechocystis PCC 6803, Synechococcus PCC 7942, Synechococcus PCC 7002, Synechococcus UTEX 2973, and the filamentous strain Nostoc (Anabaena) PCC 7120 (for reviews see [80,83,84,339,457]. Briefly, CRISPR/Cas genome editing systems exploit the Class II family of Cas endonucleases, which have a site-specific RNA-guided DNA cleavage activity. As compared to the well-established gene deletions techniques based on homologous DNA recombination, the interest of the CRISPR/Cas system are (i) CRISPR/Cas systems can allow the engineering of nontransformable cyanobacteria, providing they can be manipulated by conjugation; (ii) a marker-less mutation is generated at the DNA target site; and (iii) multiple DNA loci can be modified simultaneously, by co-expressing the appropriate guide RNAs and editing templates.
The limitation of the CRISPR/Cas technology are the potential toxicity of the Cas DNase and the time required to eliminate the CRISPR/Cas plasmid vector from the generated mutant. However, this curing step can be accelerated by the presence of a negative selection marker in the CRISPR/Cas vector [80].
A variant of the CRISPR/Cas system, the CRISPRi (CRISPR interference) system that make use of DNase-inactive variants of Cas, is especially relevant to repress (fully or not) the transcription of studied genes, including the essential ones that cannot be deleted. This strategy was used for the targeted repression of vital genes to arrest growth and increase carbon partitioning and biofuel titers in Synechocystis PCC 6803 [345]. The CRISPRi technology was also employed to generate mutants with increased yields of growth and lactate secretion [458].

Responses to Stresses: The Recent Progress in Omics Technics Are Limited by the Large Number of Genes of Still Unknown Function
Because of their oxygenic photosynthesis, which triggered oxygen-promoted changes in metal availability, and the fact that they colonized most aquatic biotopes of our planet, cyanobacteria have always been challenged by changes in light, metals, and nutrients availabilities [28]. The responses to these stresses have been well studied with omics techniques that measure the changes in abundance of transcripts (transcriptomics), proteins (proteomics), or metabolites (metabolomics). The available genome sequences facilitate the use of transcriptomic and proteomic approaches.
Later, several genome-wide Synechocystis PCC 6803 microarrays were also developed, based on long oligonucleotide probes (60-to 70-mer) spotted on a glass support. Such oligonucleotides-based microarrays circumvented the labor-intensive and error-prone steps of probe amplification and purification. They were used to study cell responses to sulfur starvation [476] and the deletion of the AbrB2 transcription regulator [477].
As an alternative to microarray analyses, which are based on hybridization of mRNA to DNA probes, the direct sequencing of RNA (an approach designated as RNA-Seq) was adapted to cyanobacteria [478]. RNA-Seq rapidly became the standard method for cyanobacterial transcriptomics [250,479]. It revealed that the Synechocystis PCC 6803 transcriptome includes more than 4000 transcriptional units, half of which representing small RNAs (sRNAs), which often harbor a small protein-coding sequence of less than 100 amino acid residues, and non-coding RNAs (ncRNAs) [480]. These ncRNAs could not be detected by DNA microarrays that only comprised probes for protein-coding genes. The vast majority of ncRNAs are still uncharacterized, and most of them are antisense transcripts (asRNAs). The phylogenetic conservation of ncRNAs across genomes of relatively distant cyanobacteria and their regulated transcription in response to major stresses, such as, light, iron, carbon, or nitrogen availability, nitrogen starvation [481], and butanol or ethanol stress [482], suggest that many ncRNAs may be involved in regulation [250,479].
Proteomics was also used to study the cyanobacterial responses to stresses, again starting with Synechocystis PCC 6803 [483], the genome of which is predicted to contain 3672 putative open reading frames (ORFs, i.e., protein coding sequences). Of these, 3264 and 408 ORFs are located on the chromosome and the seven endogenous plasmids, respectively [484]. Traditionally, two-dimensional polyacrylamide gels (2D-PAGE) and utilization of different fluorescence dyes (difference gel electrophoresis; 2D-DIGE) were employed to estimate concentrations for each protein between stress versus unstressed conditions. Later strategies took advantage of the sensitivity of liquid chromatography (LC), coupled with tandem mass spectrometry (MS), known as LC-MS/MS, for quantitative proteomic analysis, using different tags such as the isobaric tags for relative and absolute quantitation (iTRAQ)-based quantitative proteomics [485,486]. This quantitative technique became the dominant proteomics method for the identification of differentially expressed proteins of Synechocystis PCC 6803 [485][486][487][488].
However, our current understanding of the transcriptome and proteome responses to various challenges is limited by the fact that a large number of the responsive genes or proteins have still an unknown function.
In addition to transcriptomics and proteomics, metabolomics that focuses on lowmolecular-weight metabolites provides the most straightforward characterization of metabolic responses to environmental changes. Compared to other omics studies, a few metabolomic research studies have been performed in cyanobacteria, and again Synechocystis PCC 6803 has been the most studied model. As many metabolites turn over quickly, fast sampling through fast culture filtration appeared to be very important in metabolomic analyses [519,520]. The combination of gas chromatography or liquid chromatography with mass spectrometry permits quantitative analysis of more than 100 metabolites in cyanobacterial cells. In addition to metabolomics, which seeks comprehensive profiling of predominantly intra-organism compounds, volatilomics assesses those compounds released by an organism: the key components of chemically mediated inter-organismal communication [521,522]. The field of volatilomics grew out as advances in collection methods of volatile organic compounds and gas chromatography coupled with mass spectrometry.
Finally, in several studies, transcriptomics, proteomics, and/or metabolomics were integrated to better analyze the responses of Synechocystis PCC 6803 to environmental conditions [492,497,[523][524][525][526][527][528][529], as well as the production of [530] or the tolerance to chemicals [531]. In some cases, it appeared that omics data at different levels do not necessarily correlate a finding that can be explained by regulations occurring at the levels of gene expression [6,93,479,[532][533][534] and/or enzyme stability and activity [40,483,527].

Conclusions
Cyanobacteria are a widely-diverse photosynthetic prokaryotes of wide interest for basic and applied sciences. So far, cyanobacterial research has focused primarily on a few models, such as the three unicellular non-nitrogen fixing species Synechocystis PCC 6803, Synechococcus PCC 7942, and Synechococcus PCC 7002, which are straightforward to culture under laboratory conditions, easily amenable to genetic modification and can be frozen for long-term storage. Extensive "omics" data sets and many genetic tools and genomescale metabolic models (GSM) have been generated to guide the engineering of these model cyanobacteria for the photosynthetic production of biotechnologically interesting chemicals. Interestingly, it has been put forward that GSM should take into account and describe photon absorption and light-shading thereby addressing the challenge of accurately modeling light as a metabolite [535,536]. In addition, GSM should be validated with experimental data obtained after measurement of metabolic fluxes and metabolic pool sizes [80,86,87,536]. However, omics data interpretation and GSM metabolic designs are based on our currently limited understanding of the genotype-phenotype relationships of cyanobacteria. Thus, to generate robust and predictive GSM models of the cyanobacterial metabolism, it is important to continue the analysis of Synechocystis PCC 6803, Synechococcus PCC 7942 and Synechococcus PCC 7002, and increase the efforts to (i) verify the function of numerous genes that have been annotated merely by sequence analogy with those genes characterized only in intensively studied nonphotosynthetic models (E. coli, yeast, etc.), which may have a different function in cyanobacteria; (ii) and analyze the specificity/redundancy of multiple gene families; (iii) characterize the function of the large number of as yet unknown genes and non-coding RNAs; (iv) identify the comprehensive set of genes that are essential to the growth of cells incubated in well-defined conditions.
Furthermore, most of the attempts to reprogram Synechocystis PCC 6803, Synechococcus PCC 7942, or Synechococcus PCC 7002 for the photoproduction of chemicals have focused on increasing product synthesis by small-scale cultures growing under laboratory conditions because most academic researchers lack access to large-scale production systems that are necessary to evaluate the potential of engineered strains under realistic industrial conditions.
Moreover Synechocystis PCC 6803, Synechococcus PCC 7942, and Synechococcus PCC 7002 represent only a limited part of the wide biodiversity of cyanobacteria. This arguably limits fundamental discovery and applied research towards wider commercialization. Thus, new phylogenetically-distant candidate cyanobacteria should be isolated and developed from diverse environments with a robust growth and high tolerance to local conditions, so as to be used as chassis for the photosynthetic production of high-value chemicals in diverse industrial sites. We think that the genetic modifiability of such candidate strains using the conjugative transfer of RSF1010-derived broad-host-range plasmids will be key for such works.
To summarize, we recommend to strengthen the communication between academic researchers, who know well cyanobacteria and can manipulate them, but have a limited access to large photobioreactors and industrial partners, who attempt to use cyanobacteria to produce interesting chemicals at reasonable costs, but often lack knowledge on cyanobacterial genetics, physiology, and metabolism. Moreover, to minimize operation costs we need to develop robust cyanobacteria capable to grow on industrial waters and fumes, in huge photobioreactors, as well as well as efficient technologies to harvest the end products.