RNA-Seq Analysis Provides Insights for Understanding Photoautotrophic Polyhydroxyalkanoate Production in Recombinant Synechocystis Sp.

The photosynthetic cyanobacterium, Synechocystis sp. strain 6803, is a potential platform for the production of various chemicals and biofuels. In this study, direct photosynthetic production of a biopolymer, polyhydroxyalkanoate (PHA), in genetically engineered Synechocystis sp. achieved as high as 14 wt%. This is the highest production reported in Synechocystis sp. under photoautotrophic cultivation conditions without the addition of a carbon source. The addition of acetate increased PHA accumulation to 41 wt%, and this value is comparable to the highest production obtained with cyanobacteria. Transcriptome analysis by RNA-seq coupled with real-time PCR was performed to understand the global changes in transcript levels of cells subjected to conditions suitable for photoautotrophic PHA biosynthesis. There was lower expression of most PHA synthesis-related genes in recombinant Synechocystis sp. with higher PHA accumulation suggesting that the concentration of these enzymes is not the limiting factor to achieving high PHA accumulation. In order to cope with the higher PHA production, cells may utilize enhanced photosynthesis to drive the product formation. Results from this study suggest that the total flux of carbon is the possible driving force for the biosynthesis of PHA and the polymerizing enzyme, PHA synthase, is not the only critical factor affecting PHA-synthesis. Knowledge of the regulation or control points of the biopolymer production pathways will facilitate the further use of cyanobacteria for biotechnological applications.


Introduction
Cyanobacteria are believed to be one of the oldest groups of photosynthetic organisms on Earth and played a significant role in the development of the oxygenic atmosphere we breathe today [1]. In modern day, cyanobacteria continue to play a pivotal role in global carbon recycling, the nitrogen cycle and most importantly, the maintenance of the composition of the atmosphere [2,3]. Cyanobacteria are considered to be ideal producers of various fine chemicals and biofuels because they fix carbon dioxide into biomass using solar energy. Fluctuations of nutrient concentrations constantly occur in natural environments and microorganisms respond to nutrient starvation by accumulating various carbon and energy storage compounds [4]. The study of these storage polymers, particularly polyhydroxyalkanoate (PHA), has gained considerable interest in recent years in an attempt to address the waste disposal problems caused by petrochemical plastics [5].
At present, the major biological processes utilized for industrial production of PHA are fermentations of heterotrophic bacteria. Nevertheless, the economic viability of PHA as a commodity polymer is limited by high production costs due to costly carbon substrates and requirements during the fermentation processes. Substantial effort has been devoted to investigating PHA production processes that are more cost-effective [6]. An interesting and promising approach is the use of photosynthetic cyanobacteria as the host for PHA production. The cyanobacteria, as 'microbial factories', can fix carbon dioxide from the atmosphere into high molecular weight PHA directly via photosynthesis. Besides being photoautotrophic, cyanobacteria require minimal nutrients for growth, eliminating the cost of carbon sources and complex growth media [7]. Thus, the application of cyanobacteria offers the potential of a cost-competitive and sustainable approach for the production of this environmentally friendly polymer.
The presence of PHA in cyanobacteria was first described by Carr whom analyzed PHA in Chloroglea fritschii based on acid hydrolysis of poly(3-hydroxybutyrate), P(3HB), to crotonic acid followed by UV spectroscopic measurement of the hydrolysis product [8]. Since then, much research has demonstrated the presence of PHA in several other cyanobacteria including Aphanothece sp. [9], Oscillatoria limosa [10], some species of the genus Spirulina [11,12] and the thermophilic strain Synechococcus sp. MA19 [13]. So far, cyanobacteria are characterized by their ability to produce PHA containing only 3-hydroxybutyrate (3HB) and/or 3-hydroxyvalerate (3HV) monomers [9,10,14]. Although there are many reports on the occurrence of PHA in cyanobacteria, most of these studies explored the physiology and fermentation aspects of PHA accumulation in cyanobacteria. The biochemical and molecular basis of PHA synthesis in cyanobacteria are not well understood.
The model cyanobacterium Synechocystis sp. strain PCC 6803 is considered as a promising candidate for various biotechnological productions because of the availability of its genome sequence information [15] and the ease of genetic manipulation of this strain due to its naturally transformable feature [16]. In this study, Synechocystis sp. was metabolically engineered by increasing the flux of intermediates to PHA biosynthesis and introducing a PHA synthase with higher activity. RNA-seq analysis was carried out to examine the differential expression involved in the global biological processes and metabolic pathways during the improved photoautotrophic production of PHA. This information will facilitate the potential use of cyanobacteria for the sustainable production of this 'green' polymer.

Enhanced PHA Production in Recombinant Cyanobacteria
In the well-studied PHA biosynthetic pathway of Cupriavidus necator, P(3HB) synthesis occurs in a three-step reaction and starts with the condensation of acetyl-CoA to acetoacetyl-CoA by bketothiolase [17]. Under photosynthetic conditions, it was hypothesized that the acetyl-CoA pool in cyanobacteria is insufficient to drive the thermodynamically unfavorable condensation reaction forward [18]. Instead of relying solely on the native b-ketothiolase-mediated condensation to form acetoacetyl-CoA, an acetoacetyl-CoA synthase (nphT7 Ss ) from Streptomyces sp. CL190 was incorporated in the P(3HB) pathway design. The nphT7 Ss gene catalyzes the irreversible condensation of acetyl-CoA and malonyl-CoA to give acetoacetyl-CoA, driving the reaction toward the formation of PHA. The evolution of carbon dioxide from the condensation reaction effectively pushes the reaction toward the formation of acetoacetyl-CoA [19]. A highly active PHA synthase from Chromobacterium sp. USM2 (phaC Cs ) [20] was co-expressed with nphT7 Ss to improve the photosynthetic production of P(3HB) in Synechocystis sp. In view of the stimulatory effects of nutrient limitation, carbon supplementation and air-exchange limitation on PHA accumulation [7,21], biosynthesis experiments were designed under these cultivation conditions: N-or P-deficient (nitrogen-or phosphorus-deficient), air-exchange limitation, and/or in the presence of carbon sources (CO 2 , acetate and/or fructose).
For the design of PHA production pathway in Synechocystis sp., the vector plasmid pTKP2031V was used for the insertion of transgenes into the genome via homologous recombination between sites slr2030 and slr2031 [22]. Synechocystis sp. was transformed with a plasmid harboring phaC Cs , nphT7 Ss and C. necator acetoacetyl-CoA reductase (phaB Cn ) genes under the control of the light-inducible psbAII promoter. The successful transformant strain C Cs NphT7B Cn was analyzed for PHA production under a two-stage culture system consisting of sequential cell growth and PHA accumulation phases. The strain C Cs NphT7B Cn achieved an encouraging direct photosynthetic production of PHA from CO 2 , with a maximum of 14 wt% P(3HB) content on day 7 of cultivation ( Fig. 1A). In comparison, strain C Cs A Cn B Cn expressing phaC Cs , C. necator b-ketothiolase (phaA Cn ) and phaB Cn recorded a reduction in P(3HB) content (7 wt%) under the same cultivation conditions. The strain pTKP2031V, with only a kanamycin resistance cassette integrated into the genome, showed the lowest P(3HB) production potential (5 wt%). Prolonged incubation until day 14, however, did not exert any significant impact on the P(3HB) accumulation potential of the cyanobacteria under photoautotrophic conditions. At higher cell densities, P(3HB) accumulation may be limited by competition for carbon dioxide and light.
In order to boost P(3HB) production, an exogenous carbon source [0.4%(w/v) acetate] was provided to the cyanobacterial cultures at the PHA accumulation phase. Strain C Cs NphT7B Cn recorded the highest PHA content of 29 wt% on day 10 of incubation (Fig. 1B). The increase in the P(3HB) pool resulting from the addition of a carbon source affirms earlier findings on the effect of external carbon source supplementation on PHA production [7,21]. In the case of air-exchange limitation effect, a significant increase in P(3HB) was observed (up to 41 wt%) for strain C Cs NphT7B Cn (Fig. 1C). These observations imply that the P(3HB) accumulation potential of Synechocystis sp. is affected by the provision of carbon source and air-exchange. Interestingly, the increase in CO 2 supply (5%) to photoautotrophic cultures of Synechocystis sp. was found to increase the PHA content up to 16 wt% in strain C Cs NphT7B Cn ( Table 1). The simultaneous addition of acetate and fructose to N-or P-deficient cultures of strain pTKP2031V showed a reduction in P(3HB) accumulation compared to photoautotrophic conditions (5% CO 2 ). There were no significant changes in the P(3HB) content of strains C Cs NphT7B Cn and C Cs A Cn B Cn under the same cultivation conditions. The order of PHA-producing potential of the recombinant Synechocystis sp. on day 7 of incubation and under the cultivation conditions tested in this study is C Cs NphT7B Cn . C Cs A Cn B Cn .pTKP2031V.

Expression Levels of PHA Synthesis-related Genes
The expression levels of native PHA biosynthetic genes in Synechocystis sp. consisting of phaC Ss , phaA Ss and phaB Ss were monitored by real-time PCR analysis ( Fig. 2A and 2B). Surprisingly, comparative quantification of phaA Ss and phaB Ss expression levels in Synechocystis sp. strain C Cs NphT7B Cn relative to pTKP2031V show expression that was approximately 2-fold lower. However, there were no significant differences in the expression level of the native phaC Ss gene in the recombinant Synechocystis sp. (pTKP2013V, C Cs A Cn B Cn and C Cs NphT7B Cn ) investigated. The expression levels of phaC Cs and phaB Cn that were introduced into the genome on the same operon showed at least 3fold lower expression in strain C Cs NphT7B Cn compared to C Cs A Cn B Cn . Despite higher levels of PHA accumulation in C Cs NphT7B Cn , the expression levels of most PHA synthesisrelated genes in this strain were relatively lower compared to C Cs A Cn B Cn and pTKP2013V.

Analysis of Synechocystis sp. Transcriptional Response Under Photoautotrophic PHA Accumulation Conditions
To gain insight into PHA accumulation in cyanobacteria, transcriptomes of recombinant Synechocystis sp. with different PHAproducing potential were analyzed. RNA-seq libraries were prepared from cells cultivated for 7 days in N-deficient BG-11 under photoautotrophic conditions. Sequencing was performed using the Illumina platform yielding a total of 93-million reads for 6 samples, with an average of 15.5-million reads per sample. Scatter plots between the two biological replicates for each recombinant Synechocystis sp. sample show correlation coefficients between 0.96-0.98, indicating the reproducibility of the sequencing data (Fig. S1). The expression levels for each gene were quantified as reads per kilobase of exon model per million mapped reads (RPKM).
The RNA-seq data provide detailed information on the genes that are regulated in response to photoautotrophic PHA accumulation conditions in recombinant Synechocystis sp. strains pTKP2013V, C Cs A Cn B Cn and C Cs NphT7B Cn . In general, the highly expressed genes in Synechocystis sp. were mainly involved in photosynthesis, the electron transport chain, protein metabolic processes and nucleic acid metabolism (Table S1). In particular, the transcript levels of genes involved in photosystem I (psaB, psaA, psaF and psaL) and photosystem II (psbA3, psbA2, psbX, psbY, psbU, psbK and psbD2) activities were among the most abundant. A comparison between gene expression in the recombinant strains that were more efficient in PHA production (C Cs A Cn B Cn andC Cs NphT7B Cn ) relative to the reference strain (pTKP2013V) was made ( Table 2). The up-regulated genes in strains C Cs A Cn B Cn and C Cs NphT7B Cn significantly enriched photosynthesis, transport and cell communication. In contrast, the down-regulated genes were found to be involved mostly in the metabolism of cofactors and vitamins, protein metabolic processes and DNAbinding. The photosystem I reaction center subunit XII gene (psaM) that is detected only in cyanobacteria was strongly upregulated in both C Cs A Cn B Cn and C Cs NphT7B Cn . Another photosystem I reaction center subunit gene, psaJ was found to be up-regulated more than 10-fold. Both of these subunits are required to form a functional photosystem I [23]. In addition to genes encoding the photosystem I subunits, photosystem IIassociated genes were among the significantly up-regulated genes. PsbX and PsbK, that have been found essential for the stability of photosystem II [24], were induced more than 5-fold in both C Cs A Cn B Cn andC Cs NphT7B Cn . Up-regulation of cytochrome B6-f complex subunits, PetG and PetL that are important for either stability or assembly of the complex, was also observed [25]. Two genes involved in porphyrin and chlorophyll metabolism, magnesium-protoporphyrin IX monomethyl ester cyclase (sll1874) and protoheme IX farnesyltransferase (sll1899) were up-regulated in both strains. Collectively, genes encoding proteins involved in several aspects of photosynthetic activity, e.g. photosystem I and II, cytochrome and chlorophyll metabolism were up-regulated in recombinant Synechocystis sp. that were actively synthesizing PHA.
A comparison between gene expression in the recombinant Synechocystis sp. strains C Cs NphT7B Cn and C Cs A Cn B Cn was made to gain substantial insights into the global responses of cyanobacteria to accommodate the extensive accumulation of PHA (Table 3). The analysis showed that strain C Cs NphT7B Cn employed a combination of induced stress response, photosynthesis, energy metabolism and transport during the PHA accumulation phase. Notably, genes encoding proteins involved in several aspects of photosynthetic activity e.g. uroporphyrinogen decarboxylase (HemE), ferredoxin component (slr1205), protochlorophilide reductase subunit (BchB), protohome IX farnesyltransferase (CtaB), photosystem II reaction center protein N (PsbN) and iron-stress chlorophyll-binding protein (IsiA) were up-regulated in strain C Cs NphT7B Cn compared to C Cs A Cn B Cn. The increased photosynthetic activity suggests that the carbon fixing capacity was enhanced to accommodate the increased diversion of carbon to polymer formation. Polyhydroxyalkanoate are bacterial storage compounds synthesized in response to conditions of physiological stress [26]. In the current study, stressrelated genes in cyanobacteria include co-chaperonin (groES), Holliday junction resolvase (ruvC), molecular chaperon (groEL), superoxide dismutase (sodB) and heat shock protein 90 (htpG) were modestly up-regulated. As it was proposed that PHA accumulation confers survival and stress tolerance in a changing environment [27], stress conditions may trigger responses that favor PHA production. In addition, the transcript level for the global nitrogen regulator, NtcA was detected at an increased level. NtcA is known to regulate the expression of a large number of genes involved in nitrogen metabolism [28] and induction of the gene encoding this protein can be related to the N-deficient cultivation conditions that were applied to increase PHA biosynthesis in Synechocystis sp. Conversely, down-regulation of the genes involved in DNAbinding, transport, translation and DNA repair were observed in strain C Cs NphT7B Cn (Table S3).

Discussion
Current limitation of direct photosynthetic production using cyanobacteria is the relatively low PHA content obtained. In this study, it was encouraging to obtain 14 wt% of P(3HB) from direct photosynthetic fixing of carbon dioxide without the addition of an external carbon source. Although cyanobacteria have simple nutrient requirements, the addition of 0.4%(w/v) acetate was found to increase P(3HB) content up to 41 wt% under airexchange limiting conditions. Previous studies suggested that enhanced P(3HB) accumulation was the result of direct metabolism of acetate for PHA synthesis by employing an existing pathway operating in cyanobacteria [7,21]. The provision of exogenous carbon was found to have a positive impact on PHA accumulation albeit at concentrations that were 10-to 20-fold lower than those required by heterotrophic bacteria. Recently, the development of new photobioreactors for mass cultivation of cyanobacteria is in progress and these findings will greatly aid the use of cyanobacteria for potential industrial applications [29,30].
Early studies indicate that the PHA biosynthetic genes of Synechocystis sp. 6803 do not co-localise together to form an operon   [31,32]. Instead, the PHA synthase of Synechocystis sp. consisting of phaC and phaE subunits are linked in the genome and coexpressed. On the other hand, the b-ketothiolase and acetoacetyl-CoA reductase of Synechocystis sp. do not map close to the PHA synthase locus but are probably clustered together and constitute an operon in a different section of the genome. The expression levels of these two genes were surprisingly lower in the recombinant Synechocystis sp. strains C Cs A Cn B Cn and C Cs NphT7B Cn that had higher PHA production potential compared to strain pTKP2013V that accumulated a lower content of PHA. These results suggest that the endogenous PHA biosynthetic pathway operating in Synechocystis sp. did not have a significant impact on the PHA-synthesizing abilities of strains C Cs A Cn B Cn and C Cs NphT7B Cn . The Chromobacterium sp. PHA synthase and C. necator acetoacetyl-CoA reductase that were introduced into the genome as an operon showed similar lower expression in strain C Cs NphT7B Cn . The observation that the expression levels of most of the PHA biosynthetic genes were lower in strain C Cs NphT7B Cn suggests that the concentration of these enzymes is not the limiting factor in achieving higher PHA accumulation. Based on the results presented here, the transcription of genes encoding enzymes involved in PHA biosynthesis is highly regulated and may be affected by the PHA content in the cells (Fig. 3). When the PHA accumulated by the cells has exceeded a certain threshold level, adequate levels of the enzymes may already be present to meet the biosynthetic demand. Thus, the PHA granule itself or some other sensing factors may exert negative feedback on the expression of these enzymes. However, the expression levels of the enzyme catalyzing the last step of PHA biosynthesis, Synechocystis sp. PHA synthase, remained grossly constant in all recombinant Synechocystis sp. because negative feedback regulations are likely exerted in the upper part of the pathway.
Previous genetic studies have focused on the engineering of various bacteria or plant hosts for PHA production, but less is known about the global transcriptional changes of the recombinant host under a PHA-synthesizing environment. A comprehensive view of the cyanobacterial transcriptome during cultivation under conditions favorable for PHA synthesis was generated using RNA-seq analysis. One particularly interesting observation is the up-regulation of photosynthetic activity in recombinants Synechocystis sp. with higher PHA-synthesizing potential (Fig. 4 A and B). In recent years, there has been tremendous interest in strategies to improve photosynthetic activity in crops [33,34]. It has been suggested that an increase in photosynthetic activity will improve the yield of crops and provide a potential solution to future food shortages [35]. In this context, the increase of photosynthetic activity in cyanobacteria may explain the higher PHA accumulation observed in recombinant Synechocystis sp. strains C Cs NphT7B Cn and C Cs A Cn B Cn .
The gene encoding one of the most important enzymes in carbon fixation, the ribulose-1,5 biphosphate carboxylase/oxygenase (RuBisCo) large subunit (rbcL), was up-regulated in both C Cs NphT7B Cn and C Cs A Cn B Cn . RuBisCo is a biologically important enzyme that catalyzes the first step of the reaction that converts atmospheric carbon dioxide into organic carbon [36]. Besides RuBisCo, genes encoding proteins involved in different  aspects of photosynthesis and electron transport chain were significantly induced in both C Cs NphT7B Cn and C Cs A Cn B Cn. In particular, the induction of photosynthesis and electron transport chain-related genes was most prominent in strain C Cs NphT7B Cn with the highest PHA accumulation, suggesting the possible correlation of photosynthetic activity with PHA content. The Synechocystis sp. cells may utilize enhanced photosynthesis, carbon fixation and electron transport chain activities as a means to provide precursors that are necessary to drive the production of PHA. The increased photosynthetic production of PHA reveals that similar metabolic engineering approaches can be applied to the production of biofuels or chemicals using this versatile organism. As cyanobacteria and plants share similar photosynthetic machinery, it is likely that the strategy can be extended in future efforts to improve PHA production in higher plants.
In living cells, catabolic reactions that produce energy and anabolic biosynthetic reactions are regulated to maintain a balance of supply and demand. To cope with the higher PHA production demand, carbon dioxide fixing was enhanced to replenish the pool of carbon that was lost to PHA formation. Concomitant with the increase in photosynthetic activity, the flow of newly fixed carbon dioxide into biosynthetic reactions other than PHA was reduced. Genes encoding metabolism of cofactors and vitamins as well as protein metabolic process were found to be down-regulated in strains C Cs NphT7B Cn and C Cs A Cn B Cn. The reduced growth of recombinant Synechocystis sp. under nutrient-deficient cultivation conditions may account for the depression of these metabolic processes. These cellular anabolic reactions were regulated to maintain the balance of resources in cells. The expression levels of genes involved in the tricarboxylic acid cycle (TCA) were shown to be down-regulated in strains C Cs NphT7B Cn and C Cs A Cn B Cn . These observations agree well with previous finding that reported on the repressed of the TCA cycle genes in C. necator H16 during PHA production [37].
RNA-seq transcriptome analysis reveals that the heterologous expression of PHA synthesis-related genes in Synechocystis sp. affect not only the regulation of PHA biosynthesis but also the preceding pathways that are involved in the provision of precursors for this biosynthesis. The direct photosynthetic production of 14 wt% of P(3HB) from strain C Cs NphT7B Cn is the highest value achieved for Synechocystis sp. 6803 so far. This work suggests the use of carbon flux as a possible driving force for the biosynthesis of intracellular inclusions e.g. PHA. Future work can be done to confirm this finding by enhancing carbon fixation in cyanobacteria through engineering or overexpressing the enzymes involved in the process.

Chemicals and Reagents
All chemicals were purchased from Nacalai Tesque (Tokyo, Japan) or Wako Pure Chemical (Tokyo, Japan) unless otherwise specified. KOD Plus high-fidelity DNA polymerase was purchased from Toyobo (Tokyo, Japan). Restriction enzymes and the DNA ligation kits used were from Takara (Shiga, Japan).

Organism and Culture Conditions
All Synechocystis sp. PCC6803 strains (Table S4)  Escherichia coli DH5a used for plasmid cloning was grown with shaking (180 r.p.m) at 37uC in Lysogeny broth. For the selection and maintenance of plasmids, kanamycin (50 mg/mL) or ampicillin (100 mg/mL) were added. To promote PHA biosynthesis in cyanobacteria, a two-stage cultivation was performed. The cultures were first grown in BG-11 medium until the late exponential phase and then harvested, washed and transferred to BG-11 medium devoid of sodium nitrate. P-deficiency was achieved by cultivating cells in BG-11 medium without K 2 HPO 4 . Different carbon sources [0.2% (w/v) and 0.4% (w/v) of fructose and/or acetate] were added to study the effects of carbon supplementation on PHA accumulation. Air-exchange limiting conditions on cultures were imposed by sealing the mouth of culture vessels with cotton plugs and covering with aluminium foil [7]. The cyanobacterial cultures were cultivated in the above  culture conditions for 7, 10 or 14 days, harvested by centrifugation (8000 g, 10 min) and then lyophilized.

Plasmid Construction and Transformation of Synechocystis sp
The constructs used for transformation of Synechocystis sp. were derived from pTKP2031V (Table S4). pTKP2031V was designed for insertion into the genome via homologous recombination between sites slr2030 and slr2031 together with a kanamycin resistance cassette [22]. The expression of all cyanobacterial constructs was under the psbAII promoter. The gene cluster containing b-ketothiolase (phaA Cn ) and acetoacetyl-CoA reductase (phaB Cn ) were amplified from chromosomal DNA of C. necator H16 using primers phaAB Cn (F; NdeI) and phaAB Cn (R; HpaI) ( Table S5). The resulting PCR product was digested with NdeI and HpaI and inserted into NdeIand HpaI-digested pTKP2031V to obtain pTKP2031V-phaAB Cn . The PHA synthase (phaC Cs ) was prepared from chromosomal DNA of Chromobacterium sp. USM2 using primers phaC Cs (F; SfuI) and phaC Cs (R; AatI). This PCR fragment was digested with SfuI and AatI and subcloned into the appropriate restriction sites of pTKP2031V-phaAB Cn by ligation to yield pTKP2031V-phaC Cs A Cn B Cn . pTKP2031V-phaC Cs nphT7phaB Cn was constructed by replacing phaA Cn in pTKP2031V-phaC Cs A Cn B Cn with nphT7. The nphT7 gene from Streptomyces sp. CL190 was amplified from a previously prepared template, pHis_nphT7 [19] using primers nphT7 (F; SfuI) and nphT7 (R; AatI). Transformation of Synechocystis sp. was performed as described previously [39]. Briefly, 100-200 mL of an exponentially growing culture were mixed with a plasmid solution to a final concentration of 1-2.5 mg/ mL. The mixture was then spread onto a nitrocellulose membrane filter placed on a BG-11 plate and incubated overnight (,12 h) at 30uC under continuous white light (75-100 mmol photons/m 2 s). The membrane filter was transferred onto a new BG-11 plate containing 50 mg/mL kanamycin. Kanamycin-resistant colonies were isolated and replated three times. The presence and complete segregation of the transgene in the cyanobacterial genome were confirmed by PCR analysis and sequencing.

Quantitative Analysis of PHA
Approximately 25-30 mg of lyophilized cyanobacterial cells were washed with methanol and dried at 65uC overnight. The dry cells were subsequently extracted with chloroform at 65uC for 48 h. The chloroform extract was subjected to methanolysis with a solution consisting of 85%(v/v) methanol and 15%(v/v) concentrated sulphuric acid at 100uC for 140 min [40]. The organic phase comprising the hydroxyacyl methyl esters was analyzed by gas chromatography-mass spectrometry (GC-MS) using the Agilent 7890A GC/5975 MSD system equipped with a HP25 column (Agilent, USA).

RNA Preparation
Total RNA was extracted from cells using Trizol reagent (Invitrogen, USA) in combination with the PureLink RNA Mini Kit (Invitrogen, USA) according to manufacturer's protocol. Any traces of DNA remaining in the RNA samples were removed by digestion with DNase I (Takara, Japan). The quality and quantity of the RNA samples were analyzed using a Bioanalyzer 2100 (Agilent, USA).

Real-time PCR Analysis
cDNA synthesis was performed with 250 ng of RNA using the QuantiTect Reverse Transcription Kit (Qiagen, USA). Real-time PCR quantification was performed using Thunderbird SYBR qPCR Mix (Toyobo, Japan) and gene-specific primers with the Mx3000P QPCR system (Agilent, USA). The cycling conditions were as follows: 95uC for 10 min, 40 cycles: 95uC for 15 s and 60uC for 1 min. A melting curve analysis (60uC-95uC) was performed after each amplification to ensure specificity of the reaction. Transcript levels were quantified based on determination of the quantification cycle (Ct). The transcript levels of genes of interest were normalized to the level of the housekeeping gene (16S rRNA) used in this study. Comparative quantification was used to compare the expression levels of genes of interest in Synechocystis sp. PCC 6803 strains C Cs NphT7B Cn and C Cs A Cn B Cn (target) relative to pTKP2031V (calibrator) and C Cs NphT7B Cn (target) relative to C Cs A Cn B Cn (calibrator).

RNA-seq Library Preparation, Illumina Sequencing and Data Analysis
For each sample, 2 mg total RNA was subjected to ribosomal RNA depletion using the Ribo-Zero rRNA removal kit (Epicentre, USA). The cDNA libraries for RNA-seq were constructed from total RNA depleted of rRNA using the Illumina TruSeq Stranded mRNA Sample Preparation Kit (Illumina, USA) according to the manufacturer's specifications. In brief, preparation of the cDNA libraries included the following steps: RNA fragmentation, cDNA synthesis, 39 ends adenylation, adapter ligation and cDNA template enrichment. Quantification of the libraries was carried out using a Bioanalyzer 2100 (Agilent, USA) and 4 to 6 pM of the template was used for cluster generation. Libraries were sequenced on a Miseq (Illumina, USA) instrument using the 26250 paired end protocol. The sequence data has been submitted to the NCBI Gene Expression Omnibus (GEO) under accession number GSE50688. The RNA-seq data analysis was performed using CLC Genomics Workbench 6 software (CLC bio, Denmark). Sequence reads were pre-processed to trim low-quality reads and filter reads shorter than 20 bp. The qualified sequence reads were mapped to the Synechocystis sp. PCC 6803 genome (NC_000911), allowing a maximum of two mismatches. The reference genome sequences and annotations were downloaded from NCBI (downloaded on May 23, 2013). Sequence reads that mapped to noncoding RNA and reads that did not map to unique positions were excluded from further analysis. The transcript levels were expressed as reads per kilobase of exon model per million mapped reads in which the read count for a gene was normalized by the length of the gene and the total number of reads mapped in the sample [41]. Statistical analysis was performed and genes with a False Discovery Rate (FDR) p-value correction ,0.05 were determined as differentially regulated genes [42]. . The scheme shows the cellular changes in recombinant Synechocystis sp. strains (a) C Cs A Cn B Cn and C Cs NphT7B Cn (compared with pTKP2031V) (b) C Cs NphT7B Cn (compared with C Cs A Cn B Cn ) under photoautotrophic PHA biosynthesis conditions. Only a selection of cellular changes is shown. The genes or pathways that are up-regulated are in red; the downregulated ones are in green. Black dashed lines indicate the engineered route. AP, allophycocyanin; PC/PEC, phycocyanin/phycoerythrocyanin; Cytb 6 /f, cytochrome b6/f complex; PQ, plastoquinone; FNR, ferredoxin-NADP(+) reductase; Pc, plastocyanin; PSI, photosystem I; PSII, photosystem II; Ndh, NADH dehydrogenase; Glc-6-P, glucose-6-phosphate; Fru-6-P, fructose-6-phosphate; Fru-1,6-bp; fructose-1,6-biphosphate; Glycerate-1,3-P 2 , 1,3-biphosphoglycerate; 3-P-Glycerate, 3phosphoglycerate; Ru-1,5-bisP, ribulose-1,5-biphosphate; PEP, phosphoenolpyruvate. doi:10.1371/journal.pone.0086368.g004 Figure S1 Correlation of RNA-Seq data between biological replicates. Normalized expression values from each sample were used. Correlation coefficients are indicated inside the plots. (TIF)