Increased Biomass and Polyhydroxybutyrate Production by Synechocystis sp. PCC 6803 Overexpressing RuBisCO Genes

The overexpression of the RuBisCO (rbc) gene has recently become an achievable strategy for increasing cyanobacterial biomass and overcoming the biocompound production restriction. We successfully constructed two rbc-overexpressing Synechocystis sp. PCC 6803 strains (OX), including a strain overexpressing a large subunit of RuBisCO (OXrbcL) and another strain overexpressing all large, chaperone, and small subunits of RuBisCO (OXrbcLXS), resulting in higher and faster growth than wild type under sodium bicarbonate supplementation. This increased biomass of OX strains significantly contributed to the higher polyhydroxybutyrate (PHB) production induced by nutrient-deprived conditions, in particular nitrogen (N) and phosphorus (P). As a result of higher PHB contents in OX strains occurring at days 7 and 9 of nutrient deprivation, this enhancement was apparently made possible by cells preferentially maintaining their internal lipids while accumulating less glycogen. The OXrbcLXS strain, with the highest level of PHB at about 39 %w/dry cell weight (DCW) during 7 days of BG11-NP treatment, contained a lower glycogen level (31.9 %w/DCW) than wild type control (40 %w/DCW). In contrast, the wild type control strain exposed to N- and NP-stresses tended to retain lipid levels and store more glycogen than PHB. In this model, we, for the first time, implemented a RuBisCO-overexpressing cyanobacterial factory for overproducing PHB, destined for biofuel and biomaterial biotechnology.


Introduction
A part of global biomass and atmospheric oxygen are reliably produced by cyanobacteria, which are photosynthetic prokaryotes capable of fixing atmospheric CO 2 and converting it to cellular carbon supply [1][2][3]. In terms of sustainable and renewable applications, cyanobacteria have established themselves as one of the most promising bioresources for alternate biofuels and bioproducts, even though their yield and efficiency have a tight correlation with their biomass [4,5]. To overcome the lower carbon constraint to the productivity of biofuel and bioproducts, many approaches for increasing algal biomass have been developed. Promising approaches to achieve this objective include nutritional adjustment or the genetic and metabolic engineering of cyanobacteria and algae. With respect to biofuel applications, the high content of stored carbohydrate means that microalgae are great candidates for bioethanol production [6][7][8][9][10]. Use of a nitrate-deprived medium for culturing the cyanobacterium Synechococcus sp. PCC 7002 resulted in a certain reduction in cell growth, but increased total carbohydrate content, herein glycogen, resulting in a higher C:N ratio of hydrolyzed biomass that could be further utilized by yeast and produced more ethanol [9]. The genetic engineering approach has also been applied to generate a higher biomass solution. Native overexpression of rbc or the RuBisCO gene operon encoding the ribulose-1,5-bisphosphate carboxylase/oxygenase (RuBisCO) enzyme in the Calvin-Benson-Bassham (CBB) cycle showed enhanced growth and photosynthesis in Synechocystis sp. PCC 6803 [11]. Each co-overexpression of the four genes Figure 1. Overview of the Calvin-Benson-Bassham (CBB) cycle and related biosynthetic pathways of polyhydroxybutyrate (PHB), membrane lipids, and glycogen in cyanobacterium Synechocystis sp. PCC 6803. Abbreviations of genes are: aas, acyl-ACP synthetase; ccaA, carbonic anhydrase; glgA, glycogen synthase; glgC, ADP-glucose pyrophosphorylase; glgP, glycogen phosphorylase; glgX, glycerol-3-phosphate dehydrogenase; lipA, a lipolytic enzyme-encoding gene; phaA, β-ketothiolase; phaB, acetoacetyl-CoA reductase; phaE and phaC, the heterodimeric PHB synthase; plsX and plsC, putative phosphate acyl-transferases; RuBisCO, the RuBisCO gene cluster, including rbcLXS, encoding RuBisCO large, chaperone, and small subunits, respectively. Abbreviations of intermediates are: FASII, fatty acid synthesis type II; fatty acyl-ACP, fatty acyl-acyl carrier protein; G1P, glucose 1phosphate; F6P, fructose 6-phosphate; G6P, glucose 6-phosphate; GAP, glyceraldehyde-3-phosphate; 3PGA, 3-phosphoglycerate; PHB, polyhydroxybutyrate; RuBP, ribulose-1,5-bisphosphate; TCA cycle, tricarboxylic acid cycle.

Overexpression of Native rbc Genes in Synechocystis sp. PCC 6803
First, three engineered Synechocystis sp. PCC 6803 strains, including wild type control (WTc), OXrbcL, and OXrbcLXS (Table 1), were constructed by double homologous recombination. The wild type control (WTc) strain was created by substituting the psbA2 gene with a Cm R cassette in the Synechocystis wild type (WT) genome ( Figure 2A). For single and triple recombinant plasmids, a native rbcL (or slr0009) gene fragment with a size of 1.4 kb, and a triple gene fragment containing rbcL-rbcX (or slr0011)-rbcS (or slr0012), with a size of about 2.4 kb, were ligated between flanking regions of the psbA2 gene of the expression vector pEERM and the upstream region of Cm R cassette, respectively (Table 1, Figure 2B,C). By PCR, using specific primer pairs (Supplementary Information Table S1), all overexpressing (OX) strains were confirmed for their complete segregation and gene location ( Figure 2B,C). PCR products with rbcL_F and CM_R primers confirmed the correct sizes of 2.3 and 3.3 kb, respectively, in OXrbcL and OXrbcLXS strains ( Figure 2B.1,C.1), compared with no band in WT and WTc. The UppsbA2_F and DSpsbA2_R primers confirmed the expected sizes of about 3.6 and 4.6 kb in OXrbcL and OXrbcLXS, respectively ( Figure 2B.2,C.2), while there were 2.4 and 2.2 kb in WT and WTc, respectively. Moreover, RT-PCR data confirmed gene overexpression in all OX strains ( Figure 2D). The oxygen evolution rate of both OXrbcL and OXrbcLXS was also increased when compared to WT and WTc ( Figure 2E).

Growth, Intracellular Pigment Contents and PHB Accumulation under Bicarbonate Supplementation
As a carbon source, NaHCO 3 (sodium bicarbonate) supplementation is substantially used for cyanobacterial cell growth instead of CO 2 gas bubbling [30], resulting in a pool of HCO 3 − , which is subsequently supplied with more CO 2 in the carboxysome [17]. We found a higher growth rate and shorter doubling time of both OXrbcL and OXrbcLXS under sodium bicarbonate supplementation when compared to WTc, whereas there were no significant changes under the normal BG 11 condition ( Figure 3A,B). Likewise, higher levels of intracellular pigments, including chlorophyll a and carotenoids, were noted in OX strains than in WTc under the sodium-bicarbonate-supplemented condition ( Figure 3C-F).
For bioplastic PHB production, higher accumulation occurred at about 7.30 and 6.40 %w/DCW in OXrbcL and OXrbcLXS, respectively, than in WTc with 4.65 %w/DCW of PHB content at day 11 under normal growth conditions ( Figure 4A). It is worth noting that the sodium-bicarbonate-supplemented BG 11 medium (BG 11 + NaHCO 3 ) could accelerate higher PHB production earlier in all strains at days 3 and 7 of cultivation ( Figure 4B). On the other hand, the lipid content of cells grown under the BG 11 + NaHCO 3 condition was slightly higher than for those cells grown in normal BG 11 medium ( Figure 4C,D). However, glycogen accumulation was lowered in all strains under the BG 11 + NaHCO 3 condition ( Figure 4E,F). In contrast, the increased glycogen production was noted at day 15 of cultivation under the normal BG 11 condition, representing the stationary phase of cells grown in nutrient-depleted medium ( Figure 4E).     (C.2) PCR products using UppsbA2_F and DSpsbA2_R primers, Lanes WT and WTc: Negative controls of a 2.4 kb fragment in WT and a 2.2 kb fragment in WTc, respectively, Lane OX1: one clone no. 1 containing a 4.6 kb fragment of UpsbA2-rbcL-rbcX-rbcS-Cm R -DpabA2. Transcript levels (D) measured by RT-PCR using specific primers, shown in Supplementary Information, Table S1, of rbc genes in WT, WTc and two overexpressing strains, including OXrbcL and OXrbcLXS. The 0.8% agarose gel electrophoresis of PCR products of all transcripts was performed from cells grown for 7 days in normal BG 11 medium. The 16s rRNA was used as reference. Oxygen evolution rates (E) of Synechocystis sp. PCC 6803 WT, WTc, OXrbcL and OXrbcLXS cells grown in normal BG 11 medium condition for 3 days (means ± S.D., n = 3). The statistical difference in the data between the values of WT and the engineered strain is represented by an asterisk at * p < 0.05. The cropped gels (in (D)) were taken from the original images of RT-PCR products on agarose gels as shown in Supplementary Information Figures S1 and S2).

Growth and PHB Accumulation of Cells Treated under Nutrient-Deprived Conditions
In this study, all Synechocystis strains were grown for 7 days in BG 11 + NaHCO 3 medium before being transferred to nutrient-deprived medium ( Figure 5). During stressed treatment, the nutrient-deprived conditions, including BG 11 -N and BG 11 -NP, caused a certain decrease in growth rate and chlorophyll a content, except for the phosphorus-deprived condition (BG 11 -P). In contrast, striking increases in PHB content were noted in OX strains treated under BG 11 -N and BG 11 -NP conditions, compared to WTc ( Figure 6). The highest level of PHB was found in the OXrbcLXS strain, with 38.9 and 38.1 %w/DCW at day 7 of treatment under BG 11 -NP and BG 11 -N, respectively ( Figure 6B,D). In addition, the PHB production of OXrbcLXS at day 9 of BG 11 -NP treatment was continuously induced up to 39.7 %w/DCW, followed by OXrbcL, which contained 34.6 %w/DCW of PHB ( Figure 6D). Nile-red-stained cells showed higher amounts of PHB granules in OX strains than WTc at day 7 of BG 11 -NP treatment ( Figure 6E). On the other hand, total lipid contents were lowered in all strains under the BG 11 -P condition, whereas insignificant amounts appeared under the BG 11 , BG 11 -N and BG 11 -NP conditions ( Figure 7A). The unsaturated lipid contents showed slight increases when compared to those under normal BG 11 condition ( Figure 7B). For glycogen accumulation, a high content of glycogen was noted when cells were exposed to nutrient-deprived conditions, in particular, BG 11 -N and BG 11 -NP ( Figure 7C).
The pha transcript levels, related to PHB biosynthesis, were upregulated under nutrient deprivation, except for the phaE transcript under the BG 11 -NP condition, compared to those under the normal BG 11 condition ( Figure 8). For membrane lipid synthesis, plsX and plsC transcript levels were increased by BG 11 -N and BG 11 -P treatments, whereas they were downregulated under the BG 11 -NP condition. On the other hand, the increased glgX transcript levels, which involve glycogen degradation, were noted in all nutrient-deprived treatments when compared to those under the normal BG 11 condition.
24, x FOR PEER REVIEW 7 of 20

Growth and PHB Accumulation of Cells Treated under Nutrient-Deprived Conditions
In this study, all Synechocystis strains were grown for 7 days in BG11 + NaHCO3 m dium before being transferred to nutrient-deprived medium ( Figure 5). During stress appeared under the BG11, BG11-N and BG11-NP conditions ( Figure 7A). The unsaturate lipid contents showed slight increases when compared to those under normal BG11 cond tion ( Figure 7B). For glycogen accumulation, a high content of glycogen was noted whe cells were exposed to nutrient-deprived conditions, in particular, BG11-N and BG11-N ( Figure 7C).    The pha transcript levels, related to PHB biosynthesis, were upregulated u ent deprivation, except for the phaE transcript under the BG11-NP condition, c those under the normal BG11 condition (Figure 8). For membrane lipid synthes  Table S2. All cropped gels were taken from the original images of RT-PCR products on agarose gels as shown in Supplementary Information Figure S3).

Discussion
In this study, we highlight the pragmatic finding of increased PHB production in the cyanobacterium Synechocystis sp. PCC 6803 overexpressing native RuBisCO genes and the relationship with lipid and glycogen metabolism (Figure 1). The primary CBB cycle converts atmospheric CO2 into carbon skeletons that are crucial for the production of valuable biocompounds. Recently, either native or heterologous expression of RuBisCO in cyanobacteria and higher plants was shown to cause increased growth and biomass yield, as well as increased photosynthesis efficiency [31][32][33]. Here, we introduced native RuBisCO or rbc genes, including rbcL and rbcLXS into Synechocystis, resulting in rbcL-and rbcLXSoverexpressing strains (OXrbcL and OXrbcLXS, respectively). Although we did not monitor the RuBisCO protein or its activity, the two OX strains had oxygen evolution rates that were 118-125 % higher than those of either WT or WTc under normal growth conditions ( Figure 2E). However, it is worth noting that the post-translational regulation at the Ru-BisCO level may not be expressed in accordance with its transcriptional regulation in an overexpressing strain [33]. On the other hand, we discovered that the addition of NaHCO3 (20 mM) significantly accelerated the growth rate of all strains, resulting in the mid-long phase emerging after 3 days of culture compared to 5 days in the typical BG11 medium ( Figure 3A,B). A low amount of NaHCO3 in a range of 5-20 mM concentrations was recommended for enhancing cyanobacterial growth due to excess NaHCO3 causing culture alkalinity [30,34,35]. When we measured chlorophyll a and carotenoids contents, there were no significant differences between WTc and OX strains grown in the normal BG11 medium ( Figure 3C,E). The observed differences in these two pigments among all strains occurred in sodium bicarbonate-supplemented conditions, particularly after day 9 of culture, when the growth phase had reached the stationary phase ( Figure 3D,F). These results suggested that a lower cell growth of WTc during the stationary phase in sodium bicarbonate-supplemented conditions in comparison with OX strains was partly caused by Figure 8. Relative transcript levels of phaA, phaB, phaC, phaE, plsX, plsC, and glgX performed by RT-PCR in Synechocystis WTc, OXrbcL, and OXrbcLXS strains after adapting in normal BG 11 medium, BG 11 -N, BG 11 -P, and BG 11 -NP for 7 days. The 16s rRNA was used as reference control. The band intensity data were shown in Supplementary Information Table S2. All cropped gels were taken from the original images of RT-PCR products on agarose gels as shown in Supplementary Information Figure S3).

Discussion
In this study, we highlight the pragmatic finding of increased PHB production in the cyanobacterium Synechocystis sp. PCC 6803 overexpressing native RuBisCO genes and the relationship with lipid and glycogen metabolism ( Figure 1). The primary CBB cycle converts atmospheric CO 2 into carbon skeletons that are crucial for the production of valuable biocompounds. Recently, either native or heterologous expression of RuBisCO in cyanobacteria and higher plants was shown to cause increased growth and biomass yield, as well as increased photosynthesis efficiency [31][32][33]. Here, we introduced native RuBisCO or rbc genes, including rbcL and rbcLXS into Synechocystis, resulting in rbcLand rbcLXS-overexpressing strains (OXrbcL and OXrbcLXS, respectively). Although we did not monitor the RuBisCO protein or its activity, the two OX strains had oxygen evolution rates that were 118-125 % higher than those of either WT or WTc under normal growth conditions ( Figure 2E). However, it is worth noting that the post-translational regulation at the RuBisCO level may not be expressed in accordance with its transcriptional regulation in an overexpressing strain [33]. On the other hand, we discovered that the addition of NaHCO 3 (20 mM) significantly accelerated the growth rate of all strains, resulting in the mid-long phase emerging after 3 days of culture compared to 5 days in the typical BG 11 medium ( Figure 3A,B). A low amount of NaHCO 3 in a range of 5-20 mM concentrations was recommended for enhancing cyanobacterial growth due to excess NaHCO 3 causing culture alkalinity [30,34,35]. When we measured chlorophyll a and carotenoids contents, there were no significant differences between WTc and OX strains grown in the normal BG 11 medium ( Figure 3C,E). The observed differences in these two pigments among all strains occurred in sodium bicarbonate-supplemented conditions, particularly after day 9 of culture, when the growth phase had reached the stationary phase ( Figure 3D,F). These results suggested that a lower cell growth of WTc during the stationary phase in sodium bicarbonate-supplemented conditions in comparison with OX strains was partly caused by lower intracellular pigment accumulations. For PHB production, it was discovered that, although cells could produce PHB throughout the growth phase, the amount of PHB accumulated was only increased when cells entered the late-log and early-stationary phases of growth, either at day 11 under the normal condition or day 7 under the sodium bicarbonate condition (Figure 4A,B). The PHB contents were from 6-8 %w/DCW, with OX strains having higher levels than WTc. Moreover, an increased accumulation of total lipids was noted in all strains caused by sodium bicarbonate supplementation at day 3 and day 7 of cultivation ( Figure 4C,D). It is important to point out that, in comparison to the condition for normal growth, glycogen accumulation was reduced in all strains under the sodium bicarbonate condition ( Figure 4E,F). This result showed that RuBisCO overexpression had little impact on glycogen levels compared to WTc under both normal and sodium bicarbonate conditions, while the flow of important intermediates from the CBB cycle under sodium bicarbonate conditions preferentially reached acetyl-CoA, a precursor for the TCA cycle, fatty acid synthesis, and PHB production. Accordingly, a strategy to increase metabolic flow towards acetyl-CoA level in order to promote fatty acid synthesis and PHB accumulation was previously carried out either by imposing environmental challenges, such as an N-starved condition [23,36], or by genetic manipulation, such as through heterologous expression of the xfpk gene, encoding phosphoketolase, which catalyzes the cleavage of xylulose 5-phosphate (Xu5P) or fructose 6-phosphate (F6P) to acetyl-P and glyceraldehyde 3-phosphate or erythrose 4-phosphate [37,38].
To obtain high levels of PHB in cyanobacteria, a combination strategy comprising genetic alteration and environmental challenges seems more promising than a single strategy [23,24,26]. In this study, the adaptation of cultivation to nutrient starvation was practically applied for inducing PHB production ( Figure 5). Although the growth rates of OX strains were higher than that of WTc under normal growth condition, no striking increases were found under all the nutrient starvation conditions studied. In contrast, an increase in PHB level was substantially induced by all nutrient-starved conditions, in particular BG 11 -N and BG 11 -NP treatments (Figures 6 and 9). Under these BG 11 -N and BG 11 -NP conditions, the OXrbcLXS strain exhibited the highest level of PHB by approximately 38.1-39.7 %w/DCW at days 7-9 of treatment ( Figure 6). A high PHB content of between about 34.5 and 34.6 %w/DCW was similarly produced by OXrbcL on day 9 of the BG 11 -N and BG 11 -NP treatments. Chlorotic cells caused by nitrogen deprivation accumulate glycogen and PHB as carbon sources for cellular acclimation and survival [23,[39][40][41], activated by the nitrogen regulators NtcA and PirC, DNA-transcriptional activators, which consequently trigger physiological changes, such as growth restriction, degradation of phycobilisomes, and chlorophyll a reduction [42][43][44][45][46]. Glycogen, which serves as the major carbon source, could subsequently contribute to increased PHB production during nitrogen depletion [23,46]. Our findings demonstrate that the amount of glycogen in WTc increased by 8-to 9-fold throughout 7 days of N-and NP-starvation compared to that under normal BG 11 , but the induced glycogen in OXrbcL and OXrbcLXS exhibited lower content than WTc ( Figure 7C). On the other hand, the phosphorus deprivation gradually induced a slight increase in glycogen and PHB throughout the 9 days of treatment ( Figures 6C and 7C). In contrast to WTc, it was interesting to find out that high PHB production started in both OX strains on day 9 of P-starved treatment ( Figure 6C). During the initial 7 to 9 days of the BG 11 -NP condition, the minor synergistic effect of P-starvation was shown to increase PHB content ( Figure 6D). Phosphorus deprivation lowers ATP production, while noncyclic photosynthetic electrons continuously flow and yield an increased NADPH pool, which is beneficial in PHB synthesis [47,48]. Figure 9. Summary of obtained results, including products and gene transcript levels from two overexpressing strains, including comparisons to Synechocystis sp. PCC6803 WTc after adapting cells in BG11, BG11-N, BG11-P, and BG11-NP at day 7 of treatment. In each box and bar graph, the number and bar represent the fold change of that value of each strain under stress condition divided by that value of that strain under normal BG11 condition. The statistical difference in the data between those values of WT and the engineered strain is represented by an asterisk at * p < 0.05.
In Figure 9, a summary of WTc and all engineered strains is shown under all nutrientdeprived conditions for 7 days. Glycogen and PHB accumulation in WTc were unquestionably raised by approximately 8-to 9-fold under N-and NP-starvation compared to that under the normal BG11 condition. It is important to note that RuBisCO gene overexpression at day 7 of treatment under both N-and NP-starvation conditions drastically lowered glycogen content but increased PHB content. This phenomenon was confirmed by the 12-to 15-fold increase in glgX transcript levels in the OXrbcL and OXrbcLXS strains compared to those under the normal condition, which are implicated in higher glycogen breakdown. In addition, phaA, phaB, and phaC transcript levels increased more than 8-fold in response to N-and NP-deprivation, but not the phaE transcript, which was dominant in increasing PHB synthesis in OX strains. In contrast, the total and unsaturated lipid levels did not change substantially during N-and NP-starvation, and cells responded to N- Figure 9. Summary of obtained results, including products and gene transcript levels from two overexpressing strains, including comparisons to Synechocystis sp. PCC6803 WTc after adapting cells in BG 11 , BG 11 -N, BG 11 -P, and BG 11 -NP at day 7 of treatment. In each box and bar graph, the number and bar represent the fold change of that value of each strain under stress condition divided by that value of that strain under normal BG 11 condition. The statistical difference in the data between those values of WT and the engineered strain is represented by an asterisk at * p < 0.05.
In Figure 9, a summary of WTc and all engineered strains is shown under all nutrientdeprived conditions for 7 days. Glycogen and PHB accumulation in WTc were unquestionably raised by approximately 8-to 9-fold under N-and NP-starvation compared to that under the normal BG 11 condition. It is important to note that RuBisCO gene overexpression at day 7 of treatment under both N-and NP-starvation conditions drastically lowered glycogen content but increased PHB content. This phenomenon was confirmed by the 12-to 15-fold increase in glgX transcript levels in the OXrbcL and OXrbcLXS strains compared to those under the normal condition, which are implicated in higher glycogen breakdown. In addition, phaA, phaB, and phaC transcript levels increased more than 8-fold in response to N-and NP-deprivation, but not the phaE transcript, which was dominant in increasing PHB synthesis in OX strains. In contrast, the total and unsaturated lipid levels did not change substantially during N-and NP-starvation, and cells responded to N-stress by preferentially maintaining intracellular lipid homeostasis. In contrast, phosphorus (P) deprivation raised PHB by 4-fold compared to that under the normal BG 11 condition but had no effect on glycogen deposition in RuBisCO-overexpressing strains compared to WTc. The phaA and phaC transcript levels were highly upregulated by more than 4-to 10-fold by P-stress when compared to the normal BG 11 condition, whereas the phaB and phaE transcript levels were only slightly altered.

Construction of rbc-Overexpressing Synechocystis sp. PCC 6803
Initially, the recombinant plasmids, including pEERM_rbcL and pEERM_rbcLXS (Table 1), were constructed to generate all overexpressing (OX) strains, including rbcLand rbcLXS-overexpressing Synechocystis strains (OXrbcL and OXrbcLXS, respectively). The construction of the pEERM_rbcL and pEERM_rbcLXS was performed by ligating the rbcL and rbcLXS genes amplified by PCR using a pair of rbcL_F and rbcL_R primers and another pair of rbcL_F and rbcS_R primers, respectively, as listed in Supplementary Information, Table S1, in between the restriction sites of SpeI and PstI in the pEERM vector [29]. After that, the recombinant plasmids were confirmed by double digestion with the SpeI and PstI restriction enzymes. Therefore, the correct recombinant plasmids (pEERM_rbcL and pEERM_rbcLXS) were transformed into Synechocystis sp. PCC 6803 WT cells by natural transformation, thereby generating OXrbcL and OXrbcLXS, respectively. For construction of the Synechocystis sp. PCC wild type control (WTc), the empty pEERM vector was transformed into Synechocystis WT cells, represented as Synechocystis WT containing the Cm R cassette gene. For host cell suspension preparation, the Synechocystis sp. PCC 6803 host cells were cultured in BG 11 medium until OD at 730 nm increased by about 0.3-0.5. Then, the cell culture (50 mL) was harvested by centrifugation at 5500 rpm (3505× g), 25 • C for 10 min, and cell pellets were resuspended in fresh BG 11 medium (500 µL). After this, at least 10 µL of constructed recombinant plasmids were added and mixed with the Synechocystis host cell suspension and incubated in the culture room with continuous white light illumination at 27-30 • C for 16-18 h. Then, the sample mixture was spread on a BG 11 agar plate containing 10 µg/mL of chloramphenicol. All agar plates were incubated in the culture room until surviving colonies occurred (for about 2-3 weeks). Each single colony was picked and streaked on a new BG 11 agar plate containing chloramphenicol up to 20 and 30 µg/mL concentration. The obtained transformants were confirmed for gene size, location, and segregation by a PCR method using many specific pairs of primers, as shown in Supplementary Information, Table S1.

Strains and Culture Conditions
Synechocystis sp. PCC 6803 wild type (WT) was derived from the Berkeley strain 6803, which was isolated from fresh water in California, USA [2], and all engineered strains (WTc, OXrbcL, and OXrbcLXS) were grown in normal BG 11 medium and BG 11 medium supplemented with 20 mM NaHCO 3 . The normal growth condition was performed at 27-30 • C, under white light illumination with 40-50 µmole photon/m 2 /s intensity. The culture flask with an initial cell density at 730 nm (OD730) of about 0.1 was placed on a shaker (160 rpm). Cell growth was determined by spectrophotometrically measuring OD730. For the nutrient-deprived treatment, all Synechocystis strains were initially grown in BG 11 containing 20 mM NaHCO 3 for 7 days before treating the harvested cell pellets with nutrient starvation. The nutrient-deprived media included (1) BG 11 without NaNO 3 with added FeSO 4 instead of ferric ammonium citrate (BG 11 -N), (2) BG 11 with KCl addition instead of KH 2 PO 4 (BG 11 -P), and (3) BG 11 without nitrogen and phosphorus (BG 11 -NP). The initial OD730 of the cell culture under nutrient-stressed treatment was adjusted to about 0.45.

Determinations of Intracellular Pigments and Oxygen Evolution Rate
One mL of cell culture was harvested by centrifugation at 5500 rpm (3505× g) for 10 min. The intracellular chlorophyll a and carotenoids in the cell pellets were extracted by N,N-dimethylformamide (DMF), vortexed, and incubated in the dark for 10 min. After quickly spinning to remove debris, the absorbance of the yellowish supernatant was spectrophotometrically measured for its absorbances at 461, 625 and 664 nm. The contents of chlorophyll a and carotenoids were calculated according to [49,50]. The photosynthetic efficiency was determined in terms of the oxygen evolution rate. The cell culture (5-10 mL) was harvested by centrifugation at 5500 rpm (3505× g) for 10 min. Cell pellets were resuspended in fresh BG 11 medium (1 mL) and incubated under darkness for 30 min before measuring the oxygen evolution rate. Saturated white light was used as a light source at 25 • C using a Clark-type oxygen electrode (Hansatech instruments Ltd., King's Lynn, UK). The unit of oxygen evolution rate was represented as µmol O 2 /mg chlorophyll a/h [24].

Total RNAs Extraction and Reverse Transcription-Polymerase Chain Reaction (RT-PCR)
Total RNAs were isolated from each strain using the TRIzol ® Reagent (Invitrogen, Life Technologies Corporation, Carlsbad, CA, USA). The purified RNAs (1 µg) were converted to cDNA by reverse transcription using ReverTra ACE ® qPCR RT Master Mix Kit (TOYOBO Co., Ltd., Osaka, Japan). Then, the cDNA was further used as the template for PCR amplification with different pairs of primers, as listed in Supplementary Information ,  Table S1. The PCR conditions were performed by initial denaturing at 94 • C for 3 min, followed by suitable cycles of each gene ( Supplementary Information, Table S1) at 94 • C for 30 s, primer melting temperature (Tm, Supplementary Information, Table S1) for 30 s, and 72 • C for 20 s, and then final extension at 72 • C for 7 min. After that, the PCR products were checked by 1% (w/v) agarose gel electrophoresis. Quantification of band intensity was determined by Syngene ® Gel Documentation (Syngene, Frederick, MD, USA).

Quantitative Analysis of PHB Contents
The PHB content was detected by high-performance liquid chromatography (HPLC) (Shimadzu HPLC LGE System, Kyoto, Japan). Cell cultures (15-30 mL) were harvested by centrifugation at 5500 rpm (3505× g) for 10 min. To prepare samples, 100 µL of internal standard (20 mg/mL adipic acid) and 800 µL of 98% (v/v) sulfuric acid were added into cell pellets. After mixing, the sample mixture was boiled at 100 • C for 1 h for digesting PHB to crotonic acid (modified from [24]). Then, the reaction mixture was filtered using a 0.45 µm polypropylene membrane filter, and detected using a carbon-18 column, Inert Sustain 3-µm (GL Sciences, Tokyo, Japan) and UV detector at 210 nm. The running buffer was 30% (v/v) acetonitrile in 70% (v/v) of 10 mM KH 2 PO 4 (pH 2.3), with a flow rate of about 1.0 mL/minute. Authentic commercial PHB (Sigma-Aldrich, Inc., St. Louis, MO, USA) was used as standard, which was prepared using the same method as the samples. In addition, the dry cell weight (DCW) was determined by incubating cell pellets at 60 • C until obtaining a constant weight [26].

Lipid Extraction and Determinations of Total Lipid and Unsaturated Lipid Contents
Cell culture (25 mL) was harvested by centrifugation at 5500 rpm (3505× g) for 10 min. The obtained cell pellet was mixed with 1 mL of CHCl 3 :MeOH (2:1 ratio) solution and incubated at 55 • C for 2 h. Then, distilled water (0.5 mL) was added and mixed by vortexing for 2 min. After leaving at room temperature for 10 min, the sample mixture was centrifuged at 5500 rpm (3505× g) for 10 min. The chloroform phase (lower layer) containing lipids was collected, while the upper aqueous phase was re-extracted by addition of chloroform (500 µL) (modified from [12,51]). The lipid-containing chloroform phase was pooled and evaporated in a fume hood at room temperature overnight to obtain total lipids. The total lipid content was determined by an acid-dichromate oxidation method (modified from [28,52]. The lipid extract was dissolved in 1 mL of 98% H 2 SO 4 and later mixed by vortexing. After that, 1 mL of potassium dichromate (K 2 CrO 7 ) solution was added to the sample and boiled at 100 • C for 30 min, followed by cooling down to room temperature. Then, 1 mL of distilled water was added and mixed. The sample was measured for absorbance at 600 nm by a spectrophotometer. Commercial canola oil was used as standard. The unit of total lipid content was %w/DCW.
The unsaturated lipid content was measured using the colorimetric sulfo-phosphovanilin (SPV) method, with commercial linoleic acid serving as the standard (modified from [53]). Extracted lipids were dissolved in 1 mL of 98% H 2 SO 4 and shaken vigorously before boiling at 100 • C for 30 min. After that, the boiled samples were cooled down to room temperature. One mL of a solution of 17% (v/v) H 3 PO 4 and 0.2 mg/mL vanillin was added and incubated for 10 min at room temperature. The absorbance at 540 nm was spectrophotometrically measured to determine the unsaturated lipid content. The unit of total unsaturated lipid content was %w/DCW.

Glycogen Extraction and Determination of Glycogen Content
Glycogen was extracted by alkaline hydrolysis [54,55]. The cell culture (15-30 mL) was harvested by centrifugation at 5500 rpm (3505× g), at 25 • C for 10 min. The cell pellets were resuspended in 400 µL of 30 %(v/v) KOH and boiled at 100 • C for 1 h. After centrifugation at the same speed to collect the supernatant, 900 µL of cold absolute ethanol was added and then incubated at −20 • C overnight to precipitate glycogen. To obtain the glycogen pellets, the sample mixture was centrifuged at 12,000 rpm (14,489× g), at 4 • C for 30 min, and dried at 60 • C for overnight. The anthrone assay was used for determining the glycogen content, and commercial oyster glycogen (Sigma-Aldrich, St. Louis, MO, USA) served as the standard (modified from [56,57]). Glycogen pellets were dissolved with 1 mL of 10% H 2 SO 4 . Then, 0.2 mL of dissolved sample was mixed with 0.2 mL of 10% H 2 SO 4 and 0.8 mL of anthrone reagent. The reaction mixture was then boiled at 100 • C for 10 min. After cooling down the samples to room temperature, their absorbance at 625 nm was measured by spectrophotometer. The unit of glycogen content was %w/DCW

Conclusions
Enhanced levels of PHB accumulation were attained in engineered strains of Synechocystis sp. PCC 6803 (OXrbcL and OXrbcLXS, involved in CBB cycle) under nutrientdeprived conditions. A faster growth rate was found in OX strains under the normal growth condition. Following the application of N-and NP-stresses, a significant carbon flux from glycogen storage flowing to PHB synthesis was definitely triggered as a stress acclimation response. In Figure 9, the mechanistic perspective provided indicates that RuBisCO overexpression responded to nutrient deprivation, in particular N-and NPstresses, by driving cells to accumulate PHB rather than glycogen and lipids when compared to wild type control cells, which preferred to store more glycogen than PHB and lipids. These RuBisCO-overexpressing Synechocystis strains are potential cell factories for the production of value-added chemicals, such as biomaterials and biofuels.