Heterologous Expression of Various PHA Synthase Genes in Rhodospirillum rubrum

The economic importance of recovery and regeneration of alternative energy and carbon sources has evolved into an ever-growing global challenge in the recent decades. Polyhydroxyalkanoates (PHAs) could be a forward-looking alternative to petrochemical-based plastics since they are biobased, biodegradable, biocompatible, and derived from renewable resources1,2,3,4. PHAs are synthesized by more than 300 different microorganisms as storage of carbon and energy, if an excess of carbon source is available but an essential nutrient is limited at the same time5,6,7. The most commonly accumulated and best studied PHA is polyhydroxybutyrate (PHB)7. Due to their elastomeric and thermoplastic character, PHAs have attracted much industrial attention for economic usage8. Various applications for PHAs as packaging material, medical implant material9, textile fibers10, feed supplements or even as precursor feedstock for biofuels have been reported4. Although there is intensive research on bacterial PHAs, their production costs are still not competitive to conventional plastics6,11. One of the main contributors to these costs arises from pure carbon sources used as feedstock for bacterial fermentations, which can comprise up to 50 % of the total production costs12. Therefore, bulk production of PHA requires feasible and economical fermentation processes from cheap carbon sources. One promising and widely investigated approach is the utilization of industrial waste products13. Besides complex industrial wastes, gasification processes enable the conversion of any carbonaceous material into synthesis gas (syngas)13,14,15. Syngas mainly consists of Heterologous Expression of Various PHA Synthase Genes in Rhodospirillum rubrum


Introduction
The economic importance of recovery and regeneration of alternative energy and carbon sources has evolved into an ever-growing global challenge in the recent decades.Polyhydroxyalkanoates (PHAs) could be a forward-looking alternative to petrochemical-based plastics since they are biobased, biodegradable, biocompatible, and derived from renewable resources 1,2,3,4 .PHAs are synthesized by more than 300 different microorganisms as storage of carbon and energy, if an excess of carbon source is available but an essential nutrient is limited at the same time 5,6,7 .The most commonly accumulated and best studied PHA is polyhydroxybutyrate (PHB) 7 .Due to their elastomeric and thermoplastic character, PHAs have attracted much industrial at-tention for economic usage 8 .Various applications for PHAs as packaging material, medical implant material 9 , textile fibers 10 , feed supplements or even as precursor feedstock for biofuels have been reported 4 .
Although there is intensive research on bacterial PHAs, their production costs are still not competitive to conventional plastics 6,11 .One of the main contributors to these costs arises from pure carbon sources used as feedstock for bacterial fermentations, which can comprise up to 50 % of the total production costs 12 .Therefore, bulk production of PHA requires feasible and economical fermentation processes from cheap carbon sources.One promising and widely investigated approach is the utilization of industrial waste products 13 .Besides complex industrial wastes, gasification processes enable the conversion of any carbonaceous material into synthesis gas (syngas) 13,14,15 .Syngas mainly consists of The phototrophic non-sulfur purple bacterium Rhodospirillum rubrum is known for its metabolic versatility.Particularly, R. rubrum is able to synthesize PHA under heterotrophic or even autotrophic growth with carbon monoxide as carbon and energy source.R. rubrum has therefore become a promising candidate for future cheap PHA production.However, R. rubrum synthesizes lower amounts of PHAs in comparison to well-known PHA producers like Ralstonia eutropha H16 or recombinant Escherichia coli strains.Since the PHA synthase is the key enzyme of PHA biosynthesis, genes encoding for twelve different PHA synthases were heterologously expressed in two generated phaC deletion mutants of R. rubrum in this study.To clearly see the effect of the foreign PHA synthases, PHA-negative mutants were required.The single mutant R. rubrum ∆phaC2 showed a PHA-leaky phenotype (< 1 % PHA, wt/wt, of CDW), while the double mutant R. rubrum ∆phaC1∆phaC2 accumulated no measurable PHA.Eight different PHA synthase genes of class I, and four of class IV were chosen for heterologous expression.All recombinant R. rubrum strains showed significant PHA synthesis and accumulation, although PHA contents in the recombinant strains of the single mutant R. rubrum ∆phaC2 were generally higher in comparison to those of the double mutant R. rubrum ∆phaC1∆phaC2.Recombinant strains of the single mutant could be divided into two groups according to the accumulation of PHA in the cells.While recombinant strains dedicated to group one showed an increased PHA synthesis when compared to the wild type carrying an empty vector, strains of group two accumulated less PHA than the wild type.Finally, it was possible to increase the accumulation of PHA by up to 25 % due to heterologous expression of PHA synthase genes compared to the wild type.

Key words:
Rhodospirillum rubrum, Polyhydroxybutyrate (PHB), PHA-synthase genes, heterologous expression, PHB-negative deletion mutants hydrogen (H 2 ), carbon monoxide (CO) and carbon dioxide (CO 2 ), and has attracted great biotechnical interest as a cheap but undefined feedstock for the fermentation of various microorganisms in the last years 14 .
The phototrophic non-sulfur purple bacterium Rhodospirillum rubrum is able to synthesize PHA under heterotrophic and even autotrophic growth with carbon monoxide (CO) and carbon dioxide (CO 2 ) as carbon and energy source 14,15 .Growth of Rhodospirillum rubrum utilizing syngas has already been intensively investigated 14,15,16 .In 2008, a profitable and technically feasible concept for PHA accumulation of R. rubrum based on the utilization of synthesis gas was presented for the first time 16 .PHA biosynthesis through syngas fermentation was carried out, and the costs were $ 2 -4 per kg less than in the case of using sugar fermentation and recombinant E. coli strains.Furthermore, H 2 was produced by R. rubrum under these conditions, which is also of great industrial interest 16 .These properties indicate the great biotechnological potential of R. rubrum.However, R. rubrum synthesizes lower amounts of PHAs than well-known PHA producers like Ralsto nia eutropha or recombinant E. coli strains.
Since the PHA synthase is known to be the key enzyme of PHA biosynthesis, many studies aimed at heterologous expression of foreign PHA-synthase genes to increase PHA accumulation in the cells.Based on their substrate specificity and subunits, PHA synthases are divided into four classes 17 .Class I PHA synthases utilize short chain length hydroxyalkanoate (HA SCL ) monomers with three to five carbon atoms and are found in R. eutropha H16, R. rubrum S1 and many other α-and b-proteobacteria 18,19  PHB biosynthesis in R. rubrum has already been investigated intensively, and it shares some general similarities with PHB biosynthesis in R. eu tropha H16, which represents the best studied model organism 23 .Basically, the PHB synthesis is divided into three steps 24,25,26 .First, two molecules of acetyl-CoA are converted by a b-ketothiolase (PhaA) in a claisen condensation to yield acetoacetyl-CoA.The NADPH-dependent acetoacetyl-CoA reductase (PhaB) subsequently reduces acetoacetyl-CoA stereospecifically to R-(-)-3-hydroxybutyryl-CoA.In a last step, the PHA synthase (PhaC) polymerizes R-(-)-3-hydroxybutyryl-CoA to PHB, and CoA is released again 17,27 .However, PhaB of R. rubrum has been described in literature as an NADH-dependent isoenzyme, which converts acetoacetyl-CoA to S-(-)-3-hydroxybutyryl-CoA 28 .As PHA synthases are known to be stereospecific for R-(-)-3-hydroxybutyryl-CoA, two enoyl-CoA hydratases obviously convert S-(-)-3-hydroxybutyryl-CoA via crotonyl-CoA to R-(-)-3-hydroxybutyrate in R. rubrum S1 28,29 .Since the entire genome sequence of R. rubrum S1 was published in 2011, a first basis for specific manipulations of the genome of R. rubrum S1 was achieved 30 .
Since 2000, it has been known that R. rubrum S1 carries at least two genes encoding for PHA synthases 31,32 .However, another homologous PHA synthase gene (phaC3) was identified recently 32 .All three PHA synthase genes were characterized, and it turned out that PhaC2 had a significantly lower enzyme activity than PhaC1 and PhaC3 in vivo.PhaC2 however, was detected in ten-fold higher concentration in the cytoplasm of the cell.Furthermore, single and multiple deletion mutants of the genes phaC1, phaC2 and phaC3 were generated and investigated by Jin and Nikolau (2012) 32 .It was shown that PhaC1 and PhaC3 were only marginally involved in the PHB synthesis, although only gene phaC1 (Rru_0275) is located in a putative PHB operon.
In this study, two phaC deletion mutants of R. rubrum S1 were generated.Twelve different PHA synthase genes were expressed heterologously in both mutants to investigate PHB biosynthesis.Appropriate synthases were identified by an intensive literature research with special emphasis on high enzyme activities, successful previous heterologous expression, and strong PHB accumulation in host strains or in the wild type, respectively.This is the first detailed study dealing with the effect of foreign PHA synthase genes on PHB biosynthesis in R. ru brum S1.
Cultivation of microorganisms.E. coli cells were grown in LB-Medium at 37 °C.Cultivation of R. rubrum strains was carried out in supplemented succinate fructose nitrogen (SSFN) medium (modified from Bose et al. 33 ) with the omission of malate and a reduced concentration of ammonium sulfate (0.5 g L -1 ).An amount of 10 g L -1 of fructose, 2 g L -1 of di-sodium succinate hexahydrate, and 0.3 g L -1 of yeast extract and casamino acids were added to the media.For cultivation, baffled flasks were used with a medium volume/flask volume ratio of 1:5 -1:10.Precultures of recombinant R. ru brum strains were incubated for 48 h in 20 mL SSFN medium at 125 rpm on a gyratory shaker.For main cultures, 100 mL SSFN medium with reduced concentration of yeast extract, casamino acids, and ammonium sulfate was inoculated with 10 % (vol/vol) of precultures in 1000 mL baffled KLETT flasks.Cultivation was carried out for 50 h at 30 °C and 125 rpm.Growth was measured via optical density (OD) measurement using a KLETT-Summerson photometer (Manostat) at 520 to 580 nm.Cultures with a volume of 5 mL were centrifuged in 1.5 mL reaction tubes at 10.000 × g for 5 minutes.Cultures with a volume of 50 mL and more were centrifuged in 50 mL reaction tubes for 15 minutes at 4.000 x g and 4 °C.The amount of 25 µg mL -1 kanamycin was applied for recombinant R. rubrum strains, and 50 µg mL -1 kanamycin for recombinant E. coli strains.

Cloning of PHA synthase genes
The chosen PHA synthase genes were amplified from genomic DNA of the respective host strain (Table 1) by PCR applying proofreading Phusion polymerase (Thermo Fischer Scientific).Primers used for amplification are shown in Table 2.After subcloning PCR fragments into vector pJET1.2/blunt,fragments were excised and ligated into the broad host range expression vector pBBR1MCS-2 34 .Recombinant vectors (Table 3) were controlled to harbor correct PHA synthase gene inserts applying sequence analysis.Obtained sequences were analyzed using Chromas software (version 1.45, Technelysium Pty.Ltd.), Genamics Expression software (version 1.100 [http://genamics.com/expression/index.htm]),BLAST online service available on NCBI (National Center for Biotechnology Information [http://blast.ncbi.nlm.nih.gov/Blast.cgi]), and BioEdit 35 .Genomic DNA was isolated using the "DNeasy Blood&Tissue Kit" (QIAGEN; Hilden, G.).Plasmid DNA was isolated by the method of Birnboim and Doly 36 or highly purified using the "peqGOLD Plasmid Miniprep Kit" (Peqlab Biotechnologie GmbH, Erlangen, G). Isolation of DNA-fragments from agarose gels was carried out with the "High Pure PCR Cleanup Micro Kit" (Hoffmann-La Roche Ltd., S).For agarose gel electrophoreses, 0.8 % (wt/vt) agarose gels were prepared.Tris-Borate-EDTA-Buffer (TBE-Buffer) was used as buffer.Electrophoresis was carried out at 100 to 170 V, and 80 mA for 50 to 90 min.Competent cells of E. coli were prepared and transformed by the CaCl 2 procedure as described by Hanahan 37 .Generation of deletion mutants of R. rubrum S1 Flanking regions of 700 to 1000 bp upstream and downstream of the target genes phaC1 (Rru_A0275) and phaC2 (Rru_A2413) were amplified by PCR, adding XbaI and EcoRI restriction sites to the resulting fragments.The fragments were then restricted with XbaI and EcoRI and ligated, forming fragments of approximately 1800 (for ΔphaC1) and 2000 (for ΔphaC2) bp.The fragments were again amplified by PCR, digested with XbaI and ligated into an XbaI digested pJQ200mp18 vector, yielding gene replacement vectors pJQ200mp18::∆phaC1 and pJQ200mp18::∆phaC2.The plasmids pJQ200mp18::∆phaC1 and pJQ200mp18: :∆phaC2 were mobilized from E. coli S17-1 as the donor strain to the corresponding recipient R. rubrum strain by the spot agar mating technique 38 .Mutants of R. rubrum were identified on SSN solid media (SSFN media without fructose) with 10 % (wt/vol) sucrose and SSN solid media containing 20 µg mL -1 gentamycine 39 .Successful gene replacement was confirmed by PCR analyses and DNA sequencing, with the primers used for generating the flanking region fragments of phaC1 and phaC2.R. rubrum ΔphaC2 was used as the initial strain for generating the double deletion mutant R. rubrum ΔphaC2ΔphaC1.Electroporation of R. rubrum cells.The transfer of generated vectors to cells of R. rubrum was carried out by electroporation 40 .However, a modified Super Optimal Catabolite Repression-Medium (SOC medium) containing fructose instead of glucose was used.Regeneration of the cells was carried out for 5 -6 h at 125 rpm and 30 °C.Cells were harvested, resuspended in 100 µL sterile saline, and transferred to SSFN solid medium containing kanamycin.Single colonies of electroporated R. rubrum cells were obtained after 3 -5 days at 30 °C.
PHB Analysis.Samples of 25 and 50 mL were collected from liquid cultures by centrifugation at 4.000 × g and 4 °C for 15 minutes.An amount of 5 -10 mg of dried cells were subjected to acid methanolysis in the presence of 85 % (vol/vol) methanol and 15 % (vol/vol) sulfuric acid for 3 h at 100 °C.The resulting methyl esters of 3-hydroxybutyrate were characterized by gas chromatography according to 20,41 .For gas-chromatographic analysis, an Agilent 6850 gas chromatograph (Agilent Technologies) equipped with a BP21 capillary column (50 m by 0.22 mm; film thickness, 250 nm; stationary phase: polyethylene glycol [PEG]), and a flame ionization detector (Agilent Technologies) were used.The evaluation of the data analysis was based on the Agilent Cerity QA-QC software.As reference, retention times of commercial 3-hydroxy fatty acid standards were used for the identification and quantification of the fatty acids present in the samples.The PHB content in % (wt/wt) of the CDW was determined from a standard curve generated, based on using purified PHB samples.
Growth experiments.For each recombinant strain, two independent biological experiments were performed.Growth experiments started with pre-cultures (20 mL SSFN supplemented with 1.5 g L -1 casamino acids/ yeast extract/ 0.5 g L -1 ammonium sulfate) in baffled flasks with a total volume of 100 mL, which were grown for 48 h at 30 °C and 125 rpm.Main cultures (two replicate flasks for each strain in each biological experiment) were cultivated in baffled KLETT flasks with a total volume of 1000 mL filled with 100 ml of SSFN medium (0.3 g L -1 casamino acids/yeast extract) for 50 h at 30 °C and 125 rpm.To inoculate main cultures, pre-cultures were harvested by centrifugation (15 min at 4000 rpm, 4 °C), and resuspended in 4 mL medium.Each replicate main culture was inoculated with 2 mL of this suspension.Samples for PHB content determination were taken after 27 h and 50 h, representing cells of the early and late stationary growth phase, respectively.Cells were harvested by centrifugation (15 min at 4000 rpm, 4 °C), frozen, and used for GC analysis as described previously.Therefore, all recombinant strains of R. rubrum were tested four times for PHB accumulation in the early, and four times in the late stationary growth phase in this study.

Heterologous expression of foreign PHA synthases in the single mutant R. rubrum ∆phaC2.
To characterize possible differences in growth and other properties of the recombinant R. rubrum strains, cultivations were performed in SSFN liquid media (Bose et al., 1961)  33 , which was modified according to the requirements of this study.The purpose was to apply media that allow good growth as well as high PHB synthesis rates.To reach this goal, 2 g L -1 di-sodium succinate hexahydrate was added to the medium as additional and rapidly convertible carbon source beside fructose (10 g L -1 ) to support cell growth.In order to promote PHB synthesis by limiting the N-sources and providing the mentioned C-sources in excess, limiting amounts of (NH 4 ) 2 SO 4 and only low amounts of the supplements yeast extract and casamino acids were provided.This adjusted media composition showed best results in initial growth experiments (data not shown).
Growth curves were generated based on average values of measured optical cell densities (OD in Klett-units, KU).Growth curves for recombinant strains of R. rubrum ∆phaC2 are shown in Fig. 1.

F i g . 1 -Growth curves of all tested recombinant strains of R. rubrum ΔphaC2. Cells were cultivated for 50 h at 30 °C and 125 rpm in 100 mL SSFN-Medium at conditions promoting PHA accu mulation. Pre-cultures were grown in SSFN-Medium for 48 h. Op tical cell density was measured in Klettunits [KU]. All strains were tested in two independent biological experiments in dupli cate. All data shown here are based on mean values. The strains R. rubrum S1 + MCS-2 (-) and R. rubrum ΔphaC2 + MCS-2 (-) were cultivated as reference strains. MCS-2 = pBBR1MCS-2; MCS-2 (-) = pBBR1MCS-2 without insert; h = hour; KU = Klett-units.
An exponential growth phase (0 -27 h) and a stationary growth phase starting after 27 h of cultivation were observed in all cultures, whereas lagphases were not detected due to the inoculation of the cultures with actively growing cells.No significant differences with regard to growth were identified during the exponential growth phase.All strains reached the stationary growth phase after 27 h but showed different maximum optical cell densities (OD).According to this, the recombinant strains could be subdivided into two groups based on the shapes of their growth curves and the maximum OD.The highest maximum OD was reached by the reference strain R. rubrum S1 pBBR1MCS-2 (-) and amounted to 600 KU.In contrast, with reference strain R. rubrum ΔphaC2 pBBR1MCS-2 (-), the lowest maximum OD (480 KU) was obtained.All other generated recombinant strains behaved quite similar concerning growth when compared to R. rubrum S1 pBBR1MCS-2 (-) (group 1) or R. ru brum ΔphaC2 pBBR1MCS-2 (-) (group 2), respectively.Therefore, the recombinant strains carrying the PHA synthase genes phaC1 R. rubrum , phaC P. denitrificans and phaC B. megaterium ::phaR B. cereus were referred to group 1.With the exception of the strain R. rubrum ΔphaC2 pBBR1MCS-2:: phaC2 R. rubrum :: phaC1 R. rubrum , all other strains exhibited similar growth curves like R. rubrum S1 pBBR1MCS-2 (-).Representative for group 1, R. rubrum S1 pBBR1MCS-2(-) showed a growth rate of µ = 0.1935 h -1 and a doubling time of t d = 3.58 h.For R. rubrum ΔphaC2 pBBR1MCS-2(-) (group 2), a growth rate of µ = 0.1898 h -1 and a doubling time of t d = 3.65 h was determined.Maximum OD values and growth rates of recombinant R. rubrum ΔphaC2 strains are shown in Table 4. Deviations in OD for the repeated cultivations of recombinant strains were low (see error bars Fig. 1).
Nearly all strains showed higher PHB content in the late stationary growth phase than in the early stationary phase (Fig. 2).While the wild type strain accumulated a maximum of 26 % PHB (CDW), the complemented strain R. rubrum ∆phaC2 pBBR1MCS-2::phaC2 R. rubrum showed the highest PHB synthesis of all strains and accumulated PHB to 32 % of CDW.Expression of several other foreign phaC genes resulted in comparable cellular PHB content.Expression of gene phaC1 R. rubrum was not able to restore the phenotype of the wild type.Expression of genes phaC1 R. rubrum and phaC2 R. rubrum likewise did not result in a positive effect on PHB accumulation.While three recombinant strains harboring vectors with shuffled phaC and phaR genes encoding, the different subunits of type IV PHA synthases of B. megaterium and B. cereus stored more PHB than the wild type, strain R. rubrum ∆phaC2 pBBR1MCS-2:  :pBBR1MCS-2::phaC B. megaterium ::phaR B. cereus interestingly stored significantly lower amounts of PHB.
Another surprising finding was that the heterologous expression of phaC C.sp. resulted in higher PHB content in the early stationary growth phase than in the late stationary growth phase.The deletion mutant R. rubrum ΔphaC2 pBBR1MCS-2 (-) accumulated low content of PHB (< 1 %, wt/wt, of the CDW) after 50 h cultivation, however, no measureable PHB was found in the early stationary growth phase.
Heterologous expression of foreign PHA synthases in the double mutant R. rubrum ∆phaC1∆phaC2.In contrast to the growth of the recombinant R. rubrum ∆phaC2 strains (Fig. 1), lower maximum ODs were measured in cells of the recombinant strains of the double mutant R. rubrum ∆phaC1∆phaC2 (Fig. 3).The reference strain R. ru brum S1 pBBR1MCS-2 (-) again showed the highest maximum OD amounting to 600 KU.Lowest maximum ODs were observed in cultures of R. ru brum ΔphaC1ΔphaC2 pBBR1MCS-2 (-) and R. ru brum ΔphaC1ΔphaC2 pBBR1MCS-2::phaC1 R. rubrum , where only 480 KU and 450 KU were measured, respectively.All other strains showed maximum ODs between 480 and 600 KU.The deviations in OD in the repeated different cultivations were likewise low (see error bars Fig. 3).
In contrast to the heterologous expression of phaC genes in the single mutant (Fig. 1 and Fig. 2), plasmid based phaC expression in the double mutant R. rubrum ∆phaC1∆phaC2 did not result in significantly increased PHB content in cells of the recombinant strains when compared to the wild type S1 harboring the empty pBBR1MCS-2 vector (Fig. 4).As expected, the recombinant strains R. ru brum ∆phaC1∆phaC2 carrying an empty vector, showed no measurable PHB content, while the plasmid-based expression of phaC1 R. rubrum resulted in a PHB accumulation of ~ 2 % (wt/wt) in the early and late stationary growth phase.Heterologous expression of pBBR1MCS-2::phaC2 R. rubrum :: phaC1 R. rubrum did not fully complement the wild type phenotype.Expression of the B. cereus PHA synthase genes phaC and phaR showed slightly higher PHB content of + 1 % (wt/wt) of CDW.Interestingly, the PHB content of nearly all recombinant strains of R. rubrum ∆phaC1∆phaC2 showed 90 to 100 % higher PHB content in the late stationary growth phase compared to the early stationary growth phase.This is not the case for the heterologous expressions in the single mutant, where PHB content was only 20 -25 % higher in the late stationary phase.

Discussion
Several previous studies dealt with the growth of the wild type R. rubrum S1 under various cultivation conditions and the used carbon sources.Cultivations were carried out under heterotrophic and autotrophic conditions as well as in the absence or presence of light 14,15,16,18,42,43 .In most preceding studies, R. rubrum showed slow growth rates leading to long and inefficient cultivations.In order to acquire fast growth of the cultures and high content of accumulated PHB in the cells, cultivation and medium components were successfully optimized in our study.

PHB accumulation of recombinant R. rubrum strains
The heterologous expression of PHA synthase genes in the single and double deletion mutants R. rubrum ΔphaC2 and R. rubrum ΔphaC1ΔphaC2, respectively, resulted in a significant accumulation of PHB in all generated strains (Fig. 2 and Fig. 4).Six of the thirteen generated phaC hybrid vectors led to an increased PHB accumulation in cells of R. rubrum ΔphaC2 in comparison to the wild type strain carrying an empty vector (Table 5).In contrast, all recombinant strains of R. rubrum ΔphaC1ΔphaC2 har boring heterologous phaC genes accumulated significantly less PHB than the recombinant single mutants and the wild type strain.Therefore, deleting the chromosomal phaC1 R. rubrum in R. rubrum might have caused polar effects on other PHB biosynthesis genes.The phaC1 R. rubrum [Rru_A0275] is located adjacent to phaA [Rru_ A0274] and to phaB [Rru_A0273] in a putative PHB operon 32 .In addition, a putative phaR gene [Rru_A0276] was observed by in silico studies in close proximity (data not shown).
Growth experiments in this study (Fig. 1 and Fig. 3) showed that recombinant strains with higher content of accumulated PHB were linked to higher growth rates and higher maximum ODs, which is due to the light scattering effect of PHB granules.All recombinant strains of R. rubrum ΔphaC1ΔphaC2 showed significantly decreased growth rates and maximum ODs when compared to the wild type carrying an empty vector.As the reductive step in PHB-synthesis is also known to act as an electron sink, the reduced ability to synthesize PHB might also lead to an imbalanced level of reducing agents, resulting in slower growth of the cells 44 .Our results are in line with alternative phaC deletion mutants of R. rubrum generated by Jin and Nikolau (2012) 33 , which likewise showed that the deletion of phaC2 resulted in decreased growth and strongly reduced PHB accumulation.
The characteristic of recombinant PHA accumulation by each synthase is reflected in both mutants.In most cases, the heterologously expressed phaC genes, which led to the highest amount of accumulated PHB (wt/wt of the CDW) in the single deletion mutant R. rubrum ΔphaC2, also resulted in the highest PHB synthesis in the double deletion mutant R. rubrum ΔphaC1ΔphaC2.However, PHB content in recombinant strains of the double mutant was generally lower than in the wild type (see Fig. 2 and Fig. 4).The hetero logous expression of the PHA synthase gene of Chromo bacterium sp., phaC C. sp., showed unexpected low PHB content in both mutants.This PHA synthase was studied intensively in previous studies 45,46 .In these studies, the heterologous expression of pBBR1MCS-2::phaC C. sp in the PHB negative strain R. eutropha PHB -4 resulted in a maximum PHB accumulation, and in vivo enzyme activity tests showed a high activity of 2462 ± 80 U g -1 .This was about 8-fold higher than the PHA synthase activity from R. eutropha H16 46 .Additionally, the heterologous expression of phaC C. sp. in R. ru brum ΔphaC2 surprisingly showed a significantly higher PHB content in the early stationary growth phase than in the late stationary growth phase in our study (see Fig. 2).This is unusual, since PHB accumulation usually reaches its maximum in the late stationary growth phase 5 .The reasons for these results with phaC C. sp. in R. rubrum remain unclear.Heterologous expression of class IV PHA synthases has been studied intensively in the past 22,47,48 .In our study, class IV PHA synthases of the PHB accumulating strains Bacillus megaterium and Bacillus cereus were investigated.In addition to the he terologous expression of phaC B. megaterium ::phaR B. megaterium and phaC B. cereus ::phaR B. cereus , newly combined hybrid PHA synthases were generated.For this, the phaR genes were replaced and two further active PHA synthases encoded by phaC B. megaterium ::phaR B. cereus and phaC B. cereus ::phaR B. megaterium were generated.Similar hybrid vectors with shuffled phaC and phaR genes were heterologously expressed in E. coli 48 .Up to 80 % (wt/wt) PHB of the CDW were synthesized by the B. cereus PHA synthase, while the PHA synthase of B. megaterium showed less PHB accumulation of 26 % (wt/wt) of the CDW.Coexpression of phaC B. megaterium and phaR B. cereus resulted in a PHB accumulation of 39 % (wt/wt).Generated hybrid synthases with rearranged phaR and phaC genes resulted in significant differences in PHB content and molecular weight 48 .In our work, the heterologous expression showed a significant PHB content in each generated recombinant strain, however, the expression of phaC B. megaterium ::phaR B. cereus resulted in a reduced PHB content of 4.8 and 1.3 % (wt/wt) of the CDW in the single mutant and double mutant compared to the co-expression in E. coli 48 .Possibly, an impaired interaction of the subunits PhaC B. megaterium and PhaR B. cereus might be responsible for the observed reduced PHB content and cell growth.However, an experimentally proven ex planation for the phenotype of R. rubrum pBBR1MCS-2::phaC B. megaterium ::phaR B. cereus is not yet available.
In summary, we engineered different strains of the metabolically versatile PHB-producing bacterium R. rubrum to heterologously express a number of twelve different phaC-Genes, thereby increasing the amount of accumulated PHB by up to 25 % compared to the wild type strain.Our study proves the PHA synthase to be a significant bottleneck in the biosynthesis of PHB, as well as evaluates the potential of several promising PHA synthases for application in PHB production in recombinant microorganisms for the first time in a single host strain.

ACKNOWLEDGMENTS
The research leading to these results has re ceived funding from the European Union's Seventh Framework Programme for research, technological development and demonstration under grant agree ment n° 311815, which is gratefully acknowledged.

F i g . 4 -
PHB determinations of all recombinant strains of the double mutant R. rubrum ΔphaC1ΔphaC2.The maximum PHB content in % (wt/wt) of the cell dry weight (CDW) is shown as mean value of all tested samples.Striped bars indicate the PHB content after 27 h of cultivation and grey bars indicate the PHB content after 50 h of cultivation.The strains R. rubrum S1 + MCS-2 (-) and R. rubrum ΔphaC2 + MCS-2 (-) were cultivated as reference strains.MCS-2 = pBBR1MCS-2; MCS-2 (-) = pBBR1MCS-2 without insert.Allrecombinant strains were examined at identical conditions in two independent biological experiments.In addition, each strain was cultivated in duplicate in each experiment.There fore, each strain was tested for PHBaccumulation four times after 27 h and 50 h of cultivation.MCS-2 = pBBR1MCS-2; * n.d.= not detectable.

L
i s t o f A b b r e v i a t i o n s a n d S y m b o l s *n.d.-not detectable ∆ -delta, deletion °C -Degrees Celsius µ -micro growth rate, h -1 µL -microliter A. -Alcaligenes ATCC -American type-and culture collection B. -Bacillus B. -Burkholderia bp -base pair C -carbon C. -Chromobacterium CDW -cell dry weight CO -carbon monoxide CO 2 -carbon dioxide CoA -Coenzyme A C x -carbon atom number x D. -Delftia DNA -deoxyribonucleic acid E. -Escherichia Fig. R. -Rhodospirillum, Ralstonia rpm -revolutions per minute SSFN -Supplemented Succinate Fructose Nitrogen Syngas -Synthesis gas TBE -Tris-Bora-EDTA td -doubling time V -volt vol/vol -volume per volume wt/vol -weight per volume wt/wt -weight per weight 22Medium chain length hydroxyalkanoates (HA MCL ) with a carbon backbone of five to fourteen carbon atoms are polymerized by PHA synthases of class II.Representative bacteria of class II PHA synthases are pseudomonads20.PHA synthases of class III and IV consist of two different subunits.Class III PHA synthases consist of two subunits which are encoded by the genes phaC and phaE.Enzymes of this class are known from Allochro matium vinosum and cyanobacteria.As precursors, substrate HA SCL and HA MCL are accepted21.Class IV PHA synthases are based on the subunits PhaC and PhaR.Until now, this class has been found only in Bacillus strains, such as Bacillus megaterium or Ba cillus cereus22.PHA synthases of class IV are described to polymerize HA SCL 19.
Ta b l e 1 -Evaluated phaC genes for heterologous expression in R. rubrum strains.Donor strain, PHAsynthase class, annotation, gene size and references dealing with the respective phaC gene are mentioned.