Rational flux-tuning of Halomonas bluephagenesis for co-production of bioplastic PHB and ectoine

Ectoine, a compatible solute synthesized by many halophiles for hypersalinity resistance, has been successfully produced by metabolically engineered Halomonas bluephagenesis, which is a bioplastic poly(3-hydroxybutyrate) producer allowing open unsterile and continuous conditions. Here we report a de novo synthesis pathway for ectoine constructed into the chromosome of H. bluephagenesis utilizing two inducible systems, which serve to fine-tune the transcription levels of three clusters related to ectoine synthesis, including ectABC, lysC and asd based on a GFP-mediated transcriptional tuning approach. Combined with bypasses deletion, the resulting recombinant H. bluephagenesis TD-ADEL-58 is able to produce 28 g L−1 ectoine during a 28 h fed-batch growth process. Co-production of ectoine and PHB is achieved to 8 g L−1 ectoine and 32 g L−1 dry cell mass containing 75% PHB after a 44 h growth. H. bluephagenesis demonstrates to be a suitable co-production chassis for polyhydroxyalkanoates and non-polymer chemicals such as ectoine.

H. bluephagenesis could be a suitable host to engineer for ectoine production due to its fast growth and high PHB accumulation under high-salt conditions. So far H. bluephagenesis has been manipulated via various approaches including CRISPR/ Cas9 34 , stimulus response-based flux-tuning (SR-FT) 38 , genomic integrations, inducible promoter mining 35 , and promoter engineering 39 . These concreted efforts enable rational strain engineering of complexity for synthesizing diverse metabolic targets, strongly expanding the capability of H. bluephagenesis as a versatile production chassis. Generally, static optimization is a widely used strategy to fine-tune the pathways involved in multi-gene regulation largely dependent on high throughput constructs and machine learning-based analysis [40][41][42] , leading to high cost of clone construction with a long R&D cycle.
This study aims to chromosomally engineer H. bluephagenesis for enhanced production of ectoine from glucose using "stimulus response-based flux-tuning" method ( Fig. 1). Specifically, the ectoine titer reaches 6.3 and 28 g L −1 during a 48 h growth in 500 mL shake flask and 28 h fed-batch cultivation in 7-L bioreactor, respectively, by chromosomally engineered H. bluephagenesis TD-ADEL-58 harboring fine-tuned ectoine synthesis pathway. At the same time, co-production of ectoine and PHB is also studied for possible economic improvement [43][44][45] , which results in 8 g L −1 ectoine and 32 g L −1 dry cell mass (DCM) containing 75% PHB after 44 h fed-batch fermentation.

Results
Identification of ectoine synthesis pathway. Previous studies reported that intracellular accumulation of ectoine plays the main role in hypersalinity resistance in most halophilic bacteria 46 , of which H. bluephagenesis is the one engineered as the chassis for NGIB 47 . As a metabolic target and compatible solute, cell growth characterization was performed to study the impact between ectoine and NaCl in the presence of various ectoine and NaCl concentrations, respectively (Fig. 2a). Interestingly, ectoine provides tolerance to H. bluephagenesis grown under high-and lowsalt conditions, and minor effects were observed in a 60MMG medium regardless of ectoine concentrations ( Fig. 2a and Supplementary Fig. 1).
It is important to note that the de novo synthesis pathway of ectoine has been identified, it is highly conservative across species 6 (Fig. 1). Based on the genomic analysis of H. bluephagenesis, the ectABC gene cluster encoding the major enzymes for ectoine synthesis ( Fig. 1) was annotated by several databases including RAST and KEGG. Particularly, the nucleotide blast results from NCBI database show that genes ectA, ectB, and ectC of H. bluephagenesis shared 86.46%, 86.05%, and 80.30% identity, respectively, compared with the ones from H. elongata DSM2581. Two strategies were employed to verify the functionality of ectABC cluster: (I) prototyping ectoine production by the start host, H. bluephagenesis TD1.0; (II) heterologous over-expressing ectABC for ectoine production by recombinant E. coli 22 .
Liquid chromatography-tandem mass spectrometry (LC-MS/ MS) was applied to assay the accumulation of ectoine by H. bluephagenesis TD1.0 cultivated in a 60MM medium supplemented with 30 g L −1 glucose in shake flask studies. Results showed that the biosynthesis ectoine shared the same retention time of mass spectra with ectoine standard, which is consistent with the secondary analysis results of LC-MS/MS ( Supplementary  Fig. 2). Furthermore, high performance liquid chromatography (HPLC) was used to quantify the initial production of ectoine by H. bluephagenesis TD1.0 after cell lysis, as most halophiles accumulate ectoine intracellularly for resisting high salt circumstances ( Supplementary Fig. 3). As a result, a production titer of 0.63 g L −1 ectoine was obtained from 9 g L −1 DCM containing 62 wt% PHB (Fig. 2b), confirming the function of ectABC cluster in H. bluephagenesis. Secondly, ectABC from H. bluephagenesis was cloned into the medium-copy-number plasmid pSEVA321, for heterogenous expression by recombinant E. coli DH5α under the control of P tac promoter induced by IPTG (Fig. 2c). The titer of ectoine varied from 2 to 4 g L −1 in the presence of different IPTG concentrations for 48 h cultivation in the MME medium ( Fig. 2d), further demonstrating the function of ectABC in H. bluephagenesis. Thus, more efforts would be conducted to improve the production yield of ectoine by employing various metabolic engineering approaches (Fig. 1).
Tuning the ectABC expression for enhanced ectoine synthesis. H. bluephagenesis was generally engineered for diverse PHA productions requiring a nitrogen limited medium, or a 60MM medium supplemented with 30 g L −1 glucose for shake flask cultivations. As ectoine is an aspartate derived compound heavily depending on both carbon-and nitrogen-catabolism, it is necessary to reinforce the nitrogen supply via a medium modification. Even more important, transcriptional tuning of ectABC cluster via adding L-aspartate as precursor is significant for channeling more fluxes to ectoine, while avoiding the possibility of endogenous negative feedback control on the ectABC expression.
Based on the above understanding, a T7-like induced system was constructed to overexpress ectABC on a plasmid-based system in recombinant H. bluephagenesis TD1.0 in the presence of various concentrations of IPTG (Fig. 3a) 35 . Specifically, 100 mM aspartate was added in the cultural medium, namely MME (see methods), to ensure the metabolic fluxes flowing towards ectoine synthesis. As a result, the optimal transcriptional level for ectABC expression was determined after the induction of 20 mg L −1 IPTG, the maximum production titer of ectoine reached 2.38 g L −1 (Fig. 3b), indicating that a medium optimization should be conducted in the presence of 20 mg L −1 IPTG in shake flask studies. Therefore, factorial analysis was performed to search the optimized concentration of NaCl in the MME media for enhanced ectoine production by recombinant H. bluephagenesis harboring ectABC expression module induced by 20 mg L −1 IPTG. The 60MME medium containing 60 g L −1 NaCl demonstrated the highest 2.82 g L −1 ectoine when cell growth and ectoine synthesis was balanced (Fig. 3c). The ectoine titer was comparable at a higher salt concentration, but the cell growth was inhibited. In contrast, a low NaCl concentration also significant decreases the production titer of ectoine (Fig. 3c).
However, the use of aspartate as a nitrogen source is not suitable for cost-effective industrial ectoine production due to its high cost. Therefore, urea, a low-cost nitrogen source replaced aspartate for de novo synthesis of ectoine (Fig. 1). On the basis of ectABC flux-tuning and NaCl concentration modification, a urea optimized concentration was found to be at 3 g L −1 , and was added to a 60MMG medium, termed 60 MMU (see methods). This resulted in a 2.83 g L −1 ectoine produced by recombinant H. bluephagenesis harboring ectABC expression module induced by 20 mg L −1 IPTG in shake flasks (Fig. 3d), which is close to that cultured in 60MME medium. Addition of urea had a negligible impact on cell growth, yet significantly decrease the PHB content when urea exceeded 1 g L −1 . This info is an important hint for co-production of PHB and ectoine in this study.
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-17223-3 ARTICLE to reduce ectoine accumulation. It is thus necessary to eliminate these two branches for blocking the degradation of ectoine by deleting doeA and ectD (Fig. 1). A CRISPR/Cas9 mediated gene editing tool was employed to generate single-knockout H. bluephagenesis TD-A (ΔdoeA), and double-knockout H. bluephagenesis TD-AD (ΔdoeA, ΔectD), respectively (Fig. 4a). Deletion of doeA and ectD were confirmed based on colony PCR using two primer designs (Fig. 4a). Sequences of sgRNA designed for gene deletions were listed in Supplementary Table 2, respectively. Subsequently, ectoine production by H. bluephagenesis TD01 derived strains H. bluephagenesis TD1.0, H. bluephagenesis TD-A and H. bluephagenesis TD-AD harboring ectABC expression module, respectively, induced using 20 mg L −1 IPTG were conducted in shake flasks. Recombinant cells were cultured in 60MME (Fig. 4b) and 60MMU medium (Fig. 4c), respectively. Results showed that doeA deletion led to a 35% increase on ectoine titer, reaching up to 3.1 g L −1 , and further deletion on ectD increased ectoine to only 3.2 g L −1 (Fig. 4c) when cultured in the 60MMU medium. Interestingly, only 5% and 15% increases were observed by H. bluephagenesis TD-A and H. bluephagenesis TD-AD, respectively, when cultured in the 60MME medium (Fig. 4b), probably attributed to the rich aspartate dosage. Both mutants revealed that deletions of doeA and ectD had little effect on their growth ( Supplementary  Fig. 4), this is helpful for scale-up studies on co-production of PHB and ectoine.
Chromosomally transcriptional tuning of ectABC. Chromosome integration of target expression module(s) allows robust performance of cell growth under antibiotic-free cultivations. Therefore, a GFP-mediated transcriptional mapping approach developed for fine-tuning target genes in the previous study 38 was employed to screen an appropriate constitutive promoter for chromosomal expression of the ectABC cluster (Fig. 5a). Specifically, fluorescent protein sfGFP serving as a global reporter for expression levels was applied to quantify and bridge the correlations cross various promoters and expression systems for efficiently screening the suitable promoter used to transcribe target expression modules, such as ectABC, respectively 38 ( Supplementary Fig. 5). In this study, the T7-like induced system was characterized again by expressing sfgfp 35 (Fig. 6b). Simultaneously, a porin promoter library was constructed and characterized by expressing sfgfp as described in our previous study 39 .
Since the optimized expression strength of ectABC depends on the T7-like induced system, such as the optimized 20 mg L −1 IPTG induction for the T7-like system, yet IPTG is an expensive chemical not suitable for large scale applications. Thus,  Table 3), was selected to express ectABC module on the chromosome of H. bluephagenesis TD-AD (G4 loci) (Fig. 5a), forming H. bluephagenesis TD-ADE strain. However, the ectABC controlled by a stronger promoter with a fluorescence intensity value greater than 50,000 to achieve a high expression had failed to construct in E. coli (also for conjugation purpose). Since the expression levels of target modules on plasmid pSEVA321 are reported to be 9-fold higher than that of the chromosomal expression, which can easily lead to poor-or non-growth due to the overload of target fluxes in donor cells (E. coli S17-1) 38 . Ectoine production by ectABC fine-tuned H. bluephagenesis TD-ADE was carried out in shake flasks containing both 60MME and 60MMU media, forming 4.7 g L −1 and 4.3 g L −1 ectoine, respectively ( Fig. 5b and c). In contrast, H. bluephagenesis TD1.0 harboring ectABC expression module under T7-like promoter encoded on a plasmid-based expression system used as positive control termed H. bluephagenesis TD-AD-E, produced approximately 4 g L −1 ectoine, significantly lower than 4.6 g L −1 ectoine produced by the H. bluephagenesis TD-ADE containing the chromosome-based ectABC system under the control of porin promoter mutant P 140 , yet much higher than the control H. bluephagenesis TD-AD strain forming only less than 2 g L −1 ectoine (Fig. 5b, c).
Fine-tuning of lysC and asd in genome of H. bluephagenesis. H. bluephagenesis TD-ADE containing the chromosome-based and fine-tuned ectABC system under the control of porin promoter mutant P 140 , has enhanced the production of ectoine (Fig. 5). It is expected that further flux-tuning of asd and lysC genes encoding L-aspartate-semialdehyde-dehydrogenase and aspartokinase, respectively, could channel more metabolic fluxes toward ectoine synthesis ( Fig. 1) 16,49 . Therefore, gene lysC of Corynebacterium glutamicum and asd of H. bluephagenesis were overexpressed in H. bluephagenesis TD-ADE, followed by transcription tuning to obtain better performance on ectoine accumulation.
Two inducible promoters are required including T7-like promoter to modulate the expression of lysC and asd genes, simultaneously and respectively. In this case, a quorum sensingbased regulator LuxR encoded by a luxR triggering the transcription of the corresponding promoter P lux in the presence of acyl homoserine lactone (AHL) was constructed on a plasmidbased expression system controlled by the constitutive promoter P J23110 50 , together with sfgfp controlled by P lux , for the doesresponse functional characterization in H. bluephagenesis TD-ADE using flow cytometer analysis ( Fig. 6a and Supplementary  Fig. 6). Results exhibited 86-folds dynamic range of fluorescence intensity (FI) of sfGFP in the presence of various AHL concentrations with low basal leakiness, FI = 157 ( Supplementary  Fig. 6). Subsequently, the expression module of P J23110 -luxR was integrated in G43 loci in H. bluephagenesis TD-ADE, forming H.    bluephagenesis TD-LuxR strain (Fig. 6a). 275-and 1843-fold dynamic ranges were obtained from H. bluephagenesis TD-LuxR under the LuxR-AHL and T7-like induced systems, respectively, the basal leakiness of chromosome-based LuxR-AHL induced system decreased to 10 FI value (Fig. 6b). Thus, two orthogonal induced systems were successfully constructed in H. bluephagenesis enabling double-module modulation (Fig. 6c). Ideally, expression of lysC and asd would be induced by IPTG and AHL, respectively, for enhanced ectoine production by consuming more urea. Before that, it is necessary to optimize urea concentration in the culture medium again. Addition of 5 g L −1 of urea to the 60MMG medium (namely 60MMU5) generated 17% improvement on ectoine accumulation although cell growth decreased significantly (Fig. 6d). Here, two expression modules, namely, asd and lysC controlled by P lux and P T7-like , respectively, were constructed on a plasmid-based system for transcriptional tuning (Fig. 6e). The induction concentrations of IPTG and AHL did have a significant effect on ectoine production. For example, an induction using 20 mg L −1 IPTG and 10 mg L −1 AHL generated the highest 6.6 g L −1 ectoine, 29% higher compared to that produced by H. bluephagenesis TD-ADE (Fig. 6d, f, Supplementary  Fig. 7). Using the GFP-mediated transcription-tuning approach, porin promoter mutants P 226 and P 140 were selected to control the expression of lysC, and P 58 and P 183 for asd on the chromosome of H. bluephagenesis TD-ADE. However, the expression module encoding lysC controlled by P 140 failed to construct in the donor E. coli as also mentioned previously. Yet, two strains, H. bluephagenesis TD-ADEL-58 and TD-ADEL-183, were successfully grown for shake flask studies using the 60MMU5 medium (Fig. 6g). The two strains produced almost 6.3 g L −1 ectoine after 48 h cultivation in the 60MMU5 medium, over 30% higher compared to that cultured in the 60MMU medium. Due to the failure of P 140 -lysC construction, a weaker transcription strength of lysC controlled by P 226 probably weakens the flux for L-β-aspartyl phosphate synthesis, which coupled with asd controlled by P 183 , it cannot bring significant increase on ectoine when grown in both 60MMU and 60MMU5 media, respectively. Thus, H. bluephagenesis TD-ADEL-58 became a better option for fed-batch studies in 7 L lab-scale bioreactor. Similarly, considering the fed-batch study for coproduction of PHB and ectoine, H. bluephagenesis TD-ADE with genomic expression of lysC and asd genes were performed in a 96-well plate to investigate the effect on growth scale-up ( Supplementary Fig. 4).
Fed-batch study for co-production of PHB and ectoine. H. bluephagenesis TD-ADEL-58 was grown for 28 h for ectoine production under open unsterile conditions in a 7 L bioreactor. 28 g L −1 ectoine were produced from this fed-batch study, which is over 4-folds higher than that of the shake flask results (Fig. 7a). The yield of ectoine on glucose and volumetric productivity reached approximately 0.21 g g −1 and 1.0 g L −1 h −1 , respectively. Ectoine accumulation tendency was highly consistent with cell growth in terms of DCM, indicating the growth-coupled property of ectoine biosynthesis. Meanwhile, PHB accumulation was maintained at a low level during the first 16 h of cultivation before its increase afterward (Fig. 7a). In contrast, a fed-batch fermentation was executed for 28 h using H. bluephagenesis TD-ADE to verify the effect of fine-tuning asd and lysC gene on ectoine production, leading to 17 g L −1 ectoine, which was over 30% lower than results of H. bluephagenesis TD-ADEL-58 ( Supplementary Fig. 8).
To achieve co-production of ectoine and PHB by H. bluephagenesis TD-ADEL-58, a tailor-made feeding solution was designed to enhance PHB accumulation during the stationary phase using reduced nitrogen dosage in the second phase feeding solution (Feed-II). Similarly, a fed-batch fermentation was carried out in the 7 L bioreactor to co-produce ectoine and PHB in different growth phases. After a 44 h cultivation, the DCM reached 32 g L −1 containing 75% PHB, together with 8 g L −1 ectoine (Fig. 7b). It was interesting to observe that ectoine biosynthesis maintained active in the stationary phase during the fed-batch cultivation (Fig. 7b). It is expected that more bioprocess optimizations should lead to even higher ectoine and PHB production under balancing nitrogen supply in different growth phases 51,52 .

Discussion
Next generation industrial biotechnology (NGIB) based on halophiles provides low-cost and convenient processing conditions for production of chemicals 47 . The most successful case has been exemplified by 5000 L scale fermentor for PHA production 27 . It has been interesting to look for other examples to test possibilities of NGIB for production of non-polymer products, or even high-value added example. Ectoine, a high-value added product derived from L-aspartate synthesized by some microorganisms, has attracted increasing attentions due to its outstanding functions of stabilizing proteins and nucleic acids 53 . To date, many ectoine commercial applications, such as cosmetic ingredient, have been developed with high-value. However, ectoine production by wild type and engineered microorganisms yields very low productivity. Some metabolic engineering approaches have been adopted for enhanced productions of amino acids-related compounds, including ectoine 54 , L-threonine and L-lysine 37 , including new tools for strain engineering, such as dynamic and static optimization for flux enforcement 55 : stimulus response-based transcriptional controls 56,57 and promoter and RBS optimizations et al 40 . H. bluephagenesis, a high-salt tolerant halophile able to grow to a high cell density with scalability, is a promising host for ectoine synthesis because of its being a natural ectoine producer harboring the ectABC cluster in its genome. Recently, many efforts have been made to establish various genetic manipulation tools and flux-tuning approaches for improving and diversifying this NGIB chassis 35,38 . This study has demonstrated successful and high yield on ectoine production by chromosomally engineered H. bluephagenesis utilizing static optimization strategies, reaching close to 30 g L −1 ectoine from a fed-batch fermentation process (Fig. 7a).
Ectoine biosynthesis pathway of H. bluephagenesis was clearly identified by heterologous expression in recombinant E. coli, in which strong enzymatic activities of ectABC cluster were found compared with previous studies 14,22 (Figs. 1, 3, and 4). Cell growth studies uncovered the enhanced resistance of cells under both low-and high-salt concentrations in the presence of ectoine, respectively. Furthermore, competitive ectoine degradation pathways were removed by eliminating the related genes doeA and ectD 48 , resulted in 89% improvement on ectoine production from shake flask cultivations. Two pathway branches reported by previous studies involved in L-threonine and L-lysine synthesis were retained in consideration for the benefits of ectoine coproduction with PHB 43,51 , a growth-dependent intracellular product, as the disruption of L-threonine and L-lysine will weaken cell growth and thus PHB accumulation. We thus attempted to utilize transcriptional balancing strategies to boost fluxes towards ectoine synthesis. In contrast to translational regulation, transcriptional regulation could be more efficient to be carried out with minor constructs during strain engineering by onestep fine-tuning of target expression modules based on stimulus response-based flux-tuning design 38 . Five genes involved in ectoine synthesis pathway were grouped into three individual expression units, including ectABC, lysC and asd, for transcriptional fine-tuning (Figs. 1, 4). It has to be noted that quorum-sensing-based inducible system, LuxR-AHL, was firstly constructed and employed to achieve combinatory transcriptional tuning together with T7-like system in H. bluephagenesis (Figs. 5, 6). Specifically, expressions of two simple constructs, p321-P J23110 -lacI-P T7-like -ectABC and p321-P T7-like -lysC-P lux -asd, could simulate countless transcriptional combinations of target modules replacing thousands of variants construction. Finally, 9.8-folds improvement of production titer of ectoine was achieved in a shake flask study by recombinant H. bluephagenesis harboring fine-tuned synthesis pathway of ectoine in genome compared with 0.64 g L −1 titer yielded by the start host (Fig. 6). Subsequently, an ectoine titer of approximately 30 g L −1 was achieved under open unsterile fed-batch fermentation conducted in the 7-L bioreactor.
Co-production of PHB and ectoine allows harvests of intracellular bioplastic PHB and extracellular small molecular chemical ectoine, it is thus economically interesting. However, it is not sure if the two products competing for metabolic fluxes could reduce their yields. A conceptional proven study was conducted in a lab-scale fermentation. Both ectoine and PHB are growthassociated products derived from aspartate and acetyl-CoA when nitrogen source supply is abundant, approximately 28 g L −1 ectoine and 25 g L −1 DCM together with 6% PHB were obtained, respectively, during a 28 h of growth (Fig. 7a). In contrast, under a reduced nitrogen source, PHB accumulation was increased to 75% of the DCM in the cells producing also 8 g L −1 ectoine (Fig. 7b). In both cases, the recombinant H. bluephagenesis was grown to over 30 g L −1 DCM that are not the highest reported by this organism 27,58 (Fig. 7). Further efforts to increase the recombinants' growth density should lead to much improved ectoine and PHB productions.
Etoine as a representative product of L-aspartate compound family with a high value, should be able to produce at a competitive cost after fine-tuning of every single gene of ectABC cluster employing SR-FT approach by developing more suitable inducible systems 59 . Although static optimization is a widely used strategy for enhanced production of metabolic targets, recent progresses focused on bifunctional dynamic control provide a strong hint for improving the production of ectoine. Since co-production of PHB and ectoine seems to be competing for flux, the efficiency to activate target pathways dynamically for PHB and ectoine must be considered during the fermentation processes 60,61 . In addition, the bioprocessing optimization, such as feeding solution optimization 62 and cascaded bioprocessing 52 , is a promising strategy to obtain simultaneously enhanced production yields for ectoine and PHB. Additionally, attempts of enhancing export system 48 and developing amino acids responsive regulators 57 could also be applied to achieve further improvement on ectoine and PHB production under cost-effective bio-manufactural process based on the NGIB chassis.
In conclusion, the halophile NGIB chassis can be effectively used for effective production of ectoine. Engineering the ectoine pathways improved ectoine production effectively. Co-production of intracellular PHA and non-polymer chemicals is possible using the halophilic chassis.

Methods
Strains culture and chemicals. LB medium is composed of (g L −1 ): 10 trypton, 5 yeast extract and 10 NaCl. 60LB medium is derived from LB medium supplemented with 60 g L −1 NaCl. 60MMG medium consists of (g L −1 ): glucose 30, NaCl 60, yeast extract 1, (NH 4 ) 2 SO 4 0.25, MgSO 4 0.2, Na 2 HPO 4 ·12H 2 O 9.65, KH 2 PO 4 1.5, trace element solution I 10 mL L −1 and trace element solution II 1 mL L −1 . The composition of trace element solution I consists of (g L −1 ): Fe(III)-NH 4 O 30. 60MME is derived from 60MMG medium supplemented with 100 mM sodium aspartate with additional 100 mM KCl, and 1.7 g L −1 citric acid. Alternatively, 100 mM sodium aspartate in 60MME medium was replaced by 3 and 5 g L −1 , respectively, to form 60MMU and 60MMU5 media for de novo synthesis of ectoine for shake flask studies. Ectoine of analytical purity purchased from Sigma-Aldrich (St. Louis, MO, USA) was used as standard for HPLC analysis for fermentative ectoine quantitation.
H. bluephagenesis TD1.0 derived from H. bluephagenesis TD01, isolated from Aydingol Lake of China, was used to engineer for production of ectoine in this study. Escherichia coli S17-1pir, the donor strain for conjugation, was cultivated in LB medium. Seed preparation of H. bluephagenesis and its derivates in this study were cultured in 60LB medium. Co-production of ectoine and PHA in shake flasks by recombinant H. bluephagenesis TD were cultivated in 60MMG, 60MME, 60MMU and 60MMU5 media, respectively.
All the derivates from H. bluephagenesis TD1.0, wild type H. bluephagenesis TD integrated with expression module of MmP1 RNA polymerase induced by IPTG, were listed in Supplementary Table 1. Firstly, doeA and ectD gene encoding bypass (ectoine degradation) enzymes were deleted, forming H. bluephagenesis TD-AD strain. Then, expression modules including ectABC, lysC and asd cluster were integrated into the chromosome of H. bluephagenesis one-by-one at G4, G7 and G43 locus, respectively, forming H. bluephagenesis strains TD-ADE, TD-ADEL-58, and TD-ADEL-183.
Plasmids construction. Plasmids were constructed to carry functional gene expressions encoded in vector pSEVA321 63 . All the constructs were assembled by Gibson Assembly kit, and conjugated into H. bluephagenesis derivates. DNA extraction was carried out by using Tiangen Plasmid Extraction kit (Shanghai, China) and Omega Gel Purification kit (USA). DNA oligonucleotides were synthesized by Zixi Biotechnology (Beijing, China). Genetic design and sequence reading were performed by Snap Gene (v2.3.2).
All the plasmids were constructed by Gibson Assembly kit. For gel purification, the Omega gel purification kit was used, and the Tiangen plasmid extraction kit was used to obtain the plasmids. All the plasmids used in this study and genes required fine-tuning to obtain enhanced ectoine production were listed in Supplementary Table 1. The sequences of sgRNA used for genome editing in the study were listed in Supplementary Table 2. CRSPR-based genomic editing approach was employed to construct various recombinant strains by gene knockout and gene integration 34 . Firstly, plasmid pQ08 containing Cas9 expression module was transformed into recombinant H. bluephagenesis by conjugation. Followed by the conjugation of constructs derived from pSEVA241 vector carrying gRNA and homologous arms (500 bp), positive colonies were screened by PCR test and confirmed by DNA sequencing after the discard of plasmids by passaging the cell cultures in 60 LB medium without antibiotics for 3-5 generations.
Conjugation. Firstly, plasmids were constructed by E. coli DH5α, subsequently transformed into E. coli S17-1, the donor strain for conjugation. Then, donor cells harboring target constructs were cultivated for 12-14 h in 150-mL shake flasks containing 20 mL LB medium, and recipient cells, recombinant H. bluephagenesis, were cultivated in 60LB medium overnight. Both of the donor and recipient cells were centrifugated and harvested under 4 o C, washed with a fresh medium and recovering to the same volume that sampled. Subsequently, cells were mixed (1:1 volume) and dropped on antibiotic-free 20-LB agar plates, followed by 6-8 h incubation at 37 o C for conjugation. Finally, conjugated cells from the lawn were suspended with 60LB medium and spread on 60LB agar plates containing relevant antibiotics for 36 h incubation to obtain positive colonies.
Positive fluorescent cells were recorded by flow cytometer (LSRFortessa4, BD bioscience, USA) at the rate of 0.5 µL s −1 for 20 s with at least 30,000 cell captured in total of each sample, and the start host, H. bluephagenesis TD1.0, was used as the negative control groups to modify the fluorescence levels of experimental groups. Intracellular sfGFP protein was excited under 488 nm, and cells were captured on the signal channels of FITC, FSC (forward scatter) and SSC (side scatter). The FlowJo (v7.6) software was used to process the raw data for obtaining mean value of fluorescence intensity and percentage of positive cells.
Cell growth study. To study the effects of ectoine and NaCl on cell growth, single colony of recombinant H. bluephagenesis TD1.0 was first picked as an inoculum for a 12 h incubation in 1 mL 60LB medium in a 96-deep-well plate (Corning/USA, 1 mL 60LB medium) at 37°C. Subsequently, seed cultures were diluted 200-folds with fresh MM medium supplemented with 20 g L −1 in the presence of various ectoine and NaCl concentrations in a new 96-well plate (flat bottom, 250 µL medium per well) for online recording of optical density at 600 nm (OD 600 ) for 48 h.
To characterize cell growth of recombinant H. bluephagenesis in this study, chromosomally engineered strains were pre-cultured in a 60-LB medium in the 96deep-well plate for 12 h by picking a single colony as an inoculum. Then, seed cultures were diluted 200-folds with a fresh 60LB medium in a new 96-well plate (flat bottom, 250 µL medium per well) for online monitoring of optical density at 600 nm (OD 600 ) for 24 h. The cell growth characterization was carried out using a micro-plate reader (Varioskan Flash, Thermo Scientific).
Shake flask and fed-batch studies. For shake flask studies, recombinant H. bluephagenesis harboring target expression modules on plasmid-or chromosomebased systems were pre-cultured overnight to obtain seed cultures (OD 600 reached 2.5 ± 0.2) from a single colony, respectively. Subsequently, 2.5 mL seed culture were inoculated into 50 mL 60MM medium supplemented with 30 g L −1 glucose and nitrogen source (urea or sodium aspartate) whenever necessary. 25 µg mL −1 chloramphenicol was added to the medium grown with H. bluephagenesis harboring the plasmid-carried expression systems. Cell cultures were harvested for dry cell mass (DCM), ectoine and PHB content assays after 48 h cultivation. Additionally, functional verification of gene cluster ectABC from H. bluephagenesis was performed using the same procedure of recombinant H. bluephagenesis grown in LB medium and recombinant E. coli DH5α harboring plasmid p321-J 23110 -lacI-Ptac-ectABC.
For fed-batch studies, cells were grown overnight in 60LB medium in 500 mL shake flasks to obtain 300 mL seed cultures (OD 600 reached 2.5 ± 0.2) as inoculums in a fed-batch study conducted in a 7-L bioreactor (NBS Bioflo3000, New Brunswick, USA). A 3 L 60 MM medium supplemented with 3 g L −1 CO(NH 2 ) 2 , 10 g L −1 yeast extract and 20 g L −1 glucose was used as an initial medium for fermentation. During the fermentation process, dissolved oxygen (DO%) was maintained at~30% of air saturation by injecting air with a maximum flow rate of 1 VVM (air volume per culture volume per min) by coupling the agitation of less than 800 rpm. pH was adjusted using 5 M NaOH solution. Feed solution of the first phase (Feed-I) is composed of 200 g glucose, 8 g yeast extract and 32 g urea (24 g of urea for co-production purpose). For ectoine production, the second phase feeding solution (Feed-II) consists of 200 g glucose, 4 g yeast extract and 28 g urea, while fed nutrients composed of 200 g glucose, 4 g yeast extract and 8 g urea for coproduction of ectoine and PHB. And Feed-III containing only 800 g L −1 glucose was fed when Feed-II ran out during the fermentation for co-production of ectoine and PHB. Feed solutions were fed to maintain the residual glucose at 6-10 g L −1 (measured by SANNUO medical glucometer, China). All batches of 7-L fermentation were conducted under non-sterilized open condition at 37°C. Then, cells sampled of every time point were harvested for cell mass, ectoine and PHB assays.
Assays of dry cell mass and PHB content. First, cells were harvested from 30 mL fermentative broth during a fed-batch study or at the end of the shake flask study via centrifugation at 12,200g (CR 21GIII, HITACHI, Japan) for 15 min, subsequently suspended and washed twice with distilled water. A dry cell mass was calculated by measuring the mass of harvested cells after 24 h of lyophilization under −60°C and 100 Pa. For PHB content assays, 30-40 mg lyophilized cells in powder form were sampled for methanolysis under 100°C for 3.5 h supplemented with 2 mL chloroform and 2 mL methanolysis solution (85 wt% methanol, 15 wt% H 2 SO 4 , and 1 g L −1 benzoic acid) mixture 48 . Then, 1 mL water was added to the methanolysis solution after cooling down to room temperature (25°C) for extraction and phase separation. 1 mL of the heavy (low) phase was sampled for PHB content calculation using GC (gas chromatography, GC-2014, SHIMADZU, Japan) analysis. 15 and 20 mg pure PHB (3HB standard from Sigma-Aldrich) were used as standards.
Ectoine assays. Five millimeters of fermentative broth from 30 mL cell cultures during fed-batch study or at the end of shake flasks were diluted 10-folds (50-folds for the fed-batch study) in a 50 mL tube. Subsequently, cells were disrupted by a high pressure homogenizer and centrifugated at 12,200g for 10 min. Then, the supernatants were filtrated once by 0.22 μm filter membrane for ectoine quantification using high-performance liquid chromatography (HPLC, SHIMADZU, Japan) equipped with a C18 column based on acetonitrile/water mixture (70:30, v/ v) at a flow rate of 1 mL per min as the mobile phase. UV (ultraviolet) detector (210 nm) was used for ectoine detection. The HPLC-MS analysis was carried out using OrbiTrap Q-Exactive LC-MS (Thermo Scientific) under the same conditions.
Reporting summary. Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Data availability
The datasets generated and analyzed during the current study are available from the corresponding authors upon request. A reporting summary for this Article is available as a Supplementary Information file. The datasets generated and analyzed during the current study are available from the corresponding author upon request. Plasmids sequences are available in Supplementary Data 1-13. Source data are provided with this paper.