Engineering the bacterial shapes for enhanced inclusion bodies accumulation

Metabolic Engineering 29 (2015) 227–237

Contents lists available at ScienceDirect

Metabolic Engineering journal homepage: www.elsevier.com/locate/ymben

Engineering the bacterial shapes for enhanced inclusion bodies accumulation Xiao-Ran Jiang a, Huan Wang a,b, Rui Shen a, Guo-Qiang Chen a,c,n a MOE Key Lab of Bioinformatics, Department of Biological Science and Biotechnology, School of Life Science, Tsinghua-Peking Center for Life Sciences, Tsinghua University, Beijing 100084, China b College of Eco-Environmental Engineering, Qinghai University, Xining 810016, China c Center for Nano and Micro Mechanics, Tsinghua University, Beijing 100084, China

art ic l e i nf o

a b s t r a c t

Article history: Received 9 August 2014 Received in revised form 17 February 2015 Accepted 31 March 2015 Available online 11 April 2015

Many bacteria can accumulate inclusion bodies such as sulfur, polyphosphate, glycogen, proteins or polyhydroxyalkanoates. To exploit bacteria as factories for effective production of inclusion bodies, a larger intracellular space is needed for more inclusion body accumulation. In this study, polyhydroxybutyrate (PHB) was investigated as an inclusion bodies representative to be accumulated by Escherichia coli JM109SG. Various approaches were taken to increase the bacterial cell sizes including deletion on actin-like protein gene mreB, weak expression of mreB in mreB deletion mutant, and weak expression of mreB in mreB deletion mutant under inducible expression of SulA, the inhibitor of division ring protein FtsZ. All of the methods resulted in different levels of increases in bacterial sizes and PHB granules accumulation. Remarkably, an increase of over 100% PHB accumulation was observed in recombinant E. coli overexpressing mreB in an mreB deletion mutant under inducible expression of FtsZ inhibiting protein SulA. The molecular mechanism of enlarged bacterial size was found to be directly relate to weakened cytoskeleton which was the result of broken skeleton helix. & 2015 International Metabolic Engineering Society. Published by Elsevier Inc. All rights reserved.

Keywords: Cell morphology PHB mreB Escherichia coli Cell size Inclusion bodies

1. Introduction Many bacteria are able to accumulate inclusion bodies for various purposes, containing different materials such as elemental sulfur, polyphosphate, glycogen, magnetosomes, triacylglycerols, wax esters, cyanophicin and polyhydroxyalkanoates (PHA) etc. (Alvarez and Steinbuchel, 2002; Carrio and Villaverde, 2002; Garcia-Fruitos et al., 2012; Shively, 2006). Many of these inclusion bodies have various industrial applications (Rehm, 2007; Rodriguez-Carmona and Villaverde, 2010). It has thus become increasingly important to exploit bacteria as microbial factories to produce these inclusion bodies effectively (Chen, 2009; Han et al., 2008; Kalscheuer et al., 2006). Taking microbial polyhydroxyalkanoates (PHA) as an example, PHA have been studied as biodegradable plastics for the past many years with limited market success to date (Chen and Patel, 2012; Gao et al., 2011; Laycock et al., 2013). High cost of PHA has been a key limiting factor (Park et al., 2012; Wang et al., 2014b). Many efforts have been made to reduce PHA production cost such as process optimization, use of low cost substrates (such as glycerol,

n Corresponding author at: School of Life Sciences, Tsinghua University, Beijing 100084, China. Fax: þ86 10 62788784. E-mail address: [email protected] (G.-Q. Chen).

cellulose and starch et al.), increasing the substrate to PHA conversion efficiency, pathway engineering, and most recently synthetic biology approaches (Choi and Lee, 1999a; Koutinas et al., 2014; Steinbüchel and Lütke-Eversloh, 2003; Wang et al., 2014b). All of these efforts have led to different levels of PHA cost reduction (Keshavarz and Roy, 2010; Li et al., 2010). However, the production cost of PHA is still significantly higher than that of the commonly used petrochemical plastics such as polyethylene (PE) or polypropylene (PP) (Meng et al., 2014). More innovations are needed to further reduce PHA production cost further (Wang et al., 2014b). Therefore, PHA in general or its representative polyhydroxybutyrate (PHB), serves as a good example for studying new approaches on enhanced inclusion body production. Since PHA are produced by bacteria as inclusion bodies, the small bacterial cell size limits the amount of PHA granules and the quantity of PHA that can be stored in each cell. At the same time, small size also increases the difficulty of separating cells from fermentation broth due to the very small difference in specific density of bacterial cells and the surrounding fermentation liquid (Serafim et al., 2008; Sudesh et al., 2000). We believe that if the size of bacteria can be increased, allowing more inclusion bodies such as PHA to be accumulated as intracellular granules, inclusion body production and extraction could become more efficient, leading to a significant reduction in production costs. Previous attempts by Wang et al. (2014a) led to limited size expansion

http://dx.doi.org/10.1016/j.ymben.2015.03.017 1096-7176/& 2015 International Metabolic Engineering Society. Published by Elsevier Inc. All rights reserved.

228

X.-R. Jiang et al. / Metabolic Engineering 29 (2015) 227–237

of Escherichia coli, resulting in some improvements on PHA accumulation. However, this size enlargement was not stable over a 48 h of cell growth period. A better way to stabilize large cells must be found. To increase the bacterial cell size through engineering, we must first consider how bacteria maintain their cell shapes. It was found that the bacterial peptidoglycan cell wall and the actin-like protein MreB cytoskeleton are major determinants of cell shape in rodshaped bacteria such as E. coli (Dominguez-Escobar et al., 2011; Garner et al., 2011; Henderson et al., 1997; Kocaoglu and Carlson, 2013). Furthermore, cell elongation specific peptidoglycan synthesizing enzyme complexes are organized by the actin homolog MreB (Rueff et al., 2014). Interestingly, when grown in rich media, mreB is an essential gene (Kruse et al., 2003). On the other hand, mreB is not an essential gene when E. coli cells are grown in a mineral media (Kruse et al., 2003). Deletions on MreB or MreBassociated morphogenetic determinants, such as MreC, MreD or RodZ, led to changing cell shape from rods to spheres (Bendezú et al., 2009; Wachi et al., 1989). Since spherical shapes have the largest ratios of volume to size, we speculated that a change from rod to sphere may increase PHA accumulation in bacterial cells. FtsZ plays an important role in the bacterial cell division process as a tubulin-like protein. Z rings formed by FtsZ assembly, is very dynamic process (Bi and Lutkenhaus, 1993; Loose and Mitchison, 2014; Margolin, 2005). FtsZ interacts with at least ten other proteins in the progression and completion of cytokinesis (Margolin, 2005). Many FtsZ inhibitors can directly interact with FtsZ protein and inhibit cell division in various manners, thus leading to filamentous cells. Examples of these include SulA, Noc, SlmA and MinCDE family (Cho et al., 2011; Dai et al., 1994; Huang et al., 1996). SulA overexpression leads to the reduction on the amount of FtsZ in the division ring and thus disrupts the Z ring, and SulA has no effect on FtsZ ring formation when they are not hydrolyzing GTP (Dajkovic et al., 2008; Higashitani et al., 1995). The inhibition of FtsZ by SulA leads to the formation of filamentary cells (Bi and Lutkenhaus, 1993; Fenton and Gerdes, 2013), which increases the cell size and intracellular space for more PHA accumulation. Thus, sulA gene was chosen to achieve a larger intracellular space. On the other hand, FtsZ overproduction accelerates cell division, resulting in high cell density growth using recombinant E. coli overexpressing ftsZ gene (Choi and Lee, 1999b; Lee, 1994; Wang and Lee, 1997). At the same time, a certain degree of the filamentary E. coli resulted from sulA over-expression could create more space for more intracellular PHA accumulation. Both methods benefit PHA accumulation from two different aspects. Finally, the combination of cell elongation with changing cell shape from rod to sphere should form large spherical cells that have a very large intracellular volume for storages of inclusion bodies such as PHA granules.

2. Materials and methods 2.1. Microorganisms, plasmids and culture conditions Plasmid isolation and DNA purification kits were purchased from Qiagen (Shanghai, China). Restriction enzymes and DNA manipulation enzymes were provided by MBI Fermentas (Vilnius, Lithuania). All synthetic oligonucleotides were obtained from Life Technologies (Carlsbad, USA). All other chemicals in analytical purity were purchased from Sigma Aldrich (St. Louis, MO). All plasmids and oligonucleotides are listed in Table 1 and Supplementary Table S1, respectively. Plasmids were verified by colony PCR, by digestion with restriction enzymes, and by sequencing. All plasmids were constructed using the Gibson assembly method (Gibson et al., 2009). All bacterial strains used in this study are listed in

Table 1. E. coli JM109SG was the wild type for all manipulations (Li et al., 2010), it is a derivative of the a recA deficiency in common E. coli JM109 strains with a recA deficiency to maintain plasmid stability and stable expression of heterologous genes. Overnight cultures were grown in 20 mL Luria-Broth (LB) containing appropriate antibiotics. Antibiotic concentrations were prepared as follows: kanamycin (50 μg/mL), chloramphenicol (25 μg/mL), ampicillin (100 μg/mL). Cultivation was carried out in MM medium consisting of 0.1% (NH4)2SO4, 0.02% MgSO4, 1.0% Na2HPO4  12H2O, 0.15% KH2PO4, 2 g L  1 yeast extract and 20 g L  1 glucose. 2.2. Microbial production of PHB from glucose in shake flasks or fermentor 5% overnight cultures were inoculated in 50 mL MMG in 500 mL flasks. Cells were grown to an OD600 of  0.6 at 30 1C on a rotary shaker (200 rpm), followed by addition of a 0.2% arabinose. The cultures were allowed to grow at 30 1C on a rotary shaker (HNY-2112B, Tianjin Honour Instrument Co., Ltd, Tianjin, China) (200 rpm) for 48 h. 30 mL of culture was taken for analysis after 48 h of cultivation. For fermentor studies, the seed culture was grown at 30 1C in Luria-Bertani medium for 12 h at 200 rpm on the rotary shaker, it was then inoculated into a 6-liter fermentor (NBS 3000, New Brunswick, USA) at 10% inoculation volume with an operating volume of 3 l. The starting fermentation medium was the same as that of shake flask except higher yeast extract concentration (15 g L  1) was added. The pH in the fermentor culture was maintained at 7.0 by automatic addition of 5 M NaOH and 5 M H3PO4. Dissolved oxygen (DO) was provided by injecting filtered air at a flow rate of 3 l min  1 and was maintained at 20% of air saturation by automatically adjusting the agitation rate from 200 to 700 rpm (Li et al., 2010). In all cases, a final concentration of 50 μg/mL kanamycin and 100 μg/mL ampicillin were added to the medium to maintain the plasmid stability of pBHR68 and pTK or pTK-mreB. 2.3. Genetic methods 2.3.1. Gene deletions and constructions of recombinant strains PCR-mediated gene deletions in E. coli JM109SG were performed according to the method reported previously (Datsenko and Wanner, 2000). The isolation and manipulation of recombinant DNA was carried out using standard protocols. E. coli transformation was performed via electroporation. Oligonucleotides used for the generation of gene deletion fragments are shown in Supplementary Table S1, comprising of 57-nt-long homology extensions and 20-nt primer sequences for the template pKD13. PCR fragments were generated containing KanR gene flanked by FLP recognition target (FRT) sites and 57-bp homologous to respective chromosomal sequences adjacent to the target gene. PCR reactions were conducted in 50 μl mixtures containing 2.5 U of TransStarts FastPfu DNA polymerase, 1 pg pKD13 DNA, 2.0 mM of each primer, 1X FastPfu buffer and 400 mM of each of the four deoxyribonucleotide. The mixtures were incubated comprising 95 1C for 5 min, followed by 30 cycles at 95 1C (30 s), 58 1C (30 s), and 72 1C (1 min 30 s); and subsequently incubated at 72 1C for an additional 10 min. Recombinants carrying pKD46RecA were grown in LuriaBertani medium at 30 1C to an OD600 ¼0.4–0.6, then transferred to 37 1C and induced with 0.2% L-arabinose for 1 h. Cells were made electro-competent by concentrating 100-folds and washing twice with ice-cold 10% glycerol. For electroporation, 50 μl of competent cells were mixed with 200 ng of the PCR products, and then electroporated, followed by addition of 1 ml of minimal

X.-R. Jiang et al. / Metabolic Engineering 29 (2015) 227–237

229

Table 1 Strains and plasmids used in this study. Name E. coli strains JM109SG JM109SG (ΔmreB) Plasmid pKD46RecA

Descriptions

References

E. coli JM109 Δsad ΔgabD JM109SGΔmreB::frt

Li et al. (2010) This work

λ-Red recombinase expression helper plasmid, oriR101, repA101(ts), AmpR

Datsenko and Wanner (2000) Datsenko and Wanner (2000) Datsenko and Wanner (2000) Spiekermann et al. (1999) Kuhlman and Cox (2010) This work This work This work

pKD13

Template plasmid with KanR gene and FLP recognition target

pCP20

FLP recombinase helper plasmid, ts-rep, AmpR, CmR

pBHR68

phaCAB expression plasmid, AmpR

pTKRED pTK-mreB pTK pTK-mreB-PBAD:sulA

Cloning vector, SpeR pTKRED derivate constitutive expression endogenous mreB gene, KanR pTKRED derivate blank vector, KanR pTKRED derivative constitutive expression of endogenous mreB gene and arabinose promoter inducible expression of sulA gene, KanR pTK-mreB-PBAD:sulA derivate constitutive expression of mreB and meos3.2 fusion gene

pTK-mreB-meos3.2SWPBAD:sulA p15A-PBAD:sulA p15A- mreB

R

p15A ori, arabinose promoter inducible expression of sulA gene, Cm p15A ori, endogenous mreB gene constitutive expression plasmid, CmR

media containing 0.1% (NH4)2SO4, 0.02% MgSO4, 1.0% Na2HPO4  12H2O, 0.15% KH2PO4, and 4 g L  1 glucose. After incubation at 30 1C for 2 h, cells were spread on selection plates containing 50 μg/mL kanamycin and incubated at 30 1C for 72 h. Positive colonies were selected via PCR verification. Subsequently, the positive colonies were transferred to the Luria-Bertani medium for cultivation at 37 1C. The KanR gene was eliminated using the plasmid pCP20 expressing the FLP recombinase, which acts on the directly repeated FRT sites. The KanR mutants were transformed with pCP20, spread on CmR plates and incubated at 30 1C for 24 h. Colonies were passaged without antibiotics at 42 1C and then tested for loss of all resistance markers. The majority lost the FRTflanked KanR gene and the FLP helper plasmid simultaneously. All gene deletions were confirmed by PCR analysis and sequencing. The plasmid pTK-mreB-PBAD:sulA was transformed into E. coli JM109SG (ΔmreB), forming JM109SG (ΔmreB/pTK-mreB-PBAD:sulA). Since the pSC101-based plasmid pTK-mreB-PBAD:sulA is compatible with f1-based plasmids, the plasmids pBHR68 was transformed into JM109SG (ΔmreB/pTK-mreB-PBAD:sulA), resulting in JM109SG (ΔmreB/pTK-mreB-PBAD:sulA/pBHR68).

2.4. Analytical methods 2.4.1. PHB assay using gas chromatography (GC) Bacteria were harvested by centrifugation at 10,000g for 10 min and then washed with distilled water. Cell dry weight (CDW) was measured after vacuum lyophilization (LGJ-10C, Beijing Sihuan Scientific Instrument Factory Co., LTD, Beijing, China). PHB contents were analyzed via gas chromatography (Hewlett-Packard model 6890) after methanolysis of lyophilized cells in chloroform (Li et al., 2010).

2.4.2. Electron microscopy analysis Cells were harvested by centrifugation at 5000g for 3 min, washed three times with PBS (phosphate buffer pH 7.2). Subsequently, the cells were fixed with 2% glutaraldehyde (pH 7.2) at room temperature for 2 h and washed again as above. The cells were prepared for scanning or transmission electron microscopy as described previously (Denner et al., 1994).

This work This work This work

2.4.3. Glucose analysis using high-performance liquid chromatography (HPLC) Final glucose concentration were analyzed after 48 h and 72 h of cultivation (Table 3) by using 1 mL of cell supernatant for HPLC analysis to obtain the final glucose concentration. The HPLC system (P2000, AS3000, Thermo Spectra System, USA) was equipped with an anion exchange column (Aminexs HPX-87H, 7.8  300 mm2, BioRad) and a refractive index detector (RI-150, Thermo Spectra System, USA). A mobile phase of 2 mM H2SO4 was used at a 0.5 ml min  1 flow rate (Li et al., 2010). The standard curve for glucose was established by measuring standard solutions containing 2.5, 5, 10, 20 and 30 g L  1 glucose in pure water, respectively. 2.4.4. Fluorescence microscopy analysis of intracellular PHB granules For fluorescence microscopy analysis, cells were stained with BODIPY fluorescent dyes (Lee et al., 2013). The fermentation cultures were cooled on ice for 10 min, and the cells harvested by centrifugation (5 min, 1000g, 4 1C). The cell pellets were first resuspended (the first suspension) in 1 mL of ice-cold TSE buffer containing 10% (w/v) sucrose, 10 mM Tris–HCl (pH 7.5) and 2.5 mM Na–EDTA, and incubated on ice for an additional 10 min. After centrifugation for 5 min at 3000g and 4 1C, cells were resuspended (the second suspension) in deionized water (DDW). Subsequently, the cells were mixed with 5 μl of BODIPY (Invitrogen, Eugene, OR, USA) dissolved in DMSO (1 μg/μl). Mixtures were incubated in the dark for 5 min at room temperature after vigorous mixing. Finally, Stained cells were washed three times with DDW to remove any residual fluorescent dye prior to visualization by fluorescence microscope (OLYMPUS IX83, Tokyo, Japan). Cell excitation was accomplished using a 488 nm argon laser. Photographs were captured with an OLYMPUS software (CellSens Standard 1.9). 2.4.5. Super-resolution imaging of MreB First, mEos3.2, a recently developed photo-convertible fluorescent protein (Zhang et al., 2012), is truly monomeric. A functional sandwich fusion (MreB–mEos3.2SW) was constructed by inserting mEos3.2 between helices 6 and 7 of the protein MreB (Bendezú et al., 2009). To observe the dynamics of the MreB cytoskeleton and its relationship with the cell wall growth, total

230

X.-R. Jiang et al. / Metabolic Engineering 29 (2015) 227–237

internal reflection fluorescence microscopy (TIRFM), a sensitive method for studying events at cell surfaces (Domínguez-Escobar et al., 2011), was employed. mEos3.2 channel was collected with a continuous constant 561 nm irradiation (  2 kW cm  2) and a continuous 405 nm laser, which was slowly adjusted for optimal photo-conversion rates. 100 nm Tetraspeck beads (Invitrogen) were used as fiducial markers for drift correction and alignment between the two channels.

3. Results 3.1. Deletion on actin-like protein gene mreB resulted in larger and spherical cells MreB is the bacterial actin homolog (van den Ent et al., 2001). The gene mreB in was deleted from the genome of E. coli JM109SG using a PCR-mediated gene deletions method (Datsenko and Wanner, 2000) (See “Section 2” and Supplementary Table S1), generating E. coli JM109SG (ΔmreB), which displays a spherical cell shape (Fig. 1a and b) compared with rod-shape E. coli JM109SG (Fig. 1g). When grown in a minimal medium, cells of E. coli JM109SG (ΔmreB) showed a smaller size (Fig. 1a). In contrast, the cells became much larger when grown in a Luria-Bertani medium (Fig. 1b–d), some cells even reached a diameter of 5–10 μm (Fig. 1c and d). The gene sulA encodes a protein inhibitor of cell division (SulA) (Chen et al., 2012). By interacting with cell division FtsZ ring, SulA overexpression blocks binary division, turning the normally rodshape E. coli into filamentary cells (Bi and Lutkenhaus, 1993). To enlarge the above spherical cells of E. coli JM109SG (ΔmreB), sulA gene under the PBAD promoter encoded on a p15A-based plasmid was introduced into E. coli JM109SG (ΔmreB). Cells of E. coli JM109SG (p15A-PBAD:sulA) at first elongated to filamentary shape, but subsequently became shortened to original rods during the long course of incubation for 72 h (Supplementary Fig. S1). However, most cells of E. coli JM109SG (ΔmreB/p15A-PBAD: sulA) did not become elongated (Fig. 1e). Instead, they ruptured after 12 h of arabinose induction (Fig. 1f), possibly due to weakened cytoskeleton in the absence of MreB protein. When the plasmid pBHR68 containing the PHB synthesis operon phbCAB from Ralstonia eutropha (Spiekermann et al., 1999) was introduced into E. coli JM109SG (ΔmreB), the modified cells of E. coli JM109SG (ΔmreB/pBHR68) provided a much large space for the accumulation of PHB granules (Fig. 1h) compared with rod-shapes E. coli JM109SG (pBHR68) (Fig. 1g). However, E. coli JM109SG (ΔmreB/pBHR68) grew poorly compared with the parent strain, which is generally not desirable for more cell growth and more PHB accumulation that are important for industrial production. 3.2. PHB production in mreB overexpressing E. coli strain Although the deletion on mreB led to larger volumes for PHB accumulation by E. coli JM109SG (ΔmreB/pBHR68) (Fig. 1h), a lot of cells lost their viability, resulting in reduced PHB production. Based on this result (Fig. 1), it thus became interesting to know what happens when mreB is over-expressed. To understand the effect of mreB over-expression in E. coli JM109SG, we introduced native promoters controlling mreB encoded in a medium copy number plasmid termed p15A (  15 copies) into E. coli JM109SG. The resulting strain E. coli JM109SG (p15A-mreB) formed irregular morphologies (Fig. 2b). Furthermore, these irregular cells were commonly larger than the wild type E. coli JM109SG (Fig. 2a).

The larger E. coli JM109SG (p15A-mreB) was transformed with pBHR68, and was cultivated in Luria-Bertani medium containing 20 g L  1glucose at 37 1C for 48 h. The resulting E. coli JM109SG (p15A-mreB/pBHR68) produced 7.12 g L  1 PHB, an increase of 23% compared to the parent strain E. coli JM109SG (pBHR68) (Fig. 2c). Since the cell dry weights (CDW) were similar between E. coli JM109SG (p15A-mreB/pBHR68) and E. coli JM109SG (pBHR68), the non-PHB compositions (mostly related to cell proteins and DNA) indicated more cells of E. coli JM109SG (pBHR68) compared with E. coli JM109SG (p15A-mreB/pBHR68). This demonstrated that enlarged E. coli JM109SG (p15A-mreB/pBHR68) allowed more PHB accumulation compared with smaller E. coli JM109SG (pBHR68), even though the larger E. coli JM109SG (p15A-mreB/ pBHR68) grew slower (Fig. 2c). 3.3. Enhanced PHB production by E. coli JM109SG (ΔmreB) overexpressing mreB gene Since E. coli JM109SG (ΔmreB) has a spherical shape, the volume of which is larger than the volume of the rod-shaped cells of its parent strain, E. coli JM109SG (ΔmreB) was capable of accumulating more PHB granules than its parent strain (Fig. 1g and f). However, E. coli JM109SG (ΔmreB) grew poorly compared with its parent strain, resulting in less PHA accumulation. To circumvent this problem, wild-type mreB allele was cloned and inserted into the low-copy-number constitutive expression vector pTKRED (Kuhlman and Cox, 2010). This weakly compensated expression of mreB in E. coli JM109SG (ΔmreB) restored the rod shapes of most cells (Supplementary Fig. S2b). When the plasmid pBHR68 containing the PHB synthesis operon was transformed into E. coli JM109SG (ΔmreB/pTK-mreB), the resulting E. coli JM109SG (ΔmreB/pTK-mreB/pBHR68) produced 9.29 g L  1 PHB after 50 h, an increase of 62% compared with that of the wildtype JM109SG (pTK/pBHR68) (Table 2). When accumulating PHB granules, E. coli JM109SG (ΔmreB/pTK-mreB/pBHR68) had inflated cell sizes (Fig. 3a). Some of the E. coli JM109SG (ΔmreB/pTK-mreB/ pBHR68) accumulated so many PHB granules that the cells ruptured and released their PHB granules (Fig. 3d). TEM study on sections of cells of control strain E. coli JM109SG (pTK/pBHR68) (Fig. 3e) and E. coli JM109SG (ΔmreB/pTK-mreB/pBHR68) (Fig. 3f) also confirmed much larger cell size for the later with much more PHB granules accumulation. E. coli JM109SG (ΔmreB/pTK-mreB/pBHR68) accumulated 8.11 g L  1 PHB when grown in a minimal medium containing 20 g L  1 glucose, compared with its control E. coli JM109SG (pTK/ pBHR68) which made only 4.44 g L  1 PHB (Fig. 3b). It also exhibited significantly higher glucose to PHB conversion efficiency or yield (g PHB/g glucose) (Table 3), namely, 0.40 g/g for E. coli JM109SG (ΔmreB/pTK-mreB/pBHR68) compared with only 0.23 g/g for E. coli JM109SG (pTK/pBHR68). These results indicated that a large cell size resulted from various manipulations of genes related to cell morphology are beneficial for more accumulation of inclusion bodies, such as the PHB granules in this study. This strategy should also be feasible for accumulation of other inclusion bodies such as elemental sulfur, polyphosphate, glycogen and waxes as well as fatty acid substances. In addition, it is better for a high final product accumulation if the shape change is induced after cells have already grown to a high density. 3.4. Enhanced PHA production by E. coli JM109SG (ΔmreB) overexpressing mreB gene under inducible expression of sulA gene Filamentary cells formed by SulA inhibition on the formation of FtsZ ring led to an enhanced PHA production due to an enlarged space for more PHA granules (Wang et al., 2014a). In order to

X.-R. Jiang et al. / Metabolic Engineering 29 (2015) 227–237

231

Fig. 1. Electron microscopy study on morphology of E. coli JM109SG (ΔmreB) expressing PHB synthesis genes. (a) Genomic mreB in E. coli JM109SG was deleted by insertion of kanamycin resistant gene, yielding E. coli JM109SG (ΔmreB::KanR). Cells with a small size were grown in a minimal medium containing 4 g L  1 Glucose at 30 1C. (b–d) Cells with a large size were grown in LB medium at 37 1C. Arrows: inflated bacteria cells with a very large size. (e–f) SEM morphology studies on E. coli JM109SG (ΔmreB) expressing sulA gene under induction. Recombinant E. coli JM109SG (ΔmreB/p15A-PBAD:sulA) was grown in a Luria-Bertani medium at 30 1C to an OD600 of 0.4–0.6. Subsequently, 0.2% L-arabinose was added to induce sulA expression for 6 h (e) and 12 h (f). Arrows: ruptures of bacterial cells. (g–h) TEM morphology studies on PHB granules accumulated in E. coli JM109SG (pBHR68) (g) and E. coli JM109SG (ΔmreB/pBHR68) (h). Cells were cultivated in a LB medium supplemented with 20 g/L glucose under 37 1C for 48 h. Scale bar, 5 μm.

232

X.-R. Jiang et al. / Metabolic Engineering 29 (2015) 227–237

Fig. 2. SEM morphology studies on E. coli JM109SG overexpressing mreB. (a) E. coli JM109SG. (b) E. coli JM109SG (p15A-mreB). Scale bars, 10 μm. (c) E. coli JM109SG overexpressing mreB and harboring pBHR68 were cultivated in LB medium containing 20 g L  1 glucose at 37 1C for 48 h. 100 μg/mL Ampicillin and 25 μg/mL chloramphenicol were added for plasmid retention. Error bars are s.d. (n¼ 3).

Table 2 Production of PHB from glucose by E. coli JM109SG (ΔmreB/pTK-mreB/pBHR68), JM109SG (ΔmreB/pTK-mreB-PBAD:sulA/pBHR68) and its wild type E. coli JM109SG (pBHR68) grown in shake flasks under various conditions, respectively. Strains JM109SG JM109SG JM109SG JM109SG

(pTK/pBHR68)a (ΔmreB/pTK-mreB /pBHR68)a (pTK/pBHR68)b (ΔmreB/pTK-mreB-PBAD:sulA/pBHR68)b

CDW (g L  1)

PHB (wt%)

PHB (g L  1)

10.75 7 0.03 12.63 7 0.19 10.25 7 0.15 13.34 7 2.32

53.197 0.99 73.53 7 2.72 49.727 0.68 80.417 2.23

5.727 0.12 9.29 7 0.46 5.107 0.12 10.677 1.6

Abbreviations: CDW, cell dry weight. Errors are s.d. (n¼ 3). a E. coli JM109SG (ΔmreB/pTK-mreB/pBHR68) and it control E. coli JM109SG (pTK/pBHR68) were cultivated in LB medium at 30 1C for 10 h, then added with 20 g L  1 glucose for another 40 h. 250 μg/mL ampicillin were added to maintain stability of pBHR68. b E. coli JM109SG (ΔmreB/pTK-mreB-PBAD:sulA/pBHR68) and it control E. coli JM109SG (pTK/pBHR68) were cultivated in LB medium at 30 1C to an OD600 ¼ 0.4–0.6, then induced with 0.2% L-arabinose for 6 h. Subsequently, 20 g L  1 glucose was added, respectively, to each culture for another 40 h. 250 μg/mL ampicillin were added to maintain stability of pBHR68. Data shown are the average and standard deviation of three parallel experiments.

enhanced PHA production further, we conducted PHB accumulation in E. coli JM109SG (ΔmreB) overexpressing mreB in combination with inducible expression of sulA gene (Supplementary Fig. S3). Since pSC101-based pTK-mreB-PBAD:sulA, which replicates at 30 1C, is a temperature sensitive plasmid. It was expected that the loss of this plasmid after arabinose induced expression of sulA allowed the E. coli cells to change the shape from rod to larger spheres. Such a process should generate a sufficiently high number of rod-shaped cells that later change to elongated and later large spherical shapes. E. coli JM109SG (pTK-mreB-PBAD:sulA) and E. coli JM109SG (ΔmreB/pTK-mreB-PBAD:sulA) were grown in a Luria-Bertani medium at 30 1C to an OD600 of 0.4–0.6, followed by induction using

0.2% L-arabinose for 72 h at 30 1C. In contrast to E. coli JM109SG (p15A-PBAD:sulA) shortened to original rods during the entire incubation of 72 h (Supplementary Fig. S1), E. coli JM109SG (pTKmreB-PBAD:sulA) (Fig. 4a) and E. coli JM109SG (ΔmreB/pTK-mreBPBAD:sulA) (Fig. 4b) maintained the longer filamentary shapes during the entire incubation of 72 h, this compensated the shortages of the previous studies (Wang et al., 2014a). To investigate effects of cell elongation on PHB accumulation, E. coli JM109SG (ΔmreB/pTK-mreB-PBAD:sulA/pBHR68) was grown in Luria-Bertani medium at 30 1C to an OD600 of 0.4–0.6 and, subsequently induced with 0.2% L-arabinose for 6 h, followed by addition of 20 g L  1 glucose and an additional 40 h cultivation. The strain produced 10.67 g L  1 PHB after 50 h growth, an increase of 109%

X.-R. Jiang et al. / Metabolic Engineering 29 (2015) 227–237

233

Fig. 3. Electron microscopy studies on morphology and PHB production by E. coli JM109SG (ΔmreB) overexpressing mreB. (a) Schematic of PHB accumulation in E. coli JM109SG (ΔmreB) overexpressing mreB. Scale bar, 0.5 μm. (b) Growth and PHB accumulation by recombinants harboring pBHR68 cultivated in minimal medium at 30 1C for 10 h followed by addition of 20 g L  1 glucose and another 40 h of growth. Error bars are s.d. (n ¼3). E. coli JM109SG (pTK/pBHR68) (c) and E. coli JM109SG (ΔmreB/pTK-mreB/ pBHR68) (d) were grown in LB medium at 30 1C for 10 h followed by addition of 20 g L  1 glucose and another 40 h of growth. TEM on Sections of cells of control E. coli JM109SG (pTK/pBHR68) (e) and E. coli JM109SG (ΔmreB/pTK-mreB/pBHR68) (f) cultivated in the LB medium at 30 1C for 10 h, followed by addition of 20 g L  1 glucose and another 40 h of growth. Scale bar, 5 μm.

Table 3 Substrate (glucose) to PHB conversion efficiencies by E. coli JM109SG (ΔmreB/pTK-mreB/pBHR68) and its control E. coli JM109SG (pTK/pBHR68), respectively. Strains

Glucose consumed (g L  1)

PHB (g L  1)

Yield (g PHB/g Glucose)

% of Max (0.48 g/g)

JM109SG (pTK/pBHR68) JM109SG (ΔmreB/pTK-mreB/pBHR68)

19.57 70.096 19.95 70.087

4.447 0.06 8.117 0.43

0.23 0.40

47.9 84.6

The recombinants were cultivated in the minimal medium at 30 1C for 10 h followed by addition of 20 g L  1 glucose for another 40 h growth. Errors are s.d. (n¼ 3).

compared to E. coli JM109SG (pTK/pBHR68) obviously due to the difference in shape and size (Fig. 4c and d). TEM study on sections of E. coli JM109SG (ΔmreB/pTK-mreB-PBAD:sulA/pBHR68) and its control E. coli JM109SG (pTK/pBHR68) confirmed that much more PHB granules were accumulated in the longer and larger E. coli JM109SG (ΔmreB/pTK-mreB-PBAD:sulA/pBHR68) compared to the short rodshaped E. coli JM109SG (pTK/pBHR68) (Fig. 4e and f). When grown in minimal medium at 30 1C to an OD600 of 0.4–0.6 followed by induction with 0.2% L-arabinose for 6 h and a subsequent addition of 20 g L  1 glucose and additional 42 h cultivation, E. coli JM109SG (ΔmreB/pTK-mreB-PBAD:sulA/pBHR68) produced approximately 9 g L  1 PHB after 48 h, an increase of 87% compared to E. coli JM109SG (pTK/pBHR68) (Fig. 4g). Importantly, the cell number of E. coli JM109SG (ΔmreB/pTK-mreB-PBAD:sulA/ pBHR68) reflected by non-PHB cell mass was lower than E. coli JM109SG (pTK/pBHR68) (Fig. 4g). It was thus very clear that PHA accumulation per individual cell of E. coli JM109SG (ΔmreB/pTKmreB-PBAD:sulA/pBHR68) was much more higher than per individual cell of E. coli JM109SG (pTK/pBHR68) due to the larger average size of E. coli JM109SG (ΔmreB/pTK-mreB-PBAD:sulA/ pBHR68). Such larger size cells containing more PHB granules are heavier, resulting in rapid spontaneous sedimentation (Fig. 4g). Taken together, the larger and longer E. coli JM109SG (ΔmreB/pTKmreB-PBAD:sulA/pBHR68) produced more PHB, had higher glucose

to PHB conversion efficiency, and have the potential for easier separation from fermentation broth. The product of kil gene specifically inhibits ftsZ gene expression in E. coli (Conter et al., 1996). When comparing effects of overexpressions of kil and sulA on PHB production, it was found that the cell dry weight and PHB content of kil overexpressing E. coli were significantly decreased compared with sulA overexpression (Supplementary Table S2). After 6 h induction, kil overexpressing E. coli became enlarged and often had an irregular shape. While most sulA overexpressing E. coli had a filamentary shape. Gene kil differs from previously described inhibitors on its ability to abolish rod shape when strongly expressed (Conter et al., 1996). Probably kil overexpression at high levels affected the viability of the bacteria. As a conclusion, gene sulA overexpression was observed to be better for changing the cell shape without negatively affecting the cell viability compared with kil gene. 3.5. Enhanced PHA production by E. coli JM109SG (ΔmreB) overexpressing mreB gene in a 6-liter fermentor The morphology engineered E. coli JM109SG (ΔmreB/pTKmreB/pBHR68) and its control E. coli JM109SG (pTK/pBHR68) were grown in a 6-liter NBS fermentor using a fed-batch process, respectively. Glucose was supplied as a carbon source. The

234

X.-R. Jiang et al. / Metabolic Engineering 29 (2015) 227–237

Fig. 4. Effect of sulA over expression on cell morphology and PHB accumulation in E. coli JM109SG (ΔmreB/pTK-mreB-PBAD:sulA). E. coli JM109SG (pTK-mreB-PBAD:sulA) (a) and E. coli JM109SG (ΔmreB/pTK-mreB-PBAD:sulA) (b) were grown in a LB medium at 30 1C to an OD600 of 0.4–0.6, followed by induction using 0.2% L-arabinose for 24 h at 30 1C. Control strain E. coli JM109SG (pTK/pBHR68) (c and e), E. coli JM109SG (ΔmreB/pTK-mreB-PBAD:sulA /pBHR68) (d and f) were grown in LB medium at 30 1C to an OD600 of 0.4–0.6, followed by induction with 0.2% L-arabinose for 6 h, then addition of 20 g L  1 glucose and another 40 h of cultivation. Cell dry weight (CDW), PHB production (PHA) and non-PHB cell residual weight (CRW) (g left) and gravity precipitation (g right) of elongated E. coli JM109SG (ΔmreB/pTK-mreB-PBAD:sulA/pBHR68) compared with that of the control E. coli JM109SG (pTK/pBHR68) (NC). Cells were grown for 50 h in a minimal medium at 30 1C. Gravity precipitation of elongated E. coli JM109SG (ΔmreB/pTKmreB-PBAD:sulA/pBHR68) was observed after 30 min without agitation. Scale bar, 5 μm. Errors are s.d. (n¼ 3).

Fig. 5. Cell dry weight, PHB content and PHB granule accumulation of the morphology engineered E. coli JM109SG (ΔmreB/pTK-mreB/pBHR68) grown in a 6-l fermenter. (a) Time profiles of cell dry weight and PHB content during the fed-batch culture of morphology engineered E. coli JM109SG (ΔmreB/pTK-mreB/pBHR68) and its control E. coli JM109SG (pTK/pBHR68). Microscopic images of PHB granules in E. coli JM109SG (pTK/pBHR68) (b) and in E. coli JM109SG (ΔmreB/pTK-mreB/pBHR68) (c). After 40 h cultivation, cells were stained with BODIPY, and their fluorescence image analyzed using fluorescence microscope. Excitation wavelength was 488 nm. Scale bar, 5 μm.

morphology engineered E. coli produced 34 g l  1 cell dry weight containing 86% PHB after 90 g l  1 glucose was consumed in 38 h (Fig. 5a). In contrast, control E. coli produced 24 g l  1 cell dry weight containing 72% PHB under the same conditions. Both the CDW and final PHB content enhancement were similar with that of the shake flask cultures (Fig. 5a). Process optimization including enriched oxygen supply has yet to be done to reach very high cell density as reported by Lee's Group (Choi and Lee, 1999b; Lee, 1994; Wang and Lee, 1997) To learn the accumulation of PHB granules under fermentor conditions. E. coli containing PHB were stained with BODIPY (Lee

et al., 2013). Since the morphology engineered E. coli was able to grow without enriched oxygen to a higher density (with higher PHB content) compared with its control, much more PHB granules were observed in engineered larger spherical cells (Fig. 5c) compared with its control (Fig. 5b) as a result. 3.6. Mechanism study on enhanced PHA production by E. coli JM109SG (ΔmreB) weakly expressing mreB gene Weakly compensated expression of mreB in E. coli JM109SG (ΔmreB/pTK-mreB) provided more bacterial viability, leading also

X.-R. Jiang et al. / Metabolic Engineering 29 (2015) 227–237

to a longer and larger shapes compared with its control (Supplementary Fig. S2). The cytoskeleton of E. coli JM109SG (ΔmreB/pTKmreB) was weaker than its control E. coli JM109SG (pTK) as evidenced by an additional study using Atomic Force Microscope (AFM) (supplementary Fig. S4A and B, supplementary Table S3): the Young's modulus of the control E. coli JM109SG (pTK) was approximately 217 kPa compared with only 25 kPa of the morphology engineered E. coli JM109SG (ΔmreB/pTK-mreB). That is to say, a weaker cell strength is beneficial to reduce the space limitation on PHB granule accumulation. Thus, morphology engineered E. coli can be enlarged during PHB accumulation due to its weak strength and thus flexibility (Fig. 3d). A video provided evidences that morphology engineered E. coli JM109SG (ΔmreB/ pTK-mreB) ruptured during the process of growth (Supplementary video and supplementary Fig. S4C). Supplementary material related to this article can be found online at http://dx.doi.org/10.1016/j.ymben.2015.03.017. To explain the mechanism for enhancement on PHB accumulation by E. coli JM109SG (ΔmreB) overexpressing mreB in combination with inducible expression of sulA gene, it was believed that MreB was involved in orientating the FtsZ rings (Fenton and Gerdes, 2013; Jones et al., 2001), reduced MreB disoriented the proteins on FtsZ ring. Unlike a wild-type that comes back to its original short rod shape after division, E. coli JM109SG (ΔmreB) with weakly complemented mreB expression and inducible sulA expression, was found capable of maintaining the longer elongated cells after 24 h (Fig. 4b). The pattern of mreB location was observed to be different between E. coli JM109SG (pTK-mreB-meos3.2SW-PBAD:sulA) (Fig. 6a) and E. coli JM109SG (ΔmreB/pTK-mreB-meos3.2SW-PBAD:sulA) (Fig. 6b). The pattern of mreB location in E. coli JM109SG (pTKmreB-meos3.2SW-PBAD:sulA) was a helix (Fig. 6a), which can normally orient the FtsZ ring formation. In comparison, the pattern of mreB location in E. coli JM109SG (pTK-mreB-meos3.2SW-PBAD:sulA) was observed to be a broken helix (Fig. 6b), which provides a weak strength allowing flexible expansion of the cells.

4. Discussion and conclusion Many bacteria contain various inclusion bodies made of different materials such as elemental sulfur, polyphosphate, glycogen,

235

magnetosomes, carboxysomes, polyamino acids and polyhydroxyalkanoates (PHA) et al. (Shively et al., 2001). These inclusion bodies offer the promise of exciting new industrial applications. Therefore, it is important to exploit bacteria as microbial factories to produce these inclusion bodies as efficiently as possible. However, most bacteria have a small size ranging from 0.5 to 2 μm, preventing the bacterial cells from accumulating large amounts of inclusion bodies intracellularly, even though the bacteria are able to grow very fast. To overcome the size limitation, it is important to make bacterial cells larger. That is to say, a larger intracellular space is needed for more inclusion body accumulation. Consequently, MreB, the actin-like bacterial cytoskeletons (van den Ent et al., 2001), which also affects bacterial morphology, was considered a suitable engineering target for expanding the cell volumes. When mreB was deleted, E. coli changed from rods to spherical shapes (Fig. 1a and b), some cells even increased their sizes to diameters of around 10 μm (Fig. 1c and d). More PHB granules were accumulated in the large E. coli JM109SG (ΔmreB) cells (Fig. 1h). However, E. coli JM109SG (ΔmreB) also appeared to be fragile and a fraction of cells ruptured during the growth stage (Fig. 1f). This phenomenon showed that MreB may provide critical support for maintaining the cell shape. Ectopic expression of MreB in a wild-type bacterium was found to interfere with normal MreB cytoskeleton formation, resulting larger cell size compared with that of a wild type as observed in our study (Fig. 2b). To further increase the cell size, mreB gene was constitutively and compensated expressed in a weaker manner in MreB deleted E. coli JM109SG together with an arabinose inducible sulA gene encoding an inhibitory protein for the formation of the cell division ring (FtsZ ring), the overexpression of which leads to elongated cells. It was expected that E. coli JM109SG (ΔmreB/pTK-mreB-PBAD: sulA) was able to reverse back to its original rod-shaped firstly, after which the cells were elongated by inducible expression of sulA. When cultured at 37 1C, E. coli JM109SG (ΔmreB/pTK-mreBPBAD:sulA) should lose its plasmid, leading to a larger spherical cell. However, complementary expression of mreB encoded in plasmid pTK-mreB-PBAD:sulA reversed E. coli JM109SG (ΔmreB/pTK-mreBPBAD:sulA) back to the original rod-shape (Supplementary Fig. S2b). After induced expression of sulA, E. coli JM109SG (ΔmreB/p TK-mreB-PBAD:sulA) elongated to become long filamentary cells

Fig. 6. Photo-activation localization microscopy (PALM) images of representative fixed elongated E. coli cells expressing MreB-mEos3.2. E. coli JM109SG (pTK-mreBmeos3.2SW-PBAD:sulA) (a) and E. coli JM109SG (ΔmreB/pTK-mreB-meos3.2SW-PBAD:sulA) (b) were grown in a LB medium at 30 1C to an OD600 ¼ 0.4–0.6, followed by induction using 0.2% L-arabinose for12 h at 30 1C. Color represents Z-ring position. Scale bar, 500 nm.

236

X.-R. Jiang et al. / Metabolic Engineering 29 (2015) 227–237

(Fig. 4b). When cultured at 30 1C for 72 h, E. coli JM109SG (ΔmreB/ pTK-mreB-PBAD:sulA) maintained its original long filamentary shapes. Due to weak mreB compensated expression, MreB located differently for for its control E. coli JM109SG and for E. coli JM109SG (ΔmreB), which cannot normally orient the FtsZ ring formation. Therefore, FtsZ rings cannot contract as usually occurring in a mreB intact strain during a growth process. Even though the cells reversed back to rod-shapes via complementary expression of mreB, E. coli JM109SG (ΔmreB) might be expected to have a weakened cytoskeletons compared to that of the wild-type) (supplementary Fig. S4A and B, supplementary Table S3) (Kruse et al., 2003). This weakened cytoskeleton allowed inflating cell sizes when PHB granules filled the intracellular space (Fig. 3d and f). As a result, higher PHA production was observed in larger and elongated E. coli JM109SG (ΔmreB/pTK-mreB-PBAD:sulA/ pBHR68) even though the total cell number was decreased compared with their control strain E. coli JM109SG (pTK/pBHR68) (Fig. 4d, f and g). Under fed-batch fermentation conditions, the morphology engineered E. coli JM109SG (ΔmreB/pTK-mreB/pBHR68) and its control E. coli JM109SG (pTK/pBHR68) were able to grow to cell dry weights of 34 g l  1 and 24 g l  1 containing 86% PHB and 72% PHB after 90 g l  1 glucose was consumed in 38 h, respectively (Fig. 5a). E. coli JM109SG (ΔmreB/pTK-mreB/pBHR68) was also capable of maintaining their large shapes during their entire fedbatch growth phase (Fig. 5c). This marked a significant improvement over the previously report results that the elongated shapes were unstable over the growth phase (Wang et al., 2014a). Biosynthesis of the peptidoglycan cell wall is organized by the actin homolog MreB (Rueff et al., 2014). Deletions on MreB led to changing cell shape from rods to spheres (Bendezú et al., 2009; Wachi et al., 1989). However, the spherical bacteria formed by MreB inactivation have very poor viability. Therefore, weakly compensated expression of mreB in E. coli JM109SG (ΔmreB/pTKmreB) was provided to maintain the bacterial viability, resulting in changing original short rod shape to a longer and larger shapes compared with its control (Supplementary Fig. S2). We hypothesize that: E. coli JM109SG (ΔmreB/pTK-mreB) has abnormal cytoskeleton even though they had recovered to its original rod shape. The cytoskeleton of E. coli JM109SG (ΔmreB/ pTK-mreB) was found weaker than its control E. coli JM109SG (pTK) (supplementary Fig. S4A and B, supplementary Table S3) as supported by a higher Young's modulus of the control E. coli JM109SG (pTK) compared with the morphology engineered E. coli JM109SG (ΔmreB/pTK-mreB). That is to say, weaker cell strength provided the cells with expansion flexibility for more PHB granule accumulation (Fig. 3d). This hypothesis was also confirmed by a video (supplementary info), some cells even ruptured during the process of growth (supplementary Fig. S4C). Regarding the molecular insights for enhancement on PHB accumulation by E. coli JM109SG (ΔmreB) overexpressing mreB in combination with inducible expression of sulA gene, it was believed that FtsZ structure around the regular cylinder of the wild-type cell envelope allows formation of a closed, circular FtsZ ring. However, weaker mreB expression did not lead to the complete curving Z structure, resulting in a failure to close on itself. Therefore, instead of forming a spiral or helix, FtsZ rings were distorted (Jones et al., 2001). Also, since MreB was involved in orientating the FtsZ rings (Fenton and Gerdes, 2013; Jones et al., 2001), reduced MreB disoriented the proteins on FtsZ ring. Based on these views, we could further hypothesize that: due to weak mreB compensated expression, MreB may be located differently for E. coli JM109SG (ΔmreB/pTK-mreB-meos3.2SW-PBAD:sulA) and for its control E. coli JM109SG (pTK-mreB-meos3.2SW-PBAD:sulA). Therefore, FtsZ rings cannot contract as usually occurring in a mreB intact strain during a growth process. Therefore, unlike a wild-type

that comes back to its original short rod shape after its binary division, E. coli JM109SG (ΔmreB) with weakly complemented mreB expression and inducible sulA expression, was maintained as longer elongated cells after 24 h (Fig. 4b). This hypothesis was also supported by the pattern of mreB location observed to be different between E. coli JM109SG (pTKmreB-meos3.2SW-PBAD:sulA) (Fig. 6a) and E. coli JM109SG (ΔmreB/ pTK-mreB-meos3.2SW-PBAD:sulA) (Fig. 6b). The pattern of mreB location in E. coli JM109SG (pTK-mreB-meos3.2SW-PBAD:sulA) was a helix (Fig. 6a). In comparison, the pattern of mreB location in E. coli JM109SG (pTK-mreB-meos3.2SW-PBAD:sulA) was observed to be a broken helix (Fig. 6b). The broken helix could not function fully maintain the cell as a rigid integraty. In conclusion, this study demonstrated that accumulation of bacterial inclusion bodies such as the PHB granules in this study, can be significantly improved by manipulating the bacterial cell morphology related genes such as mreB and/or sulA genes. Stable enlarged bacterial shapes were maintained over a period of at least 72 h both in shake flasks and in fermentor studies. The molecular mechanism related to the enlarged shape was found to be the results of a broken cytoskeleton helix in a cell that cannot support the rigid integrity of a cell, thus allowing the cell to expand its size for more inclusion body accumulation.

Acknowledgments We are grateful to the Center of Biomedical Analysis, Tsinghua University for the SEM and TEM studies, and to the State Key Laboratory of Biomembrane and Membrane Biotechnology, Biodynamic Optical Imaging Center (BIOPIC), Peking University for the super-resolution studies. This project was supported by the State Basic Science Foundation 973 (Grant no. 2012CB725201) and National Natural Science Foundation of China (Grant no. 31430003).

Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.ymben.2015.03.017. References Alvarez, H.M., Steinbuchel, A., 2002. Triacylglycerols in prokaryotic microorganisms. Appl. Microbiol. Biotechnol. 60, 367–376. Bendezú, F., Hale, C., Bernhardt, T., de Boer, P., 2009. RodZ (YfgA) is required for proper assembly of the MreB actin cytoskeleton and cell shape in E. coli. EMBO J. 28, 193–204. Bi, E., Lutkenhaus, J., 1993. Cell division inhibitors SulA and MinCD prevent formation of the FtsZ ring. J. Bacteriol. 175, 1118–1125. Carrio, M., Villaverde, A., 2002. Construction and deconstruction of bacterial inclusion bodies. J. Biotechnol. 96, 3–12. Chen, G.Q., 2009. A microbial polyhydroxyalkanoates (PHA) based bio- and materials industry. Chem. Soc. Rev. 38, 2434–2446. Chen, G.Q., Patel, M.K., 2012. Plastics derived from biological sources: present and future: a technical and environmental review. Chem. Rev. 112, 2082–2099. Chen, Y., Milam, S., Erickson, H., 2012. SulA inhibits assembly of FtsZ by a simple sequestration mechanism. Biochemistry 51, 3100–3109. Cho, H., McManus, H.R., Dove, S.L., Bernhardt, T.G., 2011. Nucleoid occlusion factor SlmA is a DNA-activated FtsZ polymerization antagonist. Proc. Natl. Acad. Sci. USA 108, 3773–3778. Choi, J., Lee, S.Y., 1999a. Factors affecting the economics of polyhydroxyalkanoate production by bacterial fermentation. Appl. Microbiol. Biotechnol. 51, 13–21. Choi, J., Lee, S.Y., 1999b. Production of poly(3-hydroxybutyrate) with high P(3HB) content by recombinant Escherichia coli harboring the Alcaligenes latus P(3HB) biosynthesis genes and the Escherichia coli ftsZ gene. J. Microbiol. Biotechnol. 9, 722–725. Conter, A., Bouche, J.-P., Dassain, M., 1996. Identification of a new inhibitor of essential division gene ftsZ as the kil gene of defective prophage Rac. J. Bacteriol. 178, 5100–5104. Dai, K., Mukherjee, A., Xu, Y., Lutkenhaus, J., 1994. Mutations in ftsZ that confer resistance to SulA affect the interaction of FtsZ with GTP. J. Bacteriol. 176, 130–136.

X.-R. Jiang et al. / Metabolic Engineering 29 (2015) 227–237

Dajkovic, A., Mukherjee, A., Lutkenhaus, J., 2008. Investigation of regulation of FtsZ assembly by SulA and development of a model for FtsZ polymerization. J. Biotechnol. 190, 2513–2526. Datsenko, K., Wanner, B., 2000. One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proc. Natl. Acad. Sci. USA 97, 6640–6645. Denner, E.B.M., Mcgenity, T.J., Busse, H.J., Grant, W.D., Wanner, G., Stanlotter, H., 1994. Halococcus Salifodinae Sp-Nov, an Archaeal Isolate from an Austrian Salt Mine. Int. J. Syst. Bacteriol. 44, 774–780. Domínguez-Escobar, J., Chastanet, A., Crevenna, A., Fromion, V., Wedlich-Söldner, R., Carballido-López, R., 2011. Processive movement of MreB-associated cell wall biosynthetic complexes in bacteria. Science 333, 225–228. Fenton, A.K., Gerdes, K., 2013. Direct interaction of FtsZ and MreB is required for septum synthesis and cell division in Escherichia coli. EMBO J. 32, 1953–1965. Gao, X., Chen, J.C., Wu, Q., Chen, G.Q., 2011. Polyhydroxyalkanoates as a source of chemicals, polymers, and biofuels. Curr. Opin. Biotechnol. 22, 768–774. Garcia-Fruitos, E., Vazquez, E., Diez-Gil, C., Corchero, J.L., Seras-Franzoso, J., Ratera, I., Veciana, J., Villaverde, A., 2012. Bacterial inclusion bodies: making gold from waste. Trends Biotechnol. 30, 65–70. Garner, E.C., Bernard, R., Wang, W., Zhuang, X., Rudner, D.Z., Mitchison, T., 2011. Coupled, circumferential motions of the cell wall synthesis machinery and MreB filaments in B. subtilis. Science 333, 222–225. Gibson, D., Young, L., Chuang, R.-Y., Venter, J., Hutchison, C., Smith, H., 2009. Enzymatic assembly of DNA molecules up to several hundred kilobases. Nat. Methods 6, 343–345. Han, M.J., Yun, H., Lee, S.Y., 2008. Microbial small heat shock proteins and their use in biotechnology. Biotechnol. Adv. 26, 591–609. Henderson, T.A., Young, K.D., Denome, S.A., Elf, P.K., 1997. AmpC and AmpH, Proteins related to the class C beta-lactamases, bind penicillin and contribute to normal morphology of E. coli. J. Bacteriol. 179, 6112–6121. Higashitani, A., Higashitani, N., Horiuchi, K., 1995. A cell division inhibitor SulA of Escherichia coli directly interacts with FtsZ through GTP hydrolysis. Biochem. Biophys. Res. Commun. 209, 198–204. Huang, J., Cao, C., Lutkenhaus, J., 1996. Interaction between FtsZ and inhibitors of cell division. J. Bacteriol. 178, 5080–5085. Jones, L.J., Carballido-López, R., Errington, J., 2001. Control of cell shape in bacteria: helical, actin-like filaments in Bacillus subtilis. Cell 104, 913–922. Kalscheuer, R., Stoveken, T., Luftmann, H., Malkus, U., Reichelt, R., Steinbuchel, A., 2006. Neutral lipid biosynthesis in engineered Escherichia coli: jojoba oil-like wax esters and fatty acid butyl esters. Appl. Environ. Microbiol. 72, 1373–1379. Keshavarz, T., Roy, I., 2010. Polyhydroxyalkanoates: bioplastics with a green agenda. Curr. Opin. Microbiol. 13, 321–326. Kocaoglu, O., Carlson, E.E., 2013. Penicillin-binding protein imaging probes. Curr. Protoc. Chem. Biol. 5, 239–250. Koutinas, A.A., Vlysidis, A., Pleissner, D., Kopsahelis, N., Lopez Garcia, I., Kookos, I.K., Papanikolaou, S., Kwan, T.H., Lin, C.S., 2014. Valorization of industrial waste and by-product streams via fermentation for the production of chemicals and biopolymers. Chem. Soc. Rev. 43, 2587–2627. Kruse, T., Møller-Jensen, J., Løbner-Olesen, A., Gerdes, K., 2003. Dysfunctional MreB inhibits chromosome segregation in Escherichia coli. EMBO J. 22, 5283–5292. Kuhlman, T., Cox, E., 2010. Site-specific chromosomal integration of large synthetic constructs. Nucleic Acids Res. 38, e92. Laycock, B., Halley, P., Pratt, S., Werker, A., Lant, P., 2013. The chemomechanical properties of microbial polyhydroxyalkanoates. Prog. Polym. Sci. 38, 536–583. Lee, J.H., Lee, S.H., Yim, S.S., Kang, K.-H., Lee, S.Y., Park, S.J., Jeong, K.J., 2013. Quantified High-Throughput Screening of Escherichia coli Producing Poly(3hydroxybutyrate) Based on FACS. Appl. Biochem. Biotechnol. 170, 1767–1779.

237

Lee, S.Y., 1994. Suppression of filamentation in recombinant Escherichia coli by amplified FtsZ activity. Biotechnol. Lett. 16, 1247–1252. Li, Z.-J., Shi, Z.-Y., Jian, J., Guo, Y.-Y., Wu, Q., Chen, G.-Q., 2010. Production of poly(3hydroxybutyrate-co-4-hydroxybutyrate) from unrelated carbon sources by metabolically engineered Escherichia coli. Metab. Eng. 12, 352–359. Loose, M., Mitchison, T.J., 2014. The bacterial cell division proteins FtsA and FtsZ self-organize into dynamic cytoskeletal patterns. Nat. Cell Biol. 16, 38–46. Margolin, W., 2005. FtsZ and the division of prokaryotic cells and organelles. Nat. Rev. Mol. Cell Biol. 6, 862–871. Meng, D.C., Shen, R., Yao, H., Chen, J.C., Wu, Q., Chen, G.Q., 2014. Engineering the diversity of polyesters. Curr. Opin. Biotechnol. 29C, 24–33. Park, S.J., Kim, T.W., Kim, M.K., Lee, S.Y., Lim, S.C., 2012. Advanced bacterial polyhydroxyalkanoates: towards a versatile and sustainable platform for unnatural tailor-made polyesters. Biotechnol. Adv. 30, 1196–1206. Rehm, B.H.A., 2007. Biogenesis of microbial polyhydroxyalkanoate granules: a platform technology for the production of tailor-made bioparticles. Curr. Issues Mol. Biol. 9, 41–62. Rodriguez-Carmona, E., Villaverde, A., 2010. Nanostructured bacterial materials for innovative medicines. Trends Microbiol. 18, 423–430. Rueff, A.S., Chastanet, A., Dominguez-Escobar, J., Yao, Z., Yates, J., Prejean, M.V., Delumeau, O., Noirot, P., Wedlich-Soldner, R., Filipe, S.R., Carballido-Lopez, R., 2014. An early cytoplasmic step of peptidoglycan synthesis is associated to MreB in Bacillus subtilis. Mol. Microbiol. 91, 348–362. Serafim, L.S., Lemos, P.C., Albuquerque, M.G., Reis, M.A., 2008. Strategies for PHA production by mixed cultures and renewable waste materials. Appl. Microbiol. Biotechnol. 81, 615–628. Shively J.M. (ed.), Microbiology Monographs Inclusions in Prokaryotes, Vol. 1, 2006, 349pp. Berlin, Heidelberg: Springer, Verlag. Shively, J. M., Cannon, G. C., Heinhorst, S., Bryant, D. A., DasSarma, S., Bazylinski, D., Preiss, J., Steinbüchel, A., Docampo, R., Dahl, C., 2001. Bacterial and archaeal inclusions. eLS. Spiekermann, P, Rehm, BH, Kalscheuer, R, Baumeister, D., Steinbüchel, A, 1999. A sensitive, viable-colony staining method using Nile red for direct screening of bacteria that accumulate polyhydroxyalkanoic acids and other lipid storage compounds. Arch. Microbiol. 171, 73–80. Steinbüchel, A., Lütke-Eversloh, T., 2003. Metabolic engineering and pathway construction for biotechnological production of relevant polyhydroxyalkanoates in microorganisms. Biochem. Eng. J. 16, 81–96. Sudesh, K., Abe, H., Doi, Y., 2000. Synthesis, structure and properties of polyhydroxyalkanoates: biological polyesters. Prog. Polym. Sci. 25, 1503–1555. van den Ent, F., Amos, L.A., Löwe, J., 2001. Prokaryotic origin of the actin cytoskeleton. Nature 413, 39–44. Wachi, M., Doi, M., Okada, Y., Matsuhashi, M., 1989. New mre genes mreC and mreD, responsible for formation of the rod shape of Escherichia coli cells. J. Bacteriol. 171, 6511–6516. Wang, F., Lee, S.Y., 1997. Production of poly(3-hydroxybutyrate) by fed-batch culture of filamentation-suppressed recombinant Escherichia coli. Appl. Environ. Microbiol. 63, 4765–4769. Wang, Y., Wu, H., Jiang, X., Chen, G.Q., 2014a. Engineering Escherichia coli for enhanced production of poly(3-hydroxybutyrate-co-4-hydroxybutyrate) in larger cellular space. Metab. Eng. 25, 183–193. Wang, Y., Yin, J., Chen, G.Q., 2014b. Microbial polyhydroxyalkanoates, Challenges and opportunities. Curr. Opin. Biotechnol. 30C, 59–65. Zhang, M., Chang, H., Zhang, Y., Yu, J., Wu, L., Ji, W., Chen, J., Liu, B., Lu, J., Liu, Y., Zhang, J., Xu, P., Xu, T., 2012. Rational design of true monomeric and bright photoactivatable fluorescent proteins. Nat. Methods 9, 727–729.