Increased Lysine Content Is the Main Characteristic of the Soluble Form of the Polyamide Cyanophycin Synthesized by Recombinant Escherichia coli

Abstract

Cyanophycin, a polyamide of cyanobacterial or noncyanobacterial origin consisting of aspartate, arginine, and lysine, was synthesized in different recombinant strains of Escherichia coli expressing cphA from Synechocystis sp. strain PCC 6308 or PCC 6803, Anabaena sp. strain PCC 7120, or Acinetobacter calcoaceticus ADP1. The molar aspartate/arginine/lysine ratio of the water-soluble form isolated from a recombinant strain expressing CphA₆₃₀₈ was 1:0.5:0.5, with a lysine content higher than any ever described before. The water-insoluble form consisted instead of mainly aspartate and arginine residues and had a lower proportion of lysine, amounting to a maximum of only 5 mol%. It could be confirmed that the synthesis of soluble cyanobacterial granule polypeptide (CGP) is independent of the origin of cphA. Soluble CGP isolated from all recombinant strains contained a least 17 mol% lysine. The total CGP portion of cell dry matter synthesized by CphA₆₃₀₈ from recombinant E. coli was about 30% (wt/wt), including 23% (wt/wt) soluble CGP, by using terrific broth complex medium for cultivation at 30°C for 72 h. Enhanced production of soluble CGP instead of its insoluble form is interesting for further application and makes recombinant E. coli more attractive as a suitable source for the production of polyaspartic acid or dipeptides. In addition, a new low-cost, time-saving, effective, and common isolation procedure for mainly soluble CGP, suitable for large-scale application, was established in this study.

INTRODUCTION

Cyanophycin, a highly branched polypeptide that is contained in the cyanobacterial granule, consists of a polyaspartic acid backbone to which arginine residues are linked by isopeptide bonds at free carboxylate groups, and the polymer is therefore called multi-l-arginyl-polyaspartic acid and cyanobacterial granule polypeptide (CGP) (1, 2). The polypeptide is synthesized via ATP-, MgCl₂-, KCl-, and sulfhydryl reagent-dependent nonribosomal protein biosynthesis by the cphA-encoded cyanophycin synthetase CphA (1, 3). The incorporation of l-arginine is dependent on the presence of l-aspartic acid and results in a branched copolymer with molecular masses ranging from 25 to 100 kDa (1, 4). The accumulation of CGP in cyanobacteria is, for example, triggered by phosphate starvation (5) and serves the cells as a nitrogen and energy storage compound (3). The ability to synthesize CGP is not a unique feature of cyanobacteria but also occurs in other groups of bacteria (6, 7). The properties of CGP concerning solubility are interesting, because it is insoluble at neutral pH but soluble at pHs of <2 and >9 (8). This makes it easy to isolate CGP with 0.1 N HCl, because after the treatment of cyanobacterial cells with acid, the only polymer isolated from extracts is a polymer consisting of arginine and aspartic acid at a molar ratio of about 1:1 (3).

Heterologously synthesized CGP, e.g., from bacterial hosts like Escherichia coli, Pseudomonas putida, and Ralstonia eutropha or from yeasts, is quite different (9, 10, 11). The water-insoluble cyanophycin-like material isolated from recombinant E. coli expressing cphA from Synechocystis sp. strain PCC 6803 exhibits an electrophoretic mobility higher than that of cyanophycin isolated from cyanobacteria by the same procedure, and it contains l-lysine or other basic amino acids besides the principal amino acids arginine and aspartate (9). It was demonstrated by Krehenbrink et al. (6) that up to 10 mol% lysine can be incorporated in vivo and up to 15 mol% can be incorporated in vitro by using purified CphA from Synechocystis sp. strain PCC 6308. Heterologous expression of cphA₆₃₀₈ in Corynebacterium glutamicum, R. eutropha, and P. putida, respectively, in contrast, resulted in insoluble polymeric material that consists of equimolar amounts of aspartate and arginine. Examination of the molecular masses of the isolated cyanophycin showed that the polymer strands produced by recombinant CphA-harboring cells, respectively, exhibit a lower mass (25 to 30 kDa) and polydispersity than the authentic polymer from cyanobacteria. This was explained by the absence of an unknown cyanobacterial factor (10). In 2002, Ziegler et al. (7) published results of the heterologous expression of a noncyanobacterial cphA gene in E. coli. The genome of Desulfitobacterium hafniense harbors genes encoding putative proteins with considerable sequence homology to cphA of cyanobacteria, and the heterologous expression of cphA_Dh resulted in the formation of a polydisperse copolymer consisting of aspartate and arginine and a minor amount of lysine. The molecular mass of the polymer was 30 kDa, and it was water soluble, probably because of the differences in structure and/or composition. Füser and Steinbüchel (12) reconfirmed the results of Ziegler et al. (7) and proved with different experiments that water-insoluble CGP can be solubilized in vitro to extents of up to about 80% (wt/wt) in salt solutions of different inorganic salts such as LiCl, NaCl, KCl, RbCl, KBr, MgCl₂, or CaCl₂. Experiments with yeasts maintained the idea of a water-soluble form and a modified composition of CGP synthesized by CphA₆₃₀₈. In 2008, Steinle et al. (11) showed that in Saccharomyces cerevisiae expressing cphA₆₃₀₈, two forms of CGP are available. The major part accounts for a water-soluble form, and a minor part accounts for an insoluble form. Metabolic engineering of S. cerevisiae confirmed the results of Krehenbrink et al. (6) concerning the incorporation of lysine (16 mol%), citrulline (20 mol%), and ornithine (8 mol%) in CGP synthesized by CphA₆₃₀₈ (13). Wiefel et al. (14) also demonstrated that in recombinant P. putida ATCC 4359 expressing cphA₆₃₀₈, a soluble polymer could be isolated consisting of 8 mol% lysine and 8 mol% citrulline instead of arginine. Tseng et al. (15) also showed that a soluble CGP can be isolated from recombinant E. coli expressing the cyanobacterial cphA₆₈₀₃ gene. Without any metabolic engineering, Tseng et al. (15) demonstrated the incorporation of 25 mol% lysine in soluble CGP isolated from cells cultivated in terrific broth (TB) containing ampicillin and chloramphenicol at 26°C.

CGP is interesting for different applications because, on the one hand, it is possible to get β-dipeptides consisting of aspartate-arginine or aspartate-lysine by hydrolysis of the polymer with cyanophycinases as described by Sallam et al. (16). On the other hand, cyanophycin and its derivatives are good candidates as starting material for the production of nitrogen-containing bulk chemicals and are also a source of a polyaspartic acid-like polymer with a reduced arginine content that may find applications in different areas (17, 18).

In this study, we demonstrated that recombinant E. coli BL21(DE) expressing cphA₆₃₀₈ synthesizes a major part of soluble CGP with an average lysine content of 25 mol% and a minor part of insoluble CGP with a lower lysine content. In contrast to the study by Tseng et al. (15), optimized cultivation conditions resulted in a total CGP polymer content of over 30% (wt/wt) of the cell dry matter (CDM). Furthermore, it became clear that the soluble form of recombinant CGP is not a unique feature of the recombinant CphA proteins from D. hafniense and Synechocystis sp. strains PCC 6803 and PCC 6308 because the heterologous expression of cphA from Anabaena sp. strain PCC 7120 in E. coli and of the noncyanobacterial cphA gene from A. calcoaceticus ADP1 in E. coli, respectively, also yielded soluble CGP with an altered composition in comparison to the insoluble form. A molar lysine fraction of more than 15% seems to be crucial for CGP to become soluble in vivo. In addition to these findings, a new time-saving and efficient method for the isolation of primary soluble CGP, inspired by the results of Steinle and Steinbüchel (19), was established for E. coli. The results of this study substantiated that a lysine-rich derivative of cyanophycin is a water-soluble and heat-resistant polymer that can be isolated from E. coli with a greater yield than described before by using adjusted parameters and culture conditions and an efficient isolation method.

MATERIALS AND METHODS

Strains, media, and growth conditions.

The bacterial strains and plasmids used in this study are listed in Table 1. E. coli TOP10 was used for cloning experiments and plasmid conservation, and the cells were cultivated in Luria-Bertani (LB) complex medium (20) containing kanamycin (50 μg/ml). Cultivation was done overnight at 37°C and with agitation at 300 rpm. E. coli BL21(DE3) was used for the expression of cphA genes from different sources. A 20-ml LB preculture of the recombinant strain was inoculated from glycerol stock and cultivated overnight as described before. TB complex medium was supplemented with 1 mM MgSO₄ and 5052 solution containing 0.5% glycerol, 0.05% glucose, and 0.2% α-lactose for autoinduction (21). Kanamycin (100 μg/ml) was used for plasmid maintenance during main cultivation experiments (100, 200, or 400 ml) at 30 or 37°C and agitation at 300 rpm for 24, 48, 72, or 120 h.

Table 1.

Bacterial strains and plasmids used in this study

Strain or plasmid	Relevant characteristics	Source
Escherichia coli strains
TOP10	recA1 endA1 gyrA96 thi-1 hsdR17 (rK− mK+) supE44 relA1 DlacU169	Invitrogen
BL21(DE3)	F− ompT hsdSB(rB− mB−) gal dcm(DE3)	Novagena
Plasmids
pBBR1MCS2::cphA6308/lysC	Kmr, cphA from Synechocystis sp. strain PCC 6308, lysC from Corynebacterium glutamicum	SK 09990b
pET23a::cphA6803	Apr, T7 promoter, cphA from Synechocystis sp. strain PCC 6803	SK 11951b
pET23a::cphA7120	Apr, T7 promoter, cphA from Anabaena sp. strain PCC 7120	SK 11835b
pET23a::cphAADP1	Apr, T7 promoter, cphA from Acinetobacter calcoaceticus ADP1	SK 11944b
pCOLADuet-1	Kmr, T7/lac promoter	Novagen
pCOLADuet::cphA6308	Kmr, T7/lac promoter, cphA from Synechocystis sp. strain PCC 6308 cloned as 2.6-kbp NotI/PstI fragment into pCOLADuet-1	This study
pCOLADuet::cphA6803	Kmr, T7/lac promoter, cphA from Synechocystis sp. strain PCC 6803 cloned as 2.6-kbp NdeI/XhoI fragment into pCOLADuet-1	This study
pCOLADuet::cphA7120	Kmr, T7/lac promoter, cphA from Anabaena sp. strain PCC 7120 cloned as 2.6-kbp NdeI/XhoI fragment into pCOLADuet-1	This study
pCOLADuet::cphAADP1	Kmr, T7/lac promoter, cphA from Acinetobacter calcoaceticus ADP1 cloned as 2.6-kbp NdeI/XhoI fragment into pCOLADuet-1	This study

^aDarmstadt, Germany.

^bSteinbüchel laboratory strain collection.

Transfer, isolation, and cloning of DNA.

Competent cells of E. coli were prepared by the CaCl₂ method of Sambrook et al. (20). Transformation of E. coli was done by following the procedure of Hanahan (22). Plasmid isolation was accomplished by the alkaline lysis method described by Sambrook et al. (20). For cloning of cphA₆₃₀₈ into the vector pCOLADuet, providing two multiple cloning sites, PCR was done with Herculase II Fusion DNA polymerase (Agilent Technologies) by following the manufacturer's manual. Plasmid pBBR1-MCS2::cphA₆₃₀₈/lysC (Table 1) was used as the template for PCRs with the oligonucleotides 5′-AAACTGCAGATGAAAATCCTCAAAACACAAAC and 5′-AAAGCGGCCGCTATTCACTACTGAGATGATATTTC, harboring PstI and NotI restriction sites (underlined), respectively, as primers. The PCR product was cut out from standard agarose gel and purified by using the gel extraction kit from Peqlab by following the manufacturer's instructions. Cloning of the PCR product into the linearized plasmid with T4 DNA ligase (NEB) followed. Clones were obtained by using solid LB medium with kanamycin. Plasmids were isolated, and the correct sequence and orientation of cphA₆₃₀₈ were verified by sequencing. For cloning of cphA₆₈₀₃, cphA₇₁₂₀, and cphA_ADP1 into pCOLADuet, respectively, available plasmids carrying the designated genes (Table 1) were cut with NdeI and XhoI, purified, and ligated into the plasmid as described above.

Cell harvesting, pH measurement, determination of CDM, and cell disruption.

Cells of E. coli were harvested by using a bench centrifuge (15 min, 4,000 rpm, 4°C), and cell pellets were washed once with saline (0.9%, wt/vol). The pH values of culture supernatants were measured with a pH electrode (neoLab). For determination of CDM, pellets were lyophilized and the cell mass was determined gravimetrically. Cell pellets were resuspended in 20 mM Tris-HCl buffer (pH 7.5) and disrupted by sonication (1 min/ml; MS 73; Bandelin Electronic) or by using a French pressure cell (Amicon, Silver Spring, MD) at 1,000 mPa.

Isolation of CGP.

Cell disruption was performed as described above. Isolation of soluble CGP was performed by a modified form of the method of Ziegler et al. (7). Cell extracts were digested by overnight incubation with proteinase K (200 μg/ml) at 60°C. For analysis of the soluble fraction, the suspension was centrifuged afterward in a bench centrifuge (15 min, 4,000 rpm, 20°C). The supernatant was treated with ice-cold ethanol (96%), and precipitated material was collected by a centrifugation step and a subsequent washing step with acetone. To obtain a highly purified polymer, the sample was repeatedly precipitated with ethanol and washed with acetone; finally, the polymer was dried in vacuo (Vacutherm; Heraeus). Water-insoluble CGP was isolated as described by Frey et al. (23), with 0.1 N HCl for resuspension of cell debris and shaking overnight. After a centrifugation step (15 min, 4,000 rpm, 4°C), the soluble fraction was neutralized by the addition of 1 N NaOH, and the precipitate was isolated by a further centrifugation step. The isolated insoluble polymer was then washed with water and lyophilized.

Heat-induced isolation of CGP.

Dried cells were resuspended in 20 mM Tris-HCl buffer (pH 7.5). The cell suspensions were incubated for 20 min at 120°C in an autoclave or in a water bath for 5, 10, 15, 30, 45, or 60 min at 100°C, respectively. After this procedure, the chilled samples were centrifuged for 15 min in a bench centrifuge at 4,000 rpm at 4°C. The sedimented cells were treated with 0.1 N HCl for isolation of insoluble CGP as described by Frey et al. (23). In contrast, soluble CGP was directly isolated from clear supernatant by precipitation with 2 volumes of ice-cold 96% ethanol, washing with acetone as described above, and then drying in vacuo. The isolated CGP was analyzed by different analytical methods.

SDS-PAGE.

Sodium dodecyl sulfate (SDS)-polyacrylamide electrophoresis (PAGE) was performed with 11.5% (wt/vol) polyacrylamide gels as described by Laemmli (24) to analyze purified CGP. A prestained molecular mass marker from Fermentas, which is a three-color ladder with 10 recombinant prokaryotic proteins covering a molecular mass range of 10 to 170 kDa, was used to determine molecular masses. A 3-fold-concentrated denaturation buffer consisting of β-mercaptoethanol, glycerol, SDS, and bromophenol blue was used for sample preparation. Proteins were stained with Serva Blue R.

HPLC for amino acid analysis.

Amino acid constituents of isolated CGP were determined by high-performance liquid chromatography (HPLC) with a Waters B801 column (300 by 4 mm) as described by Aboulmagd et al. (25) and Steinle et al. (13). Precolumn o-phthalaldehyde derivatization was carried out in a Smartline Autosampler 3900 (Knauer GmbH, Berlin, Germany) as described in the manual. Calibration was done with samples from a reference kit for aspartate, arginine, and lysine (Kollektion AS-10; Serva Feinbiochemica, Heidelberg, Germany).

RESULTS

Selection of CphA proteins.

In this study, the four cyanophycin synthetases CphA₆₃₀₈, CphA₆₈₀₃, CphA₇₁₂₀, and CphA_ADP1 were chosen for heterologous expression in E. coli and for subsequent characterizations of the synthesized CGP. CphA from Synechocystis sp. strain PCC 6308 was shown to be unspecific for substrates because it incorporates other amino acids besides arginine, such as lysine, ornithine, and citrulline, into the polymer (6, 13, 14). The altered composition of CGP obviously leads to changes in solubility behavior, as soluble CGP was isolated from recombinant yeast (13, 26) and P. putida ATCC 4359 (14) expressing cphA₆₃₀₈. Until now, studies of the heterologous expression of cphA₆₃₀₈ in E. coli resulted in insoluble polymers with lysine contents of only about 10 mol% and no soluble CGP was detected (6).

CphA from Synechocystis sp. strain PCC 6803 was chosen for this study because previous studies demonstrated that isolated recombinantly synthesized insoluble polymer contained up to 4 mol% lysine in addition to arginine (9, 27). Furthermore, Tseng et al. (15) have described the isolation of soluble and lysine-rich CGP from recombinant E. coli harboring cphA₆₈₀₃.

For CphA₇₁₂₀, no hints of the synthesis of soluble CGP after heterologous expression in different bacteria have existed until now. Voss et al. (28) detected a minor amount of lysine in insoluble CGP isolated from recombinant E. coli and about 10 mol% in the insoluble polymer isolated from recombinant cells of R. eutropha and P. putida.

CphA_ADP1 was chosen for this study because the noncyanobacterial synthetase was reported to be very specific for aspartate and arginine and an insoluble polymer without any detectable lysine was synthesized by recombinant strains of E. coli expressing this CphA protein (29).

A new approach to the isolation of CGP.

The isolation of insoluble CGP from recombinant E. coli by the acid extraction method published by Frey et al. in 2002 (23) is well established. The isolation of soluble CGP from recombinant E. coli was established in the same year by Ziegler et al. (7). Nevertheless, both methods are time-consuming, and in particular, the method described for isolation of soluble CGP is inefficient and cost-intensive because of an overnight incubation step with proteinase K. Hence, this method should be improved for possible technical-scale production or for simultaneous handling of different samples.

In 2010, Steinle and Steinbüchel (19) published a new and alternative method for the isolation of soluble CGP from yeast cells at different temperatures. This method seemed to be very efficient and beneficial. CGP yields of 21% (wt/wt) of the CDM were obtained with three cycles of 10 min of incubation at 70 or 80°C and a 5-min cooling step, followed by precipitation with 2 volumes of ethanol. In addition, it was shown by SDS-PAGE that thermal treatment had no influence on the molecular mass of the isolated polymers. On the other hand, until now, no soluble CGP could be isolated from cells of recombinant E. coli by this method (19). An optimized and simplified protocol for the isolation of soluble CGP from E. coli based on the publications of Ziegler et al. (7) and Steinle and Steinbüchel (19) is presented in this report. First, a sample of freeze-dried cells of E. coli BL21(DE3)/pCOLADuet::cphA₆₃₀₈ was resuspended in 20 mM Tris-HCl buffer (pH 7.5) and autoclaved for 20 min at 121°C. Thereafter, the sample was centrifuged to remove insoluble cell matter and ethanol was added to the clear, slightly yellowish supernatant. The white precipitate was isolated and analyzed by SDS-PAGE. An influence of temperature on the size of the isolated polymer was not observed by SDS-PAGE (result not shown) and confirmed the results obtained at a lower temperature by Steinle and Steinbüchel (19). It was shown by this first experiment that (i) soluble CGP can be isolated from recombinant E. coli cells expressing cphA₆₃₀₈ for the first time, (ii) CGP exhibits considerable heat resistance under neutral conditions, and (iii) this approach seems to be very suitable for the isolation of soluble CGP from recombinant E. coli.

For better control of time and temperature, the heat stress method with a water bath at 100°C for 60 min was also used and this approach was compared to the method of Frey et al. (23) (Fig. 1), which comprises no heating steps.

Fig 1.

SDS-PAGE analysis of insoluble (A, C) and soluble (B, C) CGP synthesized by different CphA proteins. One milligram of isolated insoluble polymer was solubilized in 1 ml of 0.1 N HCl, and 1 mg of isolated soluble polymer was solubilized in 1 ml of 20 mM Tris-HCl (pH 7.5). Eighty microliters of each solution was applied to an 11.5% (wt/vol) SDS gel. CGP was isolated by different methods in order to determine if 1 h of treatment at 100°C under acidic or neutral conditions has an influence on the molecular mass. CGP isolated as described by Frey et al. (23) was used as a control and for comparison. Recombinant strains of E. coli BL21(DE3) were grown in TB complex medium with 50-fold-concentrated 5052 solution for induction at 30°C for 72 h. Lanes M, prestained molecular mass markers (10 to 170 kDa; Fermentas). (A, B) Lanes 1, CphA₆₃₀₈ (control); lanes 2, CphA₆₈₀₃ (control); lanes 3, CphA₇₁₂₀ (control); lanes 4, CphA₆₃₀₈ (0.1 N HCl, 100°C, 1 h); lanes 5, CphA₆₈₀₃ (0.1 N HCl, 100°C, 1 h); lanes 6, CphA₇₁₂₀ (0.1 N HCl, 100°C, 1 h); lanes 7, CphA₆₃₀₈ (20 mM Tris-HCl, 100°C, 1 h); lanes 8, CphA₆₈₀₃ (20 mM Tris-HCl, 100°C, 1 h); lanes 9, CphA₇₁₂₀ (20 mM Tris-HCl, 100°C, 1 h). (C) Lane 1, insoluble CphA_ADP1 (control); lane 2, insoluble CphA_ADP1 (0.1 N HCl, 100°C, 1 h); lane 3, insoluble CphA_ADP1 (20 mM Tris-HCl, 100°C, 1 h); lane 4, soluble CphA_ADP1 (control); lane 5, soluble CphA_ADP1 (0.1 N HCl, 100°C, 1 h); lane 6, soluble CphA₇₁₂₀ (20 mM Tris-HCl, 100°C, 1 h).

Aside from this experiment, it would be interesting to unravel if the synthesis of soluble CGP is a unique feature of recombinant E. coli expressing cphA₆₃₀₈, as shown above, or if the heterologous expression of other CphA proteins also leads to the synthesis of soluble CGP in recombinant E. coli. Therefore, the cphA₆₈₀₃, cphA₇₁₂₀, and cphA_ADP1 genes were cloned into the vector pCOLADuet for better comparability, and competent E. coli BL21(DE3) cells were transformed and further investigated. The synthesis of both soluble and insoluble CGP in four recombinant strains of E. coli heterologously expressing the cphA₆₃₀₈, cphA₆₈₀₃, cphA₇₁₂₀, and cphA_ADP1 genes could be demonstrated for the first time by three different methods (Fig. 1). As outlined in the next sections, the results of the synthesis of soluble CGP in recombinant E. coli could be confirmed by the following additional experiments (see below).

Isolation of CGP from CDM and fresh cells with heat induction.

Insoluble CGP from recombinant bacterial strains was always isolated from lyophilized cells to allow a conclusion about yield with regard to the CDM and to compare the productivities of different strains or different culture conditions. But the question of whether lyophilization of E. coli cells is necessary to improve the yield of soluble CGP in comparison to that of fresh cells, as it was demonstrated for yeasts (19), may be allowed. Furthermore, the omission of this step could reduce the period of downstream processing. To investigate the correlation between sample condition (lyophilized versus fresh) and the yield of isolated CGP, cells of recombinant E. coli BL21(DE3) were cultivated at 30°C for 72 h. As shown above, there was a first hint of the synthesis of soluble CGP in E. coli BL21(DE3)/pCOLADuet::cphA₆₃₀₈, which was confirmed by the following experiments. E. coli BL21(DE3) harboring pCOLADuet::cphA₆₃₀₈, pCOLADuet::cphA₆₈₀₃, pCOLADuet::cphA₇₁₂₀, or pCOLADuet::cphA_ADP1 was cultivated as described above, harvested, and washed with saline. In each case, one sample was lyophilized and the other sample was immediately prepared for isolation of soluble CGP by applying heat treatment for 60 min at 100°C. After the process described was finished, the results obtained with freeze-dried and fresh cells and the four different CphA proteins were compared with regard to CGP yield, polymer size, and amino acid composition. With recombinant CphA₆₃₀₈, CphA₆₈₀₃, and CphA₇₁₂₀, the yield of insoluble CGP was lower than 5% (wt/wt) of the freeze-dried cells, and with CphA_ADP1, it was between 5 and 10% (wt/wt) of the CDM (Fig. 2). There were no significant differences, irrespective of whether the cells were previously lyophilized or not. For soluble CGP, different results were observed regarding the yield of the polymer and the origin of the CphA protein used. From recombinant E. coli expressing cyanobacterial cphA₆₃₀₈, up to 25% (wt/wt) soluble polymer could be isolated. This was a very high yield in comparison to those of the other three CphA proteins, because their yields were only between 5 and 10% (wt/wt) (Fig. 2). The differences between the yields of soluble CGP from freeze-dried cells and fresh cells were noticeable but minor. In the end, it became clear that a recombinant strain of E. coli expressing CphA₆₃₀₈ is a good source for the synthesis of soluble CGP.

Fig 2.

CGP contents of fresh and freeze-dried cells obtained from recombinant E. coli strains expressing different CphA proteins. Fresh and freeze-dried cells were resuspended in 20 mM Tris-HCl (pH 7.5) and incubated for 1 h at 100°C. Soluble CGP was isolated from the supernatants, whereas insoluble CGP was isolated from cell pellets after overnight treatment with 0.1 N HCl. Graph A shows the results for insoluble CGP isolated from CphA₆₃₀₈, CphA₆₈₀₃, CphA₇₁₂₀, or CphA_ADP1. Graph B shows the results for soluble CGP, and graph C shows the results for total CGP as percentages (wt/wt) for the same strains as in panel A. The results obtained with freeze-dried cells are shown as black columns; the results obtained with fresh cells are shown as light gray columns. Standard errors calculated from at least three experiments are shown as error bars.

With these experiments, it could be confirmed that the synthesis and isolation of soluble polymer from recombinant E. coli strains are possible not only with CGP synthetases CphA_Dh and CphA₆₈₀₃ (7, 15) but also with CGP synthetases CphA₆₃₀₈, CphA₇₁₂₀, and CphA_ADP1. Furthermore, in contrast to the results obtained by Steinle and Steinbüchel (19) with yeast, it also became clear that freeze-drying of E. coli cells does not have a positive effect on the yield of soluble CGP.

Further analysis of the isolated polymers showed, as described above, that the molecular masses of the soluble and insoluble CGPs differed from each other. The molecular mass of the insoluble CGP isolated from recombinant E. coli was 20 to 35 kDa in the cases of recombinant cyanobacterial CphA₆₃₀₈, CphA₆₈₀₃, and CphA₇₁₂₀. In the case of noncyanobacterial CphA_ADP1, the isolated polymer seemed to be more polydisperse, with a range of about 25 to 170 kDa (Fig. 3). The molecular masses of soluble CGP isolated from the four different recombinant strains were almost identical and ranged from approximately 12 to 40 kDa (Fig. 3). These results were identical for freeze-dried and fresh cells. The results confirmed the assumption that two forms of CGP are present in recombinant E. coli, and further analyses were conducted to investigate the causes of this difference.

Fig 3.

SDS-PAGE analysis of insoluble and soluble CGP synthesized by different CphA proteins. Fresh (A) and freeze-dried (B) cells were resuspended in 20 mM Tris-HCl (pH 7.5) and treated for 1 h at 100°C. Soluble CGP was isolated from supernatant, and insoluble CGP was isolated from the cell pellet after overnight treatment with 0.1 N HCl. Whereas 1 mg of isolated insoluble polymer was solubilized in 1 ml of 0.1 N HCl, 1 mg of isolated soluble polymer was solubilized in 1 ml of 20 mM Tris-HCl (pH 7.5). Eighty microliters of each sample was applied to an SDS (11.5%) gel. Lanes 1, CphA₆₃₀₈ (insoluble); lanes 2, CphA₆₈₀₃ (insoluble); lanes 3, CphA₇₁₂₀ (insoluble); lanes 4, CphA_ADP1 (insoluble); lanes 5, CphA₆₃₀₈ (soluble); lanes 6, CphA₆₈₀₃ (soluble); lanes 7, CphA₇₁₂₀ (soluble); lanes 8, CphA_ADP1 (soluble).

Optimization for isolation of soluble CGP.

After comparison of four different CphA proteins, the recombinant strain E. coli BL21(DE3)/pCOLADuet::cphA₆₃₀₈ seemed to be a prime example for further optimization and analysis of soluble CGP and its isolation. Since 60 min is a very long period of incubation at 100°C, it was investigated whether shorter incubation periods are also sufficient with regard to the yield and purity of CGP. After a temperature of 100°C was applied for 0, 5, 10, 15, 30, or 45 min to a constant amount of freeze-dried cells, CGP was isolated and the yields were compared (Fig. 4). An incubation period of 5 min resulted in soluble CGP that was 20% (wt/wt) of the CDM; the yield of soluble polymer could be further increased to 23% (wt/wt) by extension of the incubation period to 15 min (Fig. 4). From the control (0 min at 100°C), 6% (wt/wt) soluble polymer and 13% (wt/wt) total CGP could at least be isolated. However, it was also shown that incubation at 100°C for more than 10 min gave a nearly 4-fold higher yield of soluble polymer in comparison to the control. In each case, the molecular mass of soluble CGP ranged from approximately 12 to 40 kDa and that of insoluble CGP ranged from approximately 20 to 35 kDa (Fig. 5). A total CGP content of ≥25% (wt/wt) of the CDM, including ≥23% (wt/wt) soluble CGP and ≥3% (wt/wt) insoluble CGP, could be demonstrated for the first time. For this reason, the isolation of soluble CGP from recombinant E. coli with 15 min of heat stress seems to be very efficient and also saves time.

Fig 4.

CGP contents of freeze-dried cells after incubation at 100°C for 5, 10, 15, 30, or 45 min. Recombinant cells of E. coli BL21(DE3)/pCOLADuet::cphA₆₃₀₈ were cultivated for 3 days at 30°C in complex TB medium. Three hundred milligrams of dry cells was resuspended in 15 ml of Tris-HCl buffer and heated for the indicated periods at 100°C. One sample was not heated as a control (0 min). Soluble CGP was isolated from the supernatant and insoluble CGP was isolated from the cell pellet after treatment with 0.1 N HCl overnight. The fractions of insoluble CGP are shown as black columns, and the fractions of soluble CGP are shown as white columns. The fractions of total CGP (soluble plus insoluble) as percentages (wt/wt) of the CDM are represented by gray columns. Standard errors of at least three experiments are shown as error bars.

Fig 5.

SDS-PAGE analysis of insoluble (A) and soluble (B) CGP after heat treatment for different periods. Cells of recombinant strains of E. coli BL21(DE3)/pCOLADuet::cphA₆₃₀₈ were cultivated for 3 days at 30°C in complex TB medium. Three hundred milligrams of freeze-dried cells was resuspended in 15 ml of Tris-HCl buffer and heated for different periods at 100°C. As a control, one sample was not incubated at the increased temperature (0 min). One milligram of isolated insoluble polymer was solubilized in 1 ml of 0.1 N HCl, and 1 mg of isolated soluble polymer was solubilized in 1 ml of 20 mM Tris-HCl (pH 7.5). Eighty microliters of each of these solutions was applied to an SDS-polyacrylamide gel (11.5%). Lanes M, molecular mass markers; lanes 1, 0 min; lanes 2, 5 min; lanes 3, 10 min; lanes 4, 15 min; lanes 5, 30 min; lanes 6, 45 min. Panel A, insoluble samples; panel B, soluble samples.

Optimization for insoluble CGP.

The isolation of soluble CGP was good when cells were resuspended in 20 mM Tris-HCl buffer (pH 7.5) and heated for 15 min at 100°C as described above. A bottleneck seems to be the isolation of insoluble CGP with regard to the length of the procedure. Frey et al. (23) used a period of 6 h for stirring of the cells in a suspension at pH 1. Another strategy was chosen that produced better yields of both forms of CGP. Cells were resuspended in Tris-HCl and heated for 15 min as described above. After a centrifugation step, soluble CGP was isolated from the supernatant. Insoluble CGP was isolated from the pellet as done before after treatment with 0.1 M HCl and heating again for 15 min at 100°C (23). An increased yield of total CGP was the result of using this modified isolation procedure.

Effects of cultivation temperature and duration on CGP yield.

The previous results were obtained with cells that were cultivated for 72 h at 30°C. Data about the course of CDM and yields of CGP as a function of time were still missing. For a possible optimization of the parameters, experiments with cultivation at 30 and 37°C for 24, 48, 72, and 120 h in 400 ml TB complex medium were done with strain BL21(DE3)/pCOLADuet::cphA₆₃₀₈. At each time point, the yield of total CGP as a percentage of the CDM and the ratio of soluble to insoluble CGP were determined to get more information about the best conditions for cultivation. After 24 h of incubation at 30°C, 0.27 g of CDM/100 ml of cell culture was obtained; after 48 and 72 h, the CDM increased to 0.41 g/100 ml and 0.47 g/100 ml, respectively. After 120 h, the CDM decreased to 0.36 g/100 ml because of cell lysis. The results obtained with cultivation at 37°C were different from the results of cultivation at 30°C, because the CDM amounts were 25% (wt/wt) lower in the case of the first three time points (Fig. 6). Interesting differences were obtained and countersigned for the yields of the two isolated forms of CGP from cells cultivated at 30 or 37°C. After 24 h, an insoluble CGP yield of nearly 15% (wt/wt) of the CDM was isolated from cells cultivated at 30 or 37°C by the modified method (0.1 N HCl, 15 min, 100°C) as described above. The percentage of insoluble CGP isolated from cells cultivated at 30°C decreased slightly during the cultivation process from 14.8% to 12.6, 11.5, and 9.3% (wt/wt) of the CDM, respectively. The percentage of insoluble CGP isolated from cells cultivated at 37°C was 14.7% (wt/wt) of the CDM for the first 3 days. After 120 h, cell lysis occurred and the amount of isolated insoluble CGP decreased (Fig. 6).

Fig 6.

Cell densities and CGP contents obtained after cultivation of cells of recombinant E. coli BL21(DE3) at 30 or 37°C for 24 to 120 h, respectively. E. coli BL21(DE3)/pCOLADuet::cphA₆₃₀₈ was cultivated at 30 or 37°C in 400 ml of complex TB medium, and four samples each of 100 ml were harvested at different times during the course of cultivation and lyophilized. Soluble and insoluble CGPs were isolated as described in Materials and Methods. Graph A shows the CDM obtained after 24, 48, 72, and 120 h. Graph B shows the fraction of the total CDM that was CGP (insoluble and soluble forms). Graphs C and D show the fractions of the CDM that were insoluble and soluble CGP, respectively. Data obtained from cell cultivation at 30°C are shown in black, and data from cultivation at 37°C are shown in gray. Standard errors of at least three experiments are indicated by the error bars.

The cultivation temperature is crucial for the extraction of soluble CGP and its yield. Whereas nearly 6% (wt/wt) of the soluble polymer was isolated after 24 h from cells cultivated at 30°C, no soluble polymer could be obtained from cells cultivated at 37°C. During the time course of cultivation, the difference between cells cultivated at 30 and 37°C and the yield of soluble CGP became obvious. After 48 h, a 3-fold larger amount of soluble polymer could be isolated from cells cultivated at 30°C than from cells cultivated at 37°C. After 72 h, the largest amount of soluble CGP was isolated by both approaches. The amount isolated from cells cultivated at 30°C corresponded to nearly 25% (wt/wt) of the CDM, and in the case of cells cultivated at 37°C, it was only about 10% (wt/wt) of the CDM (Fig. 6). These results made it clear that a cultivation period of 72 h at 30°C provides the best conditions for a good soluble CGP yield of more than 22% (wt/wt) of the CDM. Also, a comparison of total CGP yields showed that using the optimized method for soluble CGP (15 min, 100°C), followed by the modified procedure for insoluble CGP, as described above is a good strategy. CGP yields of 33% (wt/wt) of the CDM after 72 h and 30°C and 23% (wt/wt) of the CDM after 72 h and 37°C were obtained (Fig. 6) by this process.

Composition of soluble CGP encoded by different cphA genes.

A comparison of four different CphA proteins (CphA₆₃₀₈, CphA₆₈₀₃, CphA₇₁₂₀, and CphA_ADP1) regarding the synthesis and isolation of soluble CGP from recombinant E. coli had never been done before. Surprisingly, in the present study, insoluble CGP and soluble CGP were isolated from all four recombinant E. coli strains by using adjusted parameters. The analyses with regard to the yields and molecular masses of both forms of CGP are shown above. Further studies should analyze whether there are differences in the polymers' amino acid composition based on the differing CphA proteins. HPLC analysis revealed a lysine content of about 5 mol% in all insoluble CGP samples, as well in samples from recombinant E. coli expressing CphA_ADP1, which is considered substrate specific (6) (Fig. 7). Astonishingly, HPLC analysis demonstrated that the amino acid mixtures of isolated soluble polymers varied between one CphA protein and another, especially with respect to the lysine portion (Fig. 7). The highest lysine content in this study was found in soluble CGP synthesized by CphA₆₃₀₈ from recombinant E. coli, at an average of 25 mol%. For soluble CGP synthesized by CphA₆₈₀₃ from recombinant E. coli, an average lysine fraction of 20 mol% was determined by HPLC analysis. For CphA₇₁₂₀ and CphA_ADP1, the lysine content was, on average, 17 mol%. After analysis of isolated polymers by HPLC, it also became clear that especially the recombinant strain expressing the cphA₆₃₀₈ gene is suitable for the production of great amounts (total CGP content, >30% of the CDM; soluble CGP content, 23% of the CDM) of a soluble and lysine-rich cyanophycin.

Fig 7.

Amino acid composition of CGP isolated from different CphA proteins. CGP was isolated by heat treatment from cells of recombinant E. coli cultivated at 30°C for 72 h as described in Materials and Methods. Both forms of cyanophycin were isolated from identical cultures. The molar fractions of aspartate (black columns), arginine (white columns), and lysine (gray columns) in insoluble (A) and soluble (B) CGP isolated from recombinant strains expressing CphA₆₃₀₈, CphA₆₈₀₃, CphA₇₁₂₀, or CphA_ADP1 are shown. Standard errors of at least three experiments are indicated by error bars.

In a previous cultivation experiment, it was demonstrated that 30°C and a cultivation duration of 72 h were good parameters for a high yield of soluble CGP. A further cultivation experiment should demonstrate whether the amino acid composition of soluble CGP alters as a function of time and temperature. HPLC measurements (Fig. 8) showed that there was a slight shift of the amino acid composition of isolated polymers from cells cultivated at 30°C during the time course of the experiment. After 24 h, an average arginine content of 28 mol% and an average lysine content of 22 mol% were detected. From 48 to 72 h, the average arginine content decreased from 24 to 21 mol%, and vice versa, the average lysine content increased from 25 to 26 mol% (Fig. 8). In comparison to the samples isolated from cells cultivated at 37°C, the contents of arginine and lysine changed not as clearly as at 30°C. Furthermore, at no time was an average lysine content higher than 21 mol% detected. In summary, the most effective cultivation conditions for recombinant E. coli for the synthesis of large amounts of lysine-rich soluble cyanophycin polymers are consequently cultivation for longer than 24 h at 30°C in modified TB complex medium.

Fig 8.

Amino acid composition of CGP isolated from cells of recombinant E. coli expressing CphA₆₃₀₈ at different times during the course of cultivation. Cells of E. coli BL21(DE3)/pCOLADuet::cphA₆₃₀₈ were cultivated at 30 or 37°C in 400 ml of complex TB medium. The molar fractions of aspartate (black columns), arginine (white columns), and lysine (gray columns) in soluble CGP isolated from cells cultivated at 30°C (A) or 37°C (B), respectively, are shown. Standard errors of at least three experiments are indicated by error bars.

DISCUSSION

Tseng et al. (15) published the first results of experiments with a recombinant E. coli strain expressing the cyanobacterial cphA₆₈₀₃ gene including the formation of a soluble polymer with molecular masses ranging from 14 to 25 kDa and with a lysine fraction of 25 mol%. The yield of total CGP (12% [wt/wt] of the CDM) was not as high as that of insoluble CGP (24% [wt/wt] of the CDM) described before by Frey et al. (23) for recombinant E. coli expressing cphA₆₈₀₃ in TB complex medium. However, Tseng et al. (15) emphasized the changed solubility behavior of CGP in recombinant E. coli as a result of an increased lysine fraction. For a conclusion about soluble polymer synthesized by recombinant E. coli BL21(DE3) strains expressing cphA genes, additional experiments were needed.

Three CphA proteins in addition to CphA₆₈₀₃ were chosen for this study to get more information about the formation of soluble CGP in recombinant E. coli. The use of TB complex medium supplemented with MgSO₄ and a 50-fold-concentrated 5052 solution introduced by Studier et al. (21) for autoinduction seemed to be a good combination to obtain a high cell density. After 72 h of cultivation at 30°C, 0.4 to 0.5 g of CDM/100 ml of recombinant E. coli BL21(DE3)/pCOLADuet::cphA_xxxx was obtained.

It was an interesting observation that soluble CGP can be isolated from all four reviewed recombinant strains expressing cphA₆₃₀₈, cphA₆₈₀₃, cphA₇₁₂₀, and cphA_ADP1 under these cultivation conditions. For the recombinant E. coli expressing cphA₆₈₀₃, our data regarding CDM, lysine content, and molecular mass were marginally lower than those described by Tseng et al. (15). This could be due to the application of different strains or the use of different cultivation conditions. The results of recombinant E. coli expressing cphA₆₃₀₈ in our study are remarkable regarding the amounts of isolated polymers and the average lysine content of 25 mol% in soluble CGP. Insoluble CGP amounts of about 5% (wt/wt) and soluble CGP amounts of about 25% (wt/wt) of the CDM were measured in this recombinant E. coli strain, which is higher than all other values reported so far. A novelty is that in the case of the recombinant E. coli strain expressing cphA₇₁₂₀ constructed in this study, soluble CGP contents of about 10% (wt/wt) and soluble CGP contents of more than 5% (wt/wt) of the CDM in the case of recombinant E. coli expressing cphA_ADP1 were measured.

Analysis of the isolated polymers by SDS-PAGE yielded data that are in agreement with those described by Wiefel et al. (14) and Steinle et al. (13, 26) because they observed average molecular masses of the polydisperse soluble polymers of 14 to 66, 17 to 30, and 19 to 40 kDa, respectively. In our study, the soluble polymers isolated from recombinant E. coli exhibited an average molecular mass of 20 kDa with a mass distribution between about 12 and 40 kDa for cyanobacterial CphA proteins and an average molecular mass of around 35 kDa with a distribution between about 15 and 42 kDa for noncyanobacterial CphA.

HPLC analysis of the isolated polymers revealed different results. On the one hand, it could be determined that insoluble CGP isolated from recombinant E. coli expressing cphA_ADP1 already contained 8 mol% lysine. This result is notable because CphA_ADP1 is characterized as being very specific (29). On the other hand, solubility is conferred in spite of the different lysine contents. The calculated molar aspartate/arginine/lysine ratio was 1:0.5:0.5 for soluble CGP isolated from recombinant E. coli expressing cphA₆₃₀₈, and soluble CGP isolated from recombinant E. coli expressing cphA₆₈₀₃ contained, on average, 20 mol% lysine. The soluble CGP synthesized by CphA₇₁₂₀ and CphA_ADP1 from recombinant E. coli contained, on average, 17 mol% lysine, a remarkable result for the noncyanobacterial CphA_ADP1 protein, which was previously described as being rather specific. Soluble CGP isolated from recombinant P. putida (14) or engineered yeast (13) expressing cphA₆₃₀₈ showed that, in both studies, lysine plus citrulline contents of 16 mol% (together) or 16 mol% lysine (alone), respectively, occurred on average. As described before, Krehenbrink et al. (6) observed the incorporation of up to 10 mol% lysine into insoluble CGP synthesized by CphA₆₃₀₈. In this study, it was demonstrated that the isolated soluble CGP contained 17 to 25 mol% lysine. We might conclude that there must be an incorporated lysine limit between 11 and 16 mol%, which leads to a change in solubility behavior in recombinant E. coli.

The comparison of another biopolymer and its derivatives could maintain this conclusion. Cellulose is one of the most abundant biopolymers on earth and has one property in common with cyanophycin; i.e., both polymers are insoluble in water. In a review in 1991, Ross et al. (30) stated that a cellulose fibril may be idealized as a cable in which the lengthwise strand consists of long polymer chains composed solely of d-glucose and linked by β-1,4-glucosidic bonds. The special geometry of this unbranched covalent arrangement gives rise to extended fibril structures because all of the available hydroxyl groups of adjacent aligned β-1,4-glucan chains participate in inter- and intrachain hydrogen bonding (30). Such inter- and intrachain hydrogen bonding could also be the reason why CGP is insoluble in water because the side chain arginine offers the guanidinium group for hydrogen bonding. Borders et al. (31) proposed that arginine side chains often play a previously unappreciated general structural role in the maintenance of tertiary structure in proteins, wherein the positively charged guanidinium groups form multiple hydrogen bonds with backbone carbonyl oxygens. They even gave an account of this side chain, which almost always has a pK_a of ≥12 and the positively charged planar guanidinium groups may be hydrogen donors in the formation of up to five hydrogen bonds. Cummings et al. (32) reported that modified celluloses such as ethyl and carboxymethyl cellulose, as well as methyl cellulose, have properties that are very different from those of pure cellulose. Their substituent groups disrupt the hydrogen bonding, and the resulting polymers are more soluble. This could also occur if lysine substitutes for arginine in CGP because of altered properties introduced by another functional group.

In conclusion, the coexistence of two different forms of CGP in recombinant E. coli expressing any cphA gene should be considered during the isolation process. The established isolation method is efficient and can be applied to freeze-dried or fresh cells with similar yields. The solubility of CGP in recombinant E. coli is dependent on increased amounts of lysine side chains. Soluble CGP amounts of more than 23% (wt/wt) of the CDM and average lysine portions of 25 mol% strengthen the position of recombinant E. coli as a production organism.