Orthologs of Branching Enzymes from Cyanobacteria Accumulating Distinct Types of α-Glucans Share Common Reaction Product Specificity

Abstract

Cyanobacteria generally accumulate glycogen in their cells as a photosynthetic product. Interestingly, several unicellular diazotrophic species accumulate insoluble branched polysaccharide called cyanobacterial starch. Branching enzymes (BEs) belonging to glycoside hydrolase family 13 are universally found in the phylum cyanobacteria and are key enzymes in determining the branching pattern of polysaccharides. Many of the glycogen-producing cyanobacteria possess a single BE isozyme (BE1), while multiple BE isozymes (BE1, BE2, and BE3) are present in cyanobacterial starch-producing strains. A previous study suggested that the coexistence of three BE isozymes is essential for the trait of cyanobacterial starch-production. In this study, to obtain clues regarding the significance of the coexistence of the multiple isozymes, biochemical characterization using 11 purified recombinant BEs from both glycogen- and cyanobacterial starch-producing strains was performed. The BE1 and BE2 isozymes produced glucan chains with degree of polymerization (DP) 6 and 7 specifically, while BE3 isozymes produced short (DP 5-12) and long chains (DP 30-40) slightly. The BE1 and BE2 isozymes showed high activity, but those of BE3 isozymes were significantly low. The BE1 isozyme from cyanobacterial starch-producing Cyanobacterium sp. CLg1 showed markedly low activity. The BE1 and BE2 isozymes form cyanobacterial starch-producing Rippkaea orientalis PCC 8802 lacking BE3 isozyme shared similar reaction product specificity. These results suggested that the presence of the three isozymes is not essential and the roles of BE isozymes may vary depending on cyanobacterial species. These findings should deepen our understanding of the significance of BE isozymes in the biosynthesis of cyanobacterial starch.

Abbreviations

Amylose, synthetic amylose; BE, branching enzyme; cceBE1, BE1 from Crocosphaera subtropica ATCC 51142; Csp102756BE, BE from Cyanobacterium sp. NBRC 102756; Ca10605BE, BE from Cyanobacterium aponinum PCC 10605; CspCLg1BE, BE from Cyanobacterium sp. CLg1; DP, degree of polymerization; G6, maltohexaose; Gc7424BE, BE from Gloeothece citriformis PCC 7424; GH, glycoside hydrolase; Glc, glucose; OsBEI, BEI from Oryza sativa; OsBEIIa, BEIIa from Oryza sativa; OsBEIIb, BEIIb from Oryza sativa; Ro8802BE, BE from Rippkaea orientalis PCC 8802; SBS, surface binding site; Se7942BE, BE from Synechococcus elongatus PCC 7942; Ssp6803BE, BE from Synechocystis sp. PCC 6803

INTRODUCTION

Starch and glycogen are the major storage polysaccharides consisting of glucose units. Amylopectin, a major component of starch, is an orderly branched polysaccharide built up by α-1,4-linked glucose backbones containing branched chains linked by α-1,6-linkage at a rate of 4-6 %. The polymer consists of highly and scarcely branched regions that are tandemly repeated. The glucan chains in the scarcely branched regions form a double helix structure to show crystallinity and become water-insoluble. In contrast, glycogen is a water-soluble and randomly branched polysaccharide containing a high degree of α-1,6-linked branch points (7-10 %) [1]. Starch is accumulated by a limited number of eukaryotes having plastids originating from cyanobacteria (green algae, terrestrial plants evolved from green algae, red algae, and glaucophytes as well as secondary endosymbiosis derivatives such as alveolates and cryptophytes) [1, 2, 3]. In contrast, glycogen-accumulating organisms are widespread in the three domains of life [3].

Cyanobacteria are oxygenic photosynthetic prokaryotes and are predicted to exist in several thousand species [4]. They generally accumulate glycogen in their cells as a photosynthetic product, but surprisingly, despite being prokaryotes, several species (≤ 10) that accumulate semi-crystalline amylopectin-like polysaccharides, designated as cyanobacterial starch, have been discovered [5, 6, 7]. The physicochemical properties of cyanobacterial starch such as X-ray diffraction and gelatinization properties are almost identical to those of the starch accumulated by green plants [7]. The chloroplasts of green plants are derived from cyanobacteria through primary endosymbiosis, and thus their starch biosynthesis systems have much in common with each other, although the systems are considered to have evolved independently [1, 2, 3]. Biosynthesis of amylopectin in green plants requires the cooperative action of at least four types of enzymes: ADP-glucose pyrophosphorylase (AGPase, EC 2.7.7.27), starch synthase (EC 2.4.1.21), debranching enzymes (isoamylase, EC 3.2.1.68; pullulanase, EC 3.2.1.41), and branching enzyme (BE, EC 2.4.1.18) [8, 9]. Cyanobacteria have the same suite of enzymes as green plants. In bacteria, enzymes corresponding to starch synthase and isoamylase are generally called glycogen synthase (GS, EC 2.4.1.21) and GlgX (EC 3.2.1.68), respectively, and the latter naming is in line with the glycogen metabolism operon in Escherichia coli [10]. Physiological analyses using deficient mutants demonstrated that GS and GlgX participate in cyanobacterial starch biosynthesis [2, 3]. However, these analyses using BE-deficient mutants of cyanobacterial starch-producing strains have not yet been achieved.

BE catalyzes the intra- and/or inter-transglucosylation reaction to form a new branch point consisting of α-1,6-linkage by cleaving the glucosidic bond of α-1,4-linked backbone in the α-glucan [11, 12]. BE is a key enzyme in branched α-glucan biosynthesis because it determines the branching pattern. In the CAZy database (http://www.cazy.org/), BEs are classified into glycoside hydrolase (GH) family 13 (GH13) based on their primary structures [13]. Eukaryotic and prokaryotic GH13-type BEs, with minor exceptions, are further divided into GH13 subfamilies 8 (GH13_8) and 9 (GH13_9), respectively (Fig. 1) [12, 14, 15]. In addition to GH13-type BEs, some bacteria and archaea were found to have BEs belonging to GH family 57 [15, 16, 17, 18]. In the rice endosperm, two functionally distinct GH13_8-type BEs, BEI and BEII, the latter further diversified into BEIIa and BEIIb, are known to be involved in amylopectin biosynthesis [9].

Fig. 1. Phylogenetic tree of BEs from eukaryotes (GH13_8) and prokaryotes (GH13_9).

　The scientific names of the organisms are shown and the accession numbers for their amino acid sequences in NCBI are in parentheses. The 11 BE isozymes characterized in this study were indicated by black arrowhead. GlgX from E. coli (GH13_11) was used as an outgroup. The tree was constructed by the maximum likelihood method using programs from the Phylip package [40] and MEGA [41].

In cyanobacteria, glycogen-producing strains usually have a single GH13_9-type BE (BE1), with the exception of the glycogen-producing Gloeothece citriformis (formerly Cyanothece sp. [19]) PCC 7424, which has two GH13_9-type BE isozymes (BE1 and BE2). By contrast, cyanobacterial starch-producing strains have three GH13_9-type BE isozymes (BE1, BE2, and BE3) (Table 1, Fig. 1). The BE1 isozymes have N-terminal domain N, while it is absent in the BE2 and BE3 isozymes (Fig. 2). The BE1 and BE2 isozymes share high sequence identity (about 60 %), but the amino acid sequence of the BE3 isozyme is markedly different from them (< 30 % identity) (Fig. 1). The BE1 and BE2 isozymes have catalytic properties in common, but the BE3 isozyme has distinct properties [20, 21]. It is thought that amylopectin-producing trait of cyanobacteria is attributable to the coexistence of multiple isozymes with distinct catalytic properties, namely, BE1/BE2 and BE3 [20]. However, the relationship between BE orthologs and their catalytic properties remains obscure because BEs from only three cyanobacterial strains (one and two from glycogen- and cyanobacterial starch-producing strains, respectively) have been biochemically characterized to date (Table 1) [20, 21]. Recently, Rippkaea orientalis (formerly Cyanothece sp. [19]) PCC 8802 harboring only closely related isozymes, BE1 (Ro8802BE1) and BE2 (Ro8802BE2) (Fig. 1), was found to accumulate cyanobacterial starch [22, 23], implying that the coexistence of distantly related isozymes does not contribute to the structure of storage α-glucan in cyanobacteria.

Table 1. Cyanobacterial BE isozymes characterized to date and those used in this study.

Cyanobacterial strain	Polysaccharides	BE isozyme(s)	Reference
Synechococcus elongatus PCC 7942	Glycogen	Se7942BE1	Suzuki et al. 2015 [20]
Synechocystis sp. PCC 6803		Ssp6803BE1	El Mannai et al. 2021 [24]1 This work
Gloeothece citriformis (Cyanothece sp.) PCC 7424		Gc7424BE1, Gc7424BE2	This work
Rippkaea orientalis (Cyanothece sp.) PCC 8802	Cyanobacterial starch	Ro8802BE1, Ro8802BE2	El Mannai et al. 2021 [24]1 This work
Crocosphaera subtropica (Cyanothece sp.) ATCC 51142		cceBE1, cceBE2, cceBE3	Hayashi et al. 2015 [21]
Cyanobacterium sp. NBRC 102756		Csp102756BE1, Csp102756BE2, Csp102756BE3	Suzuki et al. 2015 [20]
Cyanobacterium sp. CLg1		CspCLg1BE1, CspCLg1BE2, CspCLg1BE3	This work
Cyanobacterium aponinum PCC 10605		Ca10605BE1, Ca10605BE2, Ca10605BE3	This work

¹ Catalytic properties are not characterized.

Fig. 2. (A) Domain structure of the cyanobacterial BE isozymes. (B) SDS-PAGE analysis (7.5 % acrylamide) of the purified recombinant BE isozymes.

　Lane M indicates the molecular mass markers. One microgram of samples was applied per lane.

In this study, a comprehensive characterization of orthologous BEs from both glycogen- and cyanobacterial starch-producing strains was performed. Purified recombinant BEs, three from glycogen-producing strains (Gc7424BE1 and Gc7424BE2 from G. citriformis strain 7424; Ssp6803BE from Synechocystis sp. PCC 6803) and eight from cyanobacterial starch-producing strains (Ro8802BE1 and Ro8802BE2 from R. orientalis strain 8802; CspCLg1BE1, CspCLg1BE2, and CspCLg1BE3 from Cyanobacterium sp. CLg1; and Ca10605BE1, Ca10605BE2, and Ca10605BE3 from Cyanobacterium aponinum PCC 10605), were prepared and characterized (Table 1). The catalytic properties of all recombinant BE isozymes characterized in this study were shared among the orthologs and were in accordance with the findings in previous studies [20, 21].

MATERIALS AND METHODS

Construction of expression plasmids.

The construction of expression plasmids harboring the genes sll0158 (NCBI accession No., AGF52369), Cyan8802_0465 (WP_012593825), and Cyan8802_2403 (WP_012595634), encoding Ssp6803BE, Ro8802BE1, and Ro8802BE2, respectively, were as described previously [24]. The Cyan7424_1226 (WP_012598620) and Cyan7424_4362 (WP_015956313) genes encoding Gc7424BE1 and Gc7424BE2, respectively, were amplified by PCR from genomic DNA isolated from G. citriformis strain 7424 as the template using the following primer sets. To obtain the Cyan7424_1226 gene (2,301 bp), the sense primer 5ʹ-GGGAATTCCATATGTCCACAACCATATCT-3ʹ and the antisense primer 5ʹ-CCCCAAGCTTTTACGAGGAAGATTGTTGTTTAACT-3ʹ were used (NdeI and HindIII sites are underlined). Throughout the study, PrimeSTAR^® HS DNA polymerase (Takara Bio Inc., Shiga, Japan) was used in PCR. The amplicons were digested with NdeI and HindIII and ligated into the corresponding restriction sites of the pColdI vector (Takara Bio Inc.). To obtain the Cyan7424_4362 gene (1,965 bp), the sense primer 5ʹ-CATATGGCTCAAAGTGAACTTATCCA-3ʹ (NdeI site is underlined) and the antisense primer 5ʹ-TCAATAGCTCATAATACCCAGT-3ʹ were used. The amplicons were cloned into pGEM^®-T Easy Vector (Promega Corporation, Madison, WI, USA) through TA cloning and transformed into E. coli DH5α cells (Toyobo Co., Ltd., Osaka, Japan). The plasmid was selected so that both ends of the Cyan7424_4362 gene were flanked by NdeI sites derived from the sense primer and the pGEM^®-T easy vector. The purified plasmid preparation (pGEM^®-T Easy/Cyan7424_4362) was digested with NdeI and ligated into the corresponding restriction site of the pColdI vector (Takara Bio Inc.).

The Cyan10605_1013 (AFZ53141), Cyan10605_0322 (AFZ52470), and Cyan10605_0759 (AFZ52892) genes encoding Ca10605BE1, Ca10605BE2, and Ca10605BE3, respectively, were amplified by PCR with the genomic DNA of C. aponinum strain 10605 as the template. The sense primer 5ʹ-CGCGGCAGCCATATGCCTAGCAAAATTACTATTGA-3ʹ and the antisense primer 5ʹ-GGCTTTGTTAGCAGCTCATTCTTCTTCTAAACT-3ʹ were used to obtain the Cyan10605_1013 gene (2,343 bp). The sense primer 5ʹ-CGCGGCAGCCATATGGTTATGGCACAAACA-3ʹ and the antisense primer 5ʹ-GGCTTTGTTAGCAGCTTATTTAACTTTTTTAAACA-3ʹ were used to obtain the Cyan10605_0322 gene (1,947 bp). The sense primer 5ʹ-CGCGGCAGCCATATGGTGAAAAAAGAAAATCAACA-3ʹ and the antisense primer 5ʹ-GGCTTTGTTAGCAGCCTAACCCTGAACTTGT-3ʹ were used to obtain the Cyan10605_0759 gene (1,953 bp). The 15 bases for In-Fusion^® cloning in these primers are underlined. Vector DNA fragment (5,696 bp) was prepared by PCR with the pET15b plasmid (Merck KGaA, Darmstadt, Germany) as the template using the following primer set: sense primer 5ʹ-GCTGCTAACAAAGCCCGA-3ʹ and antisense primer 5ʹ-CATATGGCTGCCGCGCGG-3ʹ. The amplified BE genes from C. aponinum strain 10605 were cloned into the above-mentioned pET15b vector using an In-Fusion^® HD cloning kit (Takara Bio Inc.).

The Clg1be1 (AFP43334), Clg1be2 (AFP43335), and Clg1be3 (AFP43336) genes encoding CspCLg1BE1, CspCLg1BE2, and CspCLg1BE3, respectively, were amplified by PCR with the genomic DNA of Cyanobacterium sp. strain CLg1 as the template. The sense primer 5ʹ-GAAGGTAGGCATATGACTAGTAAAATAAGCCTTGA-3ʹ and the antisense primer 5ʹ-CTTGAATTCGGATCCTTATTCTTCTTCTAAAGTAGCGA-3ʹ were used to obtain the clg1be1 gene (2,325 bp). The sense primer 5ʹ-GAAGGTAGGCATATGACTATGGCAGGA-3ʹ and the antisense primer 5ʹ-CTTGAATTCGGATCCTTATCTTTTTTTCTTTTTAAACA-3ʹ were used to obtain the clg1be2 gene (1,950 bp). The sense primer 5ʹ-GAAGGTAGGCATATGACTGTAAGCAAACA-3ʹ and the antisense primer 5ʹ-CTTGAATTCGGATCCTTACGATCGTGGCTTGAGT-3ʹ were used to obtain the clg1be3 gene (1,956 bp). The 15 bases for In-Fusion^® cloning in these primers are underlined. The vector DNA fragment (4,389 bp) was prepared by PCR with the pColdI plasmid (Takara Bio Inc.) as the template using the following primer set: sense primer 5ʹ-GGATCCGAATTCAAGCTTGT-3ʹ and antisense primer 5ʹ-CATATGCCTACCTTCGAT-3ʹ. The amplified DNA fragments carrying the BE genes from the stain CLg1 were ligated into the amplified pColdI vector fragment using the In-Fusion^® HD cloning kit (Takara Bio Inc.).

All of the plasmids constructed in this work contain a sequence encoding a His₆ tag and linker (20 and 16 amino acid residues for pET15b and pColdI, respectively) at the N-termini to facilitate the purification of recombinant proteins using immobilized metal-affinity chromatography. The nucleotide sequences of the BE genes in the constructed plasmids were confirmed using an ABI PRISM 3100xl genetic analyzer (Thermo Fisher Scientific Inc., Waltham, MA, USA).

Expression of BE genes in E. coli and purification of the gene products.

Expression of the BE genes in E. coli BL21 (DE3) cells (Merck KGaA) and purification of the gene products were performed in accordance with previously reported methods [20]. In brief, the E. coli transformants were cultured at 37 °C with agitation and expression of the genes was induced by adding 0.1 M isopropyl β-D-thiogalactopyranoside (IPTG) and further cultivated at 15 °C with the exception of CspCLg1BE1 (at 26 °C). The E. coli cells were harvested and disrupted by sonication and centrifuged. The recombinant enzymes were purified from the supernatants (crude sup) by Ni²⁺-affinity and gel-filtration column chromatography. To prevent degradation of the Gc7424BE2 by serine protease, 0.2 mM phenylmethylsulfonyl fluoride (PMSF) was added to its crude sup. The purified protein solutions were desalted and concentrated using an Amicon Ultra-15 (30,000 MWCO; Merck KGaA). The purity and homogeneity of the recombinant proteins were analyzed by SDS-PAGE. The protein concentrations were determined using a Bradford protein assay kit (Bio-Rad Laboratories, Inc., Hercules, CA, USA) with BSA (Bio-Rad Laboratories, Inc.) as the standard.

Characterization of the recombinant BEs

Enzymatic activities toward commercially available substrates amylopectin (amylopectin from potato; Merck KGaA) and amylose (synthetic amylose; Ezaki Glico Co., Ltd., Osaka, Japan) were measured by iodine-staining assay [20]. The reaction mixtures (100 μL) containing the recombinant BEs (0.5-15 μg) and 5 mg/mL amylopectin or 1 mg/mL amylose in 100 mM HEPES-NaOH buffer (pH 7.0) were incubated at 30 °C for various time periods (30-600 min). To stop the reactions, the mixtures were heated at 100 °C for 10 min and then cooled to room temperature. Subsequently, 100 μL of the iodine solutions [0.1 % (w/v) I₂/1 % (w/v) KI] were added and mixed well, and 10-μL aliquots were diluted by 10-fold with distilled water. The absorbance at 540 nm (for amylopectin) or 640 nm (for amylose) of the diluted solutions was measured using a spectrophotometer (DU 7400; Beckman Coulter, Inc., Fullerton, CA, USA). One unit of activity was defined as the number of the enzymes decreasing the absorbance by 0.1 per min at 30 °C. The thermal stability and pH optimum were measured as described previously [20]. Chain-length distribution analyses of the reaction products were performed using the P/ACE MDQ Carbohydrate System (Beckman Coulter, Inc.). The details of the methods are described in our previous paper [20].

RESULTS

Preparation of the purified recombinant enzymes.

Genes encoding Ssp6803BE, Gc7424BE1, Gc7424BE2, Ro8802BE1, Ro8802BE2, Ca10605BE1, Ca10605BE2, Ca10605BE3, CspCLgBE1, CspCLg1BE2, and CspCLg1BE3 were expressed as active enzymes in the E. coli BL21 (DE3) strain. Although the purified Gc7424BE2 undergoes proteolysis, the addition of PMSF to the crude sup enabled us to obtain the uncleaved protein. The majority of the produced CspCLg1BE1 became an inclusion body (99 %) and a meager amount of the soluble enzyme was obtained (1 %). An adequate amount of the soluble CspCLg1BE1 (about 50 %) was obtained using the pColdI vector and the E. coli BL21 (DE3) strain with IPTG induction (final concentration 0.1 mM) at 26 °C. All recombinant enzymes were purified to almost homogeneity and migrated as single bands in SDS-PAGE (Fig. 2).

Specific activities of the recombinant enzymes.

The specific activity of the cyanobacterial BEs on amylopectin and amylose was measured. The BE1 and BE2 isozymes showed remarkably higher specific activity against amylose (687-1,840 U/mg) than amylopectin (344-1,310 U/mg), with two exceptions (Table 2 and Fig. 3). One, no significant difference in the specific activities of Gc7424BE2 toward both polysaccharides was observed (687 and 660 U/mg for amylose and amylopectin, respectively). Two, CspCLg1BE1 displayed much lower specific activity values (67.9 and 38.9 U/mg for amylose and amylopectin, respectively) than the BE1 and BE2 orthologs, which had similar values to the BE3 orthologs (see below) (Table 2 and Fig. 3).

Table 2. Characteristics of cyanobacterial BE isozymes.

	Ssp6803BE	Gc7424BE1	Gc7424BE2	Ro8802BE1	Ro8802BE2	Ca10605BE1	Ca10605BE2	Ca10605BE3	CspCLg1BE1	CspCLg1BE2	CspCLg1BE3
Polysaccharides	Glycogen			Cyanobacterial starch
Optimal pH	7.0	7.0	8.0	7.0	8.0	7.5	8.0	8.0	7.5	8.0	8.0
Thermal stability (°C)	≤ 40	≤ 35	≤ 35	≤ 35	≤ 35	≤ 40	≤ 45	≤ 40	≤ 45	≤ 40	≤ 40
Specific activity (U/mg) amylopectin	344 ± 53.7	624 ± 11.3	660 ± 5.50	500 ± 8.20	1.31×103 ± 66.0	348 ± 19.9	737 ± 32.8	47.5 ± 1.90	38.9 ± 4.53	1.12×103 ± 103	47.7 ± 5.87
Specific activity (U/mg) amylose	704 ± 60.0	1.84×103 ± 118	687 ± 41.1	1.12×103 ± 102	1.63×103 ± 89.9	1.17×103 ± 101	1.32×103 ± 98.0	72.9 ± 4.05	67.9 ± 3.94	1.39×103 ± 54.5	27.0 ± 1.03
The type of chain-length distribution	BEIIb	BEIIb	BEIIb	BEIIb	BEIIb	BEIIb	BEIIb	BEI	BEIIb	BEIIb	BEI

Fig. 3. Specific activities of BE isozymes toward amylopectin from potato and synthetic amylose (both commercially available) are indicated by white and gray bars, respectively.

　Each value is the mean ± SE of at least three independent experiments.

Compared with the values for the BE1 and BE2 isozymes, values one order of magnitude lower values was observed for the BE3 isozymes: Ca10605BE3 (72.9 U/mg for amylose and 47.5 U/mg for amylopectin) and CspCLg1BE3 (27.0 U/mg for amylose and 47.7 U/mg for amylopectin) (Table 2 and Fig. 3). These values and the irregularity of the substrate specificity were in good agreement with our previous observations of BE3 isozymes from cyanobacterial starch-producing strains, Cyanobacterium sp. NBRC 102756 [20] and Crocosphaera subtropica (formerly Cyanothece sp. [19]) ATCC 51142 [21] (Table 1).

Thermal stabilities and pH optimum of the recombinant enzymes.

The thermal stabilities of the cyanobacterial BEs characterized in this study (35.0-45.0 °C) were close to the previously reported values for the BE orthologs (Table 2, Fig. S1; see J. Appl. Glycosci. Web site) [20]. These values were higher than the optimum temperatures of the BE isozymes from terrestrial plants [9, 25, 26, 27, 28, 29], indicating that cyanobacterial BEs are more thermostable. The BE1 orthologs displayed an optimum pH range from 7.0 to 7.5, but the BE2 and BE3 orthologs showed maximum activity at more alkaline conditions (pH 8.0) (Table 2, Fig. S2; see J. Appl. Glycosci. Web site). These values were in accordance with our previous report of BE isozymes from the glycogen-producing strain Synechococcus elongatus PCC 7942 and the cyanobacterial starch-producing Cyanobacterium sp. strain 102756 (Table 1) [20]. Distinct cyanobacterial BE isozymes may exhibit activities under different physiological conditions.

Chain-length distribution of the reaction product formed by the recombinant enzymes.

The chain-length distributions of reaction products formed by the cyanobacterial BE isozymes were analyzed using ae-amylopectin (amylopectin purified from rice endosperm of BEIIb-deficient amylose extender mutant [30]) as a substrate. The BE1 and BE2 isozymes mainly produced the glucan chains with degrees of polymerization (DP) 6 and 7, and a much smaller proportion of DP10 (Figs. 4A-I). Gc7424BE2 exceptionally produced proportions of glucan chains similar to those of DP6, 7, and 10 (Fig. 4F). Meanwhile, no specific production of glucan chains was observed for the BE3 isozymes and they produced a low level of long glucan chains with a DP of about 30-40 (Figs. 4J and K). These patterns were almost identical to those produced by the orthologous BE isozymes from the glycogen-producing S. elongatus strain 7942 and the cyanobacterial starch-producing strains (Cyanobacterium sp. strain 102756 and C. subtropica strain 51142) (Table 1) [20, 21].

Fig. 4. Chain-length distributions of the reaction products formed by the cyanobacterial BE isozymes [(A) Ssp6803BE, (B) Gc7424BE1, (C) Ro8802BE1, (D) Ca10605BE1, (E) CspCLg1BE1, (F) Gc7424BE2, (G) Ro8802BE2, (H) Ca10605BE2, (I) CspCLg1BE2, (J) Ca10605BE3, and (K) CspCLg1BE3] using ae-amylopectin as the substrate.

　The horizontal and vertical axes represent the DP of each chain and the molar percentage of each chain, respectively. Black bars indicate differences in the chain-length distributions of ae-amylopectin before and after the reaction with the BE isozymes. Gray lines indicate the chain-length distributions of unreacted ae-amylopectin. Representative data from three replicate experiments using independent preparations are shown. The standard errors for each molar percentage of DP5-50 between the triplicate experiments were less than 1.2 %.

DISCUSSION

This study biochemically characterized 11 cyanobacterial BE isozymes in order to understand the significance of the coexistence of the multiple isozymes. BE1 isozymes share high primary structure identity (≥ 67 %), but amino acid alignment showed that CspCLg1BE1 has specific amino acid variations (Q67, C99, R127, Y168, H356, Y383, C411, L455, E499, D534, C544, S565, and A601) (Fig. 5). The finding that CspCLg1BE1 had markedly lower specific activity than the corresponding orthologs could not be accounted for by the amino acid sequence and its 3-D structure model predicted by AlphaFold DB (UniProt: J7FLF1) [31]. The unusual property of CspCLg1BE1 may indicate that contribution of the BE1 isozyme in production of cyanobacterial starch became minor in Cyanobacterium sp. CLg1 compared with other strains. Alternatively, it is probable that the role of CspCLg1BE1 is complemented by the BE2 isozyme (CspCLg1BE2) showing nearly identical catalytic specificity (Figs. 4E and I). Together, Cyanobacterium sp. CLg1 is capable of producing cyanobacterial starch by single high activity isozyme (CspCLg1BE2) and two low activity isozymes (CspCLg1BE1 and CspCLg1BE3). Our results may imply that contribution of the BE isozymes in cyanobacterial starch biosynthesis varies depending on cyanobacterial species.

Fig. 5. Multiple sequence alignments of cyanobacterial BE1 orthologs around specific amino acid variations in CspCLg1BE1.

　The specific residues are indicated by an arrow and labeled. The alignment was generated using the programs clustalW [42] and ESPript 3.0 [43].

The specific production of glucan chains with DP6 and DP7 by the cyanobacterial BE1 and BE2 isozymes was common among the orthologs (Figs. 4A-I). Although the reaction product specificity of Gc7424BE2 was essentially the same as that of the other BE1 and BE2 isozymes, it produced DP6, 7, and 10 at similar proportions (9.3 %, 11.6 %, and 10.7 %, respectively) (Fig. 4F). This unusual property of Gc7424BE2 could be explained by the substrate binding manner of the enzyme surface. It has been proposed that the structure of the active site cleft (subsites) determines the length of the glucan chains to be transferred by BEs, as described in the crystal structure of the BE1 isozyme from C. subtropica strain 51142 (cceBE1) [32]. The cceBE1 has subsites −1 to −6 and a “stopper structure” at the nonreducing end side (subsite −6) to mainly produces DP6 (Fig. 6A). The maltohexaose (G6) molecule bound to the subsites of cceBE1 (PDB: 5GQV) was superimposed onto the structural model of Gc7424BE2 predicted by AlphaFold DB (UniProt: B7K8V9). The Gc7424BE2 predicted model also had stopper structure, and the G6 molecule was fitted without any steric hindrance (Fig. 6B). Gc7424BE2 harbored the following unique amino acid variations: V417, H468, P469, H470, Y528, and Y531 (Fig. 6C). No specific residues were found around the subsites and as such could not account for the reduction of glucan chains with DP6 and 7 produced. A plausible explanation for this reduction based on another factor is discussed below.

Fig. 6. (A) Sugar-binding structure of the active site cleft in cceBE1 (PDB: 5GQV). The bound maltohexaose (G6) molecule is shown as an orange ball and stick model. The reducing end of the glucan chain is marked with φ. Subsites (–6 to –1) are labeled. The stopper structure is depicted by a dashed circle. The figure was drawn using the program PyMOL (Schrödinger, LLC, New York, NY, USA). (B) Sugar-binding model of the active site cleft in Gc7424BE2. The structure model of Gc7424BE2 was generated by AlphaFold DB (UniProt: B7K8V9). The model was superimposed onto the sugar-binding structure of cceBE1 (PDB: 5GQV), and the G6 molecule in the active site cleft is shown. (C) Multiple alignment of the primary structures of the cyanobacterial BE isozymes. Specific amino acid variations in Gc7424BE2 are indicated by arrows and labeled. (D) Schematic drawing of the donor substrate-binding model of cceBE1. The active site cleft (subsites) and donor/acceptor substrate binding site are colored dark gray and white, respectively. Glc units bound to the subsites and those bound to the substrate binding site and SBS C1 are shown as black and gray hexagons, respectively. The Glc units at the reducing end are indicated by slash marks. The subsites are labeled from –6 to –1. The residual chain is numbered from the branch point (1 to 6) and the subsites (–6 to –1) are labeled. (E) Schematic drawing of the donor substrate binding model of Gc7424BE2. The subsites are labeled from –6 to –1 and the residual chain (1 to 10) is numbered from the branch point. (F) Sugar-binding model around the SBS C1 in Gc7424BE2. The Gc7424BE2 model is superimposed onto the sugar-binding structure of cceBE1 and the maltotetraose molecule in the SBS C1 is shown. Residues specific for Gc7424BE2 are labeled and are colored in magenta.

In the first reaction of BE (glycosylation reaction), the donor substrate is cleaved, the chain to be transferred is covalently bound to the nucleophile (covalent glycosyl-enzyme intermediate), and the residual glucan chains after cleavage by the glycosylation reaction are liberated from the enzyme [12]. The BE reaction requires longer glucan chains with DP ≥ 12 [33], so the residual chains consist of DP ≤ 11. The transferred and residual chains in the BE reaction products are reflected in the chain-length distribution pattern because the entire glucan chains in the reaction mixture are detected. It is thought that the residual chains formed by cceBE1 are mainly DP6 and DP7 (Fig. 6D) [32]. In the case of Gc7424BE2, assuming that the residual chains with DP6 and 7 are decreased and DP8-11 are abundant (especially DP10), the decrease in DP6-7 and increase in DP10 in the chain-length distribution can be explained (Figs. 4F and 6E).

Interestingly, five amino acid variations were positioned at the surface binding sites (SBSs) C1 (H468, P469, and H470) and C3 (Y528 and Y531) of Gc7424BE2 (Fig. 6C). The SBSs are located far from the active site and play multiple roles in the actions of the enzymes [34, 35, 36, 37]. In the crystal structure of cceBE1, five SBSs (A1 and A2 in domain A; C1, C2, and C3 in domain C) were found in our previous study [24, 32]. Specific residues in SBS C1 of Gc7424BE2 may affect the manner of binding of the Glc unit at the reducing end side in bound maltotetraose (Figs. 6E and F), while those in SBS C3 do not. It is possible that the decrease in the specific activity for amylose is caused by the specific residues in SBS C1 (Fig. 3), although mutational studies of the residues are needed.

The BE3 isozymes, CspCLg1BE3 and Ca10605BE3, displayed no specific production of the glucan chains (Figs. 4J and K). The amino acid sequence alignment and 3-D structural model of CspCLg1BE3 predicted by AlphaFold DB (UniProt: J7FN11) showed that the stopper structure conserved in the BE1 and BE2 isozymes was absent in the BE3 isozymes (Figs. 7A and B). We considered that the absence of the stopper structure enabled the BE3 isozymes to bind various length of glucan chains in the active site cleft, and thus glucan chains with DP5-12 were produced nonspecifically. Furthermore, branched glucan chains containing long B₂ chains were likely to be transferred since glucan chains with about DP30-40 were produced (Figs. 4J and K). The 6-OH group of Glc units bound at subsites −2, −3, −4, and −5 was in a position accessible to a solvent (Fig. 7B), supporting the assertion that the BE3 isozymes are capable of transferring the long α-1,6-branched glucan chains. Such product specificity is common among the cyanobacterial BE3 orthologs characterized to date (Table 1) [20, 21].

Fig. 7. (A) Multiple amino acid sequence alignment of cyanobacterial BE isozymes around the stopper structure. (B) Sugar-binding model of the active site cleft in CspCLg1BE3. The structural model of CspCLg1BE3 was generated by AlphaFold DB (UniProt: J7FN11), and the model was superimposed onto the sugar-binding structure of cceBE1 (PDB: 5GQV). The superimposed CspCLg1BE3 model and the G6 molecule in the subsites of cceBE1 are shown. The subsites (–6 to –1) are labeled. Solvent-accessible 6-OH groups are indicated by white triangles. The absence of the stopper structure is depicted by a dashed circle.

The properties of the GH13_8-type recombinant BEs from rice have been well studied. Recombinant rice BEI (OsBEI) produced small amounts of glucan chains with DP6-15 and DP26-39. In contrast, recombinant rice BEIIa (OsBEIIa) and BEIIb (OsBEIIb) mainly produce glucan chains with DP6 and 7 and they produce a small amount of DP8-15, although the proportion of short glucan chains (DP8-15) produced by OsBEIIa is markedly higher than that produced by OsBEIIb [9]. The chain-length distributions of the reaction products produced by the cyanobacterial BE1 and BE2 isozymes were similar to those of OsBEIIb, but the reaction products of the cyanobacterial BE3 isozymes showed a chain-length profile similar to that of OsBEI. The former and latter are therefore grouped into the BEIIb-type and BEI-type, respectively (Table 2) [20, 38].

A previous study using cyanobacterial recombinant enzymes suggested that the coexistence of multiple BE isozymes with distinct catalytic properties, BE1/BE2 (BEIIb-type) and BE3 (BEI-type), is essential for the cyanobacterial starch-producing trait [20]. The low specific activity of BE3 isozymes was remarkably similar among the presently characterized orthologs (Table 1) [20, 21]. This suggests that mutations may occurred in residues important for activity of the BE3 orthologs, resulting in minor contribution of the BE3 isozymes in cyanobacterial starch biosynthesis. Another possibility is that the nonphysiological substrates (commercially available amylopectin from potato and synthetic amylose) may lead to the low specific activities and the BE3 isozymes may act under physiological conditions that differ from those of the other paralogs. The cyanobacterial starch-producing R. orientalis strain 8802 has only two isozymes (Ro8802BE1 and Ro8802BE2) and lacks the BE3 isozyme (Table 1). Chain-length distribution analyses of the reaction products showed that Ro8802BE1 and Ro8802BE2 shared similar reaction product specificity (BEIIb-type) (Table 2, Figs. 4C and G). These findings indicate that the coexistence of both BE1/BE2 (BEIIb-type) and BE3 (BEI-type) isozymes is not always necessary for the cyanobacterial starch biosynthesis, even though green plants require both BEI and BEII isozymes to generate the fine structure of amylopectin [39]. Considering the variety of BE1 (e.g., CspCLg1BE1) and BE3 isozymes, contribution of the BE1 and BE3 isozymes in biosynthesis of cyanobacterial starch likely depends on the strains of cyanobacteria. In the cyanobacterial starch-producing strains, it remains unknown whether genes of BE isozymes are actually expressed and whether their products are functional because neither the expression levels of the genes nor the activities of their products have been analyzed. Further study, such as physiological analysis using a deficient mutant, will be needed to elucidate the role of BE isozymes in cyanobacterial starch production.

CONFLICT OF INTEREST

The authors declare no conflicts of interest.

ACKNOWLEDGMENTS

The authors would like to thank Drs. Christophe Colleoni and Steven G. Ball for kindly donating the strain Cyanobacterium sp. CLg1. The authors also thank Drs. Naoko Fujita, Naoko Crofts, and Satoko Miura for valuable discussions. The authors are also grateful to Nagase Viita Co., Ltd. (Okayama, Japan) for providing Pseudomonas isoamylase. The authors thank Enago (www.enago.jp) for the English language review. This work was supported by grants from JSPS KAKENHI 15K18685 (to R. S.), 16K07467 (to E. S.), and 18K06135 (to R. S.).

REFERENCES

• [1]  Ball S, Colleoni C, Cenci U, Raj JN, Tirtiaux C. The evolution of glycogen and starch metabolism in eukaryotes gives molecular clues to understand the establishment of plastid endosymbiosis. J Exp Bot. 2011; 62: 1775-801.
• [2]  Cenci U, Chabi M, Ducatez M, Tirtiaux C, Nirmal-Raj J, Utsumi Y, et al. Convergent evolution of polysaccharide debranching defines a common mechanism for starch accumulation in cyanobacteria and plants. Plant Cell. 2013; 25: 3961-75.
• [3]  Kadouche D, Ducatez M, Cenci U, Tirtiaux C, Suzuki E, Nakamura Y, et al. Characterization of function of the GlgA2 glycogen/starch synthase in Cyanobacterium sp. Clg1 highlights convergent evolution of glycogen metabolism into starch granule aggregation. Plant Physiol. 2016; 171: 1879-92.
• [4]  Nabout JC, da Silva Rocha B, Carneiro FM, Sant’Anna CL. How many species of cyanobacteria are there? Using a discovery curve to predict the species number. Biodivers Conserv. 2013; 22: 2907-18.
• [5]  Nakamura Y, Takahashi J, Sakurai A, Inaba Y, Suzuki E, Nihei S, et al. Some cyanobacteria synthesize semi-amylopectin type α-polyglucans instead of glycogen. Plant Cell Physiol. 2005; 46: 539-45.
• [6]  Deschamps P, Colleoni C, Nakamura Y, Suzuki E, Putaux JL, Buléon A, et al. Metabolic symbiosis and the birth of the plant kingdom. Mol Biol Evol. 2008; 25: 536-48.
• [7]  Suzuki E, Onoda M, Colleoni C, Ball S, Fujita N, Nakamura Y. Physicochemical variation of cyanobacterial starch, the insoluble α-glucans in cyanobacteria. Plant Cell Physiol. 2013; 54: 465-73.
• [8]  Nakamura Y. Towards a better understanding of the metabolic system for amylopectin biosynthesis in plants: rice endosperm as a model tissue. Plant Cell Physiol. 2002; 43: 718-25.
• [9]  Nakamura Y, Utsumi Y, Sawada T, Aihara S, Utsumi C, Yoshida M, et al. Characterization of the reactions of starch branching enzymes from rice endosperm. Plant Cell Physiol. 2010; 51: 776-94.
• [10]  Preiss J. Glycogen: biosynthesis and regulation. EcoSal Plus. 2009; 3.
• [11]  Tetlow IJ, Emes MJ. A review of starch-branching enzymes and their role in amylopectin biosynthesis. IUBMB Life. 2014; 66: 546-58.
• [12]  Suzuki R, Suzuki E. Structure and function of branching enzymes in eukaryotes. Trends Glycosci Glycotechnol. 2020; 32: E21-30.
• [13]  Drula E, Garron ML, Dogan S, Lombard V, Henrissat B, Terrapon N. The carbohydrate-active enzyme database: functions and literature. Nucleic Acids Res. 2022; 50: D571-7.
• [14]  Stam MR, Danchin EG, Rancurel C, Coutinho PM, Henrissat B. Dividing the large glycoside hydrolase family 13 into subfamilies: towards improved functional annotations of alpha-amylase-related proteins. Protein Eng Des Sel. 2006; 19: 555-62.
• [15]  Suzuki E, Suzuki R. Distribution of glucan-branching enzymes among prokaryotes. Cell Mol Life Sci. 2024; 73: 2643-60.
• [16]  Murakami T, Kanai T, Takata H, Kuriki T, Imanaka T. A novel branching enzyme of the GH-57 family in the hyperthermophilic archaeon Thermococcus kodakaraensis KOD1. J Bacteriol. 2006; 188: 5915-24.
• [17]  Santos CR, Tonoli CC, Trindade DM, Betzel C, Takata H, Kuriki T, et al. Structural basis for branching-enzyme activity of glycoside hydrolase family 57: structure and stability studies of a novel branching enzyme from the hyperthermophilic archaeon Thermococcus kodakaraensis KOD1. Proteins. 2011; 79: 547-57.
• [18]  Na S, Park M, Jo I, Cha J, Ha NC. Structural basis for the transglycosylase activity of a GH57-type glycogen branching enzyme from Pyrococcus horikoshii. Biochem Biophys Res Commun. 2017; 484: 850-6.
• [19]  Mareš J, Johansen JR, Hauer T, Zima J Jr, Ventura S, Cuzman O, et al. Taxonomic resolution of the genus Cyanothece (Chroococcales, Cyanobacteria), with a treatment on Gloeothece and three new genera, Crocosphaera, Rippkaea, and Zehria. J Phycol. 2019; 55: 578-610.
• [20]  Suzuki R, Koide K, Hayashi M, Suzuki T, Sawada T, Ohdan T, et al. Functional characterization of three (GH13) branching enzymes involved in cyanobacterial starch biosynthesis from Cyanobacterium sp. NBRC 102756. Biochim Biophys Acta. 2015; 1854: 476-84.
• [21]  Hayashi M, Suzuki R, Colleoni C, Ball S, Fujita N, Suzuki E. Crystallization and crystallographic analysis of branching enzymes from Cyanothece sp. ATCC 51142. Acta Crystallogr F Struct Biol Commun. 2015; 71: 1109-13.
• [22]  Welkie DG, Lee BH, Sherman LA. Altering the structure of carbohydrate storage granules in the cyanobacterium Synechocystis sp. strain PCC 6803 through branching-enzyme truncations. J Bacteriol. 2016; 198: 701-10.
• [23]  Bandyopadhyay A, Elvitigala T, Welsh E, Stöckel J, Liberton M, Min H, et al. Novel metabolic attributes of the genus Cyanothece, comprising a group of unicellular nitrogen-fixing cyanobacteria. mBio. 2011; 2: e00214-11.
• [24]  El Mannai Y, Deto R, Kuroki M, Suzuki R, Suzuki E. Cyanobacterial branching enzymes bind to α-glucan via surface binding sites. Arch Biochem Biophys. 2021; 702: 108821.
• [25]  Drummond GS, Smith EE, Whelan WJ. Purification and properties of potato α-1,4-glucan. α-1,4-Glucan 6-glycosyltransferase (Q-enzyme). Eur J Biochem. 1972; 26: 168-76.
• [26]  Kuriki T, Stewart DC, Preiss J. Construction of chimeric enzymes out of maize endosperm branching enzymes I and II: activity and properties. J Biol Chem. 1997; 272: 28999-9004.
• [27]  Guan HP, Li P, Imparl Radosevich J, Preiss J, Keeling P. Comparing the properties of Escherichia coli branching enzyme and maize branching enzyme. Arch Biochem Biophys. 1997; 342: 92-8.
• [28]  Wychowski A, Bompard C, Grimaud F, Potocki-Véronèse G, D’Hulst C, Wattebled F, et al. Biochemical characterization of Arabidopsis thaliana starch branching enzyme 2.2 reveals an enzymatic positive cooperativity. Biochimie. 2017; 140: 146-58.
• [29]  Hamada S, Ito H, Hiraga S, Inagaki K, Nozaki K, Isono N, et al. Differential characteristics and subcellular localization of two starch-branching enzyme isoforms encoded by a single gene in Phaseolus vulgaris L. J Biol Chem. 2002; 277: 16538-46.
• [30]  Nishi A, Nakamura Y, Tanaka N, Satoh H. Biochemical and genetic analysis of the effects of amylose-extender mutation in rice endosperm. Plant Physiol. 2001; 127: 459-72.
• [31]  Jumper J, Evans R, Pritzel A, Green T, Figurnov M, Ronneberger O, et al. Highly accurate protein structure prediction with AlphaFold. Nature. 2021; 596: 583-9.
• [32]  Hayashi M, Suzuki R, Colleoni C, Ball SG, Fujita N, Suzuki E. Bound substrate in the structure of cyanobacterial branching enzyme supports a new mechanistic model. J Biol Chem. 2017; 292: 5465-75.
• [33]  Jeon JS, Ryoo N, Hahn TR, Walia H, Nakamura Y. Starch biosynthesis in cereal endosperm. Plant Physiol Biochem. 2010; 48: 383-92.
• [34]  Nielsen MM, Bozonnet S, Seo ES, Mótyán JA, Andersen JM, Dilokpimol A, et al. Two secondary carbohydrate binding sites on the surface of barley α-amylase 1 have distinct functions and display synergy in hydrolysis of starch granules. Biochemistry. 2009; 48: 7686-97.
• [35]  Cuyvers S, Dornez E, Delcour JA, Courtin CM. Occurrence and functional significance of secondary carbohydrate binding sites in glycoside hydrolases. Crit Rev Biotechnol. 2012; 32: 93-107.
• [36]  Wilkens C, Cockburn D, Andersen S, Petersen BO, Ruzanski C, Field RA, et al. Analysis of surface binding sites (SBS) within GH62, GH13, and GH77. J Appl Glycosci. 2015; 62: 87-93.
• [37]  Cockburn D, Nielsen MM, Christiansen C, Andersen JM, Rannes JB, Blennow A, et al. Surface binding sites in amylase have distinct roles in recognition of starch structure motifs and degradation. Int J Biol Macromol. 2015; 75: 338-45.
• [38]  Sawada T, Nakamura Y, Ohdan T, Saitoh A, Francisco PB Jr, Suzuki E, et al. Diversity of reaction characteristics of glucan branching enzymes and the fine structure of α-glucan from various sources. Arch Biochem Biophys. 2014; 562: 9-21.
• [39]  Satoh H, Nishi A, Yamashita K, Takemoto Y, Tanaka Y, Hosaka Y, et al. Starch-branching enzyme I-deficient mutation specifically affects the structure and properties of starch in rice endosperm. Plant Physiol. 2003; 133: 1111-21.
• [40]  Felsenstein J. PHYLIP-Phylogeny inference package (Ver. 3.2). Cladistics. 1989; 5: 164-6.
• [41]  Tamura K, Stecher G, Kumar S. MEGA11: Molecular evolutionary genetics analysis version 11. Mol Biol Evol. 2021; 38: 3022-7.
• [42]  Larkin MA, Blackshields G, Brown NP, Chenna R, McGettigan PA, McWilliam H, et al. Clustal W and Clustal X version 2.0. Bioinformatics. 2007; 23: 2947-8.
• [43]  Robert X, Gouet P. Deciphering key features in protein structures with the new ENDscript server. Nucleic Acids Res. 2014; 42: W320-4.