Introduction

Cyanobacteria, among the Earth’s oldest organisms, have evolved sophisticated biosynthetic pathways to produce a rich source of bioactive natural products. Cyanobacterial strains belonging to the genera Microcystis, Anabaena, Nostoc, Oscillatoria, Nodularia, Aphanizomenon, and Cylindrospermum are known to produce a number of cyclic peptide hepatotoxins and alkaloid neurotoxins exhibiting algicidal, fungicidal, pesticidal, cytotoxic, immunosuppressive, and enzyme-inhibiting activities (Namikoshi and Rinehart 1996; Prasanna et al. 2010b; Jaiswal et al. 2008). A majority of these metabolites are biologically active and are products of either the nonribosomal polypeptide (NRP) or the mixed polyketide-NRP biosynthetic pathways. However, to date, limited information is available on the chemical nature of fungicidal metabolites in cyanobacteria and the gene(s) involved.

Cellulose is the major plant cell wall polysaccharide which is degraded by cellulases. This degradation is thought to be achieved by the synergistic action of three members of the cellulase complex (Henrissat et al. 1985): endoglucanases (EC 3.2.1.4; endo-β-1,4-glucanases), cellobiohydrolases (EC 3.2.1.91) and β-glucosidases (EC 3.2.1.21). The endoglucanases are ubiquitous enzymes that hydrolyze β-1,4-linkages adjacent to unsubstituted glucose residues (Henrissat et al. 1989). These are produced by a broad range of organisms, including fungi (Wood 1992), bacteria (Beguin et al. 1992), plants (Ohmiya et al. 1995), and insects (termites) (Inoue et al. 1997). Endoglucanases belong to more than 16 GH (glycoside hydrolase) families including 5, 6, 7, 8, 9, 12, 16, 17, 44, 45, 48, 51, 55, 61, 74, and 81 on the continually updated carbohydrate active enzymes (CAZy) server (http://afmb.cnrs-mrs.fr/CAZY; Coutinho and Henrisaat 1999). Watson et al. (2009) investigated the GH5-endoglucanase in the marine bacterium Saccharophagus degradans which had a high level of ability to degrade cellulose. Posta et al. (2004) identified a new endoglucanase encoding gene (cel 5b) from Thermobifida fusca which is also associated with GH5 family. Endoglucanases of the GH5 and GH61 families from Aspergillus kawachii (AkCel5A and AkCel61A; originally termed as Cel5A and Cel61A, respectively) are modular enzymes and consist of a catalytic domain (Hara et al. 2003).

Cellulases are also known as bifunctional enzymes and there have been many reports on the hydrolyzing characteristics of cellulase on chitosan (Ike et al. 2007; Liu and Xia 2005). Endo-β-1,4-glucanases have not been identified for the catalytic function in relation to fungicidal activity till now because in most of the fungi except oomycetes, fungal cell wall is composed only of β-1,3-glucan. Most of the reports have focussed on endo-1,3-β-d-glucanases for fungicidal activity against pathogenic fungi such as Rhizoctonia solani and Fusarium oxysporum (Shi et al. 2010).

Our earlier investigations revealed for the first time, the activity of several hydrolytic enzymes in Anabaena strains and their role in fungicidal activity (Prasanna et al. 2008). Chitosanase homolog(s) were also identified in Anabaena laxa (RPAN8) and Anabaena iyengarii (RPAN9) (Prasanna et al. 2010b). In the present study, we have characterized putative genes (end 1 and end 2) encoding β-1,4-endoglucanase belonging to peptidase M20 family and glycoside hydrolase family 5 (GH5) in A. laxa (RPAN8) in relation to their fungicidal activity .

Materials and methods

Anabaena laxa (RPAN8; CCC 381 in Culture Collection of Centre for Conservation and Utilization of Blue Green Algae, Indian Agricultural Research Institute, New Delhi) exhibiting fungicidal activity was used in this investigation. The fungal strain Pythium aphanidermatum (strain no. MTCC123) was obtained from the Indian Type Culture Collection, Division of Plant Pathology, Indian Agricultural Research Institute, New Delhi. The cyanobacterial strain was axenized by standard procedures employing a set of antibiotics (Kaushik 1987). Subculturing was done in nitrogen-free BG-11 medium (Stanier et al. 1971) and grown under light: dark cycle (16:8 h) under white light (50–55 µmol photons m−2 s−1) and 28 ± 2°C. The fungal cultures were grown at 28°C in a BOD incubator and maintained on Potato Dextrose Agar.

CMCase (endo-β-1,4-glucanase) activity in the culture filtrates was measured spectrophotometrically using carboxymethyl cellulose as the substrate (Ghosh et al. 1983). Reducing sugars liberated were estimated at 575 nm against the standard curve of glucose. Laminarinase (endo-β-1, 3-glucanase) activity was measured in a reaction mixture of 1 mL citrate buffer (pH 5.1) containing 1.6 mg soluble laminarin as a substrate. The samples were incubated at 30°C for 1 h and stopped by placing the reaction mixture in boiling water bath for 2 min. The reducing sugars released were estimated using the dinitrosalicylic acid (DNSA) reagent (Miller 1959). The enzyme activity was measured as µmol of reducing sugars (d-glucose) released per min per mL of culture filtrate. All values were recorded in triplicate and expressed in terms of IU mL−1. One unit of enzyme represents µmol of reducing sugars (d-glucose) released per min per mL of culture filtrate.

Genomic library was constructed using pUC19/Sma I digested vector (Fermentas, USA). The genomic DNA (10 µg) was isolated using Plant Ultra Clean DNA isolation kit (MoBio, USA) and nebulized using argon at 30 psi for 7 s. The nebulized DNA was precipitated and end-repaired using Klenow fragment and T4 DNA polymerase according to standard protocols (Sambrook et al. 1989). End-repaired DNA was run on a 0.8% gel and gel containing 2–4-kb fragment was excised out. The DNA fragments were eluted out using Gel/PCR DNA fragment extraction kit (Geneaid, India) and ligated into pUC19/Sma I vector. Transformation was carried out into E. coli JM109 (DE3). All positive colonies were picked and maintained in Luria Bertani freeze buffer. Plasmid isolation was done using High-Speed Plasmid Mini kit (Geneaid, India).

Fungicidal activity of positive clones

The positive clones were induced using 0.4 mM IPTG in the LB medium. The overnight-grown cultures of positive clones were directly used to evaluate fungicidal activity (in terms of zone of inhibition) using disk diffusion assay (as discussed above) on the lawn of P. aphanidermatum as test organism (Prasanna et al. 2008). The activity was measured after 3–4 days of incubation at 30°C. The insertless vector transformed E. coli was used as a negative control and nystatin (100 U) was used as positive control which produced an average zone of inhibition of 20 ± 5 mm with P. aphanidermatum.

Functional analysis of clones for β-1,4 endoglucanase activity was done using 1% carboxymethyl cellulose (CMC) as a substrate (Teather and Wood 1982) by the method outlined by Geelen et al. (1995). β-1,3-endoglucanase activity was evaluated using the same method as used for β-1,4-endoglucanase except that filter-sterilized laminarin (0.2%) was used as a substrate. The clones from A. laxa RPAN8 were induced using 0.4 mM IPTG in the medium and directly used for β-1,4 and β-1,3-endoglucanase activities. The insertless vector transformed E. coli was also used as a negative control.

An overnight-grown culture of all the positive clones was diluted 1:20 into fresh Luria–Bertani medium and grown until the A 660 was 0.6. Synthesis of β-1,4 and β-1,3-endoglucanases were induced with 1 mM IPTG for 3 h. E. coli cells were harvested by centrifugation, resuspended in 50 mM sodium acetate buffer (pH 5.9), and sonicated. Cell debris was removed by centrifugation (10,000×g for 15 min), and the supernatant was used as the enzyme source to evaluate the fungicidal activity, β-1,4 and β-1,3-endoglucanase activities using the same methods outlined above. The insertless vector transformed E. coli served as a negative control.

Bioinformatic and statistical tools

The fungicidal and endoglucanase activity was recorded in triplicate and subjected to ANOVA in accordance with the experimental design (completely randomized block design) using MSTAT-C statistical package to quantify and evaluate the source of variation. BLASTN and BLASTP were used to evaluate nucleotide identity and amino acid sequence similarity, respectively (Altschul et al. 1990). ORFs were identified using GENDB (Meyer et al. 2003). Nucleotide and amino acid sequences were aligned using Clustal W2 (Larkin et al. 2007). The conserved domains and motifs were identified using the NCBI Conserved Domain Tool and Motif search tool respectively. The prediction of signal peptides and cleavage sites in the translated sequence was obtained using Neural Network and HMM (Hidden Markov Model) algorithms (Bendtsen et al. 2004).

The nucleotide sequences described in this study have been deposited in Genbank under the accession numbers: GU593982 (end 1) and GU593983 (end 2).

Results

Measurement of β-1,4 and β-1,3 endoglucanase activity

The β-1,3 and 1,4-endoglucanase activities in RPAN8 was measured using laminarin and CMC as substrates and values of 0.23 and 0.315 U mL−1, respectively, were recorded in the 28-day culture (Supplementary Fig. 1).

Screening of positive clones for fungicidal activity

The construction of genomic library of RPAN8 in pUC19/Sma I produced nearly 2,000 clones of 2–4 kb each. All the clones were spotted on the lawn of P. aphanidermatum in order to screen for fungicidal activity. Four E. coli clones showed fungicidal activity. The inhibition zones of 10 and 14 mm was recorded in clone numbers 116 and 248 and 8 and 64, respectively. No inhibition zone was observed (Table 1) in the negative control (insertless vector transformed E. coli)

Table 1 Evaluation of fungicidal activity (in terms of zone of inhibition in mm), β-1,3 and 1,4 endoglucanase activities (IU mL−1) using cell-free filtrates of clones
Full size table

Mode of action of fungicidal gene(s)

The microscopic analyses of the inhibition zone was undertaken in order to determine the mode of action of the expressed gene. Microscopic observations revealed disintegration of mycelia of P. aphanidermatum whereas abnormalities in the hyphal structure leading to the curling of mycelia were observed in the positive control (nystatin).

Identification of endoglucanase activity in genomic library clones

The biochemical assay for endoglucanase activity was undertaken by streaking each of the four E coli clones on CMC and laminarin plate in order to know the expression of the clones. The expression of E .coli clones was analyzed on CMC plate in terms of change in color of the streaked region from red to orange. The A. laxa clones showed less activity on the laminarin plate. The two clones (8 and 64) showed high expression in terms of fungicidal and β-1,4-endoglucanase activities and low activity of β-1,3 endoglucanase (Table 1). Both these activities were also measured using cell-free filtrates of the clones which revealed 58–65% higher values in terms of β-1,4 endoglucanase activity in clones 8 and 64, compared to 116 and 248. However, low levels of β-1, 3 endoglucanase activity was recorded only in clones 8 and 64 (Table 1). β-1,3 as well as 1,4-endoglucanase activities were not observed in the negative control (insertless vector transformed E. coli) in the plate assay and quantitative analyses (Table 1).

Isolation and sequence analysis of gene (end) encoding endoglucanase

The sequencing results of positive clones (116 and 248) showed an identical sequence and length of 2.156 kb where the other two positive clones (8 and 64) revealed an identical sequence and length of 2.638 kb, which included both coding and non-coding regions for β-1,4-endoglucanases. The putative genes identified were referred to as end 1 and end 2, respectively encoding for β-1,4-endoglucanases. Open reading frames of 348 and 656 amino acid residues with a predicted molecular mass of 38 kDa and 73.89 kDa were detected in end 1 and end 2, respectively. The BLASTN results of end 1 showed 100% identity (query coverage 99%) with the glucanase gene of A. variabilis and Nostoc sp. PCC 7120 while end 2 showed 100% identity (query coverage 97%) with S-layer-associated multidomain endoglucanase gene (alr0290) of Nostoc sp. PCC 7120. BLASTP results of end 1 showed 97% similarity with the glucanase of A. variabilis (protein ID CAA66983) and Nostoc sp. PCC 7120 (protein ID BAB75332) characterized in peptidase M20 family. BLASTP results of end 2 showed 97% similarity with S-layer associated multidomain endoglucanase having different domain proteins coding for BglC; Endoglucanase (COG2730), RTX toxins and related Ca2+-binding proteins (COG2931), cellulase (pfam accession number 00150; glycoside hydrolase family 5) and peptidase (pfam accession number 08548; M10 family), respectively. However, BLASTP (with protein data bank) results of corresponding amino acid sequence of end 1 revealed significant similarity (23–26%) with proteins coding for aminopeptidase/endoglucanase in Bacillus cereus, Bacillus subtilis, and Thermotoga maritima. BLASTP results of end 2 revealed significant similarity (25–28%) with β-mannose, GH5 endoglucanase (cel5a), cellulase K, cellulase cel5 etc. of Bacillus sp., Thermoascus aurantiacus, Pseudoaltermonas haloplanktis, etc. (Table 2). A non-coding sequence of 181 and 522 nucleotides at the upstream and 596 and 150 nucleotides downstream were detected in the sequences of end 1 and 2, respectively. The Shine–Dalgarno sequence of end 1 and end 2 were identified before the 5 and 6 nucleotides from the start codon, respectively. The 27 and 31 nucleotides long promoter sequences was observed from upstream of end 1 and 2, respectively, which included both -10 and -35 promoter regions (Fig. 1a and b). The distance between the -35 and -10 regions was 12 and16 bp in end 1 and 2, respectively.

Table 2 Protein similarity (PDB data base) of deduced amino acid sequences of end 1 and 2 encoding endoglucanase in A. laxa RPAN8
Full size table
Fig. 1
figure 1

Identification of ribosomal binding site (rbs), −10 and −35 promoters from the upstream region of end 1 (a) and end 2 (b). The arrows indicate the positions of promoter sequences (−10 and −35) of end. The underlined sequence in close proximity to the start codon (boxed) indicates the putative ribosomal binding site (rbs)

Full size image

Bioinformatic tools guided comparative analyses of end 1 and 2 with known glucanases of cyanobacteria

The N-terminal region of the deduced amino acid sequence in end 1 and end 2 exhibited a putative peptide signal of 24 and 20 residues long, respectively. Unlike RPAN8, Anabaena and Nostoc glucanases available in the database did not show any signal peptide and cleavage site (Supplementary Fig. S2 a,b,c and d). Alignment of corresponding amino acid sequences of end 1 with available sequences of glucanase of Nostoc and Anabaena revealed one deletion and two insertions at different positions in the few starting amino acid residues (Fig. 2a). However, the comparison of end 2 with available endoglucanase from Nostoc revealed 15 substitutions in the amino acid sequence of RPAN8 (Fig. 2b).

Fig. 2
figure 2

Amino acid exchanges in the sequence of endoglucanase encoding end 1 and 2 from A. laxa (RPAN8), and comparison with similar sequences of strains from NCBI data base. a end 1 and b end 2. Shaded bars indicate the conserved residues, and non-shaded bars indicate amino acid exchanges. The arrow indicates amino acid sequence of the putative signal peptide and the cross indicates the cleavage site. The mature protein is indicated by +1

Full size image

Discussion

The use of fungicidal chemicals as a control measure has far-reaching implications on the environment and public health, and definite needs exists for identifying fungicidal compounds/enzymes from biological sources and develop as biocontrol agents. A wide range of prokaryotic and eukaryotic microorganisms such as Pseudomonas, Trichoderma, and Bacillus have the potential to produce cell-wall-degrading enzymes when chitin or isolated fungal cell wall material is present in the growth medium. Such products are available commercially under various trade names. Towards this endeavor, recently, we showed the presence of hydrolytic activity in these organisms and demonstrated their role in biocidal activity against phytopathogenic fungi (Prasanna et al. 2008; Manjunath et al. 2010). From our previous investigations, we selected A. laxa RPAN8 for in-depth analyses of biochemical and molecular aspects of fungicidal activity. The chitosanase homologue responsible for fungicidal activity had been earlier identified in this strain (Prasanna et al. 2010a). This strain has been previously identified both morphologically (Nayak et al. 2009) as well as on the basis of 16S rDNA (Accession number GQ466518, GQ466546, GQ466574).

Firstly, the genomic library for RPAN8 was constructed which generated 2000 clones each having 2-5 kb fragment size. Significant expression of the four positive E. coli clones in terms of fungicidal activity against Pythium sp. provided direct evidence for the presence of fungicidal gene(s) which have their own promoters for expression. Our target was the glucanase moiety (β-1,3 or 1,4) as β-1,3-linkage is present almost in all fungi whereas β-1,4-linkage (cellulose) is only present in the Pythium sp. and other oomycetes. This was further supported by the results of CMCase and laminarinase activities of this strain which indicated the presence of β-1, 3 and 1,4-endoglucanase like enzymes in this cyanobacterial strain tested. Microscopic analyses revealed severe disintegration of the mycelia of the fungus P. aphanidermatum. The architecture of the cell wall is very similar in most fungi, therefore, the types and amounts of specific polysaccharides (e.g., chitin, cellulose, β-glucan with 1, 3- and 1, 6- linkages) represent the main targets for cidal action of most biocontrol agents (Peberdy 1990) and the fungus Pythium is known to contain cellulose instead of chitin in its cell wall. Hill et al. (2002) studied the proteolytic release of membrane-bound β-1,4 endoglucanase activity associated with cell wall softening in Achlya ambisexualis. Previously, Loprete and Hill (2002) also isolated an endoglucanase which had potency to degrade the mixed-linkage glucan components of the Achlya cell wall matrix.

In bacteria, there are several reports on lytic enzymes such as chitinases, chitosanases, proteases and β-1,3-glucanases which are known to play a key role in the biocontrol of various soil-borne fungal pathogens (Wang et al. 1999; Chang et al. 2007; Saito et al. 2009). Recently, Shi et al. (2010) characterized a β-1,3-endoglucanase for fungicidal activity in Streptomyces sp. S27 and successfully expressed the gene in E. coli BL21 (DE3). In our investigation, expression of A. laxa clones was undertaken which revealed β-1,4 endoglucanase activity in all the positive A. laxa clones for fungicidal activity, two of which also showed low β-1,3 endoglucanase activity (116 and 248). Previous studies indicated the significant expression of chitosanase recombinant protein in terms of clear zone formation when E. coli cells harboring plasmid with chitosanase gene were spotted on agar plates containing glycol or colloidal chitosan as a substrate (Park et al. 1999; Kimoto et al. 2002). In our study, the expression of E. coli-JM109 (DE3) cells harboring pUC19-end was observed on fungal lawn, laminarin, and CMC plates but the mechanism involved in the expression of recombinant protein without cell lysis is unknown. Since the endoglucanase encoded by both end 1 and 2 is secretory in nature due to the presence of a signal peptide, the intact signal peptide may potentially allow secretion of this enzyme out of or in the periplasm of E. coli cell and thereby inhibits the growth of P. aphanidermatum. The plate assay results in terms of fungicidal, β-1,4 and 1,3-endoglucanase activities was confirmed by the results obtained in the cell-free assay of filtrates of the clones.

The sequence analyses of these E. coli clones revealed two sets of clones with identical sequences. Therefore, bioinformatics tools were employed to detect the presence of multiple copies of such genes in sequenced genomes of cyanobacteria in the database. The genome of A. variabilis ATCC 29413 which has been completely sequenced (GenBank accession number CP000117) was explored; however, no second copy of ORF was found in this organism. The clones (i.e., 8 and 116) with two different ORFs and having both start and stop codons were designated as end 1 and end 2 and characterized in depth. Both ORFs contained ATG as start codon, which was preceded by a putative ribosome binding site (rbs). The identified rbs for end 1 and 2 matched with reported rbs of Zymomonas mobilis ethanologenic genes (Theberge et al. 1992; Rajnish et al. 2008) and β-1,4-endoglucanase of Bacillus pumilus strain (Lima et al. 2005), respectively.

Computational analysis of the nucleotide sequences identified putative promoter regions of 27 and 31 bp, positioned about 144 and 243 nucleotides upstream of the initiation codon in end 1 and 2, respectively. In bacteria, investigations have revealed that various types of promoter sequences are involved in the transcriptional regulation of endoglucanases (Han et al. 2003). Recently, Endo et al. (2008) investigated a novel upstream promoter sequence (−153 to −138) cellulose responsive element (CeRE) in Aspergillus nidulans, and proved that this promoter sequence is essential for inductive expression of eglA gene. In our study, no sequence similarity was observed with known promoters for β-1,4-endoglucanase in bacteria. Sode et al. (1995) analyzed a strong cyanobacterial promoter for enhanced expression of foreign gene and indicated a tenfold higher expression of CAT (chloramphenicol acetyl transferase) structural gene as compared to other cyanobacterial promoters. The expression was obtained only in the presence of putative promoter region which was 35 nucleotides long (present 115 nucleotides upstream of CAT). This promoter region included both −10 and −35 promoter regions. In our study, the putative promoter region of end 2 showed significant similarity with this marine cyanobacterial promoter (Supplementary Fig. 3). However, in end 2 the spacing between the −35 and −10 is almost the optimal spacing as observed for all promoters including those in B. subtilis and E. coli (Harley and Reynolds 1987; Chang et al. 1992). This spacing appears to be a stringent constraint on promoter function, because it is required for recognition by the RNA polymerase holoenzyme (Warne and deHaseth 1993). The close similarity of the promoter to the consensus σA sequence suggests that these promoters, if not subjected to any regulatory constraints, would act as strong promoters in vivo. Thus, the high expression of fungicidal, β-1,4 activities in end 2 may be due to the presence of strong promoter sequence as compared to end 1. However, detailed investigations in this direction are required to gain further insight.

The corresponding amino acid sequence analyses of end 1 and 2 showed 97% similarity with glucanase of A. variabilis and Nostoc sp. PCC7120 and S-layer associated multidomain endoglucanase of Nostoc sp., and placed in peptidase M20 and GH5 family, respectively. Peptidase family M20 contains exopeptidases: carboxypeptidases, dipeptidases and a specialized aminopeptidase (MEROPS Accession MER001266). On the other hand, GH5 family proteins are known for chitosanase (EC 3.2.1.132); β-mannosidase (EC 3.2.1.25); cellulase (EC 3.2.1.4); glucan 1,3-β-glucosidase (EC 3.2.1.58); licheninase (EC 3.2.1.73); glucan endo-1,6-β-glucosidase (EC 3.2.1.75); mannan endo-β-1,4-mannosidase (EC 3.2.1.78); endo-β-1,4-xylanase (EC 3.2.1.8); cellulose β-1,4-cellobiosidase (EC 3.2.1.91); endo-β-1,6-galactanase (EC 3.2.1.-); β-1,3-mannanase (EC 3.2.1.-) and xyloglucan-specific endo-β-1,4-glucanase (EC 3.2.1.151). The results of BLASTP (with PDB data base) of corresponding amino acid sequences of end 1 revealed significant similarities (23-26%) with proteins coding for aminopeptidases/endoglucanases in B. cereus, B. subtilis, Pyrococcus horikoshii, and T. maritima. Incidentally, all these enzymes also belong to peptidase M20 family. On the other hand, BLASTP (with PDB data base) results of end 2 showed significant similarities with endoglucanases and cellulase like enzymes of Bacillus and Pseudomonas which are encoded by GH-5 family. This supports earlier data from GH5 family cellulases available from earlier investigations (Ledger et al. 2006; Wong et al. 2009).

Most cellulases have a non-specific hydrolytic action on chitosan, and some are even superior to specific chitosanases in degrading chitosan. This may be derived from the structural similarity of chitin, chitosan, and cellulose, which are all polymers of d-glucose linked by β-1,4-glycosidic linkages (in chitosan, the C-2 hydroxyl groups are replaced by amino groups), and it seems that the enzyme does not always recognize stringently the group of C2 position in glucose or glucosamine residue when enzyme-substrate complex was formed. For the same reasons, chitinases, chitosanases, and cellulases also exhibit a high degree of homology and are often associated in the same microorganism. A cellulase showing bifunctional activities was first reported by Hedges and Wolfe (1974). Since then, many other cellulases with chitosanase activity have been purified from Streptomyces griseus (Tanabe et al. 2003), B. cereus S1 (Kurakake et al. 2000), Bacillus circulans (Mitsutomi et al. 1998) and Bacillus sp. 0377BP (Choi et al. 2004), Trichoderma viride (Liu and Xia 2006), Trichoderma reesei (Ike et al. 2007), B. cereus D-11 (Gao et al. 2007), which confirmed that most cellulases have a bifunctional cellulase–chitosanase activity. In our earlier investigation, we illustrated the chitosanase/fungicidal activity and identified its homolog in this strain (Prasanna et al. 2010b), but the sequenced end 1 and 2 did not show significant similarity with this chitosanase homolog in the multiple sequence alignment (data not shown). However, the reported bifunctional cellulase-chitosanases from S. griseus HUT 6037, Bacillus licheniformis NBL420, Myxobacter sp. AL-1 (Tanabe et al. 2003; Hong et al. 2003; Pedraza-Reyes 1997) are encoded by GH5 family. Therefore, there exists a probability of GH5 endoglucanase encoded by end 2 being a bifunctional enzyme for fungicidal activity. Further research is required to know the exact roles of these identified cellulases encoded by end 1 and 2.

Comparative amino acid sequence alignment of both genes with that of wild type Anabaena and Nostoc glucanases revealed few amino acid exchanges in the sequence. One deletion and two substitutions were detected in the first few amino acid residues of the translated sequence after the start codon of end 1 while 15 substitutions were detected in end 2. This leads to the presence of signal peptides, respectively in this organism RPAN8 and the presence of cleavage site. Thus, the difference in the amino acids sequence of both end during the course of evolution may be responsible for the production of an active β-1,4 endoglucanase enzyme. Also, the lack of rbs, promoter sequences, signal peptide, and cleavage sites in the available sequences of Anabaena/Nostoc glucanases may be responsible for the lack of reports on endoglucanase activity in these organisms.

In conclusion, we have characterized two different novel genes (end 1 and end 2) encoding endoglucanase associated with Peptidase M20 family and glycoside hydrolase family 5 (GH5), respectively, and showed that they are responsible for the fungicidal activity in A. laxa. Apart from this, the identification of promoter regions, rbs, signal peptide and cleavage sites in both end provide evidence for translation of end encoding endoglucanase in this cyanobacterial strain. Thus, the current study emphasizes that this endoglucanase producing cyanobacterial strain holds promise as a potential biocontrol agent against the pathogenic fungus P. aphanidermatum.