Introduction

Phytic acid is an indigestible organic form of phosphorus, which is found in cereal grains, legume plants, oilseeds, nuts, tubers and different types of seeds (Nakashima et al. 2007; Greiner and Konietzny 2016). Myo-inositol hexakisphosphate phosphohydrolases (i.e., phytases) belong to a subclass of phosphatase enzymes which break down the phytic acid by hydrolyzing the phosphate residues from phytic acid (myo-inositol hexakisphosphate). There are several types of phytases, such as 3-phytase (EC 3.1.3.8), 4-phytase (EC 3.1.3.26) and 5-phytase (EC 3.1.3.72) which are involved in the liberation of P moiety from phytic acid at the position C3, C4 and C5, respectively. These enzymes have a molecular mass between 40 and 70 kDa (Kumar et al. 2010; Afinah et al. 2010).

Ruminant animal utilized phytate through the action of phytases produced by microbial flora in the rumen. The anaerobic gut fungi and bacteria found in the ruminants are responsible for the phytate hydrolysis within the rumen. The hydrolyzed inorganic phosphate from the phytate is utilized by both the microflora and the ruminant host (Yanke et al. 1998). However, the situation is quite different with monogastric animals such as chicken, swine, fish and birds are not able to utilize phytase phosphorus because they lack phytase-degrading enzymes or lack the insufficient amount of phytases in their digestive tracts (Gupta et al. 2015; Dahiya 2016). Hence, phytate in grains passed through the gastrointestinal tract without being digested properly (Niño-Gómez et al. 2017). Since phosphorus is an essential requirement for bone formation, deficiency of this mineral would cause of osteoporosis, tooth decay, narrow jaws, mental retardation, rickets, etc. (Berlyne et al. 1973). Phytic acid can inhibit the nutrient absorption by the intestine and also prevent the different types of ion uptake such as Zn2+, Ca2+, Mg2+, Cu2+ and digestive enzymes such as pepsin, trypsin, amylase, etc. (Gupta et al. 2015). Here, phytic acid acts as an anti-nutritional agent (Escobin-Mopera et al. 2012; El-Toukhy et al. 2013). The phytase catalyzes the hydrolysis of phytic acid and converts the organic and insoluble form of phosphorus to more soluble form and ultimately improve and facilitate intestinal absorption (Wise 1982). Importance of the phytase protein can be evidenced by many earlier workers (Konietzny et al. 1994; Mullaney et al. 2000; He and Honeycutt 2001; Lei et al. 2007, 2013; Rodriguez et al. 2000; Tazisong et al. 2008; Yao et al. 2012; He et al. 2013; Gontia-Mishra and Tiwari 2013; Zając et al. 2018).

Phytases have been reported in many fungi, e.g., Aspergillus niger (George et al. 2007), Aspergillus fumigatus (Zhang et al. 2007), Peniophora lycii (George et al. 2007), Penicillium simplicissimum (Tseng et al. 2000), Candida krusei (Quan et al. 2002), Debaryomyces castellii (Ragon et al. 2009) and Fusarium oxysporum (Gontia-Mishra et al. 2014). Moreover, a number of gram-positive- and gram-negative bacteria, e.g., Bacillus subtilis (Kerovuo et al. 1998; El-Toukhy et al. 2013), Bacillus licheniformis (Kumar et al. 2014), Pseudomonas spp. (Richardson and Hadobas 1997), Klebsiella pneumoniae (Escobin-Mopera et al. 2012), Klebsiella terrigena (Greiner et al. 1997), Enterobacter sp. (Kalsi et al. 2016), Serratia sp. (Zhang et al. 2011; Kalsi et al. 2016), and Yersinia mollaretii (Shivange et al. 2012) have also been reported earlier to possess this important enzyme.

Phytase enzyme has many beneficial roles in different industries, agriculture, poultry farm, etc. In poultry farm, phytase is used for overcoming the phosphorus deficiency (Elkhalil et al. 2007). Natuphos® and Ronozyme® are the two commercially available phytases derived from Aspergillus, one of the most abundant extracellular producers of phytase (Casey and Walsh 2004). This phytase was actually a recombinant phytase produced by the expression of the phyA gene of Aspergillus ficuum in Aspergillus niger, first produced in 1994 (Casey and Walsh 2004). However, few studies reported that bacterial phytases were also promising for commercial phytase production (Konietzny and Greiner 2004; Kalsi et al. 2016; Niño-Gómez et al. 2017). In fact, bacterial phytases are reported to be more beneficial over the fungal phytases due to higher substrate specificity, better catalytic capacity and enhanced resistance to proteolysis (Konietzny and Greiner 2004). In addition, phytase-producing bacteria are known to involve in plant growth promotion also (Singh et al. 2014).

Apart from isolating and characterizing various types of phytases, extensive computational investigations of microbial phytases have been worked out to discover many unexplored features lying within the sequences (Kumar et al. 2012, 2014; Kumar and Agrawal 2014; Gontia-Mishra et al. 2014; Mathew et al. 2014; Verma et al. 2016; Ebrahimi et al. 2016; Niño-Gómez et al. 2017; Pramanik et al. 2018). The outcomes of these studies are biotechnologically advantageous to utilize them in agricultural and industrial perspectives.

This study describes the use of various bio-computational tools to perform the phylogenetic, structural and functional analyses of the phytase enzyme of Enterobacter spp. to understand its distinctive properties essential for its industrial application.

Materials and methods

Sequence retrieval and phylogenetic analysis

UniProtKB server is the knowledge base of Swiss Institute of Bioinformatics and a comprehensive source of protein sequences and annotation database (Apweiler et al. 2004). This proteomics-based source server (http://www.uniprot.org) was used for retrieval of phytase sequences of different spp. of Enterobacter (Gram-negative-, rod-shaped bacteria). For this, Enterobacter aerogenes GN = ASV18_08060 (Protein Acc. no.: A0A0M3HCJ2) was first selected as the query sequence and BLAST search was done. A total of 16 sequences of different spp. of Enterobacter were selected (based on highest sequence identity, lowest E-value, maximum query coverage and bit score) and downloaded from UniProtKB in FASTA format for progressing to the computational investigation. Furthermore, the sequences were analyzed phylogenetically to decipher the evolutionary relationship among the selected proteins of interest by using MEGA7 software (Kumar et al. 2016). Percent similarity index within the sequences was calculated by Clustal Omega (https://www.ebi.ac.uk/Tools/msa/clustalo/).

Physiochemical characterization

ExPASy—ProtParam is a tool which is used for the calculation of different physical and chemical parameters for a protein (Gasteiger et al. 2005). This tool (http://www.espasy.org/protparam) was used to calculate physiochemical properties of retrieved sequences. Various computed parameters like amino acid composition, molecular weight (MW), isoelectric point (pI), molar extinction coefficient (EC), aliphatic index (AI), instability index (II) and grand average of hydropathicities (GRAVY) were calculated using this server.

Secondary structure prediction

Secondary structure (alpha-helix, 310 helix, pi helix, beta bridge, extended strand, beta-turn, bend region, random coil etc.) of a protein refers to the interaction of H-bond donor and acceptor residues of a polypeptide chain. Practically, secondary structural arrangements can be accurately done by nuclear magnetic resonance (NMR) spectroscopy (Meiler and Baker 2003). But some useful and easy computational tools used for in silico prediction of secondary structures of the protein. Here, secondary arrangements were predicted by SOPMA web-based server (Geourjon and Deleage 1995). For this, 16 protein sequences of different Enterobacter spp. were analyzed.

Homology protein modeling, its evaluation and submission

Homology 3D protein modeling of selected 16 sequences was performed using Swiss model workspace (Biasini et al. 2014) selecting its suitable and best-matched template. Evaluation of the predicted protein model was done in both QMEAN (Benkert et al. 2009) and SAVES server (http://services.mbi.ucla.edu/SAVES/). Based on the evaluation report the best-built model was finally submitted to Protein Model Database (PMDB) (https://bioinformatics.cineca.it/PMDB/) and the PMDB ID was obtained.

Functional analysis and protein–protein interaction

To find out the functional motifs and superfamily of the protein sequence, Motif Finder (http://www.genome.jp/tools/motif/) was used for the analyses. Highly conserved domains among the sequences were analyzed by Clustal Omega (https://www.ebi.ac.uk/Tools/msa/clustalo/). To know the interaction of Enterobacter phytase with other closely associated proteins STRING v10.0 (http://string-db.org/) server was used and functional protein association network was generated.

Results and discussion

Sequence retrieval and phylogenetic analysis

Amino acid sequences of total 16 bacterial strains (Enterobacter kobei GN = ASV01_18150, Enterobacter sp. GN02225 GN = YA53_05950, Enterobacter sp. GN02186 GN = YA50_01625, Enterobacter sp. 35730 GN = SS33_10495, Enterobacter asburiae GN = ABF78_15640, Enterobacter sp. 35699 GN = SS37_20000, Enterobacter sp. 50588862 GN = APT89_19470, Enterobacter hormaechei GN = ASU65_02620, Enterobacter sp. GN02174 GN = YA49_02765, Enterobacter sp. 35027 GN = SS00_18385, Enterobacter sp. GN02616 GN = ABR35_11890, Enterobacter sp. 42324 GN = SS44_19905, Enterobacter sakazakii GN = AFK64_08585, Enterobacter agglomerans, Enterobacter cancerogenus GN = NH00_23800, Enterobacter aerogenes GN = ASV18_08060) were phylogenetically analyzed to predict their evolutionary interrelationship (Fig. 1). Enterobacter aerogenes (A0A0M3HCJ2) clustered only with E. cancerogenus (A0A0A3YJT6) and their closest neighbour is E. agglomerans (F2VRZ7) present in the same clade. Rest of the 13 taxa were in a separate clade showing interrelationships within them. This might be correlated with the percent similarity index of the phytase sequences of the strains with the selected query phytase protein, A0A0M3HCJ2 (Fig. 2). It was found (from the percent similarity index in Clustal Omega) that A0A0M3HCJ2 showed the highest similarity with A0A0A3YJT6 (68.65%) and F2VRZ7 was next to it (66.43%). These percent similarity indices were also manifested by the closest clustering pattern of these three sequences (Fig. 1). Interestingly, rest 13 sequences showed percent similarity in between 19.46–24.11% with the A0A0M3HCJ2 sequence (Fig. 2). This was the principle reason behind forming a separate clade of 13 taxa from the rest three sequences in the phylogenetic tree of phytase proteins (Fig. 1). From this analysis, it can be interpreted that there might be a correlation among the selected taxa based on their protein sequences. The similar phylogenetic analysis was also made by some earlier workers to decipher the evolutionary significance of different taxa solely based on their protein sequences (Verma et al. 2016; Pramanik et al. 2017a, b, 2018).

Fig. 1
figure 1

Phylogenetic classification of phytase proteins from 16 different Enterobacter spp

Full size image
Fig. 2
figure 2

Calculation of percent similarity index of the selected phytase proteins with the query protein (A0A0M3HCJ2)

Full size image

Physicochemical characterization

Computational-based analysis about the physicochemical behavior of the proteins gives a theoretical overview of the nature of the protein. In this study, all the 16 phytase sequences were physicochemically characterized by a number of bio-computational tools (Table 1; Fig. 3). The analysis revealed that the isoelectric point ranges from 5.03 to 8.90 which is said to be a wide range. The isoelectric point of the proteins below the neutral pH can be said to be acidic and above it can be regarded as alkaline. Here, 13 strains showed acid phytases and rest 3 phytases were basic/alkaline in nature. Acid phytases are evident in the work of Zhang et al. (2011) while basic one was reported by Tran et al. (2011) while working on other bacterial phytases. Acid phytases were also evident from the in silico analysis of Aspergillus niger (Niño-Gómez et al. 2017). Except for the four strains (A0A0W2KF14, F2VRZ7, A0A0A3YJT6 and A0A0M3HCJ2), all strains showed instability index below 40 (Table 1). Thus, the majority of the proteins can be said as stable. Besides, higher ranges of aliphatic index prove the protein as thermostable in nature (Ikai 1980). Thermostable phytase from Enterobacter sp. was reported by Kalsi et al. (2016) which corroborated the present work. Besides higher aliphatic indices were also reported in Aspergillus niger (Niño-Gómez et al. 2017) while working on the prediction study of both 3-phytase A and 3-phytase B. Assuming all pairs of Cys residues form cystines, the molar extinction coefficients were calculated in water at 280 nm. However, EC ranges from 35,535 to 81,275 M−1 cm−1 (Table 1). The average molecular weight of the proteins is around 48 kDa. The GRAVY, calculated from ExPASy showed very low (Table 1) which implies that the proteins have better interactions with water. This interpretation was also corroborated by the work of Mathew et al. (2014), Verma et al. (2016). Moreover, the strains showed the difference in composition of amino acid residues of their respective proteins as presented graphically in Fig. 3.

Table 1 Physicochemical features of 16 Enterobacter spp. with their respective protein and gene accession numbers
Full size table
Fig. 3
figure 3

Graphical overview of composition difference of amino acids among the selected 16 strains of Enterobacter spp

Full size image

Secondary structure prediction

The predicted secondary arrangements of different Enterobacter spp. revealed mainly four types of secondary elements which were alpha helices, random coils, extended strands and beta turns (Table 2). The alpha-helical content was the highest (42.76%) as found in case of E. aerogenes (A0A0M3HCJ2) among all of the Enterobacter spp. (Table 2, Suppl. 1). Niño-Gómez et al. (2017) also showed that 43 and 38% α-helices were found in the computational investigation of Aspergillus niger 3-phytase A and 3-phytase B, respectively. This indicated the thermostable nature of the protein as Kumar et al. (1999) indicated that α-helical conformations are abundant in case of thermophiles to withstand high temperatures. These types of structural findings were also performed by a number of bioinformatics researchers. A similar prediction of secondary structures of the residual interactions of Aspergillus fumigatus phytase was done by Zhang et al. (2007) and Bacillus phytases were done by Verma et al. (2016).

Table 2 Comparison of secondary and tertiary protein structures among different spp. of Enterobacter
Full size table

Homology protein modeling, its evaluation and submission

Although homology modeling was performed for all 16 proteins (Suppl. 1), Enterobacter aerogenes (A0A0M3HCJ2) was selected as representative species (Fig. 4) to elucidate phytase protein structure of Enterobacter spp. based on QMEAN score, an overall quality factor from SAVES server and Ramachandran plot (Table 2, Suppl. 2). The predicted 3D protein modeling of Enterobacter aerogenes (A0A0M3HCJ2) divulged that the protein consisted of four monomeric polypeptide chains, i.e., it was a tetrameric protein (Fig. 4a). This structural finding was also corroborated by some earlier reports. Ragon et al. (2009) reported tetrameric phytase in fungal species—Debaryomyces castellii CBS 2923 and Shivange et al. (2012) reported in the bacterium—Yersinia mollaretii. Helix, sheet and loops are differentially coloured and presented in Fig. 4b. Besides distinct disulfide bridges were demarked as green spherical (s) in Fig. 4b. These disulfide bridges which are formed by the oxidation of thiol groups of the cystine residues (Trivedi et al. 2009) are one of the factors for the stability of a protein (Cheng et al. 2007). After the construction of 3D (.pdb) model, the evaluation and quality estimation of the model was performed (Table 2, Suppl. 2–3) and Ramachandran plot was built to show the positions allocated for each amino acid residues (Table 2, Suppl. 2). Analysis of Ramachandran plot for the PDB structure of E. aerogenes (A0A0M3HCJ2) protein showed that 97.9% residues were present in the most favoured region (Table 2, Suppl. 2). Presence of more than 90% residues in the favoured region of Ramachandran plot is the characteristics of the good quality model (Yadav et al. 2013). QMEAN4, QMEAN6 and Z-score calculated were 0.25, − 0.26 and < 1, respectively (Table 2, Suppl. 2). The desirable QMEAN scores and Z-score should be within 0–1 (Berman et al. 2000) and < 1 Benkert et al. (2009), respectively, in comparison with a non-redundant set of PDB structures to obtain a high-quality model. Moreover, the overall quality factor by the SAVES ERRAT was 96.141% (i.e., > 95%) (Suppl. 3). A good, high-resolution structure should score 95% or higher as an overall quality factor (Benkert et al. 2009). However, comprehensive evaluation of the predicted model proves to be a higher resolution model (> 3 Å) as determined by both QMEAN (Table 2, Suppl. 2) and SAVES server (Table 2, Suppl. 3). A similar type of model validation was also conducted by Pramanik et al. (2018) while working with Klebsiella phytases. Finally, the model (in .pdb format) was deposited in PMDB database and its accession number obtained is PM0080561 and the said model can be retrieved for any further investigations in future.

Fig. 4
figure 4

Predicted 3D model structure of phytase of Enterobacter aerogenes (A0A0M3HCJ2) viewed by PyMol: a showing four distinct chains of the protein. b Tertiary structure showing prominent secondary elements and disulfides (red = helix, yellow = sheet, green = loop, green balls = disulfides)

Full size image

Functional analysis and protein–protein interaction

From the functional analyses of the proteins, mainly two conserved motifs, i.e., His_Phos_1 and His_Phos_2 were found (Suppl. 4) both of which belong to the Histidine phosphatase superfamily (Suppl. 4). This superfamily is a vast functionally diverse group of proteins consisting of two clades sharing a finite sequence similarity (Rigden 2008). The multiple sequence alignment among all the 16 phytase protein sequences identified several fully conserved amino acid residues (marked as * in Fig. 5) among which “DG–DP–LG” sequence was found to be the highly conserved sequences (Fig. 5). Similarly, Niño-Gómez et al. (2017) while working with 3-Phytase A and 3-Phytase B from Aspergillus niger found that RHGXRXP-HD were the highly conserved sequences as revealed from the multiple sequence alignment study. In addition, the protein–protein interaction network produced through STRING server depicted that the query protein interacts with ten different proteins directly or indirectly (Fig. 6). Occupying the central position, the query protein interacts with bifunctional riboflavin kinase/FMN adenylyltransferase, alpha subunit of riboflavin synthase, NAD(P)H-dependent FMN reductase, FMN reductase, major facilitator superfamily protein, spore coat U domain-containing protein, lipoprotein, TonB-dependent receptor, a hypothetical protein of 117 amino acid residues and putative universal stress family protein as shown in Fig. 6. STRING analysis is very important in terms of system-level understanding of cellular processes in a computer-based way. The interacting network can be used for filtering and assessing functional genomics data as well to provide an instinctive platform for annotating evolutionary properties of proteins in addition to structure–function aspects (Schwartz et al. 2008). Exploring this type of STRING generated predicted interaction networks can suggest important directions for future experimental research, e.g., prediction of the possible pathway of the protein of interest. Similar protein–protein interaction study was previously worked out by Goñi et al. (2008), Gao et al. (2012), Guney and Oliva (2012), Quan et al. (2014), Zhang et al. (2016), Pramanik et al. (2017a, b, 2018) etc.

Fig. 5
figure 5figure 5

Multiple sequence alignment of 16 phytase sequences of Enterobacter spp. showing highly conserved amino acid residues. An * (asterisk) indicates single, fully conserved residue, a : (colon) indicates conservation between groups of strongly similar properties and a. (period) indicates conservation between groups of weakly similar properties. Therefore, the hierarchy of conservation using these symbols is * (identical) > : (colon) >. (period). Highlighted area in the sequence alignment indicates highly conserved sequences

Full size image
Fig. 6
figure 6

Protein–protein interaction map including the predicted interacting proteins with phytase protein (A0A0M3HCJ2)

Full size image

Hence, this elaborative study will be very helpful in selecting commercial phytases in future derived from E. aerogenes as potential phytases in animal feed additive. In addition, direct application of E. aerogenes, in cultivated fields will help crops to grow better by soluble P uptake released due to phytase activity of the strain.

Conclusion

In silico characterization of Enterobacter phytases revealed that the 48 kDa proteins were tetrameric, thermostable and acidic in nature belonging to the histidine phosphatase superfamily. This types of thermostable, acidic phytases derived from E. aerogenes might be beneficial in terms of application in various industrial fields such as monogastric animals, poultry farms, etc., by solubilized phosphorus supplement in their food essential for their normal growth and development. Besides phytase containing E. aerogenes can be directly applied in the agricultural field to meet up the phosphorus deficiency in crop plants. Hence, this study will help researchers to understand the essential structure–function properties of Enterobacter phytases.