The production of α-amylases has been studied extensively in submerged and/or solid state fermentations. In general, extracellular α-amylase production is growth associated (Murthy et al., 2009; Asoodeh et al., 2010; Abou Dobara et al., 2011). Among the physical parameters, the temperature and pH of the medium play an important role in α-amylase production. The influence of temperature on α-amylase production is related to the growth of the organism. Amylase production in bacteria has been studied in a wide range of temperatures. α-Amylase production has been reported from thermophilic and hyperthermophilic bacteria and archaea like Pyrococcus, Thermococcus, and Sulfolobus species (Leuschner and Antranikian, 1995; Sunna et al., 1997), G. thermoleovorans (Rao and Satyanarayana, 2007a), B. acidocaldarius (Buonocore et al., 1976), and Alicyclobacillus sp. A4 (Bai et al., 2012), from mesophiles B. amyloliquifaciens (Gangadharan et al., 2008), Halomonas meridiana (Coronado et al., 2000a) B. subtilis (Ravindar and Elangovan, 2013), Lactobacillus plantarum (Kanpiengjai et al., 2015) as well as from psychrotolerants and psychrophiles Microbacterium foliorum GA2 (Roohi and Kuddus, 2014) and Aeromonas veronii NS07 (Samie et al., 2012). Among different carbon sources used, starch, fructose, glucose, and rice flour, are known to support high enzyme production (Ezeji et al., 2005; Prakash et al., 2009). Carbon sources like glucose and maltose have been used for the production of α-amylase, but the use of starch remains ubiquitous (Mamo and Gessesse, 1999; Sajedi et al., 2005; Liu and Xu, 2008; Sharma and Satyanarayana, 2010). Industrially important enzymes have traditionally been produced in submerged fermentation, but these enzymes are also produced by solid state fermentation. Sodhi et al. (2005) and Hashemi et al. (2010) explained the use of wheat bran for the production of α-amylase. Agro-residues were used for cold-active α-amylase production from M. foliorum (Roohi and Kuddus, 2014). In general α-amylase production is inducible in nature (Aiyer, 2004; Ryan et al., 2006; Asoodeh et al., 2010; Abou Dobara et al., 2011), but in few cases α-amylase production is also constitutive (Rao and Satyanarayana, 2007a). Like most other inducible enzymes, α-amylase production is subjected to catabolite repression by maltose and glucose, starch hydrolytic products (Bhella and Altosaar, 1988; Morkeberg et al., 1995) with the exception of some Bacillus strains (Kalishwaralal et al., 2010). Nitrogen source is a major factor that affects α-amylase production. Many investigators had confirmed that organic nitrogen sources support maximum α-amylase yields (Saxena et al., 2007; Aqeel and Umar, 2010; Abou Dobara et al., 2011).

The combination of low molecular weight dextran with Tween-80 increased 27-fold higher α-amylase production (Arnesen et al., 1998). Various metal ions like Ca²⁺, Fe²⁺, Mg²⁺, and K⁺ are added to the α-amylase production medium (Sajedi et al., 2005; Liu and Xu, 2008). Phosphate is a vital requirement for microbes as it regulates the synthesis of primary and secondary metabolites. Lower and higher levels of phosphate in the medium significantly affect the growth and enzyme production (Hillier et al., 1997; Sharma and Satyanarayana, 2010).

The production of amylases by microbes is affected considerably by physical and chemical parameters of the medium (Gigras et al., 2002; Rao and Satyanarayana, 2007a; Roohi and Kuddus, 2014; Sen et al., 2014). Traditionally “one-variable-at-a-time” approach has been used (Gokhale et al., 1991; Pham et al., 1998), but it is time consuming and does not permit understanding interactions among the process parameters. The statistical Plackett and Burman design, on the other hand, allows screening of critical culture variables (Sharma and Satyanarayana, 2006; Rao and Satyanarayana, 2007a; Kumar and Satyanarayana, 2008; Roohi and Kuddus, 2014), and response surface methodology (RSM) provides information about the optimum levels of each variable, interactions among them and their effects on the product yield (Gu et al., 2005; Rao and Satyanarayana, 2007a). The statistical approaches have been proved to be useful in optimizing medium components and cultural variables for maximizing amylase titres from many organisms like B. acidicola (Sharma and Satyanarayana, 2011), Bacillus sp. KR 8104 (Hashemi et al., 2010), and G. thermoleovorans (Rao and Satyanarayana, 2007a), Alcaligenes faecalis (Sen et al., 2014), and M. foliorum (Roohi and Kuddus, 2014).

Once produced and purified, an enzyme is biochemically characterized. α-Amylases show a wide range of substrate degradation. They degrade amylose, amylopectin, cyclodextrins, glycogen, and dextrins, but they possess highest specificity toward starch (Antranikian, 1992). Various metal ions influence activity of the enzymes (Vihinen and Mantasala, 1989). α-Amylase is a metal activated enzyme and has high affinity for Ca²⁺. In, general, Ca²⁺ enhances the activity and thermal stability of most of the α-amylases (Khajeh et al., 2001). The number of bound Ca²⁺ varies from 1 to 10. Dialysis against EDTA can remove the bound Ca²⁺. The Ca²⁺-free amylase can be reactivated by adding Ca²⁺. Although most of the α-amylases are Ca²⁺-dependent, there are reports of Ca²⁺-independent α-amylases too (Sajedi et al., 2005; Rao and Satyanarayana, 2007b; Hmidet et al., 2008; Asoodeh et al., 2010; Sharma and Satyanarayana, 2010). Besides Ca²⁺-independent α-amylases, there are also a few α-amylases, which are inhibited by Ca²⁺ ions (Babu and Satyanarayana, 1993; Tanaka and Hoshino, 2003; Mehta and Satyanarayana, 2013a). The metal ions, which inhibit α-amylase activity, include Hg²⁺ ions (Mamo and Gessesse, 1999; Asoodeh et al., 2010). Inhibition of α-amylase by Hg²⁺ ions indicates the presence of carboxyl groups in enzyme molecule (Dey et al., 2002). Furthermore, Hg²⁺ is also known to oxidize indole ring and to interact with aromatic ring present in tryptophan (Zhang et al., 2007; Liu et al., 2010). Various inhibitors such as dithiothreitol, β-mercaptoethanol, N-bromosuccinimide (NBS), p-hydroxy mercuribenzoic acid, iodoacetate, PMSF (phenylmethylsuphonyl fluoride), Woodward's reagent K, EDTA, and EGTA have been shown to inhibit α-amylases (Hamilton et al., 1999). Inhibition of enzyme activity by NBS demonstrates the role of tryptophan in maintaining the conformational stability of the enzyme (Rao and Satyanarayana, 2007b). Dithiothreitol and β-mercaptoethanol are the reducing agents, the effects of which suggest the role of—SH groups in the catalytic activity of enzyme. DTT may stimulate or inhibit the α-amylase (Ballschmiter et al., 2006; Rao and Satyanarayana, 2007b). In maltogenic α-amylase of Bacillus sp. WPD616, DTT had no effect on the α-amylase activity indicating that—SH groups have no role to play in the catalytic activity or these enzymes have no free and accessible—SH groups (Liu et al., 2006). The inhibition of α-amylase by PMSF indicates the role of seryl hydroxyl group in enzyme activity. Woodward's reagent K (WRK) has a role in the chemical modification of aspartic and glutamic acid residues (Paoli et al., 1997). The inactivation of amylase by WRK also indicates the involvement of acidic amino acids in the active site of the enzyme (Chauthaiwale and Rao, 1994; Komissarov et al., 1995; Sharma and Satyanarayana, 2010).

Many factors contribute to the stability of thermostable proteins including the presence of hydrogen bonds, electrostatic interactions, salt bridges, hydrophobic interactions, disulfide bonds, reduced entropy of unfolding, oligomerization, increased occurrence of proline residues, and others (Salminen et al., 1996; Russell et al., 1997; Vogt et al., 1997; Mehta and Satyanarayana, 2013b, 2015). Any of these properties of an enzyme can be modified by using methods such as site-directed mutagenesis, directed evolution, deletion mutation, and others for increasing its thermostability. Some examples of increasing thermostability of α-amylases are explained as follows. The first determinants of thermostability in bacterial α-amylases were identified by comparison of amino acid sequences of α-amylases of B. licheniformis (BLA) and B. amyloliquifaciens (BAA). Three stabilizing mutations in BAA were proposed, which caused a significant and additive thermostabilization of BAA; deletion of R176 and G177, and the substitution of K269A (Suzuki et al., 1989). This R176-G177 deletion showed similar effects on the thermostability of other α-amylases as well (Igarashi et al., 1998; Bisgaad-Frantzen et al., 1999). Conrad et al. (1995) separately determined four regions of particular importance for the thermostability of BLA, namely 34–76, 112–142, 174–179, and 263–276. Two of these coincide with the mutations identified earlier by Suzuki et al. (1989). BLA has been further mutated to withstand high temperature and acidic pH. Two positions 133 and 209 have been identified as important for its thermostability (Declerck et al., 1990; Joyet et al., 1992). Twelve different amino acids substitutions were made at H133 by a tRNA suppressor method, and H133Y substitution improved the stability the most (Declerck et al., 1990). Random mutagenesis and screening approach was used to create A209V substitution. The stabilizing effect of this mutation was an additive with the H133Y mutation, and the half-life of the double mutant was 10-fold higher at 90°C (Joyet et al., 1992). All naturally occurring amino acids were introduced in each of these two positions (Declerck et al., 1995). Of these, H133I substitution had stronger effect on thermostabilization than the previous substitution. Random mutagenesis and screening have also been used to identify M15T and N188S as stabilizing mutations of BLA (Shaw et al., 1999). A 23-fold enhancement in stability was achieved at pH 5.0 and 83°C in the presence of 5 mM CaCl₂ by combining these mutations with H133Y/A209V. Shaw et al. (1999) also predicted a stabilizing effect for additional substitutions: A33S, A52S, N96Q, S148N, and A379S. Four rounds of DNA shuffling and subsequent recombination produced a variant from Thermus sp. IM6051 maltogenic amylase with a 15°C increase in optimum temperature as compared to wild-type (Kim et al., 2003). The half-life of this variant was 172 min at 80°C as compared to the wild type, which was completely inactivated at this temperature. After three rounds of DNA shuffling of B. thermalkalophilus ET2 maltogenic amylase, followed by recombination of selected mutations, variants with optimal reaction temperatures 10°C higher than the native enzyme was achieved. One of the variants carrying mutations displayed a 20-fold longer half-life than wild-type at 78°C (Tang et al., 2006). To improve the performance of a maltogenic amylase (Novamyl) as an antistaling agent for breads made at low pH, two epPCR libraries of Novamyl were constructed. These libraries were screened for improved thermal stability at 80°C and activity at pH 4.3 for 25 min. A triple mutant was better than the wild type for antistaling activity in bread made at pH 4.3 and exhibited a 10°C increase in melting temperature at pH 4.0 as compared to the wild type (Jones et al., 2008). The effect of deletion was investigated on the half-life of Bacillus sp. TS-23 α-amylase. An increased half-life at 70°C was observed for the mutant enzyme suggesting that Arg210-Ser211 deletion leads to a conformational change of the enzyme (Lin et al., 2008). Site-directed mutagenesis at Asn-75, Ser-76, and His-77 calcium binding sites of α-amylase from B. megaterium WHO resulted in increased thermostability. In the presence of calcium, conversion of His-77 to Glu resulted in a four-fold enhancement in enzyme half-life and a 9°C upward shift in T₅₀, which was observed in comparison with the wild type. The H77E mutant was most stable with increased affinity for calcium ion and 5°C higher optimum temperature than the wild type (Ghollasi et al., 2013). The α-amylase (Ba-amy) of Bacillus acidicola was fused with gene fragments encoding partial N- and C-terminal region of thermostable α-amylase gene of G. thermoleovorans. The chimeric enzyme (Ba-Gt-amy) expressed in E. coli showed higher catalytic efficiency and thermostability than Ba-amy. The melting temperature (T_m) of Ba-Gt-amy (73.8°C) was also higher than Ba-amy (62°C). Circular dichroism (CD) spectra also revealed the stability of the chimeric enzyme (Parashar and Satyanarayana, 2016).

Site directed mutagenesis has also been employed for altering the optimum pH and increasing acid/basic stability of an enzyme. This method reduces the time and cost spent on screening libraries for improved variants. Saturated mutagenesis of the starch-binding domain of the α-amylase from B. licheniformis followed by selection for starch binding at low pH yielded a double mutant with an improved starch hydrolysis ratio at pH 4.5 as compared to the native protein (Verhaert et al., 2002). Site directed mutagenesis of soybean α-amylase led to the removal of hydrogen bond networks around the catalytic base residue (E380) of the enzyme (Hirata et al., 2004), raising the optimum from pH 5.4 to a more neutral pH range between 6.0 and 6.6. Richardson et al. (2002) used activity and sequence based screening to single and multi-organism DNA libraries from various environments to identify three different α-amylases with one or more aspects of the necessary phenotype: temperature stability, pH optimum, and lowered requirement of Ca²⁺. The genes encoding amylases with these properties were used as parental sequences for DNA shuffling in order to combine the best aspects of the three enzyme phenotypes. The two best chimeric α-amylases found by high throughput screening had 40-fold longer half-life for activity in the absence of Ca²⁺ ions at 90°C and pH 4.5 as compared to the most stable wild-type parent. The α-amylase from B. amyloliquefaciens with optimal activity at pH 6.0 was engineered by error prone PCR (epPCR) to produce a variant having optimal activity at alkaline pH for the detergent industry. The screening of an epPCR library for amylase activity at pH 7.0 and pH 10.0 yielded 26 variants with a very high activity. Sixteen of the selected variants were then randomly recombined by DNA shuffling and screened for increased activity at pH 10. The best mutant displayed a five-fold higher activity at pH 10 than the wild type (Bessler et al., 2003). Liu et al. (2008) observed that the mutations at two crucial positions Leu₁₃₄ and Ser₃₂₀ together affected the acid resistance of the α-amylase of B. licheniformis CICC 10181. Directed evolution was used to increase acid stability. In another example, protein stability and catalytic efficiency of α-amylase from B. subtilis was increased under acidic conditions by site-directed mutagenesis. Based on the three dimensional structure model analysis, four basic histidine (His) residues His²²², His²⁷⁵, His²⁹³, and His³¹⁰ in the catalytic domain were found to be important and single, double as well as triple mutants were constructed at these sites. The acidic stability of enzyme was significantly enhanced after mutation, and 45–92% of initial activity of mutants was retained after incubation at pH 4.5 and 25°C for 24 h as compared to the wild-type (39.5%). The catalytic efficiency for each active mutant was also much higher than that of wild-type at low pH. Due to increase in the hydrogen bonds and salt bridges after mutation, an obvious shift of the basic limb toward acidity was observed. These changes around the catalytic domain contributed to the significantly improved protein stability and catalytic efficiency at low pH (Yang et al., 2013).

Oxidation also has a demonstrated negative effect on the stability of α-amylases. Cysteine 362 was identified as the oxidation prone residue in BStA (Tomazic and Klibanov, 1988; Brosnan et al., 1992), and methionine 197, which is situated close to the active site, has been shown to be responsible for the inactivation of BLA (Borchert et al., 1995). The introduction of any non-sulfur-containing amino acid at position 197 was shown to greatly reduce the oxidation sensitivity of BLA. The specific activity was, however, very much dependent on the introduced side chain, with an apparent preference for the smaller side chains (Borchert et al., 1995). Bacillus sp. TS-23 α-amylase was truncated at both N and C termini, and further mutated to increase thermal and oxidative stability. Met231 was mutated to leucine and 483th codon of this enzyme was mutated to stop codon i.e., TAA by site-directed mutagenesis. The resultant engineered enzyme showed higher T_1∕2 at 70°C and showed compatibility to detergents (oxidative stability; Chi et al., 2010). In another example, oxidative stability of α-amylase (AmyC) from Thermotoga maritima was improved by site-directed mutagenesis. In this, methionine residues at positions 43 and 44, 55, and 62 were mutated to alanine, which is oxidative-resistant. The M55A mutant showed 50% residual activity in the presence of H₂O₂, while the wild-type α-amylase was inactive (Ozturk et al., 2013).

Product re-hydrolyzation is a problem during maltooligosaccharide synthesis using maltogenic amylase. Maltogenic amylase (MAG1) from Bacillus lehensis G1 was mutated using a structure-guided protein engineering approach to decrease the hydrolysis activity of the enzyme. W359F, Y377F, and M375I mutations reduced steric interference, altered the subsite occupation and increased the internal flexibility to accommodate longer donor/acceptor molecules for transglycosylation, which resulted in an increase in the transglycosylation to hydrolysis ratio of up to 4.0-fold (Manas et al., 2015). Maltogenic amylases also use donor and acceptor sugar molecules to form α-1,4 or 1,6 glycosidic linkages between them. This process, called transglycosylation, is responsible for formation of undesirable maltotriose in the high maltose syrups. Maltogenic amylases with high hydrolytic activity than transglycosylation activity are used to decrease the maltotriose content of high maltose syrups. Maltogenic amylase from B. stearothermophilus was engineered by site-directed mutagenesis of W177 to decrease its transglycosylation activity. The mutant W177S exhibited notable increased maltose production with a minimum amount of maltotriose under industrial conditions (Sun et al., 2016). Role of His and Glu in the catalytic activity of B. licheniformis α-amylase (BLA) was determined by performing site-directed mutagenesis at His235. It was observed that the mutant enzyme (H235E) displayed high substrate transglycosylation activity, while the wild-type BLA exhibited high hydrolysis activity exclusively. This shows that Glu235 located on a wide open groove near subsite +1 is most probably involved in transglycosylation and may recognize and stabilize the non-reducing end glucose of the acceptor molecule (Tran et al., 2014)

According to the Carbohydrate-Active enZYmes (CAZY) database, the α-amylases are grouped with different kinds of glycosyl hydrolases in GH-13 family, also known as α-amylase family (Henrissat, 1991). This family contains the enzymes with the following characteristics. They act on α-glycosidic bonds and hydrolyze it to produce α-anomeric mono-or oligosaccharides (hydrolysis) or form α-1,4 or 1,6 glycosidic linkages (transglycosylation), or show both activities. They possess a (β/α)₈ or TIM barrel structure that contains the catalytic site residues (Figure 1). Their primary sequences have seven highly conserved regions with catalytic amino acids and some amino acids that are essential for the stability of the topology of conserved TIM barrel (Kuriki and Imanaka, 1999). The first four conserved regions (I to IV) are found in the TIM-barrel on β-strands 3, 4, and 5 and in the loop connecting β-strand 7 to α-helix 7. Amino acid residues corresponding to region I are found in the C-terminal end of the third β-strand of the TIM barrel (β3). This region includes the three conserved amino acids Asp 100, Asn104, and His105 (BLA numbering). Asp100 is essential for the active site integrity because it forms hydrogen bonding with Arg229. Arg229 is a completely conserved residue. This is present within hydrogen bond distance of the catalytic nucleophile Asp231 and the proton donor Glu26. Asn104 coordinates the conserved calcium ion between the A and B domains and does not directly stabilize the active site (Boel et al., 1990; MacHius et al., 1995, 1998). His105 stabilizes the interaction between the C-terminal of β3 and the rest of TIM barrel by hydrogen bonding to Asn104 and to the backbone oxygen of Tyr56 (BLA numbering), which is situated in the loop connecting β2 to α2. Region II is positioned in β4 and has the catalytic nucleophile Asp231 and the invariant residue Arg229. These two residues are indispensable for catalytic activity and are found in all α-amylases (Svensson, 1994). Lys234 and His235 are also present in this region, which form part of subsite +2. These amino acids bind the reducing end of the glucose chain in the substrate-binding cleft (Svensson, 1994). Region III includes the conserved residue and catalytic proton donor Glu261, which lies in the C-terminal part of the fifth L-strand of the TIM-barrel. The residues forming region IV are situated in the loop connecting L7 to K7 and protect the active site from the solvent. This region has fully conserved residue Asp328, which is postulated to be involved in substrate binding, substrate distortion and in elevating the pKa of Glu261. There is a strong preference for His, Phe, Val and Asn are often observed at positions 327, 323, 324, and 326, respectively (Klein et al., 1992; Knegtel et al., 1995; Strokopytov et al., 1995). The number of residues between the conserved Gly and Pro of the region VI differentiates between an α-amylase (7 residues) and a CGTase (8 residues; Janecek et al., 1995). Region VII does not contain conserved residues, but it usually starts with Gly followed by proline at i + 2 position.

Comparison of the T. hydrothermalis α-amylase sequence with those of 21 other bacterial and archaeal α-amylases, which were representatives of more than 100 known α-amylase sequences, revealed that this enzyme as well as the other Thermococcales secreted α-amylases, contain most of the β-strands and the pentapeptide stretch located near the C-terminus of loop β3 → α3 that are present in family GH-13 (Janecek, 1997; Janecek et al., 1999). This indicates that α-amylase secreted by the members of Thermococcales belong to family GH-13. However, α-amylase from P. furiosus does not contain these conserved regions, so it belongs to GH-57 family instead. It was also showed that the α-amylase from the methanogenic archaeon Methanococcus jannaschii contains features of both glycosl hydrolase families, which indicates that these two families are either derived from a distant common ancestor or that one evolved from the other (Janecek, 1998). Also, the conserved regions of Thermococcale α-amylases show some characteristic sequence features (Janecek et al., 1999). A tryptophan is present in all the archaeal enzymes in the β4 strand. Moreover, a histidine, which has been suggested to be responsible for substrate binding, is replaced by glycine in archaeal enzymes (Svensson, 1994). It was suggested that the tryptophan found in β4 strand might be involved in catalysis instead of the missing histidine (Lee et al., 1996). Phylogenetic analysis revealed that archaeal α-amylases are closely related to plant α-amylases (Janecek et al., 1999). It also suggested that archaeal α-amylases are closer to the liquefying α-amylases, than to the saccharifying ones, which also agrees with the nature of the products of starch hydrolysis by the archaeal enzymes (Janecek et al., 1999). The phylogenetic analysis of various archaeal and bacterial α-amylases is shown in Figure 2.

The structure of α-amylase, in general, consists of a polypeptide chain folded into three domains called A, B, and C (Figure 3). Domain A is the catalytic domain, which consists of N-terminal (β/α)₈ barrel. This consists of a highly symmetrical fold which includes eight parallel β-strands organized in a barrel with a border of eight α-helices. This (β/α)₈ barrel was first observed in chicken muscle triosephosphate isomerase (TIM), which (Banner et al., 1975), and hence, it is called TIM barrel (Figure 1). It is present in all the members of α-amylase family (Banner et al., 1975). The overhang between the third β-strand and the third α-helix of the TIM barrel forms Domain B. Domain B has an irregular β-sheet structure, forms a large part of the substrate binding cleft and varies considerably in structure among different α-amylases. This domain is presumed to play a role in the substrate specificity differences observed in the α-amylases (Janecek et al., 1997). Domain C constitutes the C-terminal part of the sequence and is composed of β-sandwich structure (MacGregor, 1988). Both domains B and C are located at opposite sides of the TIM barrel. Mehta and Satyanarayana (2014) have recently attributed the role of raw starch binding to the C domain of G. thermoleovorans α-amylase (Gt-amyII). Whether this is generalized for all α-amylases or only for this particular enzyme is not yet clear. In this enzyme, two tryptophan residues, W204 and W205, corresponding to W201/W202 of BaqA (B. aquimaris α-amylase) and W200/W201 of Anoxybacillus sp. α-amylase are present (Chai et al., 2012; Mok et al., 2013; Puspasari et al., 2013), which precedes β4 strand covering the catalytic nucleophile. These two tryptophan residues have been postulated to bind sugar residues of the raw starch in raw starch degrading α-amylase Baq via stacking interactions (Puspasari et al., 2013). Detailed structural analysis reveals that in GTA and Gt-amyII, W205 is not solvent accessible and is buried, in comparison to Baq, where both tryptophans are accessible to solvent (Mok et al., 2013). The domain C has also been shown to play a role in raw starch binding in barley α-amylase (Robert et al., 2003). In barley α-amylase, a “pair of sugar tongs” site in the domain C, formed by Ser378 and Tyr380, had been shown to function in starch recognition and binding (Bozonnet et al., 2007). It may be possible that a similar site is present in Gt-amyII (Mehta and Satyanarayana, 2014).

In the maltogenic amylases, the C-domain is followed by a D-domain, the function of which is also unknown (Nielsen and Borchert, 2000). A recent investigation by Tan et al. (2008) on the crystal structure of a thermostable α-amylase Amy B from H. orenii showed that it also has an additional N-terminal domain called N domain that forms a large groove, the N–C groove, in addition to the typical domain organization of GH-13 family (i.e., A, B, and C). It was shown that the N domain has been shown to increase the binding of enzyme to the raw starch. This N-domain is also found in maltogenic amylases. It has been shown to be responsible for thermostabilization via oligomerization and substrate affinity modifications in maltogenic amylases of G. thermoleovorans and Thermus sp. (Kim et al., 2001; Mehta and Satyanarayana, 2013b).

Approximately 10% of the amylolytic enzymes possess a separate domain for binding to raw starch. Besides α-amylases, the starch binding function has been reported from some glycoside hydrolases, cyclodextrin glucanotransferases, and acarviose transferases from glycoside hydrolase family GH13, β-amylases from GH14, and glucoamylases from GH15. SBDs are classified under CBMs (carbohydrate binding modules), which have been classified into 67 families (http://www.cazy.org/Carbohydrate-Binding-Modules.html) on the basis of amino acid sequence similarity. Starch-binding domain belongs to seven different families of CBMs (CBM20, CBM21, CBM 25, CBM26, CBM34, CBM41, and CBM48). α-Amylase SBDs are majorly classified in CBM20 and CBM25 families. Although large number of bacterial, archaeal as well as fungal SBD sequences are available at CAZy database; only a few α-amylase SBDs have been characterized and well described (Iefuji et al., 1996; Sumitani et al., 2000; Guillén et al., 2007; Yamaguchi et al., 2012). Florencio et al. (2000) and Morlon-Guyot et al. (2001) reported the presence of starch binding domains (SBDs) in three α-amylases from Lactobacilli. The genes encoding the α-amylases have been sequenced (Giraud and Cuny, 1997; Morlon-Guyot et al., 2001), and the amino acid sequence analysis revealed that the C-terminal possesses SBD belonging to CBM 26 family formed by direct tandem repeat units; four modules have been reported in Lactobacillus plantarum and L. manihotivorans α-amylases and five in that from L. amylovorus enzyme. The enhanced affinity in the SBD of the maltohexaose-forming amylase from B. halodurans is due to the simultaneous interaction of the two tandem CBMs present in the enzyme (one from family CBM 25 and the other from family CBM 26). Structural features of SBDs indicate that they belong to type B glycan chain-binding CBMs. The members of this class comprise several subsites that are able to accommodate individual sugar units of the polymeric ligands. These CBMs show higher affinity toward oligosaccharides and higher polymers, and show negligible interaction with oligosaccharides with a degree of polymerization (DP) of three or less. These are called as “chain binders.” According to the Boraston's classification of CBMs based on the kind of fold (Boraston et al., 2004; Hashimoto, 2006), most of the SBDs are classified in this fold family 1 of CBMs.

The most intensively studied bacterial and archaeal α-amylases for which 3D structures have been resolved (Kim et al., 1999) include: Alteromonas haloplanctis (AHA), B. subtilis (BSUA), B. amyloliquefaciens (BAA), B. licheniformis (BLA), P. woesei/furiosus (PWA/PFA; Koradi et al., 1996). These α-amylases vary considerably with respect to the properties. In general, all known α-amylases contain a conserved calcium ion, which is present at the interface between domains A and B (Boel et al., 1990; MacHius et al., 1995; Linden et al., 2003). This calcium ion is indispensable for its stability and activity of α-amylases (MacHius et al., 1998). If the calcium is removed experimentally, the α-amylase loses its catalytic properties, whereas its restoration recovers the activity. Calcium ion is positioned far away from the active site to contribute directly in catalysis. Therefore, the role of calcium ion has been suggested to be structural (Hsiu et al., 1964; Larson et al., 1994). Calcium ion also plays a role in the stabilization of thermostability of α-amylases. This can be explained by the fact that salting out of hydrophobic residues by calcium ion in the protein structure leads to the formation of a compact structure (Buisson et al., 1987). Many α-amylases contain a single calcium ion, while α-amylases from Bacillus spp. contain three or four calcium ions and one sodium ion and a calcium—sodium—calcium metal triad, which bridges between domains A and B (Feller et al., 1999; Linden et al., 2003). This metal triad contributes to maintaining the conformational stability of protein as well as for resisting thermal inactivation of the enzyme (MacGregor, 1988; Khajeh et al., 2001; Goyal et al., 2005). There are conserved Ca²⁺ binding sites to which calcium ion bind in most of the α-amylases. A new type of Bacillus α-amylase (Amy K38), a calcium-free amylase has been reported, in which two sodium ions retain the structure and function of the enzyme, instead of calcium ions (Lee et al., 2006). Besides this, there are other α-amylases that are Ca²⁺-independent. These include enzymes from P. furiosus, Thermus sp., G. thermoleovorans etc (Koch et al., 1990; Malhotra et al., 2000; Nonaka et al., 2003; Prakash and Jaiswal, 2010; Mehta and Satyanarayana, 2013b, 2014). Ca²⁺, which stimulates most of the amylases, inhibits another α-amylase in G. thermoleovorans (Mehta and Satyanarayana, 2013a), and in B. coagulans B 49 (Babu and Satyanarayana, 1993). Besides the primary Ca²⁺ binding sites, calcium ions also bind to the secondary sites, which involves catalytic residues (Asp and Glu) located at the bottom of substrate binding cleft. The presence of these secondary Ca²⁺ binding sites at catalytic sites explains inhibition of α-amylase by Ca²⁺ at higher concentration (MacHius et al., 1995; Mehta and Satyanarayana, 2013a). Many α-amylases also contain a chloride ion in the active site, which enhances the catalytic efficiency of the enzyme (Levitsky and Steer, 1974; Feller et al., 1996). Although majorly found in mammalian α-amylases (Larson et al., 1994; Brayer et al., 1995; Ramasubbu et al., 1996), a chloride ion has also been reported in a α-amylase (AHA) from the psychrophilic bacterium, A. haloplanctis (Aghajari et al., 1998). Chloride binding increases the affinity of the enzyme for the calcium ion, therefore it is much likely that binding of chloride ion also induces conformational changes around the active site (Levitsky and Steer, 1974).

In total, three steps are involved in the catalytic mechanism for retaining glycosyl hydrolases (Davies and Henrissat, 1995). Firstly, the glycosidic oxygen is protonated by the proton donor (Glu261). This is followed by a nucleophilic attack on the C1 of the sugar residue in subsite-1 by Asp231 (BLA numbering; Nielsen et al., 1999). Once the glycon part of the substrate leaves, a water molecule is activated presumably by the deprotonated Glu261. This water molecule hydrolyses the covalent bond between the nucleophilic oxygen and the Cl of the sugar residue in subsite-1, thereby completing the catalytic cycle (Nielsen et al., 1999).

At the C-terminus of the β-strands in the TIM barrel, in the interface between domain A and domain B lies the active site cleft. The substrate binding cleft can accommodate from four to ten glucose units as revealed by the X-ray structures of α-amylases complexed with the α-amylase inhibitor acarbose. Certain of the amino acid residues of α-amylase binds to each glucose unit, which is said to constitute the binding subsite for that glucose unit. Subsite nomenclature was defined by Davies et al. (1997). The location of scissile bond defines the numbering of the subsites, with negative subsite numbers on the non-reducing side of the scissile bond, and positive subsite numbers on the reducing end (Figure 4). In the α-amylases the number of subsites present on the reducing end of the scissile bond are two or three (subsites +1, +2, and +3), while those on the non-reducing side of scissile bond varies between two and seven (MacGregor, 1988; Brzozowski et al., 2000). The X-ray structures of α-amylases complexed with acarbose shows that acarbose occupies subsites −1 to +4. Density for more than one sugar units is also observed in these X-ray structures in addition to the density for the four sugar units from acarbose (Brzozowski and Davies, 1997; Dauter et al., 1999). A study on a chimaeric Bacillus α-amylase (BA2) revealed that the extra density originates from a transglycosylation event that produces longer chained sugar molecules, which span the subsites −7 to +3, in comparison to an active site mutant of BA2, in which acarbose molecule spans the subsites −7 to −4. This observation suggests that the catalytic activity of the α-amylases is necessary for converting acarbose into a more potent longer-chained inhibitor by transglycosylation (Brzozowski et al., 2000).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.