Enhancement of Characteristics and Potential Applications of Amylases: A Brief Review

Starch is the major storage carbohydrate of plant products. Amylases are the group of enzymes hydrolyzes starch and related polymers to smaller oligosaccharides and less amount of monosaccharide. Microbes are the major sources of amylases, exploited for large scale production in different industries. Recently, protein engineering has been applied to improve the structural and physicochemical properties of the enzyme for its potential applications. Amylases are mostly used for liquefaction of starch in the purpose of glucose, maltose, and high fructose containing syrup preparation, malto-oligosaccharides production, desizing, production of biofuel, detergent preparation, waste management, and preparation of digestive aids.


INTRODUCTION
Starch is the major storage plant polysaccharide, cleaved by a group of hydrolytic enzymes known as amylases. Amylolytic enzymes are widely distributed in microbial, plant and animal kingdoms (Saini et al., 2017). Several enzymes like α-amylase, pullulanase, isoamylase, CGTase (Cyclodextrin glycosyl transferase), branching enzymes, α-glucosidase, amylopullulanase, and neopullulanase belong to amylase family. The production of amylases was carried out by submerged fermentation or solid state fermentation; however, later has several potential benefits including cost effectiveness, easy to operate, and better for downstream processing (Sharif et al., 2019). Use of agricultural products in solid sate fermentation reduces the pollution levels as they are commonly destroyed by open burning. Protein engineering is the most powerful approach to increase the physicochemical properties of the enzymes. Enzyme engineering technique is applied for preparation of high efficiency, oxidation insensitive, and chelator resistance enzyme for industrial uses (Uddin et al., 2017). Starch-degrading amylolytic enzymes have great significance in biotechnological applications in different sectors. The commercial application of amylase was started from 1894, when Japanese scientist Jokichi Takamine prepared 'Taka Diastase' from Aspergillus oryzae. The era of bacterial amylases was stared form the year 1917 when amylase was isolated from Bacillus subtilis by the French scientists Boidin and Effront (Md et al., 2014). like liquefaction of starch, food preparation, baking, brewing, detergent preparation, desizing, paper making, bio-fuel production, and digestive medicine preparation (Yan and Wu, 2016).

Types of amylases
Initially, the term amylase was used to describe the enzyme hydrolyses -1,4 glycosidic bond of amylose, amylopectin, glycogen and related products. However, microbial amylases are commonly classified in different groups.
b. -Amylases: Cleave α-1,4 glucosidic bond from the non-reducing ends of the polysaccharide chains. They produce maltose from amylose, and maltose and limit dextrin from amylopectin and glycogen.

Production of amylases
Several biotech industries are using large amount of various microbial enzymes like amylase, cellulose, pectinases, protease, lipase, and others. Among these, amylases are maximally used for owing their multipurpose applications. Different microbes are the good sources of amylases ( Table 1). Large-scale production of amylases by using microbes is advantageous for their high yielding capacity that minimizes cost/production ratio. At the industrial level, amylase production is carried out in both solid state fermentation (SSF) and submerged fermentation (SmF). Several parameters including nutrient supplementation, pH of the medium, temperature, water activity, and aeration are being considered for optimum growth of the microorganisms as well as enzyme production (Shahen et al., 2019).
Among the different factors, pH and temperature have greater impact on microbial growth and enzyme production. Fungi are commonly grown in slightly acidic medium; whereas bacteria needs neutral to slight alkaline medium for their growth and enzyme production. Most of the organism can grow in particular temperature. Mesophilic organisms are able to grow in wide range of temperature 30-45 C, but give thermostable enzymes. Supply of nutrients like carbohydrate, protein, ammonium salts, vitamins (B 1 , B 2 , B 6 , biotin, folic acid, etc.) and minerals (calcium, magnesium, zinc, copper, sulfate, phosphate, etc.) provide energy for cellular growth, supply essential components for protoplasm formation, and create favorable conditions for enzyme synthesis (Samanta et al., 2014). Surface area is most important to provide much amount of oxygen. In SSF, surface area is greatly increased, which influences the rate of oxygen transfer, aerobic metabolism, and enzyme production.
Initially, SmF was more popular for its easy handling and optimization. Later, SSF has gained more importance for large-scale production of enzymes in view of several economical and engineering advantages.  SSF is the useful technique for production of enzymes, pharmaceutical components, bio-bleaching agents, food ingredients (Soccol et al., 2017). Most of the time, synthetic or semi-synthetic medium was used in SmF. However, moist agricultural polymeric substrates such as wheat, rice bran, rice husk, cassava, sunflower meal, cottonseed meal, soybean meal, and pearl millet are used in SSF process (Soccol et al., 2017;Abdulaal, 2018). These substances provide solid support and nutrients (polysaccharides -cellulose, hemicellulose, pectin, starch; crude protein; lipids, ions -calcium, magnesium, phosphate) for enzyme synthesis ( Table 2).
Applications of agricultural wastes decrease production cost as well as pollution rate. Supplementation of protein, vitamins (B 2 , B 6 biotin, folic acid), and minerals (magnesium, calcium, zinc, copper, phosphorus, sulfur) can increase the rate of enzyme production many folds. Thus, SSF has several advantageous over the SmF ( Table 3).
The usual method of process-optimization was alteration of one variable at a time (OVAT). OVAT is a time-consuming, less predictable and inconvenient process as the independent variables are separately optimized. Later, Taguchi method was adopted in this purpose, which was based on computerized statistical calculation. The observed data were feed to the software that gives the indication of optimum production levels by reducing the variance.  Advantages of RSM  It deals with the multiple response problems and explores the relationships between several explanatory variables and one or more response variables.  This model is easy to estimate overall response and also be applicable even when little is known about the process.  It is applied to maximize the production of a special substance by optimization of operational factors.  It is a first-degree polynomial model which depends on factorial experiment or a fractional factorial design. This model has sufficient capacity to determine effective variables which controls the response variable(s) of interest.  This model identifies the significant explanatory variables which are used for development of central composite design as well as second-degree polynomial model.

Disadvantages of RSM
 Design of experiment is necessary to obtain an optimal response.  The response generated through this model on the basis of approximation.  It estimates minimum variance which is related to un-correlation.  The moments of the distribution of the design points are constant.  Multiple response variables in RSM create some problem because what is optimal for one response may not be optimal for other responses.  Practitioners must be aware about its approximation to reality as it is fully based no statistical methods.  There is chance of variation between estimated optimum point and optimum in reality due to errors of the estimates and inadequacies of the model.

Advantages of ANN
 It can be implemented in any application and applicable for complex linear, non-linear and complex relationships.  One or more descriptors and/or response variables can be used to develop the model. It is applicable for categorical and continuous data.  The model is less sensitive to noise than statistical regression models.  Preparation of model is possible in case of incomplete information. The processing of information occurs in a highly parallel way and can continue without any problem when an element of the neural network fails.  It can infer unseen relationships and predict on unseen data from initial inputs.  It does not impose any restrictions on the input variables. Fixed relationship in the data is not necessary.  This model is better than any other model where data is heteroskedastic with high volatility and non-constant variance relationships.  It has ability to store the information of the entire network in the form of working memory or associative memory.

Disadvantages of ANN
 The model is computer dependent and needs training to operate. It requires high processing time for large neural networks.  It is not possible to explain the total behavior of the network.  Some difficulties arise to determine the proper network structure.  It is very difficult to screening the problem present in the network. This model does not give any meaningful way how the results are calculated.  To optimize the parameters, many parameters to be set in a neural which is challenging, due to overtraining. Two types of plot i) three-dimensional space, ii) contour plots are used to visualize the shape of the response in graphical system. Analysis of variance (ANOVA) is applicable to determine the lack of significance of experimental data fitting in RSM (Most et al., 2018). RSM provides information about the optimum response where each variable interacts among them for maximum yield (Gu et al., 2005;Rao and Satyanarayana, 2007).
Artificial neural network (ANN) is the recent technique for process-optimization. Basically, it is a mathematical model having an extensive range of applications from biological system to computer engineering. ANN has three layers of connections. First layer is inputs of the independent variables (neuron), second layer is hidden neurons, and third layer represents output neurons. The neurons of the first layer are connected to hidden neurons where several calculations have been done to achieve output response. The algorithm was made on the basis of influencing input neurons and their relations to the output result. The results of ANN are also plotted through contour plots. In case of nonlinear data, ANN methodology is better than RSM as ANN does not require a design to obtain predictive models (Cheok et al., 2012). The advantages and disadvantages of RSM and ANN are given in Table 4.

Protein engineering and approach to modification of amylases
Protein engineering is the step to improve the functional characteristics and stability of the protein. It is mostly applied in fewer proteins (enzyme) those are industrially important. The protein engineering is primarily carried out by site-directed mutagenesis which makes a precise change in the sequence of encoded genes on DNA to alter the amino acid sequence of the desire enzymes. Site-directed mutagenesis technique simplifies DNA manipulations as it lowers the hazards of crude mutagenesis of cells or organisms and easy to isolate the desire mutants from thousands or millions of offspring.
Site-directed mutagenesis has several importances. It gives idea about structural-functional relationships of gene, helps to develop a mutant protein with novel properties. There are several methods for site-directed mutagenesis. The first method was primer extension (the single-primer method). It is based on in vitro DNA synthesis for preparation of oligonucleotide (7-20 nucleotides long) that carries a base mismatch with the complementary sequence (Uddin et al., 2016). This process is easy to perform in single stranded vector (e.g. M13).

Problem Expected solution
The newly formed double-stranded heteroduplex molecules are contaminated with either single-stranded non-mutant template DNA or partially double-stranded molecules. This contamination decreases the number of mutant progeny.
They can be removed by sucrose gradient centrifugation or by Agarose gel electrophoresis.
In in vivo condition, during DNA synthesis a mixed population of mutant and non-mutant progeny has been developed.
Mutant progeny have to be purified away from the wild type molecules.
In vivo DNA repair system favors the mismatch repair of nonmethylated mutant DNA and thereby eliminating a mutation which lower the yield of mutant progeny.
This problem is solved by preventing the methyldirected mismatches repair system. It is done by using host strains carrying the mutL, mutS, or mutH mutations.
Both mutant and non-mutant progeny will be grownup after upon replication.
Suppression of the growth of non-mutants is essential.
Always require a single-stranded template.
In PCR-based mutagenesis the template can be singlestranded or double-stranded, circular or linear.
Instead of its simplicity, there are numerous difficulties in single-primer method ( Table 5) and the yields of mutants depend upon several factors. Two or three adjacent nucleotides are possible to change by using site-directed mutagenesis which generates an amino acid substitution at the site of interest. However, several oligonucleotides substitutions are time consuming and laborious. Another process for alteration of one amino acid is cassette mutagenesis, a fragment of the gene is replaced by another fragments containing the desired codon. Application of doped oligonucleotides is the alternative approach for desire mutagenesis (Neylon, 2004).
The polymerase chain reaction (PCR) can efficiently be used for site-directed mutagenesis. Single base mismatched between the amplification primer and the template can include a mutation in amplified DNA. Alternatively, incorporation of mutant base is also possible in any parts of PCR-product. This method is more complicated as it requires four primers for two sets of overlapping DNA and three PCRs (two for two sets of DNA and one to fuse the overlapping segments). Using the commercial kit, PCR mediated mutagenesis is much easier. In this purpose, two methods are adopted. According to Exsite method, desire DNA is present in both strands of vector and amplified through PCR while one of the primers carries the desired mutation. After amplification, linear duplexes carry the mutated gene along with original circular template DNA which can be cleaved by using DpnI endonuclease when the template DNA is derived from E. coli. The second method is Gene Tailoring in which the target DNA is methylated before mutagenesis and the overlapping primers carry the desire mutation. Linear amplicons are produced along with methylated template DNA. After transformation, linear mutated DNA is circularized but host cell McrBC endonuclease digests the methylated template.
To improve the characteristics of a protein, it is better to mutate the gene at random instead of site-directed mutagenesis. Normally, PCR is error prone due to low fidelity of Taq polymerase and there is high probability of base alteration in amplicons. The random mutation is possible by using error-prone PCR system. This is achieved by different ways: i) introduce a small amount of Mn 2+ instead of Mg 2+ , ii) to include an excess of dGTP and dTTP relative to other two nucleoside phosphates and iii) by using nucleoside analogs. First two methods have possibilities to achieve error rates of one nucleotide per kilobase. Error-prone PCR can randomly change the one amino acid to another. For example, a single point mutation in a valine codon can change it to others encoding phenylalanine, leucine, isoleucine, alanine, aspartate, or glycine. Random insertion/ deletion (RID) process is used for random incorporation or elimination of amino acids and this technique is mediated by ligating an insertion or deletion of a cassette at nearly random locations within the desire gene (Murakami et al., 2003).
Engineering of amylase improves the physicochemical characteristics like pH profile, thermostability, chelator insensitivity, and oxidation resistance. Altered enzymes are exploited in food, textile, paper, detergent industry and clinical system. The complete three dimensional structures, active site residues, and the position of functional amino acid residues have been determined by using bioinformatics through various software's (DSMODELER; Geno3D, and Qsite Finder). Several strategies have been applied to make the better quality of enzymes. Exchange of some oxidation prone amino acids (methionine, tryptophan, cysteine, and histidine) by other amino acids (not affected by oxidizing agents) increases oxidation resistance capacity.
Maltogenic amylase, maltotetraose forming amylase, and maltohexaose forming amylase produce maltose, maltotetraose, and maltohexaose respectively from partially hydrolyzed starch. CGTase helps to form cyclodextrin. All these amylases are used in food and pharmaceutical industry, brewing process, ethanol and biofuel production, detergent preparation, animal feed preparation, and waste management (Yan and Wu 2016;Saini et al. 2017;John, 2017). Liquefaction of starch is most important for preparation of glucose and fructose/glucose syrups. Enzymatic treatment produces maltose from starch. Maltose is widely used as sweetener, used in food industries due to its low tendency of crystallization and relatively nonhygroscopic nature. Process is less efficient to remove total amount of starch Malto-oligosaccharides mixture (maltooligomer mix) is obtained from corn starch after its amylolytic digestion. Maltooligomer mix is less sweet than sucrose and low viscous than corn syrup. It is mainly used as substitute of sucrose. Maltotetraose syrup is the ingredient of geriatric and infant foods. It is used as replacement of sucrose without altering the taste and flavour of the food. This application also controls the freezing points of the frozen foods. Amylases are also used for the clarification of beer and fruit juices, which improve the quality and clarity of the products (Kumari et al., 2012; Dey and Banerjee, 2014). Application of α-amylase in desizing process is one of the oldest uses of amylase in industrial sector. Starch paste is used as sizer during weaving and enzymatic treatment removes starch sizer from cotton fabrics. Although, there are several processes of desizing, but use of thermostable α-amylase in this purpose is more convenient ( Table 6).
Alcohol and biofuel production is carried out by using α-amylase and saccharifing amylases which hydrolyze the starch into smaller saccharides followed by alcoholic fermentation. Starchy wastes promote microbial growth and environmental pollution. Amylolytic treatment of starchy wastes can produce valuable products such as microbial biomass, animal feed, and biofuel (Kumar et al., 2014).  Kumari et al., 2012). The fungal αamylase from Aspergillus oryzae is used as digestive aids for the treatment of indigestion. The liquid and capsule formulations of the digestive medicines are available in the market ( Table 7).

CONCLUSION
In the era of biotechnology, microbes are most useful for mankind. They are exploited for antibiotic production, enzyme preparation, biomass and single cell protein synthesis, production of medicinal components (amino acids, vitamins, steroid), and waste management (Rahman et al., 2019). Microbial enzymes have multidimensional uses starting from food industry to pollution control. Among the microbial enzymes, α-amylase has promising market value. The demand of microbial α-amylases is increasing day-by-day. Several agro-industrial residues are competently used in SSF to reduce production cost for large scale enzyme preparation. This step is an eco-friendly process and a great advancement towards clean world. An approach to protein engineering makes the enzyme more thermostable, oxidant resistant, chelator insensitive along with better pH profile. All these factors increase the potential uses of amylases in different sectors. Amylases, particularly α-amylase are efficiently used in food, pharmaceutical, textile, detergent, paper, and brewing industry. Thus, amylases are fairly excellent industrial enzymes and the demand would always be high in near future.