Keywords

4.1 Protein Expression: Introduction to E. coli Expression System, Yeast Expression System, Insect Expression System. Higher-Eukaryotic Expression Systems

4.1.1 Protein Expression

The main objective of recombinant protein expression is to produce a target protein in commercial quantity in a cost-effective manner without compromising its biochemical and biological activity. Several protein expression systems have been used to produce large-scale proteins for therapeutic (biopharmaceuticals) and diagnostic purpose. Expression and purification of full-length protein is sometimes unnecessary, as the desired activity or property of the protein of interest can be achieved by a specific domain(s). Once the desired protein or domain(s) is finalized, next critical step is choosing the suitable expression system.

An expression system is defined as genetic constructs (expression vector with the desired gene sequence) specifically designed to produce protein of interest at high level, inside a host cell. The fundamental criteria for expression system selection include anticipated application of desired protein; resources availability; expenditure cost; and time. The four most well established expression systems in use for pharmaceutical purposes are bacteria, yeast, baculovirus, and mammalian, and will be discussed in details.

4.1.1.1 Bacterial Expression System

Bacterial expression system is most popular and first choice for rapid and economical production of pharmaceutically important recombinant proteins [1] Escherichia coli has emerged as most commonly used industrial microorganism, as about one-third of all the pharmaceutical relevant proteins are purified from it [2]. Other bacterial strains used for pharmaceutical purpose are Lactococcus lactis and Bacillus subtilis. The major advantages of using this system are: (1) Simple and scalable, with no sophisticated equipment requirement, (2) Quickest, (doubling time of E. coli is 20 min), (3) Easily manipulated due to availability of its genetics knowledge, and (4) Economical, as one can produce tones of recombinant protein in no time [3, 4].

An expression vector is required to insert the target gene in the bacterial cells. Numerous articles are available online that extensively discuss the desired properties of an expression vector. Inexpensive usage, a little or no leaky expression with a reliable and tunable induction promoter are some of the desirable qualities of an expression vector for large-scale protein production [5, 6]. Commonly used and commercially available vectors are the pET series, pQE series (Qiagen), pGEX, etc. for single protein expression and pACYC, pBAD, and pSC101 series or single plasmid system like the Duet vectors (Novagen) for more than one proteins expression at the same time.

Despite being the most efficient and cost-effective expression system , it has several shortcomings especially when expressing mammalian proteins. These problems include stability of mRNA and protein inside the host, codon bias [7], inclusion bodies formation [8], and protein folding and solubility [9], and absence of key post-translational modifications (glycosylation, carboxylation, and amidation). Increasing demand of the pharmaceutically important proteins in the last 5 years has led to the development of new codon optimized and genetically modified bacterial strains capable of performing desired post-translational modification, and improved expression vectors to overcome above mentioned problems [10,11,12]. A list of the commercially available modified strains with their key features is given in Table 4.1. The use of other strategies like growing cells at lower temperature to prevent protein aggregation 15–23 oC [15, 16], co-expression of molecular chaperones or the post-translational modifying enzyme [15,16,19], and translocating protein to periplasmic space has helped in increasing protein yield [18,19,22]. The protein expression in periplasmic space emerged as most successful strategy to produce several pharmaceutical proteins such as scFc, growth hormone, etc. owing to several advantages over others [23]. Protein isolation from inclusion bodies is also a viable and cost-effective option for pharmaceutical companies, as the purity of the protein in IBs is ~95% [22,23,26]. It is often better to change expression system to higher hierarchy as they are better equipped in folding and complex post-translational modifications to maintain the bioactivity of desired protein.

Table 4.1 List of modified bacterial strains [10]
Full size table

4.1.1.2 Yeast Expression System

Yeast being a eukaryotic organism and acts as a connecting link between E. coli and mammalian systems. It has properties similar to both the systems thus serves as an excellent expression system for pharmaceutically relevant protein production. The major advantages it has over other eukaryotic expression systems are: (1) Ability to grow in high densities in limited time, (2) Simple and cheap media requirements, (3) Easy genetic manipulation, and (4) Safe pathogen-free production [27]. Unlike bacterial expression system, yeast expression system has the ability to accomplish proper posttranslational modifications and extracellular expressions [28, 29]. Saccharomyces cerevisiae, Pichia pastoris, Kluyveromyces lactis, and Schizosaccharomyces pombe [30] are few commonly used yeasts. S. cerevisiae and P. pastoris are used extensively for the production of several therapeutic important recombinant proteins such as human insulin, human serum albumin, alpha 2b, trypsin, collagen, hepatitis B vaccines, and human papillomavirus (HPV) vaccines [31, 32]. The expression vector is designed so that genomic integration of expression cassette could take place leading to generation of stable expression clones with multiple copies of target gene [33]. P. pastoris based vectors also haveinducible promoters (PAOX1 (most widely used), GAP, FLD1, PEX8, and YPT1) and antibiotic selection markers, for the selection of multi-copy transformants [34]. Some commercially available plasmids have these features incorporated (such as the pYEDIS, pPIC9K, pPICZα vector).

In recent times, P. pastoris being a obligate aerobic yeast has gained more prestige as it can use methanol as a carbon source which lead to the development of an expression system based on the utilization of the inducible AOX1 promoter [35]. Other benefits include correct protein folding and secretion (by Kex2, PHO1, or α-MF as signal peptidase) outside the cell [36].

Despite being an efficient expression system , yeast has drawbacks like limited secretory expression of heterologous proteins resulting in low protein yield and limits its commercial use [37]. To address this issue, strategies involving optimization of cultivation parameters (induction temperature and time, pH, oxygen, and nutrient supply), protein-based host selection, gene copy number, co-expression of secretory proteins such as chaperones, engineering of secretory pathways, have been employed to improve the expression of proteins [38, 39]. The other major drawback is hyper N- and O-linked glycosylation of proteins unlike mammalian system, which affects the protein immunogenicity [40]. To overcome this shortcoming humanization of yeast by expressing mammalian specific glucose transferase and omitting yeast specific genes involved in glycosylation has been tried and found its application in producing humanized IgGs in yeast [39,40,41,42,45]. Latest technologies such as CRISPR/Cas9 [46] and GlycoSwitch [47] are now used for yeast genome engineering for this purpose. Many improvements are still required in the yeast expression system before it can be used for therapeutic commercial purposes; therefore, companies are shifting their focus to other more complex expression systems.

4.1.1.3 Insect Cell Expression System

Baculovirus-mediated insect cell expression systems are widely used these days to produce large quantities of proteins for pharmaceutical purpose. These proteins are difficult to express either in bacteria or yeast due to improper protein folding or posttranslational modifications. The major advantage of this system remains the presence of post-translational modifications similar to mammalian system, thus avoiding the problem of immune-reactivity [48]. The other advantages include: (1) Cheap protein production cost as compared to mammalian system, (2) High capacity of expressing multiple genes at the same time due to large and flexible viral genome (130 kb), (3) Safe, as they do not infect humans, (4) High protein yield driven by the strong promoters such as polyhedrin or p10, and (5) Easy downstream purification [49, 50].

Baculovirus vectors are used to insert desired gene and transfected into cultured insect cells. The most commonly used baculovirus system are Bac-to-Bac (Invitrogen), BacPAK (Takara), and BaculoGold (BD Biosciences), they are commercially available and has been widely used [51]. Recently, two rapid and simple baculovirus expression vector systems have been developed named as MultiBac [52] and Golden Gate-based system [53], which can be used to express multiple genes at the same time. The most common insect cells used for the production of protein are Sf9, Sf21 (Spodoptera fugiperda), High5 (Trichopulsia ni), and S2 (Drosophila melanogaster embryos) cell lines [54]. These cell lines have proven themselves for the application of recombinant protein expression from a variety of expression platforms due to their ability to grow in suspension and serum-free medium [55]. Recent advancement in genetic engineering of baculovirus system, this system has been promoted from being used during pest control to production of several recombinant proteins (viral antigens) by biopharmaceutical companies for human use [56]. Several vaccines are now available for the commercial use, which include Cervarix™ (GSK, Rixensart, Belgium), FluBlok™ (Protein Sciences Corp., Connecticut, USA), Provenge (DendreonInc., Seattle, WA), and Chimigen (Virexx Medical Corp., Calgary, Canada), [54,55,58]. Several of the known subunit vaccines against Chandipura virus, hepatitis E virus, and West Nile virus are also synthesized in insect cells [59]. Since, baculovirus are known not to infect humans, they are being evaluated to be used as efficient delivery vehicles for gene and cell therapy [60].

Despite being a promising expression system , baculovirus has several drawbacks. For instance, time consuming cloning procedure to generate stable recombinant virus, expensive media requirements, and cell lysis by baculovirus infections resulting in suboptimal protein processing. The major drawback is differential glycosylation pattern in comparison to humans, thus limiting the therapeutics use of recombinant protein [50]. In recent years, efforts are directed to deal with these problems. The stable transformation of insect cell line with plasmid having early baculovirus constitutively active promoters (e.g., IE1) [61] or using pre-infected cells [baculovirus infected insect cells (BIIC)] to avoid making of stable system has significantly reduce the processing time along with improved protein production [62]. Engineering of insect cells to have mammalian glycosyltransferases enzymes is also tested [63]. Regardless of recent advancements there is a room for improvement to produce cost-effective, industrial scale, and therapeutics standard recombinant protein in insect cell lines [64]. Several good articles are available which covers the baculovirus system in more details [63,64,67].

4.1.1.4 Mammalian Expression System

Mammalian expression system is the ideal choice for production of therapeutically important proteins because they perform similar post-translational modifications along with proper protein folding, which are critical for bioactivity of the protein. Other advantages include secretion of proteins in the cell culture, preventing additional step of protein purification [68]. Some mammalian cell lines can grow in suspension culture and serum-free chemically defined media, enabling large-scale reproducible protein production [69]. Several different mammalian host cell lines, such as Chinese Hamster Ovarian (CHO) cells, baby hamster kidney (BHK21) cells, and murine myeloma-Sp2/0/NS0 cells, have been used for large-scale production of therapeutic proteins [70]. Recently, the use of human cell line (Human Embryonic Kidney (HEK 293) cells and fibro-sarcoma HT-1080 cells) has gained importance due to identical post-translational modifications of expressed proteins [71].

Like any other expression systems, plasmid based or viral based vectors (adenoviral vector, vaccinia vector, Semliki forest viral vectors) are used to transfect desired gene into cultured mammalian cells to form either transient or stable cell lines. However, efficient integration of transgenes into correct genomic loci still remains a major challenge during stable cell line production. Incorporation of chromosomal elements (nuclear scaffold/matrix attachment regions (S/MARs) and ubiquitous chromatin opening elements (UCOEs)) into plasmid vectors is found to have a positive effect on stable gene expression. Transposon based vectors and site-specific recombinase systems, such as Cre-Lox and Flp-FRT, are also found useful in targeted integration of the gene in host genome for the production of stable cell line expressing recombinant proteins [72].

In the year 1968, the production of FDA approved first recombinant glycoprotein, tissue plasminogen activator (tPA, Activase) in CHO made a revolution [69]. Subsequently, a number of vaccines like Herpes simplex virus (HSV) vaccine [73], Synagis® vaccine used against respiratory syncytial virus [74] etc. and several therapeutic proteins, Drotrecogin alfa (XIGRIS®; Eli Lilly Corporation, Indianapolis, IN), recombinant factor IX Fc and VIII Fc fusion protein (Biogen, Cambridge, MA), dulaglutide (TRULICITY®; Eli Lily, Indianapolis, IN) etc. [68] are produced in mammalian cell line. Some of them have already received FDA approval.

The major drawback of using mammalian expression system is high cost of protein production, due slow cell growth, expensive media, and culture conditions (continuous CO2 supply, expensive transfection reagents). In recent years, mammalian cells have been further developed for the commercial production of broader range therapeutic proteins by selecting high protein-producing stable cell clones using methotrexate (MTX) amplification or glutamine synthetase (GS) system technology and high-throughput fluorescence-activated cell sorting (FACS)-based screening method [73,74,77]. Other advancements include genetic modification of mammalian cells by over-expression of anti-apoptotic proteins (bcl-2 family members and Bcl-x(L)) [78] or by inducing cell cycle arrest by adding anti-mitotic agents (such as hydroxyurea, nocodazole, colchicine, paclitaxel or vinblastine) [77] to increase cell viability along with high cell density which eventually lead to elevated protein productivity.

4.2 Protein Purification: Principle of Heterologous Protein Purification following Expression. Use of His Tag, GST-Tag, MBP-Tag, TAP-Tag, Myc-Tag

4.2.1 Protein Purification

Procuring pure and biologically active desired protein after expression is a daunting task. Separating desired protein from the rest of cellular protein pool is an essential prerequisite step for commercial production of therapeutic proteins. Several different chromatographic techniques like size exclusion, ion exchange, hydrophobic interaction, affinity, and ammonium sulfate cut-off method, are widely used to isolate desired protein from other cellular impurity. These techniques rely on exploiting protein properties like, electric charge, solubility, size and hydrophobicity, therefore, optimization of the specific purification method is a time consuming and cumbersome task. Among the above mentioned techniques, affinity chromatography is the most time and cost-effective, and usually a single step purification method [79].

In affinity chromatography, protein is expressed in fusion to an affinity tag which significantly help in protein purification. Some of the known affinity tags also increase protein solubility without affecting biological activity of the protein, an additional advantage [80]. Polyhistidine tag (his-tag) is the most commonly used affinity tag. The purification is based on IMAC (immobilized metal-ion affinity chromatography) , where adsorption of protein occurs due to coordination between an immobilized metal-ion (Ni2+ or Cu2+ ion) and an electron donor groups from the protein surface (stretch of 6–10 histidine (tryptophan and cysteine)). The protein is purified using imidazole gradient [81]. The other commonly used tags are polyarginine-tag, FLAG-tag, c-Myc-, S-, and Strep-tag; they all are around the same ~10 amino acid. Due to their small size, fusing these tags either at N- or C-terminus of desired protein usually does not affect its structure or biological activity. FLAG-tag is a hydrophilic octapeptide (DYKDDDDK) recognized by the M1 mAb resin. Due to non-reusability of antibody resin, the effective cost is high thereby restricting its widespread use. Recently, the development of anti-FLAG molecularly imprinted polymers (MIPs) approach using tetrapeptide DYKD as template has provides a cost-effective alternative solution for purifying FLAG-derived recombinant proteins [82]. Strep-tag (WRHPQFGG) binds to biotin and therefore, recombinant protein can eluted using biotin as competitor in buffer. Use of desthiobiotin facilitates regeneration and repeated use of these resins. Recent development of Strep-tag®II and Strep-Tactin has enhanced the use of Strep-tag owing to their higher affinity for the biotin [83]. Next comes the large molecular weight affinity tags (more than 200 amino acids), glutathione-S-transferase (GST), maltose-binding protein (MBP), N-utilization substance protein A (NusA), thioredoxin (Trx), ubiquitin, and SUMO [84]. Some of these affinity tags not only aid in purification but also increase the solubility of protein. MBP and GST act as both solubility enhancer and affinity tag, while NusA, SUMO, and Trx only increase solubility [85, 86]. MBP is a periplasmic E. coli protein with high solubility. MBP when fused with desired protein increases fused protein solubility due to its intrinsic chaperone activity [87]. NusA is a 55 kDa, elongation factor which regulates transcription in E. coli. As a fusion partner it improves the solubility of the protein due to its intrinsic high solubility [88]. SUMO (small ubiquitin-related modifier) protein is a reversible post-translational modification at ε-NH2-group of lysine residues of target protein. It is known to increase the solubility and expression of the protein. SUMO-specific proteases removes the SUMO tag from the target proteins thereby reducing the erroneous cleavage within the target protein [80]. Thioredoxin (Trx) is a small and highly soluble E. coli protein. Like NusA tag, Trx itself does not act as an affinity tag and thus, requires fusion partners during purification step. List of affinity tags are given in Table 4.2. However, now-a-days larger soluble tags are being replaced by small soluble tags like SET tag [91] and Fh8 tag [87] to overcome the problem faced by size of large tags.

Table 4.2 Affinity and solubility tags for recombinant proteins [modified from refs [89, 90]
Full size table

Tandem affinity purification (TAP) has also gained popularity in recent times. In TAP, desired protein is fused with at least two different affinity tags [92] or one affinity and another solubility enhancer fusion tag [93], depending on need. This helps in purifying desired protein using both the tags sequentially resulting in considerable reduction of nonspecific proteins. Overall it helps in increasing the purity of the desired protein making it useful for therapeutic purpose [80]. Numerous combinations of solubility-enhancing and affinity tags have been exploited in order to enhance both protein solubility and yield of the desired protein. Some of the commercially available TAPs are S3S-TAP-tag (is a recently developed system suitable for purification of mammalian protein complexes), FF-ZZ TAP-tag, Strep/FLAG-TAP (SF-TAP), GS-tag, PTP-tag, etc. [80].

Even though affinity tags are routinely used in laboratory protein purifications, they have a limited use in commercially therapeutic applications. Several times these tags cause structural and activity changes or immunogenicity problems. Sequence specific protease or chemical cleavage methods are developed for the removal of these tags. The tobacco etch virus (TEV) protease, thrombin, Enterokinase, and factor Xa are some of the most commonly used protease to remove tags [89]. Most of the commercially available vectors have a protease cleavage site designed between fusion tags and desired proteins. The solubility of a desired protein after tag removal cannot be predicted and enzymatic cleavage might cause negative effects, such as product heterogeneity due to cleavage at multiple sites, precipitation or poor recovery [94]. Chemical cleavage with CNBr-based method has advantages over enzymatic cleavage, as it is easy to remove from the reaction mixture and is cheap. Their use is largely restricted due to their harsh reactive nature and unwanted protein modifications making purified protein unsuitable for therapeutic use [85]. The use of the poly-ionic peptide tags (addition of 3–5 charge amino acid sequence) has shown to enhance solubility of the desired protein, regardless of their position at N- or C-terminus of the protein. The tags enhance the solubility of protein by increasing repulsive electrostatic interactions between protein molecules due to additional of charge from the tags. Due to their small size, the presence of poly-ionic peptide tags do not affect the structure or biological activity of the protein, an add-on advantage [94].

4.3 Proteomics: Introduction, Protein Detection Array, Protein Informatics, Domain Analysis, and Structure Prediction

Proteomics is defined as the study of proteome or a set of proteins found in a cell, tissue, or a whole organism. The importance of proteomics lies in the fact that unlike genome, the proteome is not constant and changes from cell to cell over time, making an individual unique or different. The proteome provides a snapshot of the cell in action and the proteomics aims at understanding the proteome status at a large scale under certain physiological or diseased conditions [95]. The term “protein” was initially introduced in 1938 by the Swedish chemist JönsJakob Berzelius, working in electrochemistry while trying to describe a class of macromolecules made up of linear chains of amino acids [96]. Although proteomics research began in 1975 with the introduction of 2-Dimensional Gel Electrophoresis by O’Farrell and Klose, it was not until the early 1990’s, when the term “proteomics” was coined by Mark Wilkins, a Ph.D. student at the Macquarie University, Australia [97].

Proteomics is a rapidly growing field with cutting-edge technologies used to investigate expression of proteins, post-translational modifications, and involvement of proteins in metabolic pathways and protein interactomics. The most commonly applied are mass spectrometry (MS)-based techniques such as Tandem-MS and gel-based techniques such as differential in-gel electrophoresis (DIGE). Another technique complementing to MS is protein microarray that has been widely applied as a promising proteomic technology with great potential for protein expression profiling, biomarker screen, drug discovery, drug target identification, and analysis of signaling pathways in health and disease [98]. The recording and analyses of the enormous amount of data generated by these high-throughput technologies are facilitated by the development of databases and online servers that are critical not only for recording and storing this data but also enable structure, function, and domain prediction of a protein [99]. For example, four major databases—UniProtKB, IntAct, Reactome, and PRIDE are responsible for storing all the up-to-date information generated for a protein [98,99,102]. In addition to that, several prediction software and servers such as Phyre2, FoldX, BisKit, etc. have facilitated protein structure prediction [101,102,105].

4.3.1 Protein Detection Array

Protein array analysis is a technique by which proteins spotted in defined locations on a solid support (a protein microarray, or protein chip) are probed for interactions with a probe molecule in a high-throughput, parallel manner [104,105,108]. Protein array analysis is used to screen protein function, drug discovery, biomarker discovery, expression profiling, and antibody analysis [109, 110]. Typically, a protein microarray is prepared by immobilizing proteins onto a microscope slide using a standard contact spotter or noncontact microarrayer [111]. The microscopic slide surface can be made of aldehyde and epoxy-derivatized glass that get attached to amines and nitrocellulose or the surface could be nickel-coated that relies on more specific affinity attachment of His6-tagged proteins which results in the generation of ten-fold better signals. After proteins are immobilized on the slides, they can be probed for a variety of functions/activities [112, 113]. Finally, the resulting signals are usually measured by detecting fluorescent or radioisotope labels. Protein microarrays are majorly categorized into two classes: analytical and functional [114, 115]. In addition, tissue or cell lysates can also be fractionated and spotted on a slide to form a reverse-phase protein microarray [116].

4.3.1.1 Analytical Microarrays

Analytical microarray is majorly represented by the antibody array which primarily employs the “analyte-labeled” assay format where an array is queried with: (1) a probe (labeled antibody or lig-and) or (2) an unknown biologic sample (e.g., cell lysate or serum sample) containing analytes of interest [117]. By tagging the query, molecules with a signal-generating moiety, a pattern of positive and negative spots is generated. For each spot, the intensity of the signal is proportional to the quantity of applied query molecules bound to the bait molecules. An image of the spot pattern is captured, analyzed, and interpreted [118, 119]. This format, successfully, found alterations in protein expression in cancer cell development, epithelial and stromal cells. However, one of the major limitations of the antibody array approach is the production of specific antibodies in a high-throughput manner. In addition, targeted protein labeling may lead to epitope destruction because of some chemical reactions [120].

This assay can also be explained as the original enzyme-linked immunosorbent assay (ELISA) in a multiplexed format, but can only detect dozens to hundreds of analytes simultaneously because cross-reactivity between antibodies can occur [120]. Recombinant antibodies have become a promising means of overcoming this problem; however, their fabrication issues such as cloning and protein expression add to complexities to their practical use [121]. To improvise on the sensitivity and specificity, analytical microarrays usually employ “sandwich” assay format [106]. This format employs two different antibodies to detect the targeted protein (1) the capture antibody that immobilizes the targeted protein on the solid phase and (2) the reporter or detection antibody that generates a signal for the detection system . This format was applied to successfully detect 75 cytokines with high specificity, femtomolar sensitivity, a 3-log quantitative range, and economy of sample consumption [106, 122, 123].

4.3.1.2 Functional Microarray

Functional protein microarrays are constructed using individually purified proteins that enable the study of various biochemical properties of proteins, such as binding activities, including protein–protein, protein–DNA, protein–lipid, protein–drug, and protein–peptide interactions, and enzyme–substrate relationships via various types of biochemical reactions [106, 120]. Functional protein microarrays are constructed by printing a large number of individually purified proteins, and in principle, it is feasible to print arrays comprised of virtually all annotated proteins of a given organism, effectively comprising a whole proteome microarray [124]. Functional protein microarrays have been successfully applied to identify protein–protein, protein–lipid, protein–antibody, protein–small molecules, protein–DNA, protein–RNA, lectin–glycan, and lectin–cell interactions, and to identify substrates or enzymes in phosphorylation, ubiquitylation, acetylation, and nitrosylation, as well as to profile immune response [114]. The first use of functional protein microarrays was demonstrated by Zhu et al. [125] to determine the substrate specificity of protein kinases in yeast. Since then, reported applications of functional protein microarrays in basic research, as well as in clinical applications, have increased rapidly [126]. Significant achievements in providing the whole proteome of several organisms (i.e., human, yeast, E. coli, virus) on arrays have provided the tools for many important biological discoveries [126].

4.3.1.3 Reverse-Phase Protein Microarrays

This format immobilizes an individual complex test sample in each array spot such that an array is comprised of hundreds of different patient samples or cellular lysates. Each array is incubated with one detection protein (i.e. antibody), and a single analyte end point is measured and directly compared across multiple samples. This method allows for the analysis of many samples obtained at different states by directly spotting tissue, cell lysates, or even fractionated cell lysates on a glass slide. Many different probes can be tested to specifically identify certain proteins in lysate samples [120]. This type of microarray was first established by Paweletz and colleagues to monitor histological changes in prostate cancer patients [127]. Using this method, they successfully detected microscopic transition stages of pro-survival checkpoint protein in three different stages of prostate cancer: normal prostate epithelium, prostate intraepithelial neoplasia, and invasive prostate cancer. The high degree of sensitivity, precision, and linearity achieved by reverse-phase protein microarrays enabled this method to quantify the phosphorylation status of some proteins (such as Akts and ERKs) in these samples; phosphorylation was statistically correlated with prostate cancer progression.

4.3.2 Protein Informatics

With the advent of high-throughput technologies like Next-Generation Sequencing (NGS) , a wealth of information about genomic sequences from a variety of organisms has been amassed. This has led to a rapid buildup of protein sequence data in the form of new protein databases and updation of the existing ones. The exponential increase in the protein related data has prompted computational biologists to develop an advanced infrastructure that facilitates better organization, structural and functional annotation, and evolutionary analyses [128]. Along with a myriad of protein analysis tools, numerous protein related databases have been created that can be categorized as sequence databases, family and domain databases, 3D structure databases, gene expression databases, enzyme and pathway databases, PTM databases, protein-protein interaction databases , etc. [129]. More information can be accessed at http://www.oxfordjournals.org/our_journals/nar/database/cap/, https://proteininformationresource.org/staff/chenc/MiMB/dbSummary2015.html or Expasy, a Swiss Bioinformatics Resource Portal.

4.3.3 Domain Analysis

Protein domains are defined as basic units of structure, function, and evolution. Ranging from 30 to 600 amino acids in length, these units are able to fold independently into stable tertiary structures [128,129,132]. Compact structures with separate hydrophobic cores represent structural domains wherein contacts between residues within the domain are found to be more extensive than between domains [133, 134]. Identification of domains forms the basis of protein classification and annotation. This is exemplified by many protein sequence databases like Pfam, SMART and Interpro, and protein structure databases such as SCOP, CATH, and PALI, which consider domains as the basis for their classification of proteins. The domain-level approach has strongly influenced our understanding in the areas of evolutionary history, homology detection and modeling, protein fold recognition, etc. Interestingly, a relatively random domain shuffling process is thought to have led to domain linkages during the course of evolution, resulting in a few beneficial domain associations being selected and propagated in the interest of cell fitness [133,134,135,138].

Most sequence-based domain recognition methods rely on the conservation of contiguous homologous segments, which is complicated by domain shuffling and recombination in multidomain proteins, accessory domains, and evolutionary divergence of sequences. Therefore, to get an insight into the functional and structural interplay of domains in multidomain proteins, all domains in the full-length amino acid sequence need to be considered simultaneously for protein classification. Keeping this in view, an alignment-free tool, named CLAP (CLAssification of Proteins), was developed for effective classification of multidomain proteins, bypassing the need for identification of domains and their sequential order [135].

Proteins are generally composed of one or more domains arranged in a distinct way that largely dictates the protein function. This is referred to as domain architecture [139, 140].A limited fraction of domain combinations have been found in proteins ruling out the possibility of random combinations. In accordance with power law distribution, the covalent linkage between domains is such that most domains have few partners with a smaller fraction of abundant domains being highly connected [139,140,143]. Based on domain co-occurrence or context, a novel approach, dPUC (Domain Prediction Using Context) was developed for domain prediction and identification. The scores are assigned by analyzing whether two domain families frequently co-occur (positive context) or have never been found as a pair (negative context) [144].

As we move from prokaryotes to eukaryotes or from unicellular eukaryotes to animals, the number of unique domains and the fraction of multidomain proteins increase. This organismal complexity-associated trend, called domain accretion , is thought to play a significant role in evolution. While researchers have likened the genomes to natural language texts, protein domains are considered analogous to words, with domain architectures and amino acids representing sentences and letters, respectively [145]. So n-gram analysis, a well-known probabilistic language-modeling (linguistic) technique helpful in the identification of meaningful word combinations by treating consecutive words in sentences as a unit, is utilized to probe the rules of domain association leading to distinct domain architectures. The set of rules, termed “proteome grammar” is employed to study genome complexity and domain evolution [146]. Some domains tend to be involved in many different domain architectures, a phenomenon called protein domain promiscuity . A bigram analysis has been employed to study the evolution of this promiscuity. Bigram refers to a pair of domains on a protein sequence [139, 147]. Domain rearrangements and domain accretion are two important aspects of evolution. It has been found that there is a nearly universal value of information gain (loss of entropy) associated with a transition to the observed domain architectures from random domain combinations. This highly conserved constant value corresponds to the minimum complexity required to maintain a functioning cell and is governed by “quasi-universal grammar.” However, two major groups viz. a subset (extremely simplified cells) of Archaea and animals (extreme complexity) have deviations from this constant value [146].

4.3.3.1 Domain Parsing

The accurate prediction of domain boundaries (domain parsing) is crucial for the design of chimeric proteins with multi-functional domains and the experimental structure determination of proteins where crystallization is adversely affected by flexible regions. It also makes the multiple sequence alignments more reliable [148, 149]. The defining feature of structural domains underlies some effective algorithms that assign domain boundaries using 3D structures. Some protein features such as signal peptide, trans-membrane helices, low-complexity, and disordered regions and coiled coils, which are not found in globular domains, can be easily predicted using relevant tools. These analyses performed subsequent to a sequence search with BLAST constitute initial steps in domain prediction. The domain boundary prediction employing templates with known structure involves three steps [150]. Sequence search against protein structure databases like PDB comprises the first step and helps in retrieving alignments between target sequence and template structure. The more accurate alignments with a percentage identity of at least 30% and enough coverage (at least 100 residues) with few gaps can be used for 3D model generation. Many methods including a highly sensitive HHPred server have been developed for detecting remote structural homologs by performing a search against a wide range of databases such as PDB, SCOP, Pfam, and COGs [151]. The final step involves 3D model generation using a modeling program like MODELLER (incorporated in HHSearch server) [150,151,154]. Phyre2 [105, 155], I-TASSER [156, 157], and ROBETTA [158] are some of the methods that detect templates and develop the model automatically.

4.3.4 Structure Prediction

A few advanced techniques, viz. X-ray crystallography, nuclear magnetic resonance (NMR), and recently developed cryo-electron microscopy (cryo-EM), have been instrumental in solving the 3D structures of proteins. In spite of this, a rapid non-proportional progress in the field of genomics has widened the already existing gap between the number of protein sequences identified and the number of available protein 3D structures [159]. To address this issue, a variety of computational methods that are faster, easier, and economical have been developed. One of the methods involves sampling the conformational space (c-space) of a protein through deterministic or heuristic approaches. With deterministic methods like homology modeling, entire or part of the c-space is scanned and sub-spaces excluded based on a priori knowledge. In heuristic algorithms (ab initio modeling, Monte Carlo, and molecular dynamics simulations), only a fraction of the c-space is sampled without a priori knowledge generating a representative set of Boltzmann-weighted conformations [160, 161].

Homology modeling (also called comparative modeling) consists of predicting 3D structure from the primary sequence of the protein. It is useful in identifying therapeutic targets, studying structure and function of proteins, protein interaction networks and signaling pathways, and mutagenesis associated with certain diseases [159, 162]. It also has applications in molecular modeling of protein complexes and in the refinement of cryo-EM 3D structures [161,162,165]. It involves multiple steps starting with the identification and selection of suitable templates by searching PDB (Protein Data Bank), an online database of known crystal structures. The protein Basic Local Alignment Search Tool (BLASTp) is employed to look for templates with a sequence identity of more than 40%. There are other algorithms available including PSI-BLAST (Position-Specific Iterated BLAST), hidden Markov models (HMMs), and profile–profile alignments, for templates with low homology [166]. Subsequent to the optimization of the selected alignments, 3D model is built using rigid-body assembly method (as in 3D-JIGSAW and SWISS-MODEL programs), segmented matching method (used by SegMod/ENCAD), spatial restraint method (used by MODELLER and DRAGON), or the artificial evolution method (used by NEST). Next, loop modeling is performed either by scanning a structure database like PDB (a knowledge-based approach used by MODELLER, 3D-JIGSAW and SWISS-MODEL) or by optimizing a scoring function through Monte Carlo or molecular dynamics methods for randomly chosen conformations (an energy-based method using an ab initio fold prediction approach). Addition of side chains to the main backbone requires rotamer libraries, which contain statistical distributions of side chain and backbone orientations extracted from known crystal structures. These are tested sequentially and scored using energy functions. Some of the tools used for side chain packing include SCWRL, FASPR, and SCAP [158,159,162]. To improve the quality of the model thus generated, optimization is done by energy minimization through molecular mechanics force fields. Other ways of model refinement employ molecular dynamics and Monte Carlo simulations. The relationship between the energy of a protein and its conformation is described by the potential energy hyper-surface (PEHS) modeled with quantum or molecular mechanical methods. The native conformation of a protein in the funnel shaped PEHS is ideally represented by a global energy minimum conformation (GEMC), although a canonical ensemble of structures is required to describe the system state completely. The top portion of the PEHS funnel contains high energy conformations resulting from steric and hydrophilic/hydrophobic clashes and unoptimized bond lengths and angles, etc. These conformations are eliminated as a protein folds and GEMC is reached with the narrowing of the funnel [160, 167].

Finally, the model is evaluated and validated by considering stereochemistry, physical parameters, statistical mechanics, etc. To perform this task, Distance-matrix ALIgnment (DALI; http://ekhidna2.biocenter.helsinki.fi/dali/) or Verify3D online servers are used. Many homology modeling programs and online servers like SWISS-MODEL [166,167,170] and Phyre2 [105] have been developed that perform most of the aforementioned steps in an automated fashion. Other homology modeling programs include MODELLER [153], I-TASSER [157], Rosetta [171], Raptor X [172, 173], GalaxyTBM [174], AlphaFold [175], etc.

4.4 Expression Sequence Tags (ESTs), Application of Protein Detection Microarray with Examples

4.4.1 Expression Sequence Tags (ESTs)

Expressed sequence tags (ESTs) are partial cDNA sequences, resulting from single pass sequencing of clones obtained from cDNA libraries. They are used for decoding genome organization and determining gene expression profiles in specific tissues under different conditions. The foremost utilization of ESTs in genome organization studies is to regulate the chromosomal localization of analogous genes employing somatic hybrid cell panel. Furthermore, ESTs contribute in comparative genetics of different species to decipher their gene function. Overall the ESTs lead to integrated genomic approach by the combination of sequence, functional, and localization data. To date, over 45 million ESTs have been generated from over 1400 different eukaryotic species. They have been proven very useful in gene identification and predictions because they are low-cost alternative to whole genome sequencing [176]. This is particularly important for eukaryotes which tend to have “less gene-dense genomes” [177].

For the generation and processing of ESTs mRNA is collected either from whole organisms or specific tissues depending on the size of the organism. This is followed by the extraction of pooled mRNAs and purified typically on the basis of their poly-adenylation. Subsequently, a cDNA library is constructed from this pool and clones are randomly picked for a single pass sequencing read. The raw data are then processed to derive the underlying sequence which is followed by further processing that removes low-quality and contaminating sequences associated with the vectors. The purified sequence data is the submitted to EST database such as dbEST [176,177,180]. The continued generation of ESTs for different species along with the technological advancement has led to its exploitation in range of applications. For example, tandem mass spectrometry matches peptide fragments to known protein sequences. However, limited number of sequences in protein databases leads to computational bias against poorly characterized proteins. ESTs are beginning to have a widespread appeal in identifying and characterizing alternative spliced isoforms [181].

4.4.2 Application of Protein Detection Microarray

Microarray technology was developed in 1989 by Roger Ekins, based on ambient analyte immunoassay [182, 183]. Later, it was transformed into DNA microarray for simultaneous detection of mRNA expression levels in multiple genes. However, the mRNA expression in a cell does not always correspond to exact protein levels [184]. Therefore, to overcome these limitations of DNA microarrays, protein microarray was developed for functional analysis of proteins as they are the major driving forces behind all cellular processes. Immunoassays are first protein microarray, based on specific antigen-antibody interactions, later expanded to antibody microarray which enabled parallel detection of multiple proteins in minute sample quantity with high sensitivity and reproducibility [184]. High-throughput protein array was developed by immobilization of purified proteins on chip glass slide/bead/nitrocellulose membrane or microtiter plate chip [106].

As discussed earlier, the arrays are divided into three main categories; (1) analytical protein microarrays (2) reverse phase protein microarrays, and (3) functional protein microarrays (Fig. 4.1). Analytical protein microarrays (APMs) or capture microarrays are composed of antibodies, aptamers, or affibody libraries attached to a solid support that binds to a specific protein in cell lysate. The APMs provide information regarding protein expression, their binding affinities and specificity; however, cross-hybridization of antibodies is still the major challenge associated with these microarrays. Analytical microarrays are generally used for identification and profiling of treated/non-treated cells and diseased/non-diseased tissues. The reverse phase protein microarray (RPPA) separates complex mixture of proteins in tissue lysate; detected by fluorescent or chemiluminescent assays, and are useful to identify altered proteins or post-translational modifications in diseased cell. Unlike APMs and RPPMs, functional protein microarray (FPM) is used to study biochemical activities in entire proteome. These microarrays are composed of full-length purified functional protein or protein domain arrays immobilized on protein chip. FPMs are used to identify protein–protein, protein–DNA, protein–RNA, protein–phospholipids, and protein–small molecule interactions, detect antibodies, and determine enzyme activity and its specificity. Compared to other methods, FPMs are more capable in detecting low level of protein expression and weak interactions. On the basis of recent developments in FPMs, they are divided into four categories, namely, (1) Purified proteome microarray, (2) Purified protein family microarray, (3) Purified protein domain microarray and, (4) Cell-free protein/peptide microarray.

Fig. 4.1
figure 1

Types of protein microarrays and their application in basic and clinical research

Full size image

The purified proteome microarray, consisting of genome-wide expressed proteins immobilized on a microarray, is widely utilized in E. coli, S. cerevisiae, and human system to study their functional and biochemical properties. Two DNA repair proteins, namely YbaZ and YbcN, were identified by Chen et al. using E. coli proteome microarray consisting of 4256 unique proteins [185]. In another study, using similar microarray, Spr phase switch and DNA binding proteins were identified in type 1 fimbria [186]. Unique antimicrobial peptide and glycosaminoglycans protein targets were identified using E. coli proteome microarray [187, 188]. Additionally, CobB deacetylation enzyme was identified as a strong binder of cyclic di-GMP (bacterial second messenger) while Yojl was found to be involved in bacterial cell invasion by probing human brain microvascular endothelial cells on E. coli proteome microarray [189, 190]. Besides this, E. coli proteome microarray has also been applied for identification of glycoproteins [191], tyrosine sulfation [192], and ClpYQ protease [193] to study bacterial physiology and host–microbial interactions.

A total of 33 novel calmodulin and more than 150 phospholipid binding proteins were identified with biotinylated calmodulin and fluorescently labeled liposomes, respectively, utilizing yeast proteome microarray [125]. The same microarray was used for the identification of SMIR3 and SMIR4 rapamycin inhibitors, Arg5,6 mitochondrial enzyme, and Pus4 and App1 brome mosaic virus antiviral proteins [192,193,196]. Lin et al. further demonstrated two signaling pathways, NuA4 complex-mediated protein acetylation reactions involved in yeast aging and substrates for HECT domain ubiquitin E3 ligase Rsp5 [197]. All these studies demonstrate the usefulness of bacterial and yeast proteome microarray in basic research. However, human proteome microarray is still the most widely used in clinical research, pharmaceutical industry, and translational research. HuProt composed of ~21,000 full-length purified human proteins, ProtoArray ~9000 purified human proteins from insect cells, and NAPPA with 10,000 human proteins are the three popular human proteome microarrays. Human proteome microarray is broadly applied in five major areas: (a) diagnostics, (b) proteomics, (c) protein functional analysis, (d) antibody characterization, and (e) treatment development. Diagnostics includes profiling of sera to discover new disease biomarkers and monitoring of disease states/responses to therapy in personalized medicine. In 2010, Song et al. identified and validated three highly specific anti-hepatitis biomarkers RPS20, Alba-like, and dUTPase with 47.5%, 45.5%, and 22.7% sensitivity, respectively, using human protein microarray consisting of 5011 non-redundant proteins [198]. In another study, a microarray with 1626 purified human recombinant proteins was utilized for validation of six highly specific biomarkers against autoimmune hepatitis with 82% sensitivity and 92% specificity [199]. Also, six highly specific biomarkers, namely PTPRN2, MLH1, MTIF3, PPIL2, NUP50, QRFPR associated with type 1 diabetes were recently validated using Nucleic Acid Programmable Protein Array (NAPPA) [200]. In addition to this, Protoarray was used for validation of transglutaminase 4 (TGM4) biomarker specific infertility causing autoimmune polyendocrine syndrome type 1 (APS1) in males [201]. Furthermore, validation of RNA Polymerase II subunit A C-terminal domain phosphatase (CTDP1) biomarker specific to Behcet disease was performed using HuProt array [202]. Three highly specific p53, PTPRA, and PTGFR biomarkers were validated against ovarian cancer using NAPPA 5177 tumor antigens microarray with 98.3% specificity [203]. In other study, four SNX1, PQBP1, IGHG1, and EYA1 biomarkers specific to glioma were identified by probing ~17,000 human protein microarray [204]. HuProt array was used for identification of COPS2, NT5E, TERF1, and CTSF biomarkers for diagnosis of gastric cancer and validation of p53, HRas, and ETHE1 for early detection of lung cancer [205, 206]. Moreover, three specific FGFR2, CALM1, and COL6A1 prostate cancer biomarkers were identified and validated using 123 purified antigens microarray platform [207]. Human proteome array has been utilized to validate IGHG4, STAT6, CRYM, HDAC7A, EFCAB2, SELENBP1, and CCNB1 biomarkers against Meningiomas [208]. In a more recent study, a highly specific biomarker panel, identified and validated using protein array based approach, was employed to discriminate Zikavirus and Dengue virus infections [209]. High-Density Nucleic Acid Programmable Protein Array (HD-NAPPA) of the pathogen Mycobacterium tuberculosis (Mtb) was used in the identification of eight antibody targets, viz. Rv0054, Rv0831c, Rv2031c, Rv0222, Rv0948c, Rv2853, Rv3405c, Rv3544c, for tuberculosis serology [210].

Currently, there are two methods to detect protein signals (1) labeled and (2) label free. The ideal protein array detection method should produce low background noise and generate high signal frequency. Therefore, the most common and widely used detection method is fluorescence labeling which is highly sensitive, safe, and compatible with readily available microarray laser scanners. Other labels used are affinity, photochemical, or radioisotope tags. As these labels are attached to the probe itself which can interfere in the probe-target protein reaction; thus, a number of label free detection methods are developed such as surface Plasmon resonance (SPR), carbon nanotubes, carbon nanowire sensors, and microelectromechanical system cantilevers. Most of these methods are relatively new and not very suitable for detection of high-throughput protein interactions; however, they do offer much promise for the future.

4.5 Data Analysis and Interpretation of Protein Detection Arrays

Protein microarrays provide wealth of information regarding protein interactions, protein functions, and signaling pathways which could be used for clinical diagnosis. However, data translation requires automated data processing and interpretation for generation of meaningful information. Protein microarray data analysis mainly depends on the design of surfaces, content, detection method, data preprocessing, inference, classification, and validation (Fig. 4.2). The design of array is a crucial step as it significantly affects data analysis and its final interpretation. Inclusion of biological replicates is recommended as they provide higher statistical confidence; however, they also make results more complex to evaluate. Data preprocessing, which includes image analysis, normalization, and data transformation, also greatly affects data analysis and interpretation. Image-processing algorithms distinguish foreground and background intensities and inference based on data analysis variability [211]. Different data analysis strategies for different types of array generate variable results. These arrays provide variety of tools for disease analysis but lack standard analytical and data processing strategy which enhance complexity in data analysis.

Fig. 4.2
figure 2

Protein microarray strategy for data analysis

Full size image

The data analysis strategies like spot-finding on slide images, Z-score calculations, and significance analysis of microarrays (SAM) have their origin in DNA microarray analysis; however, concentration-dependent analysis (CDA) is specific for protein microarrays (Fig. 4.3) [212].

Fig. 4.3
figure 3

Methods for data analysis of protein microarrays

Full size image

In spot intensity determination method, microarray image analysis starts with the fixing of spot intensity. A grid of circles with adjusted position and size are placed over protein spots to get reliable intensity data. The output file is created by GenePix Pro software (Molecular Devices, CA). In Z-score analysis, Z-score equation (Zs = Ss − μ/σ) is analyzed to determine the significantly different from the expected values, where Zs is the Z-score for the sth spot, Ss is the signal for that spot’, μ is the mean signal across all spots, and σ is the standard deviation across all spots. In concentration-dependent analysis (CDA), absolute signals generated depend on protein concentration in the sample. To overcome this issue an iteration process is used to calculate Z-score (Zs = Ss − μw/σw) and remove outliers. In the equation, Zs represents Z-score for the sth spot, Ss spot signal, μw mean signal, and σw standard deviation (Fig. 4.3).

The signals produced are detected by fluorescent dyes or colorimetric assays. There are two types of assays for colorimetric detection of proteins: single color assay and dual color assay. Single color assay is a single antibody based microarray which uses internal control system based on two colors for quantification of antigen and antibody. In the dual color assay, each sample is labeled with different fluorescent dyes and their signal intensity is measured using fluorescence image scanners. Dual color assays have better reproducibility and discriminatory efficiency then single color assays [213]. To prevent undesired technical artifacts caused by electric charges, different protein sizes, hydrophobic protein interaction of proteins and antibody/antigen binding kinetics in dual color assays the data pre-processing protocols like filtering, background correction and data normalization are required [214]. Furthermore, it employs four different microarray designs: (a) Reference in which sample of interest and reference sample is labeled with different fluorescent fluorochromes. This design is generally used for comparative studies; (b) Balanced-block where two samples bearing two different fluorochromes are hybridized to make a single block; (c) Incomplete block, more than two samples bearing only two fluorescent fluorochromes are co-hybridized on microarray; (d) Loop design where samples are hybridized in different arrays using different fluorochromes which leads to duplication of arrays.

For data normalization, different algorithms, rank-invariant selection, modified rank-invariant selection and rank difference weighted global loess are used to define the set of probes. Rank-invariant selection algorithm is used in the absence of house-keeping controls. However, its major limitation is that it does not cover entire intensity range [215]. Modified Rank-Invariant Selection Algorithm corrects intensity values through extrapolation of curve to lower and upper intensity limits. Rank Difference Weighted Global Loess is applied to whole probes on array to get global normalization. Despite the differences among data processing methods in microarray analysis the general recommendations which need to be considered for data processing in microarray analysis are: Bayesian approaches to examine intersections, quality-control, validation methods, and standardized testing platforms.

4.6 Summary

Medical biotechnology has provided several products essential for research, therapeutics, and diagnostic purpose. Recombinant protein technology is the connecting link between medical biotechnology, and mass production of therapeutic and diagnostic products. For a long time, E. coli has served as the cost-effective and low-maintenance expression system for the production of pharmaceutically important recombinant proteins. The limitation of expressing several human proteins with specific post-translational modifications, essential for their biological activity, in E. coli (which lacks post-translational modification machinery), has forced biochemists to look for alternative expression systems for large recombinant protein production. Using new and improved recombinant methods, humanization of E. coli, which has failed so far, could be tried and made successful. Lately, insect cell lines, non-human mammalian cell lines, and human cell lines with engineered genome are extensively used to produce therapeutic proteins such as growth factors, vaccines against infectious diseases, monoclonal antibodies, and IFNs to treat cancers and other diseases. The advantages and disadvantages of the different expression systems are reviewed in this chapter. We have also discussed the significance of protein tags used during protein purification following its expression. Some of these tags are also known to increase the solubility of the proteins, ultimately leading to high protein yields required for commercial purpose.

Ever since informative machineries started to evolve, proteomics technologies have been aimed at the comprehensive detection of the downstream proteins to evaluate complex disease diagnosis, allied mechanism and concerned therapy for effective management of the diseases. Moreover, to understand the complex biological organization, it is imperative to understand regulatory interconnections between DNA, RNA, and protein. For instance, microarrays, automated sequencing, and mass spectrometry have significantly contributed to systems biology approach by investigating protein–protein interactions, signal pathway analysis, studies of PTMs, or/and detection of toxins. It also has a wide array of opportunities in disease biomarker discovery to enable better disease management through improved diagnostics. More importantly, various forms of protein microarray have gradually evolved for proteomics research. With the progressive development, standardization of the experimental workflow and data interpretation, protein microarray holds promises in diagnostic applications. For protein microarray data analysis, various techniques including spot intensity determination method, z-score calculation, and concentration-dependent analysis have been used. Moreover, colorimetric assays involving fluorescent dyes are used for detecting signals.

Expressed sequence tags and cDNA provide direct evidence for all sample transcripts and are the most important resources for transcriptome exploration. ESTs have proven useful in different applications along with individual tools and pipelines for EST analysis.

Finally, an increasing number of bioinformatics methods have been developed to meet the needs of researchers for rigorous analysis of a vast amount of data generated through high-throughput genomic and proteomic techniques. From sequence-based analysis to protein structure prediction, analysis tools have been developed that focus on individual steps of the process or perform the whole process in an automated way. For instance, different protein databases along with certain analysis tools have been used for protein data analysis. Correct identification of domain boundaries for elucidating domain architecture and predicting protein structure from primary sequences using homology modeling have become possible with some of the finest tools developed recently. All these methods have facilitated the research on therapeutic agents that could be used in drug designing and other areas.