identification of potential drug targets by subtractive genome analysis of bacillus anthracis a0248:...

7
Research article Identication of potential drug targets by subtractive genome analysis of Bacillus anthracis A0248: An in silico approach Md. Anisur Rahman, Md. Sanaullah Noore, Md. Anayet Hasan *, Md. Rakib Ullah, Md. Hajur Rahman, Md. Amzad Hossain, Yeasmeen Ali, Md. Saiful Islam Department of Genetic Engineering and Biotechnology, Faculty of Biological Sciences, University Of Chittagong, Chittagong-4331, Bangladesh A R T I C L E I N F O Article history: Received 28 April 2014 Received in revised form 8 September 2014 Accepted 13 September 2014 Available online xxx Keywords: Bacillus anthracis Anthrax Biological weapon Computational approach Therapeutic drug A B S T R A C T Background: Bacillus anthracis is a gram positive, spore forming, rod shaped bacteria which is the etiologic agent of anthrax cutaneous, pulmonary and gastrointestinal. A recent outbreak of anthrax in a tropical region uncovered natural and in vitro resistance against penicillin, ciprooxacin, quinolone due to over exposure of the pathogen to these antibiotics. This fact combined with the ongoing threat of using B. anthracis as a biological weapon proves that the identication of new therapeutic targets is urgently needed. Methods: In this computational approach various databases and online based servers were used to detect essential proteins of B. anthracis A0248. Protein sequences of B. anthracis A0248 strain were retrieved from the NCBI database which was then run in CD-hit suite for clustering. NCBI BlastP against the human proteome and similarity search against DEG were done to nd out essential human non-homologous proteins. Proteins involved in unique pathways were analyzed using KEGG genome database and PSORTb, CELLO v.2.5, ngLOC these three tools were used to deduce putative cell surface proteins. Results: Successive analysis revealed 116 proteins to be essential human non-homologs among which 17 were involved in unique metabolic pathways and 28 were predicted as membrane associated proteins. Both types of proteins can be exploited as they are unlikely to have homologous counterparts in the human host. Conclusion: Being human non-homologous, these proteins can be targeted for potential therapeutic drug development in future. Targets on unique metabolic and membrane-bound proteins can block cell wall synthesis, bacterial replication and signal transduction respectively. ã 2014 Elsevier Ltd. All rights reserved. 1. Introduction Bacillus anthracis, the causative agent of anthrax in both human and animal is a gram positive, spore forming, aerobic, non-motile bacterium. B. anthracis is the most renowned and obligate pathogen of the bacillus genus (Spencer, 2003). Unlike most other bacillus species, entry of B. anthracis spores into mammals through respiratory, gastrointestinal or cutaneous routes can result in systemic infection and lethal disease (Koehler, 2009). Systemic infection resulting from the inhalation of the organism has a mortality rate approaching 100 percent and death usually occurs within a few days after the onset of symptoms (Terry et al., 1999). Consumption of contaminated meat which is usually cooked insufciently causes gastrointestinal anthrax. More than 95% of incidents are cutaneous anthrax that affects 20,000100,000 human annually. Regular infection outbreaks on the wild animal caused by anthrax spores have been reported after decades of animal interment (Siddiqui et al., 2012; Chakrabortyet al., 2012). Recently an outbreak episode of anthrax created much havoc and dreadful clinical outcome from April to August, 2011 in Bangladesh. But this incident has its history back in 19801984 with a broader manifestation on human and cattle in 20092010. This outbreak hit 12 districts and affected 607 people which led Bangladesh government to declare a red alert (Siddiqui et al., 2012). Far from this tropical region, anthrax outbreak also occurred in the western territory. In Canada, anthrax outbreak had been occurred since 1962 in domestic cattle (Dragon et al., 1999). B. anthracis had also been weaponized in World War I and in World War II and threats are marked presently for possible use of B. anthracis in Bioterrorism. Anthrax mail attack in USA in 2001 is an alarming example of bioterrorism that disclosed the drastic utilization of the pathogen (Spencer, 2003; Ahn et al., 2014). * Corresponding author. Tel.: +8801717344389. E-mail address: [email protected] (M. A. Hasan). http://dx.doi.org/10.1016/j.compbiolchem.2014.09.005 1476-9271/ ã 2014 Elsevier Ltd. All rights reserved. Computational Biology and Chemistry 52 (2014) 6672 Contents lists available at ScienceDirect Computational Biology and Chemistry journal homepa ge: www.elsev ier.com/locate/compbiolchem

Upload: independent

Post on 21-Apr-2023

1 views

Category:

Documents


0 download

TRANSCRIPT

Computational Biology and Chemistry 52 (2014) 66–72

Research article

Identification of potential drug targets by subtractive genome analysisof Bacillus anthracis A0248: An in silico approach

Md. Anisur Rahman, Md. Sanaullah Noore, Md. Anayet Hasan *, Md. Rakib Ullah,Md. Hafijur Rahman, Md. Amzad Hossain, Yeasmeen Ali, Md. Saiful IslamDepartment of Genetic Engineering and Biotechnology, Faculty of Biological Sciences, University Of Chittagong, Chittagong-4331, Bangladesh

A R T I C L E I N F O

Article history:Received 28 April 2014Received in revised form 8 September 2014Accepted 13 September 2014Available online xxx

Keywords:Bacillus anthracisAnthraxBiological weaponComputational approachTherapeutic drug

A B S T R A C T

Background: Bacillus anthracis is a gram positive, spore forming, rod shaped bacteria which is the etiologicagent of anthrax – cutaneous, pulmonary and gastrointestinal. A recent outbreak of anthrax in a tropicalregion uncovered natural and in vitro resistance against penicillin, ciprofloxacin, quinolone due to overexposure of the pathogen to these antibiotics. This fact combined with the ongoing threat of using B.anthracis as a biological weapon proves that the identification of new therapeutic targets is urgentlyneeded.Methods: In this computational approach various databases and online based servers were used to detectessential proteins of B. anthracis A0248. Protein sequences of B. anthracis A0248 strain were retrievedfrom the NCBI database which was then run in CD-hit suite for clustering. NCBI BlastP against the humanproteome and similarity search against DEG were done to find out essential human non-homologousproteins. Proteins involved in unique pathways were analyzed using KEGG genome database and PSORTb,CELLO v.2.5, ngLOC – these three tools were used to deduce putative cell surface proteins.Results: Successive analysis revealed 116 proteins to be essential human non-homologs among which17 were involved in unique metabolic pathways and 28 were predicted as membrane associated proteins.Both types of proteins can be exploited as they are unlikely to have homologous counterparts in thehuman host.Conclusion: Being human non-homologous, these proteins can be targeted for potential therapeutic drugdevelopment in future. Targets on unique metabolic and membrane-bound proteins can block cell wallsynthesis, bacterial replication and signal transduction respectively.

ã 2014 Elsevier Ltd. All rights reserved.

Contents lists available at ScienceDirect

Computational Biology and Chemistry

journal homepa ge: www.elsev ier .com/locate /compbiolchem

1. Introduction

Bacillus anthracis, the causative agent of anthrax in both humanand animal is a gram positive, spore forming, aerobic, non-motilebacterium. B. anthracis is the most renowned and obligatepathogen of the bacillus genus (Spencer, 2003). Unlike most otherbacillus species, entry of B. anthracis spores into mammals throughrespiratory, gastrointestinal or cutaneous routes can result insystemic infection and lethal disease (Koehler, 2009). Systemicinfection resulting from the inhalation of the organism has amortality rate approaching 100 percent and death usually occurswithin a few days after the onset of symptoms (Terry et al., 1999).Consumption of contaminated meat which is usually cookedinsufficiently causes gastrointestinal anthrax. More than 95% ofincidents are cutaneous anthrax that affects 20,000–100,000

* Corresponding author. Tel.: +8801717344389.E-mail address: [email protected] (M. A. Hasan).

http://dx.doi.org/10.1016/j.compbiolchem.2014.09.0051476-9271/ã 2014 Elsevier Ltd. All rights reserved.

human annually. Regular infection outbreaks on the wild animalcaused by anthrax spores have been reported after decadesof animal interment (Siddiqui et al., 2012; Chakrabortyet al.,2012).

Recently an outbreak episode of anthrax created much havocand dreadful clinical outcome from April to August, 2011in Bangladesh. But this incident has its history back in1980–1984 with a broader manifestation on human and cattlein 2009–2010. This outbreak hit 12 districts and affected607 people which led Bangladesh government to declare a redalert (Siddiqui et al., 2012). Far from this tropical region, anthraxoutbreak also occurred in the western territory. In Canada,anthrax outbreak had been occurred since 1962 in domestic cattle(Dragon et al., 1999). B. anthracis had also been weaponized inWorld War I and in World War II and threats are marked presentlyfor possible use of B. anthracis in Bioterrorism. Anthrax mail attackin USA in 2001 is an alarming example of bioterrorism thatdisclosed the drastic utilization of the pathogen (Spencer, 2003;Ahn et al., 2014).

M.A. Rahman et al. / Computational Biology and Chemistry 52 (2014) 66–72 67

Presently advocated drugs for post-exposure of cutaneous andinhalation anthrax are ciprofloxacin, doxycycline, penicillin G andamoxicillin (Chen et al., 2004; Athamna et al., 2004). Ciprofloxacincarries out its function by inhibiting DNA replication enzymes,DNA gyrase and topoisomerase IV which induce negative super-coils to relieve the strain of positive supercoils and decatenation ofDNA (Choe et al., 2000; Drlica et al., 1997). Doxycycline being intetracycline antibiotic class is a broad-spectrum antibiotic thatbinds with 30s ribosomal subunit at A site in the translationalcomplex and thus inhibits the binding of aminoacyl-tRNA at the Asite. So protein synthesis is stopped due to the failure ofincorporation of amino acids (Chopra et al., 2001). On the otherhand penicillin binds to DD-transpeptidase enzyme (also calledpenicillin binding protein-PBP) through b-lactam ring that inhibitsthe cross-linking between bacterial peptidoglycan chain causingdisruption of the cell wall and cell death (Georgopapadakou andLiu, 1980; Waxman, 1980; Tipper and Strominger, 1965). Butextensive repeated exposure of bacteria to these drugs caused theevolution of resistant strains. Naturally B. anthracis exhibitsresistance to penicillin. Moreover in vitro resistance to ciprofloxa-cin, doxycycline, quinolone, and fluoroquinolone and otherb-lactam antibiotics has been reported (Chen et al., 2004;Athamna et al., 2004; Choe et al., 2000).

All these major incidences have provoked a necessity fordeveloping potential drugs against this dreadful pathogen.Identification of novel therapeutic targets ultimately leads tonovel drug designing. There are various strategies for potentialdrug designing, among them comparative genomic can be

Subtracted to

UsingPSORTb, C ELLO, n gLOC

Subtracted to

Bacillus anthraci

Run i n CD -

Selec� on of non

Proteins > 100 am

Human n on-ho

116 Essen�al Human non

28 Human non-homolog ousmembrane b ound essential proteins

Fig. 1. Schematic diagram

considered worthy as newer genes, genomes and protein databasesare available online.

In this proposed work, subtractive genome analysis has beenused where subtraction of similar gene set is done after comparinghuman and pathogen genome. With the help of entire genome of B.anthracis, a minimum set of genes required for the survival of theorganism is identified as essential genes (Zhang and Lin, 2009). Inthe consecutive work computational approach through subtractivegenomics using necessary databases has been adopted to find outunique proteins in the metabolic pathways and proteins associatedwith the membrane. Membrane bound proteins are selectedbecause 60% of the therapeutic drugs are targeted to membraneproteins (Hossain et al., 2013; Arinaminpathy et al., 2009). Thisapproach is successfully used in many other bacteria such asPseudomonas aeruginosa (Sakharkar et al., 2004), Helicobacterpylori (Dutta et al., 2006), Burkholderia pseudomalleii (Chong et al.,2006) etc. Therefore in the current study, we have simultaneouslyapproached a goal to identify the potential drug targets in twoways – screening through the essential non-human homologproteins in unique pathways and proteins characterized asmembrane-bound, which can be valid target for drug designingby further research.

2. Methodology

Selection and characterization of potential drug targets for B.anthracis A0248 were done by a certain stepwise manner which issummarized in Fig. 1 and sequentially mentioned below.

Proteins retri eved f rom N CBI

Using 60% as thres hold val ue

Using Blas tP, E-val ue >10-4

DEG, cut off E -value, <10-10 0

KAAS server at KEGG

subtrac ted to

KEGG Gen ome

KEGG Gen ome

s strai n A0248

hit suite

-dupli cate proteins

ino aci ds

mologous proteins

-homolog ous proteins

108 Ess ential protei ns inv olved in meta bolic pat hways

17 unique metabolic pathways

Human non-homolog ous 17 essential protei ns involved in unique pat hways

of brief methodology.

68 M.A. Rahman et al. / Computational Biology and Chemistry 52 (2014) 66–72

2.1. Sequence retrieval of pathogen proteins

The complete protein sequences or proteome of B. anthracisstrain A0248 were retrieved from the NCBI (http://www.ncbi.nlm.nih.gov/) protein database. NCBI provides databases of variousformats with different perspectives. Among them FASTA formatof the protein was chosen for subsequent analysis.

2.2. Identification of paralog proteins

Proteins of B. anthracis retrieved from NCBI database weresubjected to CD-hit suite using sequence identity cut off value at60% (Huang et al., 2010). Proteins being more than 60% identicalwere considered to be duplicates or paralogs. The duplicates wereexcluded and from the remaining set of non-paralog proteins,those were selected for further screening having more than100 amino acids. This was based on the assumption that proteinsless than 100 amino acids in length were most unlikelyto represent essential proteins (Kumar et al., 2010; Haaget al., 2012).

2.3. Identification of human non-homologs

The non paralog proteins were then subjected to NCBI BlastPagainst Homo sapiens using threshold expectation value 10�4 toidentify the human non-homologous proteins of B. anthracis(Hossain et al., 2013; Haag et al., 2012; Kerfeld and Scott, 2011). Atfirst, the human homologs were excluded and then the rest of thenon-homologs were unified.

2.4. Similarity search for essential protein

The selected human non-homologous proteins were thensubjected to similarity search against the Database of EssentialGenes. A maximum threshold expectation value 10�100 and aminimum bit-score cut-off of 100 were used as parameters toanalyze the proteins for identifying essential human non-homologs (Kumar et al., 2010; Luo et al., 2013; Punta et al., 2007).

2.5. Metabolic pathway analysis

The human non-homologous essential proteins of B. anthracisobtained through DEG were then subjected to metabolic pathwayanalysis that was done by KAAS server at KEGG for theidentification of potential targets (Moriya et al., 2007). KAAS(KEGG Automatic Annotation Server) provides functional annota-tion of genes through BLAST comparisons against the manuallycurated KEGG GENES database. The result contains KO (KEGGOrthology) assignments that specify the metabolic proteins. It alsoautomatically generates KEGG pathways which indicate thosemetabolic proteins.

Table 1Subtractive analytical result for Bacillus anthracis A0248.

Successive analytical steps

Proteins from NCBI database

Non-paralogs/non-duplicates found in CD-HIT suite

Duplicates (>60% identical) found in CD-HIT suite

Proteins >100 amino acids

Human non-homologous proteins (E-value 10�4)

Essential proteins found in DEG (E-value 10�10)

Essential proteins in metabolic pathways by KAAS server

Proteins involved in unique pathways

Human non-homologous membrane bound essential prot

2.6. Detection of proteins involved in unique pathways

The unique metabolic pathways of B. anthracis against Homosapiens were identified using KEGG (Kyoto Encyclopedia of Genesand Genomes) Genome Database (Kanehisa et al., 2014; Kanehisaand Got, 2000). The three letter codes specific for each of the twoorganisms (host and pathogen) were placed in the GenomeComparison and Combination box for comparing the metabolicpathways. Ultimately the KEGG Genome generated pathway maps,among which the pathways involved only in B. anthracis, wereidentified as unique pathways. Previously selected human non-homologous metabolic proteins were screened to reveal theirinvolvement in those unique pathways.

2.7. Surface protein identification

In addition to identification of proteins involved in uniquepathways, we also detected putative membrane associatedproteins. Proteins associated with the cytoplasmic membrane ofthe pathogen can be potential target for drugs. Sub-cellularlocalization of essential human non-homologous proteins wasdone by using PSORTb (Yu et al., 2010), CELLO v.2.5 (Yu et al., 2006)and ngLOC (King et al., 2012) simultaneously. These three toolswere used with the perspective to maintain the better accuracy inthe result. If any two of the three tools were unanimous to predict aspecific protein as membrane-associated, then the prediction wasadopted. But if the three tools disagreed with each other on aspecific protein, then the protein was not considered to bemembrane-associated.

2.8. Characterization of hypothetical protein

Functional family prediction of one hypothetical proteinpreviously identified as surface protein was done by Pfam server(Punta et al., 2012).

3. Result and discussion

In this study we investigated the potential target for therapeuticdrug through subtractive genome analysis. The basic of in-silicosubtractive genome analysis is that the essential proteins of apathogen should be sorted as unique which cannot be found in thehost for precise drug usage. The essential genes are the minimumset of genes which are required for the survival of an organismunder the reasonable condition (Zhang and Lin, 2009). The resultsof the successive analytical steps are presented in Table 1 and theprocedure is summarized in Fig. 1.

In the current investigation a set of total 11,000 proteins wasretrieved which were further analyzed through subtractivegenome approach. Firstly, these proteins were run in CD-hit toolusing 60% identity as threshold value which resulted in 5112

Total number of proteins

11000511258884225336811610817

eins 28

M.A. Rahman et al. / Computational Biology and Chemistry 52 (2014) 66–72 69

non-paralog proteins. Among these, 4225 proteins were sorted outhaving more than 100 amino acids in length. Because theprobability of involvement of proteins having more than 100 aminoacids in essential metabolic pathways is greater than the proteinscontaining less than 100 amino acids (Kumar et al., 2010).

Now human non-homologous proteins were identified byperforming NCBI BlastP using 10�4 as threshold value. Screeningthrough BlastP, we found considerable similarities (E-value lessthan 10�4) between pathogen (B. anthracis) and host (Human)proteins which were identified as homologous and then excluded.Hits having expectation value greater than 10�4 were selected toensure the human non-homologous proteins. We found 3368human non-homologous proteins (Hossain et al., 2013; Haag et al.,2012; Fariselli et al., 2007; Anishetty et al., 2005). Homologousproteins between bacteria and human emerged in course ofevolution and they are involved in various common cellularsystems between host and pathogen (Hediger et al., 1989; Swangoet al., 2000; Chen et al.,1986; Johnson et al.,1989). Proteins withoutsignificant similarity with human were considered as humannon-homologous (Hossain et al., 2013; Yadav et al., 2012).

Screening through DEG revealed 116 human non-homologousproteins as essential proteins. This strategy further utilized blastanalysis against bacterial genome and excluded distant humanhomologs if any were present by chance (Fariselli et al., 2007;Anishetty et al., 2005). However, out of these 116 humannon-homologous essential proteins, 108 were finalized throughKAAS server at KEGG to be involved in metabolic pathways. KAASuses Blast analysis on the prokaryotic proteome for the selectionof metabolic proteins out of 116 human non-homologous essentialquery proteins (Hossain et al., 2013; Kumar et al., 2010; Rathi et al.,2009). Drug targeting on these essential proteins can block one ormore of these metabolic pathways which can exert a vital effect onbacterial pathogenesis and its vulnerability.

17 unique metabolic pathways of B. anthracis were identified bycomparing with H. sapiens through KEGG Genome (Table 2) andamong those pathways 17 human non-homologous metabolicproteins were involved (Table 3).

The sub-cellular localization of human non-homologousessential proteins predicted 28 proteins as membrane associated.This prediction was done by PSORTb, CELLO v.2.5 and ngLOCconcurrently with a view to comparing the results with betteraccuracy. The results differed from each other in a few cases. Forexample, four proteins were not predicted as membrane associatedby CELLO but other two tools predicted them as membraneassociated and another four proteins were not predicted as

Table 2The unique metabolic pathways.

Principal pathway Corresponding pathways

Metabolism Energy metabolism

Lipid metabolism

Metabolism of other amino acids

Glycan biosynthesis and metabolismMetabolism of terpenoids and poly

Biosynthesis of other secondary me

Xenobiotics biodegradation and me

Environmental information processing Membrane transport

Cellular processes Cell motility

Human diseases Infectious diseases: bacterial

membrane associated by ngLOC but other two tools predictedthem as membrane associated. These eight proteins wereconsidered as membrane-associated as they were predicted bymaximum two tools. This strategy remarkably increased theprobability of a protein being membrane associated than usingonly one tool. The remaining 20 proteins were unanimouslypredicted as membrane-associated by the three tools. If all threepredictors disagreed with each other on a specific query proteinthen this was not considered to be membrane-associated.

Out of these 28 membrane-bound proteins, only 1 protein washypothetical which was characterized as short chain fatty acidtransporter by using Pfam server (Table 4).

3.1. Prospective drug targets from unique pathways in B. anthracisA0248

This appendage covers the interpretation of the proteins fromthe unique pathways for auspicious drug targets. We found elevenessential human non-homologous proteins involved in peptido-glycan biosynthesis, three in phosphotransferase system, two inoxidative phosphorylation and one in D-alanine metabolism.

Naturally occurring penicillin resistant B. anthracis strains havebeen reported having upto 16% resistance. In vitro resistance wasalso found to other beta-lactum antibiotics like cephalosporin andcefuroxime (Doganay and Aydin, 1991; Frean et al., 2003). This isdue to the functional b-lactamases encoded by bla genes thathydrolyze carbon-nitrogen bonds in cyclic amides. Conventionaltargets on penicillin binding proteins (PBPs) to disrupt thepeptidoglycan layer in the last step of peptidoglycan biosynthesisshould be altered to alleviate the penicillin-resistance in B.anthracis (Hossain et al., 2013).

MurA catalyzing the formation of UDP-N-acetylglucosaminee-nolpyruvate at the first step of peptidoglycan biosynthesis is aneffective target for fosfomycin. But one disadvantage of targetingmurA is the presence of two different murA1 and murA2 geneshaving less than 60% similarity and encoding the enzymes withdifferent active sites that can again substitute each other. Sodevelopment of a potential drug against murA is not so convenient(Hossain et al., 2013; Kedar et al., 2008; Du et al., 2000). Similar isthe scenario in case of murB catalyzing the synthesis of UDP-N-acetylmuramic acid. MurB has an alternate form murB2 that canfoil drug action.

But murC and murD of mur ligase family catalyzing theformation of UDP-N-acetylmuramoyl-L-alanine (3rd step) andUDP-N-acetylmuramoyl-L-alanyl-D-glutamate (4th step) have

Unique metabolic pathways

1. Photosynthesis2. Secondary bile acid biosynthesis3. D-Alanine metabolism

4. Peptidoglycan biosynthesisketides 5. Tetracycline biosynthesis

6. Nonribosomal peptide structures7. Biosynthesis of siderophore group nonribosomalpeptides

tabolites 8. Stilbenoid, diarylheptanoid and gingerol biosynthesis9. Isoflavonoid biosynthesis10. beta-Lactam resistance

tabolism 11. Xylene degradation12. Ethylbenzene degradation13. Dioxin degradation14. Phosphotransferase system (PTS)15. Bacterial chemotaxis16. Flagellar assembly17. Vibrio cholerae pathogenic cycle

Table 3Proteins involved in unique pathways.

SN Accession no. Name of protein KO number (gene name) Pathways

1 gi|229603396 F0F1 ATP synthase subunit alpha K02111(atpA) Photosynthesis2 gi|229603044 F0F1 ATP synthase subunit beta K02112(atpD) Photosynthesis3 gi|229599992 alanine racemase K01775(dal1) D-Alanine metabolism4 gi|229601214 UDP-N-acetylglucosamine 1-carboxyvinyltransferase K00790(murA1) Peptidoglycan biosynthesis5 gi|229601457 UDP-N-acetylglucosamine 1-carboxyvinyltransferase K00790(murA2) Peptidoglycan biosynthesis6 gi|229601217 UDP-N-acetylenolpyruvoylglucosaminereductase K00075 (murB) Peptidoglycan biosynthesis7 gi|229604037 UDP-N-acetylenolpyruvoylglucosaminereductase K00075 (murB2) Peptidoglycan biosynthesis8 gi|229602579 UDP-N-acetylmuramate–L-alanine ligase K01924(murC) Peptidoglycan biosynthesis9 gi|229604216 UDP-N-acetylmuramoyl-L-alanyl-D-glutamate synthetase K01925(murD) Peptidoglycan biosynthesis

10 gi|229602560 UDP-N-acetylmuramoylalanyl-D-glutamate–2,6-diaminopimelate ligase K01928(murE) Peptidoglycan biosynthesis11 gi|229604728 UDP-N-acetylmuramoyl-tripeptide–D-alanyl-D-alanine ligase K01929(murF) Peptidoglycan biosynthesis12 gi|229600324 UDP diphospho-muramoylpentapeptide beta-N acetylglucosaminyltransferase K02563 (murG) Peptidoglycan biosynthesis13 gi|229601892 Penicillin-binding protein 1A K05366 Peptidoglycan biosynthesis14 gi|229604649 Penicillin-binding protein K08724 Peptidoglycan biosynthesis15 gi|229602264 Phosphoenolpyruvate-protein phosphotransferase K08483(pstI) Phosphotransferase system (PTS)16 gi|229603629 PTS system, glucose-specific IIABC component K02764(ptsG) Phosphotransferase system (PTS)17 gi|229604345 PTS system, fructose-specific IIABC component K02770 Phosphotransferase system (PTS)

70 M.A. Rahman et al. / Computational Biology and Chemistry 52 (2014) 66–72

single copy of each gene. UDP-MuNAc-L-Ala-D-Glu is an importantprecursor for peptidoglycan biosynthesis that can serve as apropitious target point (Kanehisa et al., 2014; Nathenson et al.,1964; Gordon et al., 2001).

Another suitable pathway is PEP: carbohydrate phosphotrans-ferase system uniquely found in bacteria. Simultaneous transpor-tation and phosphorylation of carbohydrates by phosphoryl groupfrom PEP are catalyzed by two enzymes – enzyme I (phosphoenolpyruvate-protein phosphotransferase) and enzyme II (specific fordifferent carbohydrates like fructose specific IIABC component andglucose specific IIABC component) (Deutscher et al., 2006). PEPphosphotransferase, glucose and fructose specific IIABC compo-nents can be blocked to inhibit phosphorylation cascade and AtxAtranscription factor activation which is responsible for virulence inB. anthracis (Stülke, 2007; Deckers-Hebestreit and Altendorf,1996).

Table 4Membrane associated proteins.

SN. Accession No. Name of proteins

1 gi|229604843 L-Lactate permease2 gi|229604726 Penicillin-binding protein3 gi|229604649 Penicillin-binding protein4 gi|229604345 PTS system, fructose-specific IIABC component5 gi|229604029 D-alanyl-lipoteichoic acid biosynthesis protein DltB6 gi|229603992 Carbon starvation protein A7 gi|229603880 Ferredoxin, 4Fe-4S8 gi|229603726 Respiratory nitrate reductase, beta subunit9 gi|229603629 PTS system, glucose-specific IIABC component

10 gi|229603432 ResB protein11 gi|229603339 Respiratory nitrate reductase, alpha subunit12 gi|229603044 F0F1 ATP synthase subunit beta13 gi|229602974 Cation-transporting ATPase, E1-E2 family14 gi|229602903 Sensory box histidine kinase YycG15 gi|229602448 resC protein16 gi|229602369 Preproteintranslocase subunit SecY17 gi|229601997 Cation-transporting ATPase, E1-E2 family18 gi|229601955 Arylsulfatase19 gi|229601892 Penicillin-binding protein 1A20 gi|229601600 Bifunctionalpreproteintranslocase subunit SecD/SecF21 gi|229600757 Stage III sporulation protein E22 gi|229600694 Heavy metal-transporting ATPase23 gi|229600580 DHH subfamily 1 protein24 gi|229600332 Arylsulfatase25 gi|229600155 Arylsulfatase26 gi|229600142 LysylphosphatidylglycerolsynthetaseMprF27 gi|229599950 Potassium-transporting ATPase subunit A

SN. Accession No. Hypothetical protein family predicted by Pfam

28 gi|229602747 Short chain fatty acid transporter

3.2. Prospective drug targets from membrane-bound proteins

In our study we found 28 membrane bound proteins thatcontribute in communication between the internal and externalenvironment of bacterial cell. The membrane associated proteinsinteract with extracellular signaling molecule and thus play amajor role in signal transduction. The advantage of targetingmembrane proteins is that it is easier to study their function bycomputer based method than wet-lab process. They have generallya uniform structure and a tendency to form secondary structurewhich can also be predicted by computational method. Forexample ab initio modeling methods can be used even whenthere is no experimental information or high resolution 3Dstructure available (Punta et al., 2007 Vaidehi et al., 2009).Moreover, combining experimental procedures with structuralgenomic studies, it has been possible to determine the crystalstructures of more than 180 membrane proteins (Deutscher et al.,2006). As a result structure based drug design would be possible infuture (Hossain et al., 2013; Sousa et al., 2006; Morya et al., 2010).

One major target can be glucose or fructose specific IIABCcomponent of PTS system. Their hydrophilic A, B domains areexposed at outer membrane and hydrophobic C domain isintegrated in plasma membrane which can be targeted to blockuptake and phosphorylation of carbohydrates (Frean et al., 2003).Another potential target can be F0F1 ATP synthase subunit beta.ATP synthase produces ATP from ADP in the presence of a protongradient across the membrane. The b regulatory subunit resides inperipheral F1 complex of ATP synthase. If beta subunit is madeinactive then ATP synthesis and hydrolysis will be inhibited leadingto bacterial instability (Deckers-Hebestreit and ltendorf, 1996).

This subtractive genome analysis method has been effectivelyapplied in various organisms such as Mycoplasma pneumoniaeM129 (Kumar et al., 2010), Staphylococcus aureus N315 (Hossainet al., 2013), Mycobacterium tuberculosis F11 (Hosen et al., 2014),Neisseria meningitides serogroup B for drug target identification(Narayan et al., 2009). The proteins identified in our work willequally lead to constructive approach for effective therapeuticdrug designing for future researchers.

4. Conclusion

In our whole study, we have investigated through twoequidistant ways to identify potential targets for drugs. We havefound 17 unique pathways and 17 proteins in unique pathways and28 membrane-bound proteins. Unique pathways are ideal for drug

M.A. Rahman et al. / Computational Biology and Chemistry 52 (2014) 66–72 71

targeting as they are not found in host (H. sapiens) and obtainedthrough subtractive genome analysis. Membrane associatedproteins can be targeted to oppose signal transduction. Structuralmodeling through virtual simulation can manifest the domains ofmembrane proteins whether it is extracellular, trans-membrane orintracellular for drug targeting. Being verified by laboratoryexperimental analysis, our identified human non-homologousunique metabolic proteins and membrane bound proteins wouldmake it easier for future researchers to develop effective drugagainst anthrax. Hence these findings will be significantly helpfulbecause there are reports on increasing threat of resistance of B.anthracis and its utilization as biological weapon now-a-days.

Conflict of interests

Author(s) discloses no potential conflict of interests.

Funding sources

There is no funding source.

References

Ahn, Y.Y., Lee, D.S., Burd, H., Blank, W., Kapatral, V., 2014. Metabolic networkanalysis-based identification of antimicrobial drug targets in category abioterrorism agents. PLoS One 9 (1), e85195.

Anishetty, S., Pulimi, M., Pennathur, G., 2005. Potential drug targets in Mycobacteri-um tuberculosis through metabolic pathway analysis. Comput. Biol. Chem. 29(5), 368–378.

Arinaminpathy, Y., Khurana, E., Engelman, D.M., Gerstein, M.B., 2009. Computa-tional analysis of membrane proteins: the largest class of drug targets. Drug Dis.Today 14 (23–24), 1130–1135.

Athamna, A., Athamna, M., Rashed, N.A., Medlej, B., Bast, D.J., Rubinstein, E., 2004.Selection of Bacillus anthracis isolates resistant to antibiotics. J. Antimicrob.Chemother. 54, 424–428.

Chakraborty, A., Khan, S.U., Hasnat, M.A., Parveen, S., Islam, M.S., Mikolon, A., et al.,2012. Anthrax Outbreaks in Bangladesh, 2009–2010. Am. J. Trop. Med. Hyg. 86(4), 703–710.

Chen, I.T., Dixit, A., Rhoads, D.D., Roufa, D.J.,1986. Homologous ribosomal proteins inbacteria, yeast, and humans. Proc. Natl. Acad. Sci. U. S. A. 83 (September),6907–6911.

Chen, Y., Tenover, F.C., Koehler, T.M., 2004. b-Lactamase Gene Expression in apenicillin-resistant Bacillus anthracis strain. Antimicrob. Agents Chemother. 48(12), 4873–4877.

Choe, C.H., Bouhaouala, S.S., Brook, I., Elliott, T.B., Knudson, G.B., 2000. In vitrodevelopment of resistance to ofloxacin and doxycycline in Bacillus anthracissterne. Antimicrob. Agents Chemother. 44 (6), 1766.

Chong, C.E., Lim, B.S., Nathan, S., Mohamed, R., 2006. In silico analysis ofBurkholderia pseudomallei genome sequence for potential drug targets. InSilico Biol. 6, 341–346.

Chopra, I., Roberts, M., Tetracycline Antibiotics, N.H., 2001. Mode of action,applications, molecular biology, and epidemiology of bacterial resistance.Microbiol. Mol. Biol. Rev. 65 (2), 232–260.

Deckers-Hebestreit, G., Altendorf, K., 1996. The F0F1-type ATP synthases ofbacteria: structure and function of the F0 complex. Ann. Rev. Microbiol.791–824.

Deutscher, J., Francke, C., Postma, P.W., 2006. How PhosphotransferaseSystem-Related Protein Phosphorylation Regulates Carbohydrate Metabolismin Bacteria. Microbiol. Mol. Biol. Rev. 70 (4), 939–1031.

Doganay, M., Aydin, N., 1991. Antimicrobial susceptibility of Bacillus-Anthracis.Scand. J. Infect. Dis. 23 (3), 333–335.

Dragon, D.C., Elkin, B.T., Nishi, J.S., Ellsworth, T.R., 1999. A review of anthrax inCanada and implications for research on the disease in northern Bison. J. Appl.Microbiol. 87, 208–213.

Drlica, K., Zhao, X., 1997. DNA gyrase, topoisomerase IV and the 4-quinolones.Microbiol. Mol. Biol. Rev. 61 (3), 377–392.

Du, W., Brown, J.R., Sylvester, D.R., Huang, J., Chalker, A.F., So, C.Y., et al., 2000. Twoactive forms of UDP-N-acetylglucosamine enolpyruvyl transferase ingram-positive bacteria. J. Bacteriol. 182 (15), 4146–4152.

Dutta, A., Singh, S.K., Ghosh, P., Mukherjee, R., Mitter, S., et al., 2006. In silicoidentification of potential therapeutic targets in the human pathogenHelicobacter pylori. Silico Biol. 6, 43–47.

Fariselli, P., Rossi, I., Capriotti, E., 2007. CasadioR: the WWWH of remote homologdetection: the state of the art. Brief Bioinform. 8 (2), 78–87.

Frean, J., Klugman, K.P., Arntzen, L., Bukofzer, S., 2003. Susceptibility of Bacillusanthracis to eleven antimicrobial agents including novel fluoroquinolones and aketolide. J. Antimicrob. Chemother. 52, 297–299.

Georgopapadakou, N.H., Liu, F.Y., 1980. Penicillin-binding proteins in bacteria.Antimicrob. Agents Chemother. 18 (1), 148–157.

Gordon, E., Flouret, B., Chantalat, L., Heijenoort, J.V., Lecreulx, D.M., Dideberg, O.,2001. Crystal structure of UDP-N-acetylmuramoyl–alanyl–glutamate:meso-diaminopimelate ligase from Escherichia Coli. J. Biol. Chem. 276 (14),10999–110061.

Haag, N.L., Velk, K.K., Wu, C., 2012. In silico identification of drug targets inmethicillin/multidrug-resistant Staphylococcusaureus. Int. J. Adv. Life Sci. 4(1-2), 21–32.

Hediger, M.A., Turk, E., Wright, E.M., 1989. Homology of the human intestinal Na+/glucose and Escherichia coli Na/proline cotransporters. Proc. Natl. Acad. Sci. U.S. A. 86 (August), 5748–5752.

Hosen, M.I., Tanmoy, A.M., Mahbuba, D.A., Salma, U., Nazim, M., Islam, M.T.,Akhteruzzaman, S., 2014. Application of a subtractive genomics approach for insilico identification and characterization of novel drug targets in Mycobacteriumtuberculosis F11. Interdiscip. Sci. 6 (1), 48–56.

Hossain, M., Chowdhury, D.U.S., Farhana, J., Akbar, M.T., Chakraborty, A., Islam, S., etal., 2013. Identification of potential targets in Staphylococcus aureus N 315 usingcomputer aided protein data analysis. Bioinformation 187–192.

Huang, Y., Niu, B., Gao, Y., Fu, L., Li, W., 2010. CD-HIT suite: a web server for clusteringand comparing biological sequences. Bioinformatics 26, 680–682.

Johnson, R.B., Fearon, K., Mason, T., Jindal, S., 1989. Cloning and characterization ofthe yeast chaperonin HSP60 gene. Gene 84 (December (2)), 295–302.

Kanehisa, M., Got, S., 2000. KEGG: kyoto encyclopedia of genes and genomes.Nucleic Acids Res. 28, 27–30.

Kanehisa, M., Goto, S., Sato, Y., Kawashima, M., Furumichi, M., Tanabe, M., 2014. Data,information, knowledge and principle: back to metabolism in KEGG. NucleicAcids Res. 42, 199–205.

Kedar, G.C., Driver, V.B., Reyes, D.R., Hilgers, M.T., Stidham, M.A., Shaw, K.J., et al.,2008. Comparison of the essential cellular functions of the two mura genes ofBacillus anthracis. Antimicrob. Agents Chemother. 52 (6), 2009–2013.

Kerfeld, C.A., Scott, K.M., 2011. Using BLAST to teach E-value-tionary concepts. PLoSBiol. 9 (2), e1001014.

King, B.R., Vural, S., Pandey, S., Barteau, A., 2012. GudaC.ngLOC: software and webserver for predicting protein subcellular localization in prokaryotes andeukaryotes. BMC Res. Notes 5, 351.

Koehler, T.M., 2009. Bacillus anthracis physiology and genetics. Mol. Aspects Med. 30(6), 386–396.

Kumar, G.S., Sarita, S., Kumar, G.M., Pant, K.K., Seth, P.K., 2010. Definition of potentialtargets in Mycoplasma Pneumoniae through subtractive genome analysis. J.Antivir. Antiretrovir. 2, 038–041.

Luo, H., Lin, Y., Gao, F., Zhang, C.T., Zhang, R., 2013. DEG 10, an update of the databaseof essential genes that includes both protein-coding genes and non-codinggenomic elements. Nucleic Acids Res. 42, 574–580.

Moriya, Y., Itoh, M., Okuda, S., Yoshizawa, A.C., Kanehisa, M., 2007. KAAS: anautomatic genome annotation and pathway reconstruction server. NucleicAcids Res. 35, 182–185.

Morya, V.K., Dewaker, V., Mecarty, S.D., Singh, R., 2010. In silico analysis metabolicpathways for identification of putative drug targets for Staphylococcus aureus. J.Comput. Sci. Syst. Biol. 3, 062–069.

Narayan, S.A., Rakesh, A., Qamar, R., Nidhi, T., 2009. Subtractive genomics approachfor in silico identification and characterization of novel drug targets in Neisseriameningitides serogroup B. Sci. Syst. Biol. 2 (5), 255–258.

Nathenson, S.G., Strominger, J.L., Ito, E., 1964. Enzymatic synthesis of the peptide inbacterial uridine nucleotides IV. Purification and properties of -glutamicacid-adding enzyme. J. Biol. Chem. 239, 1773–1776.

Punta, M., Forrest, L.R., Bigelow, H., Kernytsky, A., Liu, J., Rost, B., 2007. Membraneprotein prediction methods. Methods 41 (4), 460–474.

Punta, M., Coggill, P.C., Eberhardt, R.Y., Mistry, J., Tate, J., Boursnell, C., et al., 2012. ThePfam protein families database. Nucleic Acids Res. 40 (D1), D290–D301.

Rathi, B., Sarangi, A.N., Trivedi, N., 2009. Genome subtraction for novel targetdefinition in Salmonella typhi. Bioinformation 4 (October (4)), 143–150.

Sakharkar, K.R., Sakharkar, M.K., Chow, V.T., 2004. A novel genomics approach forthe identification of drug targets in pathogens, with special reference toPseudomonas aeruginosa. Silico Biol. 4 (3), 355–360.

Siddiqui, M.A., Khan, M.A., Ahmed, S.S., Anwar, K.S., Akhtaruzzaman, S.M., Salam, M.A., 2012. Recent outbreak of cutaneous anthrax in Bangladesh:clinico–demographic profile and treatment outcome of cases attended atRajshahi Medical College Hospital. BMC Res. Notes 5, 464.

Sousa, S.F., Fernandes, P.A., Ramos, M.J., 2006. Protein–ligand docking: currentstatus and future challenges. Proteins 65, 15–26.

Spencer, R.C., 2003. Bacillus anthracis. J. Clin. Pathol. 56 (3), 182–187.Stülke, J., 2007. Regulation of virulence in Bacillus anthracis: the phosphotransferase

system transmits the signals. Mol. Microbiol. 63 (3), 626–628.Swango, K.L., Hymes, J., Brown, P., Wolf, B., 2000. Amino acid homologies between

human biotinidase and bacterial aliphatic amidases: putative identification ofthe active site of biotinidase. Mol. Genet. Metab. 69 (February (2)), 111–115.

Terry, C., Dixon, B.S., Meselson, M., Guillemin, J., Hanna, P.C.,1999. Anthrax. N. Engl. J.Med. 341, 815–826.

Tipper, D.J., Strominger, J.L., 1965. Mechanism of action of penicillins: a proposalbased on their structural similarity to acyl–alanyl–alanine. Proc. N. A. S. 54,1133–1141.

Vaidehi, N., Pease, J.E., Horuk, R., 2009. Modeling small molecule-compoundbinding to G-protein-coupled receptors. Methods Enzymol. 460, 263–288.

Waxman, D.J., Strominger, J.L., 1980. Sequence of active site peptides from thepenicillin-sensitive -alanine carboxypeptidase of Bacillus subtili, mechanism ofpenicillin action and sequence homology to p-lactamases. J. Biol. Chem. 255 (9),3964–3976.

72 M.A. Rahman et al. / Computational Biology and Chemistry 52 (2014) 66–72

Yadav, P.K., Singh, G., Singh, S., Gautam, B., Saad, E.I.F., 2012. Potential therapeuticdrug target identification in community acquired-methicillin resistantStaphylococcus aureus (CA-MRSA) using computational analysis. Bioinformation8 (14), 664–672.

Yu, C.S., Chen, Y.C., Lu, C.H., Hwang, J.K., 2006. Prediction of protein subcellularlocalization. Proteins Struct. Funct. Bioinf. 64, 643–651.

Yu, N.Y., Wagner, J.R., Laird, M.R., Melli, G., Rey, S., Lo, R., et al., 2010. PSORTb 3.0:improved protein subcellular localization prediction with refined localizationsubcategories and predictive capabilities for all prokaryotes. Bioinformatics 26(13), 1608–1615.

Zhang, R., Lin, Y., 2009. DEG 5.0, a database of essential genes in both prokaryotesand eukaryotes. Nucleic Acid Res. 37, 455–458.