Upload
duongngoc
View
218
Download
0
Embed Size (px)
Citation preview
Research ArticleReconstruction of the Fatty Acid BiosyntheticPathway of Exiguobacterium antarcticum B7 Based onGenomic and Bibliomic Data
Regiane Kawasaki1 Rafael A Barauacutena1 Artur Silva1 Marta S P Carepo2 Rui Oliveira2
Rodolfo Marques2 Rommel T J Ramos1 and Maria P C Schneider1
1Genomics and Systems Biology Center Institute of Biological Sciences Federal University of Para 66075-110 Belem PA Brazil2Rede de Quımica e TecnologiaCentro de Quımica Fina e Biologica Chemistry Department Universidade Nova de Lisboa2829-516 Costa da Caparica Portugal
Correspondence should be addressed to Rafael A Barauna rabaraunagmailcom
Received 10 November 2015 Accepted 16 June 2016
Academic Editor Yongsheng Bai
Copyright copy 2016 Regiane Kawasaki et al This is an open access article distributed under the Creative Commons AttributionLicense which permits unrestricted use distribution and reproduction in any medium provided the original work is properlycited
Exiguobacterium antarcticum B7 is extremophile Gram-positive bacteria able to survive in cold environments A key factorto understanding cold adaptation processes is related to the modification of fatty acids composing the cell membranes ofpsychrotrophic bacteria In our study we show the in silico reconstruction of the fatty acid biosynthesis pathway of E antarcticumB7 To build the stoichiometric model a semiautomatic procedure was applied which integrates genome information using KEGGandRASTSEED Constraint-basedmethods namely Flux Balance Analysis (FBA) and elementarymodes (EM) were applied FBAwas implemented in the sense of hexadecenoic acid production maximization To evaluate the influence of the gene expression inthe fluxome analysis FBA was also calculated using the log
2FC values obtained in the transcriptome analysis at 0∘C and 37∘C
The fatty acid biosynthesis pathway showed a total of 13 elementary flux modes four of which showed routes for the productionof hexadecenoic acid The reconstructed pathway demonstrated the capacity of E antarcticum B7 to de novo produce fatty acidmolecules Under the influence of the transcriptome the fluxome was altered promoting the production of short-chain fatty acidsThe calculated models contribute to better understanding of the bacterial adaptation at cold environments
1 Introduction
Bacteria are increasingly used in industrial processes to pro-duce chemicals foods and drugs among other products [1]The main biochemical pathways of bacteria may be manipu-lated and optimized to more efficiently produce compoundsof industrial interest in various areas for example metabolicpathways ofCorynebacterium glutamicum are rationally engi-neered to produce L-amino acids on an industrial scale[2] To accomplish this task specific tools are used suchas FMM (from metabolite to metabolite) [3] Cytoscape [4]CellDesigner [5] SBW (Systems Biology Workbench) [6]COPASI (COmplex PAthway SImulator) [7] and COBRA(COnstraints Based Reconstruction and Analysis) toolbox[8]
The genomes of several bacterial strains have beensequenced and annotated and have been used in combina-tion with biochemical and physiological data to reconstructmetabolic networks at a genome-scale [9] Recently genomicmodels were reconstructed for some bacterial species aimingto increase the amount and quality of data that has been anno-tated in either the literature or databases [10ndash12] The draft network generated from the annotated genome still requiressignificantmanual curation for a comprehensive and accuratemetabolic representation of the organism [13]
The need to develop automatic or at least semiautomaticmethods to reconstruct metabolic networks from genomeannotation is increasing because the number of completegenome sequences available is growing fast Recent stud-ies [13 14] have highlighted the problems associated with
Hindawi Publishing CorporationBioMed Research InternationalVolume 2016 Article ID 7863706 9 pageshttpdxdoiorg10115520167863706
2 BioMed Research International
genome annotations anddatabases which performautomaticreconstructions and thus require manual assessment Thecurrently available 96-step protocol proposed by Thiele andPalsson [15] is a well-established process for the assemblycuration and validation of metabolic reconstruction Thisprotocol is combined with computational tools includingthe visualization and numerical calculation software packageMATLAB (MathWorks USA)
Constraint-based modeling is frequently used as finalvalidation step of the reconstructed network This step isextremely useful for simulating the phenotypic behaviorunder different physiological environments [16ndash18] therebyallowing assessing if the reconstructed network representswell the in vivo cellular system Microbial adaptation to coldenvironments is one of the applications of these methods
Psychrotrophic microorganisms have an optimal growthtemperature higher than 15∘C but are also able to grow andadapt to extremely cold environments with temperatures ofapproximately 0∘C [19] Thus the unique physicochemicalcharacteristics of their habitat and the biological appara-tus developed by these microorganisms to survive underthese conditions render these organisms valuable sources ofbiotechnological processes The cellular response to cold bypsychrotrophic bacteriamay be studied from a general stand-point with the advent of omics methods Recently the B7strain ofExiguobacteriumantarcticumwas isolated fromGin-ger Lake sediments located in the Antarctic Peninsula Region(69∘301015840S 65∘W)This lake was formed due to the warming inthe region which led to partial melting of ice caps [20] Thegenome of this strain was sequenced [21] and its responseto cold was evaluated through differential expression of itsgenome at 37∘Cand0∘Cusing omicsmethods [22]One of themechanisms of cold adaptation of all psychrophilic or psy-chrotrophic organisms is the change in the chemical struc-tures of the membrane phospholipids The fatty acid chainsbecome shorter and unsaturated at low temperatures Ac-cordingly the fluidity of the membrane is kept intact [19 23]
Bacterial de novo synthesis of fatty acids is regulatedby the protein FapR [24] which is responsible for acti-vatingdisabling a regulon consisting of four operons in Eantarcticum B7 In cold two of these operons are repressed(fabH1-fabF and fabI) and the expression levels of the othertwo remain unaltered [22]The chemical components whichare included in this regulon must be reconstructed andevaluated and then associated with their respective geneticelements to further understand this metabolic pathway andto reach more complete conclusions about its importance foradaptation to cold [23 24] Bioinformatics methods are usedfor in silico reconstruction of metabolic pathways [25]
In this work we present the in silico reconstruction ofthe fatty acid biosynthesis pathway of the Exiguobacteriumantarcticum B7 based on linear programming (FBA) andconvex cone method (elementary modes) The influence oftranscriptome in FBA calculation was also evaluated
2 Materials and Methods
21 DataCollection Thegenomic data ofE antarcticumB7 inthe formats gbk and fasta were collected from NCBI under
accession number NC 018665 The metabolic pathway of thefatty acid biosynthesis was initially evaluated in the KEGG(Kyoto Encyclopedia of Genes and Genomes) database [26]When necessary the visualization of the genome of thebacterium was performed using Artemis software [27]
22 Preliminary Reconstruction This stepwas performed fol-lowing two methods one semiautomatic and the other auto-maticThe tools within the KEGGdatabases were used essen-tially in the semiautomatic method Thefasta file with theE antarcticum B7 genome was submitted to the online toolKAAS (KEGG Automatic Annotation Server) [28] availableat httpwwwgenomejpkaas-binkaas main The parame-ters chosen to run this software were as follows (a) bidirec-tional Best Hit (BBH) Method recommended for completegenomes performs the search for orthologous genes betweena specific group of organisms and (b) prokaryote the set ofgenes chosen should be representative of the target organismin this case the bacterium E antarcticum B7 Following theprocessing a text file was generated (queryko) Each line ofthis file is formed by two parameters the first consists of thesequence identified (gene) and the second when presentconsists of the KO assignment termed119870 number This valueindicates orthologous groups encoding the same enzymaticactivity Afterwards the file generated is passed through a fil-ter an auxiliary computer software program (script) Pythondeveloped for the present study which only selects 119870 num-bers and individually and increasingly commands per lineinto a new file (new queryko) This file was used as entry inthe option User Data Mapping of the Pathway Mapping toolof the KEGG
The automaticmethod essentially consisted of submittingthe fasta file of the E antarcticum B7 genome to the onlinetool RAST (RapidAnnotation using Subsystems Technology)[29] available at httprastnmpdrorg to generate the draftsof the metabolic network and of the fatty acid biosynthesispathway of the target microorganism The final draft of thisstep was generated from the combination of the resultingpathways of the semiautomatic and automatic models Thecommon pathways were maintained while surplus com-pounds enzymes and reactions that is present in some butabsent in others were not directly excluded but were insteadreserved for the curated step
23 Manual Curation The following steps were completed inthis manual curation stage following the protocol explainedabove (i) Draft refinement this phase beganwith the analysisof enzymes and reactions components of the fatty acidbiosynthesis pathway by reading books and articles specifi-cally on the subjectTheobjectivewas to diagnose the absenceor presence of more than one element of the study pathwayThe online databases KEGG ENZYME [30] and SEED [31]were consulted to ratify the enzymes and the structures of thereactions
(ii) Assessment of the stoichiometry and reversibility ofthe reactions in this step all model reactions were assessedand stoichiometrically corrected if necessary The biochemi-cal data on the organism are very important to determine thereversibility of the reaction For this purpose the databases
BioMed Research International 3
Table 1 Genes locus tags and EC numbers identified in the draft of the fatty acid biosynthesis pathway of E antarcticum B7 The featuresdisplayed were generated using the methods semiautomatic and automatic
Semiautomatic method Automatic methodGene Locus tag EC number Gene Locus tag EC numberaccA Eab7 2059 6412 accA Eab7 2059 6412accB Eab7 0870 6412 accB Eab7 0870 64114accC Eab7 0871 6412 accC Eab7 0871 6412accD Eab7 2060 6412 accD Eab7 2060 64114fabD Eab7 1760 23139 fabD Eab7 1760 23139fabH1 Eab7 1911 231180 fabH1 Eab7 1911 231180fabF Eab7 1910 231179 fabF Eab7 1910 231179fabG Eab7 1795 111100 fabG Eab7 1795 111100fabZ Eab7 2463 42159 fabI Eab7 1885 13110fabI Eab7 1885 13110fabK Eab7 0377 1319mdash Eab7 2235 114192
(KEGG SEED and ENZYME) and the tool eQuilibrator [32]were used to analyze the reversibility of the reactions Thusthe thermodynamic constraints were respected
(iii) Addition of gene data and reaction location theArtemis tool was used to identify the genes of the reactions(enzymes) from their locations in the genome assessed usingthe draft generated
(iv) Assessment of Gene-Protein-Reaction (GPR) associ-ations in this step the function of each gene is indicatedGPRassociations were identified using databases of the organismand specific literature
(v) Definition of constraints some constraints weredefined in the model in this manual curation step includingstoichiometric and thermodynamic constraints (through thereversibility and irreversibility of fluxes)
24 Metabolic Model Design The metabolic model designedand refined following themanual curation stepwas convertedinto a mathematical representation termed a stoichiometricmatrix which encouraged the development of a wide varietyof computational tools to analyze network properties
The constraints of capacity which are the upper and lowerlimits defining the maximum and minimum fluxes allowedfor the reactions were added in this step The inputs of thestoichiometricmatrix are the coefficients of themetabolites inthe reactions with negative values for consumed metabolites(substrates) and positive values when themetabolites are pro-duced or secreted (products) (Additional File 1 see Supple-mentaryMaterial available online at httpdxdoiorg10115520167863706)
25Metabolic PathwayValidation Thecomputationalmodelsought to examine the metabolic capabilities and to evaluatethe systemproperties theymayperformunder the constraintsimposed on the cellThus the final step in the reconstructionprocess consisted of assessing evaluating and validating thefatty acid biosynthesis pathway of E antarcticum B7 Thevalidation of that metabolic model was performed using sim-ulation and flux analysis The fatty acid biosynthesis pathway
is well described in the literature because it is a highlyconserved process among organisms which facilitated itscomplete definition Thus most gaps had already been filledduring the manual curation process
3 Results and Discussion
The expectation to understand the relationship between thegenome and the physiology of a particular organismwas a keyincentive for reconstructing metabolic networks Protocoladaptations using semiautomatic and automatic methods arenecessary to reconstruct themetabolic networks of organismswith few reported data on theirmetabolic capabilities includ-ing E antarcticum B7
31 Pathway Reconstruction Using the Semiautomatic MethodThe draft of the metabolic network of E antarcticum B7was retrieved from the KEGG database [26] The KEGGMetabolic Pathway tool was used to highlight the fattyacid biosynthesis pathway from the resulting draft of themetabolic networkThe genes annotated and identified usingKEGG and their respective enzymes are shown in Figure 1(a)in green and the others are listed in white boxes Table 1shows the genes locus tags and enzymes identified using theKEGG Metabolic Pathway tool A total of 11 locus tags asso-ciated with their respective genes and Enzyme Commission(EC) numbers were identified only the locus tag Eab7 2235has no added gene associated with it
32 Pathway Reconstruction Using the Automatic MethodThe RASTSEED tool does not provide graphic display ofthe metabolic map draft as KEGG for this purpose it usesa standard table to list the 247 metabolic pathways that com-pose the network regardless of whether they were identifiedin the genome of the microorganism RAST identified 25reactions 40 compounds and 20ECnumbers in the fatty acidbiosynthesis pathway ofE antarcticumB7 Table 1 outlines thegenes locus tags and enzymes identified at this step Anno-tated genes identified by RAST and their respective enzymes
4 BioMed Research International
FAS2
FASN
FAS1 Fas
FASN
Fas
FabY
Fas
FASN
FAS1 FabD
6412
Acetyl-[acp]
Acetyl-CoA
Citrate cycle
Malonyl-CoA
ACP
Pyruvate metabolism
120573-Alanine metabolism
Malonyl-[acp]
FabH
FabF
FabB
Acetoacetyl- [acp]
FASN
FAS2
Fas
FabG
(R)-3-Hydroxy-butyryl-[acp]
FASN
FAS1
Fas FabZ
FabA
But-2-enoyl-[acp]
FASN
FAS1
Fas
FabI
FabK
FabL
FabV
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
FabA
FabI
FabK
FabL
FabV
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
FabA
FabI
FabK
FabL
FabV
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
FabA
FabI
FabK
FabL
FabV
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
FabA
FabI
FabK
FabL
FabV
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
FabA
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
FabA
FabI
FabK
FabL
FabV
(R)-3-Hydroxy-hexanoyl-[acp]
trans-Hex-2-enoyl-[acp]
3-Oxooctanoyl-[acp]
(R)-3-Hydroxy-octanoyl-[acp]
trans-Oct-2-enoyl-[acp]
3-Oxodecanoyl-[acp]
(R)-3-Hydroxy-decanoyl-[acp]
trans-Dec-2-enoyl-[acp]
3-Oxododecanoyl-[acp]
(R)-3-Hydroxy-dodecanoyl-[acp]
trans-Dodec-2-enoyl-[acp]
3-Oxo-tetradecanoyl-[acp]
(R)-3-Hydroxy-tetradecanoyl-[acp]
trans-Tetradec-2-enoyl-[acp]
FabI
FabK
FabL
FabV
3-Oxohexadecanoyl-[acp]
(R)-3-Hydroxy-hexadecanoyl-[acp]
trans-Hexadec-2-enoyl-[acp]
FabF
FabB
FabG
FabZ
FabI
FabK
Fas
Fas
Fas
Fas
3-Oxooctadecanoyl-[acp]
(R)-3-Hydroxy-octadecanoyl-[acp]
trans-Octadec-2-enoyl-[acp]
Butyryl-[acp] Hexanoyl-[acp] Octanoyl-[acp]
31221
Octanoic
Lipoic acid metabolism
53314
cis-Dec-3-enoyl-[acp]
n-7 unsatured acyl-[acp]
Decanoyl-[acp]
31221
Decanoic acid
Dodecanoyl-[acp]
31221
Dodecanoic acid
Tetradecanoyl-[acp]
31214 FASN
Tetradecanoic acid
Hexadeca-noyl-[acp]
31214 FASN
Hexadecanoic acid
6213
FAS1 Fas
Fatty acid degradation
Hexadecanoyl-CoA
6213
31214
Hexadecanoic acid
31214
Octadecanoic acid
6213 Octadecanoyl-[acp]
Octadecenoyl-[acp]
31214
Octadecenoic acid
Glycerolipid metabolism
Glycerophospholipid metabolism
Fatty acid elongation
3-Oxohexanoyl-[acp]
(a)
Fas
FAS2
FASN
FAS1 FabH
FASN
Fas
FabD
6412
Acetyl-[acp]
Acetyl-CoA
Citrate cycle
Malonyl-CoA
Pyruvate metabolism
120573-Alanine metabolism
Malonyl-[acp]
FabF
FabB
Acetoacetyl- [acp]
FASN
FAS2
Fas
FabG
(R)-3-Hydroxy-butyryl-[acp]
FASN
FAS1
Fas FabZ
42158
But-2-enoyl-[acp]
FASN
FAS1
Fas
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
42158
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
FabA
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
42161
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
42161
FAS2
FASN
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
42161
3-Oxohexanoyl-[acp]
(R)-3-Hydroxy-hexanoyl-[acp]
trans-Hex-2-enoyl-[acp]
3-Oxooctanoyl-[acp]
(R)-3-Hydroxy-octanoyl-[acp]
trans-Oct-2-enoyl-[acp]
3-Oxodecanoyl-[acp]
(R)-3-Hydroxy-decanoyl-[acp]
trans-Dec-2-enoyl-[acp]
3-Oxododecanoyl-[acp]
(R)-3-Hydroxy-dodecanoyl-[acp]
trans-Dodec-2-enoyl-[acp]
3-Oxo-tetradecanoyl- [acp]
(R)-3-Hydroxy-tetradecanoyl-
[acp]
trans-Tetradec-
3-Oxohexadecanoyl-[acp]
(R)-3-Hydroxy-hexadecanoyl-[acp]
trans-Hexadec-2-enoyl-[acp]
FabF
FabB
FabG
FabZ
FabI
FabK
Fas
Fas
Fas
Fas
3-Oxooctadecanoyl-[acp]
(R)-3-Hydroxy-octadecanoyl-[acp]
trans-Octadec-2-enoyl-[acp]
Butyryl-[acp] Hexanoyl-[acp] Octanoyl-[acp]
FatB
Octanoic acid
Decanoyl-[acp]
FatB
Decanoic acid
Dodecanoyl-[acp]
31221
Dodecanoic acid
Tetradecanoyl-[acp]
31214 FASN
Tetradecanoic acid
Hexadecanoyl-[acp]
31214
FASN
Hexadecanoic acid
Fas
114192
Hexadecanoic acid
31214
Octadecanoic acid
114192
Octadecanoyl-[acp]
Octadecenoyl-[acp]
31214
Octadecenoic acid
Fas
BCCA
FASN
FAS1
63414
Carboxybiotin-carboxyl-carrier protein
Biotin-carboxyl-carrier protein
FabH
FabI
FabK
FabL
FabI
FabK
FabL
42158
FabI
FabK
FabL
FabI
FabK
FabL
FabI
FabK
FabL
2-enoyl-[acp]
FabI
FabK
FabL
FabI
FabK
FabL
Fas
FatB FatA FASN
Fatty acid metabolism
31214
ACP
(b)
Figure 1 Drafts of the fatty acid biosynthesis pathway of E antarcticumB7 bacteriaThe drafts were designed using the followingmethods (a)semiautomatic method generated by the KEGG database and (b) automatic method generated by online tool RAST Colored boxes indicatethe possibility of the presence of enzymes in the pathway
BioMed Research International 5
Table 2 Relationships between components of the E antarcticum B7 fatty acid biosynthesis pathway following curation The signalsrArr andhArr indicate irreversible and reversible reactions respectively
Gene Locus tag EC number Enzyme Reaction Fold change (log2FC)
accA Eab7 2059 6412 Acetyl-CoA carboxylase carboxyltransferase alpha subunit
ATP + acetyl-CoA + HCO3minusrArr ADP +
orthophosphate + malonyl-CoA 04562
accB Eab7 0870 6412 Acetyl-CoA carboxylasebiotin-carboxyl carrier protein
ATP + acetyl-CoA + HCO3minusrArr ADP +
orthophosphate + malonyl-CoA minus005773
accC Eab7 0871 641264114
Acetyl-CoA carboxylase biotincarboxylase subunit
ATP + acetyl-CoA + HCO3minusrArr ADP +
orthophosphate + malonyl-CoA minus05623
accD Eab7 2060 6412 Acetyl-CoA carboxylasecarboxyl transferase beta subunit
ATP + acetyl-CoA + HCO3minusrArr ADP +
orthophosphate + malonyl-CoA minus02811
fabD Eab7 1760 23139 ACP S-malonyl transferase Malonyl-CoA + ACPhArr CoA +malonyl-(acp) 09448
fabH1 Eab7 1911 231180 3-Oxoacyl-ACP synthase III Acetyl-CoA + malonyl-(acp)hArracetoacetyl-(acp) + CoA + CO
2
08512
fabF Eab7 1910 231179 3-Oxoacyl-ACP synthase II Acetyl-(acp) + malonyl-(acp)rArracetoacetyl-(acp) + CO
2+ ACP 06942
fabG Eab7 1895 111100 3-Oxoacyl-ACP reductase Acetoacetyl-(acp) + NADPH + H+ hArr(R)-3-hydroxybutanoyl-(acp) + NADP+ 08523
fabZ Eab7 2463 42159 3-Hydroxyacyl-ACP dehydratase (R)-3-hydroxybutanoyl-(acp)hArrbut-2-enoyl-(acp) + H
2O 012902
fabI Eab7 1885 131913110 Enoyl-ACP reductase I But-2-enoyl-(acp) + NADH + H+ hArr
butyryl-(acp) + NAD+ 02969
mdash Eab7 2235 114192 Acyl-ACP desaturaseHexadecanoyl-(acp) + acceptor reduced+ O2rArr hexadecenoyl-(acp) + acceptor +
2H2O
13768
are colored in purple and the others are listed in white boxes(Figure 1(b))
The analysis of both fatty acid biosynthetic pathway draftsshows that the draft generated using KEGG apparently hasthe most complete flux except for enzyme 63414 whichis exclusively present in the draft resulting from the RASTtool The draft generated using RAST has a gap in which theenzymes FabA and FabB are not included in the pathwayelongation process The flux for the production of hexade-cenoic acid is also absent from the pathway generated usingRAST
Artemis software was used to confirm the presence ofall genes selected through the automatic and semiautomaticmethodsThe genes accABCD fabD fabH1 fabF fabG fabZand fabI and the locus tag Eab7 2235 were described in thegenome of E antarcticum B7 except fabK gene which wasdetected only by the automatic method
The KGML file produced by KEGG was submitted to thesoftware KEGGtranslator [33] to be converted into a SBML(System Biology Markup Language) file [34] This file wasconverted into an Excel spreadsheet using MATLAB func-tions The files in SBML format and the Excel spreadsheetsare the most used formats in metabolic reconstructions Thereactions and metabolites of the preliminary network gener-ated using KEGG could be visualized in the spreadsheet
The data generated using the RASTSEED tool were ana-lyzed and added to the first step of the process supplementingthe data collected using KEGGThe files generated with bothplatforms were used to manage the manual curation data
The larger number of genes identified using KEGG (12)compared to those found using RASTSEED (9) may beexplained because the former uses orthology (KEGGOrthol-ogy (KO)) through protein homology to identify the so-calledmetabolite genes [35] in a genome which facilitates findinggene-protein-reaction (GPR) associations
33 Manual Curation of the Metabolic Pathway of De NovoFatty Acid Synthesis The reactions of the fatty acid synthesispathway were annotated and refined The metabolites wereorganized into two compartments (cytoplasm and extracellu-lar compartment) based on the location of the enzymes asso-ciated with each pathway The cofactors and the reversibilityof the reactions were compiled from the data published in theliterature and online tools (ENZYME and BRENDA)The ECnumberwas noted and the genes were identified A summaryof those results is shown in Table 2Thermodynamic analysisof the reactions revealed that malonyl-CoA synthesis fromacetyl-CoA (AcCoA) is an irreversible process similar tothe process regarding the fabF gene the Eab7 2235 locustag and the extracellular metabolites the other processes arereversible
It is very important to assess the quality of the annotatedgenome submitted to the online tools during curation Theliterature categorically states that the quality of the recon-structed network directly depends on the annotated genomeof the organism The rule is to use the latest updated versionof the annotated genome [36ndash38]
6 BioMed Research International
297
295
296294
184
288300
64
32432533299185
44257
7018
26186182177
52286283280278276270268
81921920
10321
165 170 173 178 181 187 27160
54758697108119130142
162 167 168 175 176 183
154
66
147
157
156
153
158
155
67
163
259 61 60 59 58 536263262
161 166 169 174136 125 114
28188180179172171159 164
51 23 22
Figure 2 Layout of the fatty acid biosynthesis pathway generated using Cytoscape Green vertices fewer connections Yellow vertices regularnumber of connections Red vertices large number of connections This network is a free model scale
34 Curated Metabolic Network The previous step addedconstraints regarding the stoichiometry (chemical balance)thermodynamics (reversibility of reactions) and physiology(cofactors used) of the study model The result was a systemof equations that describes the cell metabolism according tothe metabolites of interest The mathematical representationof this model essentially consisted of describing the perfor-mance of the fatty acid biosynthesis pathway using a stoichio-metric matrix This data structure consists of 54 metabolitesand 59 reactions resulting in a 54 times 59 stoichiometric matrix(Additional File 1) Additional File 2 shows the list of bio-chemical reactions identified in the curated model
The reconstructed pathway model was converted intoSBML in the used MATLAB toolbox The SBML file wasvalidated using the tool SBML validator and was then sub-mitted to the tool Cytoscape which generated the networkof Figure 2 The gene-protein-reaction (GPR) representationtherein describes the degree of connectivity of each enzyme inthe pathway Vertices with few connections are in green thevertices with regular numbers of connections are in yellowand the vertices with large numbers of connections are in redThe network connectivity obeys a scale-free model [39]
35 Flux Balance Analysis (FBA) The FBA was coded inMATLAB implementing a constrained linear program usingthe GLPK (GNU Linear Programming Kit) linear optimiza-tion library [8] All fluxes were calculated in percentage of
the input flux of AcCoA (reaction 39) which was fixed to100 Hexadecenoic acid is the key metabolic product thusthe respective flux (reaction 41) was set as the maximiza-tion target for FBA To improve convergence upper andlower bounds were [0 100] for irreversible reactions and[minus100 100] for reversible reactionThe final optimized fluxesare shown inAdditional File 3The targetmaximum reactionR41 in Additional File 3 was 769 which may be read as 769moles of hexadecenoic acid produced for every 100 moles ofAcCoA consumed per unit time per unit cell mass
Figure 3(a) shows the generated flux plot which showsthe variation occurring between the response fluxes with themajority approximately 37 showing positive values smallerthan 20 while 15 are above that range
36 Influence of Transcriptome in FBA Calculation Log base2-fold change values (log
2FC) obtained in vitro by DallrsquoAgnol
and colleagues [22] were used to evaluate the influence ofdifferential expression in the FBA calculation These valueswere obtained by comparison of RPKM (reads per kilobaseper million reads sequenced) generated in the transcriptomeof the bacterium at 0∘C and 37∘C The log
2FC of genes that
composes the fatty acid biosynthesis pathway is shown inTable 2
As presented by DallrsquoAgnol and colleagues [22] theaerobic energetic metabolism of E antarcticum B7 at 0∘C isrepressed and a fraction of the acetyl-CoA is probably used as
BioMed Research International 7
Reaction
FBA
minus40
minus20
0
20
40
60
80
100
Flux
val
ue (
)
lsquolsquoR1rsquorsquo
lsquolsquoR4rsquorsquo
lsquolsquoR7rsquorsquo
lsquolsquoR10
rsquorsquolsquolsquoR
13rsquorsquo
lsquolsquoR16
rsquorsquolsquolsquoR
19rsquorsquo
lsquolsquoR22
rsquorsquolsquolsquoR
25rsquorsquo
lsquolsquoR28
rsquorsquolsquolsquoR
31rsquorsquo
lsquolsquoR34
rsquorsquolsquolsquoR
37rsquorsquo
lsquolsquoR40
rsquorsquolsquolsquoR
43rsquorsquo
lsquolsquoR46
rsquorsquolsquolsquoR
49rsquorsquo
lsquolsquoR52
rsquorsquolsquolsquoR
55rsquorsquo
lsquolsquoR58
rsquorsquo
(a)
Transcriptome
lsquolsquoR1rsquorsquo
lsquolsquoR4rsquorsquo
lsquolsquoR7rsquorsquo
lsquolsquoR10
rsquorsquolsquolsquoR
13rsquorsquo
lsquolsquoR16
rsquorsquolsquolsquoR
19rsquorsquo
lsquolsquoR22
rsquorsquolsquolsquoR
25rsquorsquo
lsquolsquoR28
rsquorsquolsquolsquoR
31rsquorsquo
lsquolsquoR34
rsquorsquolsquolsquoR
37rsquorsquo
lsquolsquoR40
rsquorsquolsquolsquoR
43rsquorsquo
lsquolsquoR46
rsquorsquolsquolsquoR
49rsquorsquo
lsquolsquoR52
rsquorsquolsquolsquoR
55rsquorsquo
lsquolsquoR58
rsquorsquo
Reaction
02468
10121416
Flux
val
ue (
)
(b)
Figure 3 Flux graph generated using the software MATLAB which uses methods based on FBA constraints The vertical axis represents thefluxes calculated from the input of 100 moles of acetyl-CoA (AcCoA) The horizontal axis represents the reactions participating in the fattyacid biosynthesis pathwayThe blue bars determine the output percentage for each pathway reaction (a) FBA for maximizing the productionof hexadecenoic acid (b) FBA calculated with the log
2FC values obtained from the transcriptome
a substrate to synthesize short-chain fatty acids in coldThe synthesis begins with the conversion of acetyl-CoA tomalonyl-CoA catalyzed by the multimeric enzyme encodedby the genes accABCD In FBA analysis only half of the inputof acetyl-CoA (100mM) is converted tomalonyl-CoA whichbinds to the acyl carrier protein (ACP) at 0∘C (Figure 3(b)) Asstated earlier the other half of acetyl-CoA is probably used forenergy generation
The remaining route is cyclical being the reactionscatalyzed by enzymes encoded by fabF fabG fabZ and fabIgenes At each round two carbons are added to the growingchain of fatty acid In these reactions the flow of metabolitesremains unchanged until the fatty acid molecule reachesa size of eight carbons (octadecanoyl-ACP in reaction 38)where the percentage of the flow amount decreases (Fig-ure 3(b)) These results are consistent with the previouslypublished data which affirm that bacteria decrease their fattyacid chains to survive in coldThese short fatty acidmoleculeswill be converted into membrane phospholipids in order tomaintain the fluidity of this biological barrier
37 Elementary Flux Modes The calculation of elementarymodes was performed in MATLAB using the Metatooltoolbox [40] (modo elementarm code in the SupplementalInformation) A total of 13 elementary fluxmodes were foundfor the fatty acid biosynthesis pathway of E antarcticum B7(Additional File 4) Of these only 4 elementary modes (25 8 and 11) have a positive nonzero coefficient for reaction41 which indicates that the target product hexadecenoicacid may only be generated by one of these four elementarymodes The routes identified in EM2 and EM5 begin at thesecond reaction (R2) which is catalyzed by the enzyme FabDThevalue of R2 for both elementarymodes indicates a consid-erable production of malonyl-CoA The remaining reactionsof EM2 and EM5 are presented in Additional File 4 In EM8
and EM11 the routes begin from R1The value 1 for this reac-tion indicates a lower activity reflecting in a lower productionof hexadecenoic acid Regarding reaction 41 the values ofelementary modes 2 and 5 in this case 1 are higher thanthe values of 8 and 11 (0142857 each) indicating that bothelementary modes 2 and 5 produce 1mol hexadecenoic acidwhen they are active while elementary modes 8 and 11produce 0142857 moles
4 Conclusions
The first metabolic pathway of E antarcticum B7 recon-structed following the steps defined in this work suggeststhat the protocol used is a suitable tool for further metabolicreconstruction studies Almost all the first steps of the processwere automated however manual curation was as usuallaborious because it required an intensive search for availabledata
The metabolic pathway of fatty acid biosynthesis wasrepresentative and consistent under the limits and boundaryconditions setThe FBA and elementary mode methods wereused to examine the hexadecenoic acid production potentialof the reconstructed pathway The application of constraint-based modeling revealed being very useful to assess networkoperation plasticity even if the intracellular kinetics areunknownThe in silico analysis performed using FBA enableda quantitative assessment of hexadecenoic acid productionpotential
Finally a key issue involves deciding when to stop theprocess and to consider the reconstruction finalized at leasttemporarilyThis decision is usually based on the reconstruc-tion purpose The most complete metabolic model currentlyavailable is the E coli model which has been researchedand refined for over 10 years [41ndash45] Other studies con-stantly updating their models are Homo sapiens with three
8 BioMed Research International
reconstructions [46ndash48] and S cerevisiae with more than adozen reconstructions including two in 2013 [49 50] Theprotocol reported in the present studymay be used to compileseveral data pieces available in the literature aimed at propos-ing possible metabolic pathways thereby enabling deeperresearch of the metabolism under study
Competing Interests
The authors declare that they have no competing interests
Acknowledgments
This study was supported by the Conselho Nacional de Des-envolvimento Cientıfico e Tecnologico (CNPq) and Coor-denacao de Aperfeicoamento de Pessoal de Nıvel Superior(CAPES)
References
[1] JW Lee DNa JM Park J Lee S Choi and S Y Lee ldquoSystemsmetabolic engineering of microorganisms for natural and non-natural chemicalsrdquo Nature Chemical Biology vol 8 no 6 pp536ndash546 2012
[2] J Becker and C Wittmann ldquoSystems and synthetic metabolicengineering for amino acid productionmdashthe heartbeat of indus-trial strain developmentrdquoCurrent Opinion in Biotechnology vol23 no 5 pp 718ndash726 2012
[3] C-H ChouW-C Chang C-M Chiu C-C Huang andH-DHuang ldquoFMM a web server for metabolic pathway reconstruc-tion and comparative analysisrdquo Nucleic Acids Research vol 37no 2 pp W129ndashW134 2009
[4] P Shannon A Markiel O Ozier et al ldquoCytoscape a softwareEnvironment for integratedmodels of biomolecular interactionnetworksrdquo Genome Research vol 13 no 11 pp 2498ndash25042003
[5] A Funahashi Y Matsuoka A Jouraku M Morohashi NKikuchi and H Kitano ldquoCellDesigner 35 a versatile modelingtool for biochemical networksrdquo Proceedings of the IEEE vol 96no 8 pp 1254ndash1265 2008
[6] F T Bergmann and HM Sauro ldquoSBWmdashamodular frameworkfor systems biologyrdquo in Proceedings of the 37th Winter Simula-tion Conference (WSC rsquo06) pp 1637ndash1645 December 2006
[7] S Hoops S Sahle R Gauges et al ldquoCOPASImdasha COmplexPAthway SImulatorrdquo Bioinformatics vol 22 no 24 pp 3067ndash3074 2006
[8] J Schellenberger R Que R M T Fleming et al ldquoQuantitativeprediction of cellular metabolism with constraint-based mod-els the COBRA Toolbox v20rdquo Nature Protocols vol 6 no 9pp 1290ndash1307 2011
[9] M Terzer N D Maynard M W Covert and J StellingldquoGenome-scale metabolic networksrdquo Wiley InterdisciplinaryReviews Systems Biology andMedicine vol 1 no 3 pp 285ndash2972009
[10] DMcCloskey B Oslash Palsson andAM Feist ldquoBasic and applieduses of genome-scale metabolic network reconstructions ofEscherichia colirdquo Molecular Systems Biology vol 9 article 6612013
[11] M Fondi I Maida E Perrin et al ldquoGenome-scale metabolicreconstruction and constraint-basedmodelling of the Antarctic
bacterium Pseudoalteromonas haloplanktis TAC125rdquo Environ-mental Microbiology vol 17 no 3 pp 751ndash766 2015
[12] DMcCloskey B Oslash Palsson andAM Feist ldquoBasic and applieduses of genomeminusscale metabolic network reconstructions ofEscherichia colirdquo Molecular Systems Biology vol 9 article 6612013
[13] A M Feist M J Herrgard I Thiele J L Reed and B OslashPalsson ldquoReconstruction of biochemical networks in microor-ganismsrdquoNature ReviewsMicrobiology vol 7 no 2 pp 129ndash1432009
[14] J L Reed I Famili IThiele and B Oslash Palsson ldquoTowardsmulti-dimensional genome annotationrdquoNature Reviews Genetics vol7 no 2 pp 130ndash141 2006
[15] I Thiele and B Oslash Palsson ldquoA protocol for generating ahigh-quality genome-scale metabolic reconstructionrdquo NatureProtocols vol 5 no 1 pp 93ndash121 2010
[16] J S Edwards R U Ibarra and B O Palsson ldquoIn silico pre-dictions of Escherichia colimetabolic capabilities are consistentwith experimental datardquoNature Biotechnology vol 19 no 2 pp125ndash130 2001
[17] R Mahadevan B Oslash Palsson and D R Lovley ldquoIn situ toin silico and back elucidating the physiology and ecology ofGeobacter spp using genome-scale modellingrdquo Nature ReviewsMicrobiology vol 9 no 1 pp 39ndash50 2011
[18] J D Trawick andCH Schilling ldquoUse of constraint-basedmod-eling for the prediction and validation of antimicrobial targetsrdquoBiochemical Pharmacology vol 71 no 7 pp 1026ndash1035 2006
[19] R Y Morita ldquoPsychrophilic bacteriardquo Bacteriological Reviewsvol 39 no 2 pp 144ndash167 1975
[20] R Mulvaney N J Abram R C A Hindmarsh et al ldquoRecentAntarctic Peninsula warming relative to Holocene climate andice-shelf historyrdquo Nature vol 489 no 7414 pp 141ndash144 2012
[21] A R Carneiro R T J Ramos H DallrsquoAgnol et al ldquoGenomesequence of Exiguobacterium antarcticum B7 isolated from abiofilm inGinger Lake KingGeorge Island Antarcticardquo Journalof Bacteriology vol 194 no 23 pp 6689ndash6690 2012
[22] HDallrsquoAgnol R A Barauna P de Sa et al ldquoOmics profiles usedto evaluate the gene expression of Exiguobacterium antarcticumB7 during cold adaptationrdquo BMCGenomics vol 15 no 1 article986 2014
[23] A Casanueva M Tuffin C Cary and D A Cowan ldquoMolecularadaptations to psychrophily the impact of lsquoomicrsquo technologiesrdquoTrends in Microbiology vol 18 no 8 pp 374ndash381 2010
[24] G E Schujman L Paoletti A D Grossman and D de Men-doza ldquoFapR a bacterial transcription factor involved in globalregulation ofmembrane lipid biosynthesisrdquoDevelopmental Cellvol 4 no 5 pp 663ndash672 2003
[25] E Pitkanen J Rousu and E Ukkonen ldquoComputational meth-ods for metabolic reconstructionrdquo Current Opinion in Biotech-nology vol 21 no 1 pp 70ndash77 2010
[26] H Ogata S Goto K Sato W Fujibuchi H Bono and MKanehisa ldquoKEGG kyoto encyclopedia of genes and genomesrdquoNucleic Acids Research vol 27 no 1 pp 29ndash34 1999
[27] T Carver S R Harris M Berriman J Parkhill and J AMcQuillan ldquoArtemis an integrated platform for visualizationand analysis of high-throughput sequence-based experimentaldatardquo Bioinformatics vol 28 no 4 pp 464ndash469 2012
[28] Y Moriya M Itoh S Okuda A C Yoshizawa and M Kane-hisa ldquoKAAS an automatic genome annotation and pathwayreconstruction serverrdquoNucleic Acids Research vol 35 no 2 ppW182ndashW185 2007
BioMed Research International 9
[29] R K Aziz D Bartels A Best et al ldquoThe RAST Server rapidannotations using subsystems technologyrdquo BMCGenomics vol9 article 75 2008
[30] A Bairoch ldquoThe ENZYME database in 2000rdquo Nucleic AcidsResearch vol 28 no 1 pp 304ndash305 2000
[31] R Overbeek R Olson G D Pusch et al ldquoThe SEED andthe rapid annotation of microbial genomes using SubsystemsTechnology (RAST)rdquo Nucleic Acids Research vol 42 no 1 ppD206ndashD214 2014
[32] A Flamholz E Noor A Bar-Even and R Milo ldquoEQuilibra-tormdashthe biochemical thermodynamics calculatorrdquo NucleicAcids Research vol 40 no 1 pp D770ndashD775 2012
[33] C Wrzodek A Drager and A Zell ldquoKEGGtranslator visualiz-ing and converting the KEGG PATHWAY database to variousformatsrdquo Bioinformatics vol 27 no 16 pp 2314ndash2315 2011
[34] M Hucka A Finney H M Sauro et al ldquoThe systems biologymarkup language (SBML) a medium for representation andexchange of biochemical network modelsrdquo Bioinformatics vol19 no 4 pp 524ndash531 2003
[35] M Ashburner C A Ball J A Blake et al ldquoGene ontology toolfor the unification of biologyrdquoNature Genetics vol 25 no 1 pp25ndash29 2000
[36] J J Hamilton and J L Reed ldquoSoftware platforms to facilitatereconstructing genome-scale metabolic networksrdquo Environ-mental Microbiology vol 16 no 1 pp 49ndash59 2014
[37] G J E Baart and D E Martens ldquoGenome-scale metabolicmodels reconstruction and analysisrdquo Methods in MolecularBiology vol 799 pp 107ndash126 2012
[38] V Lacroix L Cottret P Thebault and M-F Sagot ldquoAn intro-duction to metabolic networks and their structural analysisrdquoIEEEACM Transactions on Computational Biology and Bioin-formatics vol 5 no 4 pp 594ndash617 2008
[39] A-L Barabasi and Z N Oltvai ldquoNetwork biology understand-ing the cellrsquos functional organizationrdquo Nature Reviews Geneticsvol 5 no 2 pp 101ndash113 2004
[40] T Pfeiffer I Sanchez-Valdenebro J C Nuno F Montero andS Schuster ldquoMETATOOL for studying metabolic networksrdquoBioinformatics vol 15 no 3 pp 251ndash257 1999
[41] E Almaas B Kovacs T Vicsek Z NOltvai andA-L BarabasildquoGlobal organization of metabolic fluxes in the bacteriumEscherichia colirdquo Nature vol 427 no 6977 pp 839ndash843 2004
[42] J S Edwards and B Oslash Palsson ldquoThe Escherichia coli MG1655in silico metabolic genotype its definition characteristics andcapabilitiesrdquo Proceedings of the National Academy of Sciences ofthe United States of America vol 97 no 10 pp 5528ndash5533 2000
[43] I M Keseler J Collado-Vides S Gama-Castro et al ldquoEcoCyca comprehensive database resource for Escherichia colirdquoNucleicAcids Research vol 33 pp D334ndashD337 2005
[44] J D Orth T M Conrad J Na et al ldquoA comprehensivegenome-scale reconstruction of Escherichia coli metabolism-2011rdquoMolecular Systems Biology vol 7 article 535 2011
[45] A M Feist and B Oslash Palsson ldquoThe growing scope of appli-cations of genome-scale metabolic reconstructions usingEscherichia colirdquo Nature Biotechnology vol 26 no 6 pp 659ndash667 2008
[46] N C Duarte S A Becker N Jamshidi et al ldquoGlobal recon-struction of the human metabolic network based on genomicand bibliomic datardquo Proceedings of the National Academy ofSciences of the United States of America vol 104 no 6 pp 1777ndash1782 2007
[47] T Y Kim S B Sohn Y B KimW J Kim and S Y Lee ldquoRecentadvances in reconstruction and applications of genome-scalemetabolic modelsrdquo Current Opinion in Biotechnology vol 23no 4 pp 617ndash623 2012
[48] I Thiele N Swainston R M Fleming et al ldquoA community-driven global reconstruction of human metabolismrdquo NatureBiotechnology vol 31 no 5 pp 419ndash425 2013
[49] B D Heavner K Smallbone N D Price and L P WalkerldquoVersion 6 of the consensus yeast metabolic network refinesbiochemical coverage and improves model performancerdquoDatabase vol 2013 Article ID bat059 2013
[50] T Osterlund I Nookaew S Bordel and J Nielsen ldquoMap-ping condition-dependent regulation of metabolism in yeastthrough genome-scale modelingrdquo BMC Systems Biology vol 7article 36 2013
Submit your manuscripts athttpwwwhindawicom
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Anatomy Research International
PeptidesInternational Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
International Journal of
Volume 2014
Zoology
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Molecular Biology International
GenomicsInternational Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
BioinformaticsAdvances in
Marine BiologyJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Signal TransductionJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
BioMed Research International
Evolutionary BiologyInternational Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Biochemistry Research International
ArchaeaHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Genetics Research International
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Advances in
Virolog y
Hindawi Publishing Corporationhttpwwwhindawicom
Nucleic AcidsJournal of
Volume 2014
Stem CellsInternational
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Enzyme Research
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
Microbiology
2 BioMed Research International
genome annotations anddatabases which performautomaticreconstructions and thus require manual assessment Thecurrently available 96-step protocol proposed by Thiele andPalsson [15] is a well-established process for the assemblycuration and validation of metabolic reconstruction Thisprotocol is combined with computational tools includingthe visualization and numerical calculation software packageMATLAB (MathWorks USA)
Constraint-based modeling is frequently used as finalvalidation step of the reconstructed network This step isextremely useful for simulating the phenotypic behaviorunder different physiological environments [16ndash18] therebyallowing assessing if the reconstructed network representswell the in vivo cellular system Microbial adaptation to coldenvironments is one of the applications of these methods
Psychrotrophic microorganisms have an optimal growthtemperature higher than 15∘C but are also able to grow andadapt to extremely cold environments with temperatures ofapproximately 0∘C [19] Thus the unique physicochemicalcharacteristics of their habitat and the biological appara-tus developed by these microorganisms to survive underthese conditions render these organisms valuable sources ofbiotechnological processes The cellular response to cold bypsychrotrophic bacteriamay be studied from a general stand-point with the advent of omics methods Recently the B7strain ofExiguobacteriumantarcticumwas isolated fromGin-ger Lake sediments located in the Antarctic Peninsula Region(69∘301015840S 65∘W)This lake was formed due to the warming inthe region which led to partial melting of ice caps [20] Thegenome of this strain was sequenced [21] and its responseto cold was evaluated through differential expression of itsgenome at 37∘Cand0∘Cusing omicsmethods [22]One of themechanisms of cold adaptation of all psychrophilic or psy-chrotrophic organisms is the change in the chemical struc-tures of the membrane phospholipids The fatty acid chainsbecome shorter and unsaturated at low temperatures Ac-cordingly the fluidity of the membrane is kept intact [19 23]
Bacterial de novo synthesis of fatty acids is regulatedby the protein FapR [24] which is responsible for acti-vatingdisabling a regulon consisting of four operons in Eantarcticum B7 In cold two of these operons are repressed(fabH1-fabF and fabI) and the expression levels of the othertwo remain unaltered [22]The chemical components whichare included in this regulon must be reconstructed andevaluated and then associated with their respective geneticelements to further understand this metabolic pathway andto reach more complete conclusions about its importance foradaptation to cold [23 24] Bioinformatics methods are usedfor in silico reconstruction of metabolic pathways [25]
In this work we present the in silico reconstruction ofthe fatty acid biosynthesis pathway of the Exiguobacteriumantarcticum B7 based on linear programming (FBA) andconvex cone method (elementary modes) The influence oftranscriptome in FBA calculation was also evaluated
2 Materials and Methods
21 DataCollection Thegenomic data ofE antarcticumB7 inthe formats gbk and fasta were collected from NCBI under
accession number NC 018665 The metabolic pathway of thefatty acid biosynthesis was initially evaluated in the KEGG(Kyoto Encyclopedia of Genes and Genomes) database [26]When necessary the visualization of the genome of thebacterium was performed using Artemis software [27]
22 Preliminary Reconstruction This stepwas performed fol-lowing two methods one semiautomatic and the other auto-maticThe tools within the KEGGdatabases were used essen-tially in the semiautomatic method Thefasta file with theE antarcticum B7 genome was submitted to the online toolKAAS (KEGG Automatic Annotation Server) [28] availableat httpwwwgenomejpkaas-binkaas main The parame-ters chosen to run this software were as follows (a) bidirec-tional Best Hit (BBH) Method recommended for completegenomes performs the search for orthologous genes betweena specific group of organisms and (b) prokaryote the set ofgenes chosen should be representative of the target organismin this case the bacterium E antarcticum B7 Following theprocessing a text file was generated (queryko) Each line ofthis file is formed by two parameters the first consists of thesequence identified (gene) and the second when presentconsists of the KO assignment termed119870 number This valueindicates orthologous groups encoding the same enzymaticactivity Afterwards the file generated is passed through a fil-ter an auxiliary computer software program (script) Pythondeveloped for the present study which only selects 119870 num-bers and individually and increasingly commands per lineinto a new file (new queryko) This file was used as entry inthe option User Data Mapping of the Pathway Mapping toolof the KEGG
The automaticmethod essentially consisted of submittingthe fasta file of the E antarcticum B7 genome to the onlinetool RAST (RapidAnnotation using Subsystems Technology)[29] available at httprastnmpdrorg to generate the draftsof the metabolic network and of the fatty acid biosynthesispathway of the target microorganism The final draft of thisstep was generated from the combination of the resultingpathways of the semiautomatic and automatic models Thecommon pathways were maintained while surplus com-pounds enzymes and reactions that is present in some butabsent in others were not directly excluded but were insteadreserved for the curated step
23 Manual Curation The following steps were completed inthis manual curation stage following the protocol explainedabove (i) Draft refinement this phase beganwith the analysisof enzymes and reactions components of the fatty acidbiosynthesis pathway by reading books and articles specifi-cally on the subjectTheobjectivewas to diagnose the absenceor presence of more than one element of the study pathwayThe online databases KEGG ENZYME [30] and SEED [31]were consulted to ratify the enzymes and the structures of thereactions
(ii) Assessment of the stoichiometry and reversibility ofthe reactions in this step all model reactions were assessedand stoichiometrically corrected if necessary The biochemi-cal data on the organism are very important to determine thereversibility of the reaction For this purpose the databases
BioMed Research International 3
Table 1 Genes locus tags and EC numbers identified in the draft of the fatty acid biosynthesis pathway of E antarcticum B7 The featuresdisplayed were generated using the methods semiautomatic and automatic
Semiautomatic method Automatic methodGene Locus tag EC number Gene Locus tag EC numberaccA Eab7 2059 6412 accA Eab7 2059 6412accB Eab7 0870 6412 accB Eab7 0870 64114accC Eab7 0871 6412 accC Eab7 0871 6412accD Eab7 2060 6412 accD Eab7 2060 64114fabD Eab7 1760 23139 fabD Eab7 1760 23139fabH1 Eab7 1911 231180 fabH1 Eab7 1911 231180fabF Eab7 1910 231179 fabF Eab7 1910 231179fabG Eab7 1795 111100 fabG Eab7 1795 111100fabZ Eab7 2463 42159 fabI Eab7 1885 13110fabI Eab7 1885 13110fabK Eab7 0377 1319mdash Eab7 2235 114192
(KEGG SEED and ENZYME) and the tool eQuilibrator [32]were used to analyze the reversibility of the reactions Thusthe thermodynamic constraints were respected
(iii) Addition of gene data and reaction location theArtemis tool was used to identify the genes of the reactions(enzymes) from their locations in the genome assessed usingthe draft generated
(iv) Assessment of Gene-Protein-Reaction (GPR) associ-ations in this step the function of each gene is indicatedGPRassociations were identified using databases of the organismand specific literature
(v) Definition of constraints some constraints weredefined in the model in this manual curation step includingstoichiometric and thermodynamic constraints (through thereversibility and irreversibility of fluxes)
24 Metabolic Model Design The metabolic model designedand refined following themanual curation stepwas convertedinto a mathematical representation termed a stoichiometricmatrix which encouraged the development of a wide varietyof computational tools to analyze network properties
The constraints of capacity which are the upper and lowerlimits defining the maximum and minimum fluxes allowedfor the reactions were added in this step The inputs of thestoichiometricmatrix are the coefficients of themetabolites inthe reactions with negative values for consumed metabolites(substrates) and positive values when themetabolites are pro-duced or secreted (products) (Additional File 1 see Supple-mentaryMaterial available online at httpdxdoiorg10115520167863706)
25Metabolic PathwayValidation Thecomputationalmodelsought to examine the metabolic capabilities and to evaluatethe systemproperties theymayperformunder the constraintsimposed on the cellThus the final step in the reconstructionprocess consisted of assessing evaluating and validating thefatty acid biosynthesis pathway of E antarcticum B7 Thevalidation of that metabolic model was performed using sim-ulation and flux analysis The fatty acid biosynthesis pathway
is well described in the literature because it is a highlyconserved process among organisms which facilitated itscomplete definition Thus most gaps had already been filledduring the manual curation process
3 Results and Discussion
The expectation to understand the relationship between thegenome and the physiology of a particular organismwas a keyincentive for reconstructing metabolic networks Protocoladaptations using semiautomatic and automatic methods arenecessary to reconstruct themetabolic networks of organismswith few reported data on theirmetabolic capabilities includ-ing E antarcticum B7
31 Pathway Reconstruction Using the Semiautomatic MethodThe draft of the metabolic network of E antarcticum B7was retrieved from the KEGG database [26] The KEGGMetabolic Pathway tool was used to highlight the fattyacid biosynthesis pathway from the resulting draft of themetabolic networkThe genes annotated and identified usingKEGG and their respective enzymes are shown in Figure 1(a)in green and the others are listed in white boxes Table 1shows the genes locus tags and enzymes identified using theKEGG Metabolic Pathway tool A total of 11 locus tags asso-ciated with their respective genes and Enzyme Commission(EC) numbers were identified only the locus tag Eab7 2235has no added gene associated with it
32 Pathway Reconstruction Using the Automatic MethodThe RASTSEED tool does not provide graphic display ofthe metabolic map draft as KEGG for this purpose it usesa standard table to list the 247 metabolic pathways that com-pose the network regardless of whether they were identifiedin the genome of the microorganism RAST identified 25reactions 40 compounds and 20ECnumbers in the fatty acidbiosynthesis pathway ofE antarcticumB7 Table 1 outlines thegenes locus tags and enzymes identified at this step Anno-tated genes identified by RAST and their respective enzymes
4 BioMed Research International
FAS2
FASN
FAS1 Fas
FASN
Fas
FabY
Fas
FASN
FAS1 FabD
6412
Acetyl-[acp]
Acetyl-CoA
Citrate cycle
Malonyl-CoA
ACP
Pyruvate metabolism
120573-Alanine metabolism
Malonyl-[acp]
FabH
FabF
FabB
Acetoacetyl- [acp]
FASN
FAS2
Fas
FabG
(R)-3-Hydroxy-butyryl-[acp]
FASN
FAS1
Fas FabZ
FabA
But-2-enoyl-[acp]
FASN
FAS1
Fas
FabI
FabK
FabL
FabV
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
FabA
FabI
FabK
FabL
FabV
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
FabA
FabI
FabK
FabL
FabV
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
FabA
FabI
FabK
FabL
FabV
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
FabA
FabI
FabK
FabL
FabV
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
FabA
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
FabA
FabI
FabK
FabL
FabV
(R)-3-Hydroxy-hexanoyl-[acp]
trans-Hex-2-enoyl-[acp]
3-Oxooctanoyl-[acp]
(R)-3-Hydroxy-octanoyl-[acp]
trans-Oct-2-enoyl-[acp]
3-Oxodecanoyl-[acp]
(R)-3-Hydroxy-decanoyl-[acp]
trans-Dec-2-enoyl-[acp]
3-Oxododecanoyl-[acp]
(R)-3-Hydroxy-dodecanoyl-[acp]
trans-Dodec-2-enoyl-[acp]
3-Oxo-tetradecanoyl-[acp]
(R)-3-Hydroxy-tetradecanoyl-[acp]
trans-Tetradec-2-enoyl-[acp]
FabI
FabK
FabL
FabV
3-Oxohexadecanoyl-[acp]
(R)-3-Hydroxy-hexadecanoyl-[acp]
trans-Hexadec-2-enoyl-[acp]
FabF
FabB
FabG
FabZ
FabI
FabK
Fas
Fas
Fas
Fas
3-Oxooctadecanoyl-[acp]
(R)-3-Hydroxy-octadecanoyl-[acp]
trans-Octadec-2-enoyl-[acp]
Butyryl-[acp] Hexanoyl-[acp] Octanoyl-[acp]
31221
Octanoic
Lipoic acid metabolism
53314
cis-Dec-3-enoyl-[acp]
n-7 unsatured acyl-[acp]
Decanoyl-[acp]
31221
Decanoic acid
Dodecanoyl-[acp]
31221
Dodecanoic acid
Tetradecanoyl-[acp]
31214 FASN
Tetradecanoic acid
Hexadeca-noyl-[acp]
31214 FASN
Hexadecanoic acid
6213
FAS1 Fas
Fatty acid degradation
Hexadecanoyl-CoA
6213
31214
Hexadecanoic acid
31214
Octadecanoic acid
6213 Octadecanoyl-[acp]
Octadecenoyl-[acp]
31214
Octadecenoic acid
Glycerolipid metabolism
Glycerophospholipid metabolism
Fatty acid elongation
3-Oxohexanoyl-[acp]
(a)
Fas
FAS2
FASN
FAS1 FabH
FASN
Fas
FabD
6412
Acetyl-[acp]
Acetyl-CoA
Citrate cycle
Malonyl-CoA
Pyruvate metabolism
120573-Alanine metabolism
Malonyl-[acp]
FabF
FabB
Acetoacetyl- [acp]
FASN
FAS2
Fas
FabG
(R)-3-Hydroxy-butyryl-[acp]
FASN
FAS1
Fas FabZ
42158
But-2-enoyl-[acp]
FASN
FAS1
Fas
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
42158
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
FabA
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
42161
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
42161
FAS2
FASN
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
42161
3-Oxohexanoyl-[acp]
(R)-3-Hydroxy-hexanoyl-[acp]
trans-Hex-2-enoyl-[acp]
3-Oxooctanoyl-[acp]
(R)-3-Hydroxy-octanoyl-[acp]
trans-Oct-2-enoyl-[acp]
3-Oxodecanoyl-[acp]
(R)-3-Hydroxy-decanoyl-[acp]
trans-Dec-2-enoyl-[acp]
3-Oxododecanoyl-[acp]
(R)-3-Hydroxy-dodecanoyl-[acp]
trans-Dodec-2-enoyl-[acp]
3-Oxo-tetradecanoyl- [acp]
(R)-3-Hydroxy-tetradecanoyl-
[acp]
trans-Tetradec-
3-Oxohexadecanoyl-[acp]
(R)-3-Hydroxy-hexadecanoyl-[acp]
trans-Hexadec-2-enoyl-[acp]
FabF
FabB
FabG
FabZ
FabI
FabK
Fas
Fas
Fas
Fas
3-Oxooctadecanoyl-[acp]
(R)-3-Hydroxy-octadecanoyl-[acp]
trans-Octadec-2-enoyl-[acp]
Butyryl-[acp] Hexanoyl-[acp] Octanoyl-[acp]
FatB
Octanoic acid
Decanoyl-[acp]
FatB
Decanoic acid
Dodecanoyl-[acp]
31221
Dodecanoic acid
Tetradecanoyl-[acp]
31214 FASN
Tetradecanoic acid
Hexadecanoyl-[acp]
31214
FASN
Hexadecanoic acid
Fas
114192
Hexadecanoic acid
31214
Octadecanoic acid
114192
Octadecanoyl-[acp]
Octadecenoyl-[acp]
31214
Octadecenoic acid
Fas
BCCA
FASN
FAS1
63414
Carboxybiotin-carboxyl-carrier protein
Biotin-carboxyl-carrier protein
FabH
FabI
FabK
FabL
FabI
FabK
FabL
42158
FabI
FabK
FabL
FabI
FabK
FabL
FabI
FabK
FabL
2-enoyl-[acp]
FabI
FabK
FabL
FabI
FabK
FabL
Fas
FatB FatA FASN
Fatty acid metabolism
31214
ACP
(b)
Figure 1 Drafts of the fatty acid biosynthesis pathway of E antarcticumB7 bacteriaThe drafts were designed using the followingmethods (a)semiautomatic method generated by the KEGG database and (b) automatic method generated by online tool RAST Colored boxes indicatethe possibility of the presence of enzymes in the pathway
BioMed Research International 5
Table 2 Relationships between components of the E antarcticum B7 fatty acid biosynthesis pathway following curation The signalsrArr andhArr indicate irreversible and reversible reactions respectively
Gene Locus tag EC number Enzyme Reaction Fold change (log2FC)
accA Eab7 2059 6412 Acetyl-CoA carboxylase carboxyltransferase alpha subunit
ATP + acetyl-CoA + HCO3minusrArr ADP +
orthophosphate + malonyl-CoA 04562
accB Eab7 0870 6412 Acetyl-CoA carboxylasebiotin-carboxyl carrier protein
ATP + acetyl-CoA + HCO3minusrArr ADP +
orthophosphate + malonyl-CoA minus005773
accC Eab7 0871 641264114
Acetyl-CoA carboxylase biotincarboxylase subunit
ATP + acetyl-CoA + HCO3minusrArr ADP +
orthophosphate + malonyl-CoA minus05623
accD Eab7 2060 6412 Acetyl-CoA carboxylasecarboxyl transferase beta subunit
ATP + acetyl-CoA + HCO3minusrArr ADP +
orthophosphate + malonyl-CoA minus02811
fabD Eab7 1760 23139 ACP S-malonyl transferase Malonyl-CoA + ACPhArr CoA +malonyl-(acp) 09448
fabH1 Eab7 1911 231180 3-Oxoacyl-ACP synthase III Acetyl-CoA + malonyl-(acp)hArracetoacetyl-(acp) + CoA + CO
2
08512
fabF Eab7 1910 231179 3-Oxoacyl-ACP synthase II Acetyl-(acp) + malonyl-(acp)rArracetoacetyl-(acp) + CO
2+ ACP 06942
fabG Eab7 1895 111100 3-Oxoacyl-ACP reductase Acetoacetyl-(acp) + NADPH + H+ hArr(R)-3-hydroxybutanoyl-(acp) + NADP+ 08523
fabZ Eab7 2463 42159 3-Hydroxyacyl-ACP dehydratase (R)-3-hydroxybutanoyl-(acp)hArrbut-2-enoyl-(acp) + H
2O 012902
fabI Eab7 1885 131913110 Enoyl-ACP reductase I But-2-enoyl-(acp) + NADH + H+ hArr
butyryl-(acp) + NAD+ 02969
mdash Eab7 2235 114192 Acyl-ACP desaturaseHexadecanoyl-(acp) + acceptor reduced+ O2rArr hexadecenoyl-(acp) + acceptor +
2H2O
13768
are colored in purple and the others are listed in white boxes(Figure 1(b))
The analysis of both fatty acid biosynthetic pathway draftsshows that the draft generated using KEGG apparently hasthe most complete flux except for enzyme 63414 whichis exclusively present in the draft resulting from the RASTtool The draft generated using RAST has a gap in which theenzymes FabA and FabB are not included in the pathwayelongation process The flux for the production of hexade-cenoic acid is also absent from the pathway generated usingRAST
Artemis software was used to confirm the presence ofall genes selected through the automatic and semiautomaticmethodsThe genes accABCD fabD fabH1 fabF fabG fabZand fabI and the locus tag Eab7 2235 were described in thegenome of E antarcticum B7 except fabK gene which wasdetected only by the automatic method
The KGML file produced by KEGG was submitted to thesoftware KEGGtranslator [33] to be converted into a SBML(System Biology Markup Language) file [34] This file wasconverted into an Excel spreadsheet using MATLAB func-tions The files in SBML format and the Excel spreadsheetsare the most used formats in metabolic reconstructions Thereactions and metabolites of the preliminary network gener-ated using KEGG could be visualized in the spreadsheet
The data generated using the RASTSEED tool were ana-lyzed and added to the first step of the process supplementingthe data collected using KEGGThe files generated with bothplatforms were used to manage the manual curation data
The larger number of genes identified using KEGG (12)compared to those found using RASTSEED (9) may beexplained because the former uses orthology (KEGGOrthol-ogy (KO)) through protein homology to identify the so-calledmetabolite genes [35] in a genome which facilitates findinggene-protein-reaction (GPR) associations
33 Manual Curation of the Metabolic Pathway of De NovoFatty Acid Synthesis The reactions of the fatty acid synthesispathway were annotated and refined The metabolites wereorganized into two compartments (cytoplasm and extracellu-lar compartment) based on the location of the enzymes asso-ciated with each pathway The cofactors and the reversibilityof the reactions were compiled from the data published in theliterature and online tools (ENZYME and BRENDA)The ECnumberwas noted and the genes were identified A summaryof those results is shown in Table 2Thermodynamic analysisof the reactions revealed that malonyl-CoA synthesis fromacetyl-CoA (AcCoA) is an irreversible process similar tothe process regarding the fabF gene the Eab7 2235 locustag and the extracellular metabolites the other processes arereversible
It is very important to assess the quality of the annotatedgenome submitted to the online tools during curation Theliterature categorically states that the quality of the recon-structed network directly depends on the annotated genomeof the organism The rule is to use the latest updated versionof the annotated genome [36ndash38]
6 BioMed Research International
297
295
296294
184
288300
64
32432533299185
44257
7018
26186182177
52286283280278276270268
81921920
10321
165 170 173 178 181 187 27160
54758697108119130142
162 167 168 175 176 183
154
66
147
157
156
153
158
155
67
163
259 61 60 59 58 536263262
161 166 169 174136 125 114
28188180179172171159 164
51 23 22
Figure 2 Layout of the fatty acid biosynthesis pathway generated using Cytoscape Green vertices fewer connections Yellow vertices regularnumber of connections Red vertices large number of connections This network is a free model scale
34 Curated Metabolic Network The previous step addedconstraints regarding the stoichiometry (chemical balance)thermodynamics (reversibility of reactions) and physiology(cofactors used) of the study model The result was a systemof equations that describes the cell metabolism according tothe metabolites of interest The mathematical representationof this model essentially consisted of describing the perfor-mance of the fatty acid biosynthesis pathway using a stoichio-metric matrix This data structure consists of 54 metabolitesand 59 reactions resulting in a 54 times 59 stoichiometric matrix(Additional File 1) Additional File 2 shows the list of bio-chemical reactions identified in the curated model
The reconstructed pathway model was converted intoSBML in the used MATLAB toolbox The SBML file wasvalidated using the tool SBML validator and was then sub-mitted to the tool Cytoscape which generated the networkof Figure 2 The gene-protein-reaction (GPR) representationtherein describes the degree of connectivity of each enzyme inthe pathway Vertices with few connections are in green thevertices with regular numbers of connections are in yellowand the vertices with large numbers of connections are in redThe network connectivity obeys a scale-free model [39]
35 Flux Balance Analysis (FBA) The FBA was coded inMATLAB implementing a constrained linear program usingthe GLPK (GNU Linear Programming Kit) linear optimiza-tion library [8] All fluxes were calculated in percentage of
the input flux of AcCoA (reaction 39) which was fixed to100 Hexadecenoic acid is the key metabolic product thusthe respective flux (reaction 41) was set as the maximiza-tion target for FBA To improve convergence upper andlower bounds were [0 100] for irreversible reactions and[minus100 100] for reversible reactionThe final optimized fluxesare shown inAdditional File 3The targetmaximum reactionR41 in Additional File 3 was 769 which may be read as 769moles of hexadecenoic acid produced for every 100 moles ofAcCoA consumed per unit time per unit cell mass
Figure 3(a) shows the generated flux plot which showsthe variation occurring between the response fluxes with themajority approximately 37 showing positive values smallerthan 20 while 15 are above that range
36 Influence of Transcriptome in FBA Calculation Log base2-fold change values (log
2FC) obtained in vitro by DallrsquoAgnol
and colleagues [22] were used to evaluate the influence ofdifferential expression in the FBA calculation These valueswere obtained by comparison of RPKM (reads per kilobaseper million reads sequenced) generated in the transcriptomeof the bacterium at 0∘C and 37∘C The log
2FC of genes that
composes the fatty acid biosynthesis pathway is shown inTable 2
As presented by DallrsquoAgnol and colleagues [22] theaerobic energetic metabolism of E antarcticum B7 at 0∘C isrepressed and a fraction of the acetyl-CoA is probably used as
BioMed Research International 7
Reaction
FBA
minus40
minus20
0
20
40
60
80
100
Flux
val
ue (
)
lsquolsquoR1rsquorsquo
lsquolsquoR4rsquorsquo
lsquolsquoR7rsquorsquo
lsquolsquoR10
rsquorsquolsquolsquoR
13rsquorsquo
lsquolsquoR16
rsquorsquolsquolsquoR
19rsquorsquo
lsquolsquoR22
rsquorsquolsquolsquoR
25rsquorsquo
lsquolsquoR28
rsquorsquolsquolsquoR
31rsquorsquo
lsquolsquoR34
rsquorsquolsquolsquoR
37rsquorsquo
lsquolsquoR40
rsquorsquolsquolsquoR
43rsquorsquo
lsquolsquoR46
rsquorsquolsquolsquoR
49rsquorsquo
lsquolsquoR52
rsquorsquolsquolsquoR
55rsquorsquo
lsquolsquoR58
rsquorsquo
(a)
Transcriptome
lsquolsquoR1rsquorsquo
lsquolsquoR4rsquorsquo
lsquolsquoR7rsquorsquo
lsquolsquoR10
rsquorsquolsquolsquoR
13rsquorsquo
lsquolsquoR16
rsquorsquolsquolsquoR
19rsquorsquo
lsquolsquoR22
rsquorsquolsquolsquoR
25rsquorsquo
lsquolsquoR28
rsquorsquolsquolsquoR
31rsquorsquo
lsquolsquoR34
rsquorsquolsquolsquoR
37rsquorsquo
lsquolsquoR40
rsquorsquolsquolsquoR
43rsquorsquo
lsquolsquoR46
rsquorsquolsquolsquoR
49rsquorsquo
lsquolsquoR52
rsquorsquolsquolsquoR
55rsquorsquo
lsquolsquoR58
rsquorsquo
Reaction
02468
10121416
Flux
val
ue (
)
(b)
Figure 3 Flux graph generated using the software MATLAB which uses methods based on FBA constraints The vertical axis represents thefluxes calculated from the input of 100 moles of acetyl-CoA (AcCoA) The horizontal axis represents the reactions participating in the fattyacid biosynthesis pathwayThe blue bars determine the output percentage for each pathway reaction (a) FBA for maximizing the productionof hexadecenoic acid (b) FBA calculated with the log
2FC values obtained from the transcriptome
a substrate to synthesize short-chain fatty acids in coldThe synthesis begins with the conversion of acetyl-CoA tomalonyl-CoA catalyzed by the multimeric enzyme encodedby the genes accABCD In FBA analysis only half of the inputof acetyl-CoA (100mM) is converted tomalonyl-CoA whichbinds to the acyl carrier protein (ACP) at 0∘C (Figure 3(b)) Asstated earlier the other half of acetyl-CoA is probably used forenergy generation
The remaining route is cyclical being the reactionscatalyzed by enzymes encoded by fabF fabG fabZ and fabIgenes At each round two carbons are added to the growingchain of fatty acid In these reactions the flow of metabolitesremains unchanged until the fatty acid molecule reachesa size of eight carbons (octadecanoyl-ACP in reaction 38)where the percentage of the flow amount decreases (Fig-ure 3(b)) These results are consistent with the previouslypublished data which affirm that bacteria decrease their fattyacid chains to survive in coldThese short fatty acidmoleculeswill be converted into membrane phospholipids in order tomaintain the fluidity of this biological barrier
37 Elementary Flux Modes The calculation of elementarymodes was performed in MATLAB using the Metatooltoolbox [40] (modo elementarm code in the SupplementalInformation) A total of 13 elementary fluxmodes were foundfor the fatty acid biosynthesis pathway of E antarcticum B7(Additional File 4) Of these only 4 elementary modes (25 8 and 11) have a positive nonzero coefficient for reaction41 which indicates that the target product hexadecenoicacid may only be generated by one of these four elementarymodes The routes identified in EM2 and EM5 begin at thesecond reaction (R2) which is catalyzed by the enzyme FabDThevalue of R2 for both elementarymodes indicates a consid-erable production of malonyl-CoA The remaining reactionsof EM2 and EM5 are presented in Additional File 4 In EM8
and EM11 the routes begin from R1The value 1 for this reac-tion indicates a lower activity reflecting in a lower productionof hexadecenoic acid Regarding reaction 41 the values ofelementary modes 2 and 5 in this case 1 are higher thanthe values of 8 and 11 (0142857 each) indicating that bothelementary modes 2 and 5 produce 1mol hexadecenoic acidwhen they are active while elementary modes 8 and 11produce 0142857 moles
4 Conclusions
The first metabolic pathway of E antarcticum B7 recon-structed following the steps defined in this work suggeststhat the protocol used is a suitable tool for further metabolicreconstruction studies Almost all the first steps of the processwere automated however manual curation was as usuallaborious because it required an intensive search for availabledata
The metabolic pathway of fatty acid biosynthesis wasrepresentative and consistent under the limits and boundaryconditions setThe FBA and elementary mode methods wereused to examine the hexadecenoic acid production potentialof the reconstructed pathway The application of constraint-based modeling revealed being very useful to assess networkoperation plasticity even if the intracellular kinetics areunknownThe in silico analysis performed using FBA enableda quantitative assessment of hexadecenoic acid productionpotential
Finally a key issue involves deciding when to stop theprocess and to consider the reconstruction finalized at leasttemporarilyThis decision is usually based on the reconstruc-tion purpose The most complete metabolic model currentlyavailable is the E coli model which has been researchedand refined for over 10 years [41ndash45] Other studies con-stantly updating their models are Homo sapiens with three
8 BioMed Research International
reconstructions [46ndash48] and S cerevisiae with more than adozen reconstructions including two in 2013 [49 50] Theprotocol reported in the present studymay be used to compileseveral data pieces available in the literature aimed at propos-ing possible metabolic pathways thereby enabling deeperresearch of the metabolism under study
Competing Interests
The authors declare that they have no competing interests
Acknowledgments
This study was supported by the Conselho Nacional de Des-envolvimento Cientıfico e Tecnologico (CNPq) and Coor-denacao de Aperfeicoamento de Pessoal de Nıvel Superior(CAPES)
References
[1] JW Lee DNa JM Park J Lee S Choi and S Y Lee ldquoSystemsmetabolic engineering of microorganisms for natural and non-natural chemicalsrdquo Nature Chemical Biology vol 8 no 6 pp536ndash546 2012
[2] J Becker and C Wittmann ldquoSystems and synthetic metabolicengineering for amino acid productionmdashthe heartbeat of indus-trial strain developmentrdquoCurrent Opinion in Biotechnology vol23 no 5 pp 718ndash726 2012
[3] C-H ChouW-C Chang C-M Chiu C-C Huang andH-DHuang ldquoFMM a web server for metabolic pathway reconstruc-tion and comparative analysisrdquo Nucleic Acids Research vol 37no 2 pp W129ndashW134 2009
[4] P Shannon A Markiel O Ozier et al ldquoCytoscape a softwareEnvironment for integratedmodels of biomolecular interactionnetworksrdquo Genome Research vol 13 no 11 pp 2498ndash25042003
[5] A Funahashi Y Matsuoka A Jouraku M Morohashi NKikuchi and H Kitano ldquoCellDesigner 35 a versatile modelingtool for biochemical networksrdquo Proceedings of the IEEE vol 96no 8 pp 1254ndash1265 2008
[6] F T Bergmann and HM Sauro ldquoSBWmdashamodular frameworkfor systems biologyrdquo in Proceedings of the 37th Winter Simula-tion Conference (WSC rsquo06) pp 1637ndash1645 December 2006
[7] S Hoops S Sahle R Gauges et al ldquoCOPASImdasha COmplexPAthway SImulatorrdquo Bioinformatics vol 22 no 24 pp 3067ndash3074 2006
[8] J Schellenberger R Que R M T Fleming et al ldquoQuantitativeprediction of cellular metabolism with constraint-based mod-els the COBRA Toolbox v20rdquo Nature Protocols vol 6 no 9pp 1290ndash1307 2011
[9] M Terzer N D Maynard M W Covert and J StellingldquoGenome-scale metabolic networksrdquo Wiley InterdisciplinaryReviews Systems Biology andMedicine vol 1 no 3 pp 285ndash2972009
[10] DMcCloskey B Oslash Palsson andAM Feist ldquoBasic and applieduses of genome-scale metabolic network reconstructions ofEscherichia colirdquo Molecular Systems Biology vol 9 article 6612013
[11] M Fondi I Maida E Perrin et al ldquoGenome-scale metabolicreconstruction and constraint-basedmodelling of the Antarctic
bacterium Pseudoalteromonas haloplanktis TAC125rdquo Environ-mental Microbiology vol 17 no 3 pp 751ndash766 2015
[12] DMcCloskey B Oslash Palsson andAM Feist ldquoBasic and applieduses of genomeminusscale metabolic network reconstructions ofEscherichia colirdquo Molecular Systems Biology vol 9 article 6612013
[13] A M Feist M J Herrgard I Thiele J L Reed and B OslashPalsson ldquoReconstruction of biochemical networks in microor-ganismsrdquoNature ReviewsMicrobiology vol 7 no 2 pp 129ndash1432009
[14] J L Reed I Famili IThiele and B Oslash Palsson ldquoTowardsmulti-dimensional genome annotationrdquoNature Reviews Genetics vol7 no 2 pp 130ndash141 2006
[15] I Thiele and B Oslash Palsson ldquoA protocol for generating ahigh-quality genome-scale metabolic reconstructionrdquo NatureProtocols vol 5 no 1 pp 93ndash121 2010
[16] J S Edwards R U Ibarra and B O Palsson ldquoIn silico pre-dictions of Escherichia colimetabolic capabilities are consistentwith experimental datardquoNature Biotechnology vol 19 no 2 pp125ndash130 2001
[17] R Mahadevan B Oslash Palsson and D R Lovley ldquoIn situ toin silico and back elucidating the physiology and ecology ofGeobacter spp using genome-scale modellingrdquo Nature ReviewsMicrobiology vol 9 no 1 pp 39ndash50 2011
[18] J D Trawick andCH Schilling ldquoUse of constraint-basedmod-eling for the prediction and validation of antimicrobial targetsrdquoBiochemical Pharmacology vol 71 no 7 pp 1026ndash1035 2006
[19] R Y Morita ldquoPsychrophilic bacteriardquo Bacteriological Reviewsvol 39 no 2 pp 144ndash167 1975
[20] R Mulvaney N J Abram R C A Hindmarsh et al ldquoRecentAntarctic Peninsula warming relative to Holocene climate andice-shelf historyrdquo Nature vol 489 no 7414 pp 141ndash144 2012
[21] A R Carneiro R T J Ramos H DallrsquoAgnol et al ldquoGenomesequence of Exiguobacterium antarcticum B7 isolated from abiofilm inGinger Lake KingGeorge Island Antarcticardquo Journalof Bacteriology vol 194 no 23 pp 6689ndash6690 2012
[22] HDallrsquoAgnol R A Barauna P de Sa et al ldquoOmics profiles usedto evaluate the gene expression of Exiguobacterium antarcticumB7 during cold adaptationrdquo BMCGenomics vol 15 no 1 article986 2014
[23] A Casanueva M Tuffin C Cary and D A Cowan ldquoMolecularadaptations to psychrophily the impact of lsquoomicrsquo technologiesrdquoTrends in Microbiology vol 18 no 8 pp 374ndash381 2010
[24] G E Schujman L Paoletti A D Grossman and D de Men-doza ldquoFapR a bacterial transcription factor involved in globalregulation ofmembrane lipid biosynthesisrdquoDevelopmental Cellvol 4 no 5 pp 663ndash672 2003
[25] E Pitkanen J Rousu and E Ukkonen ldquoComputational meth-ods for metabolic reconstructionrdquo Current Opinion in Biotech-nology vol 21 no 1 pp 70ndash77 2010
[26] H Ogata S Goto K Sato W Fujibuchi H Bono and MKanehisa ldquoKEGG kyoto encyclopedia of genes and genomesrdquoNucleic Acids Research vol 27 no 1 pp 29ndash34 1999
[27] T Carver S R Harris M Berriman J Parkhill and J AMcQuillan ldquoArtemis an integrated platform for visualizationand analysis of high-throughput sequence-based experimentaldatardquo Bioinformatics vol 28 no 4 pp 464ndash469 2012
[28] Y Moriya M Itoh S Okuda A C Yoshizawa and M Kane-hisa ldquoKAAS an automatic genome annotation and pathwayreconstruction serverrdquoNucleic Acids Research vol 35 no 2 ppW182ndashW185 2007
BioMed Research International 9
[29] R K Aziz D Bartels A Best et al ldquoThe RAST Server rapidannotations using subsystems technologyrdquo BMCGenomics vol9 article 75 2008
[30] A Bairoch ldquoThe ENZYME database in 2000rdquo Nucleic AcidsResearch vol 28 no 1 pp 304ndash305 2000
[31] R Overbeek R Olson G D Pusch et al ldquoThe SEED andthe rapid annotation of microbial genomes using SubsystemsTechnology (RAST)rdquo Nucleic Acids Research vol 42 no 1 ppD206ndashD214 2014
[32] A Flamholz E Noor A Bar-Even and R Milo ldquoEQuilibra-tormdashthe biochemical thermodynamics calculatorrdquo NucleicAcids Research vol 40 no 1 pp D770ndashD775 2012
[33] C Wrzodek A Drager and A Zell ldquoKEGGtranslator visualiz-ing and converting the KEGG PATHWAY database to variousformatsrdquo Bioinformatics vol 27 no 16 pp 2314ndash2315 2011
[34] M Hucka A Finney H M Sauro et al ldquoThe systems biologymarkup language (SBML) a medium for representation andexchange of biochemical network modelsrdquo Bioinformatics vol19 no 4 pp 524ndash531 2003
[35] M Ashburner C A Ball J A Blake et al ldquoGene ontology toolfor the unification of biologyrdquoNature Genetics vol 25 no 1 pp25ndash29 2000
[36] J J Hamilton and J L Reed ldquoSoftware platforms to facilitatereconstructing genome-scale metabolic networksrdquo Environ-mental Microbiology vol 16 no 1 pp 49ndash59 2014
[37] G J E Baart and D E Martens ldquoGenome-scale metabolicmodels reconstruction and analysisrdquo Methods in MolecularBiology vol 799 pp 107ndash126 2012
[38] V Lacroix L Cottret P Thebault and M-F Sagot ldquoAn intro-duction to metabolic networks and their structural analysisrdquoIEEEACM Transactions on Computational Biology and Bioin-formatics vol 5 no 4 pp 594ndash617 2008
[39] A-L Barabasi and Z N Oltvai ldquoNetwork biology understand-ing the cellrsquos functional organizationrdquo Nature Reviews Geneticsvol 5 no 2 pp 101ndash113 2004
[40] T Pfeiffer I Sanchez-Valdenebro J C Nuno F Montero andS Schuster ldquoMETATOOL for studying metabolic networksrdquoBioinformatics vol 15 no 3 pp 251ndash257 1999
[41] E Almaas B Kovacs T Vicsek Z NOltvai andA-L BarabasildquoGlobal organization of metabolic fluxes in the bacteriumEscherichia colirdquo Nature vol 427 no 6977 pp 839ndash843 2004
[42] J S Edwards and B Oslash Palsson ldquoThe Escherichia coli MG1655in silico metabolic genotype its definition characteristics andcapabilitiesrdquo Proceedings of the National Academy of Sciences ofthe United States of America vol 97 no 10 pp 5528ndash5533 2000
[43] I M Keseler J Collado-Vides S Gama-Castro et al ldquoEcoCyca comprehensive database resource for Escherichia colirdquoNucleicAcids Research vol 33 pp D334ndashD337 2005
[44] J D Orth T M Conrad J Na et al ldquoA comprehensivegenome-scale reconstruction of Escherichia coli metabolism-2011rdquoMolecular Systems Biology vol 7 article 535 2011
[45] A M Feist and B Oslash Palsson ldquoThe growing scope of appli-cations of genome-scale metabolic reconstructions usingEscherichia colirdquo Nature Biotechnology vol 26 no 6 pp 659ndash667 2008
[46] N C Duarte S A Becker N Jamshidi et al ldquoGlobal recon-struction of the human metabolic network based on genomicand bibliomic datardquo Proceedings of the National Academy ofSciences of the United States of America vol 104 no 6 pp 1777ndash1782 2007
[47] T Y Kim S B Sohn Y B KimW J Kim and S Y Lee ldquoRecentadvances in reconstruction and applications of genome-scalemetabolic modelsrdquo Current Opinion in Biotechnology vol 23no 4 pp 617ndash623 2012
[48] I Thiele N Swainston R M Fleming et al ldquoA community-driven global reconstruction of human metabolismrdquo NatureBiotechnology vol 31 no 5 pp 419ndash425 2013
[49] B D Heavner K Smallbone N D Price and L P WalkerldquoVersion 6 of the consensus yeast metabolic network refinesbiochemical coverage and improves model performancerdquoDatabase vol 2013 Article ID bat059 2013
[50] T Osterlund I Nookaew S Bordel and J Nielsen ldquoMap-ping condition-dependent regulation of metabolism in yeastthrough genome-scale modelingrdquo BMC Systems Biology vol 7article 36 2013
Submit your manuscripts athttpwwwhindawicom
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Anatomy Research International
PeptidesInternational Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
International Journal of
Volume 2014
Zoology
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Molecular Biology International
GenomicsInternational Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
BioinformaticsAdvances in
Marine BiologyJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Signal TransductionJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
BioMed Research International
Evolutionary BiologyInternational Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Biochemistry Research International
ArchaeaHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Genetics Research International
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Advances in
Virolog y
Hindawi Publishing Corporationhttpwwwhindawicom
Nucleic AcidsJournal of
Volume 2014
Stem CellsInternational
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Enzyme Research
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
Microbiology
BioMed Research International 3
Table 1 Genes locus tags and EC numbers identified in the draft of the fatty acid biosynthesis pathway of E antarcticum B7 The featuresdisplayed were generated using the methods semiautomatic and automatic
Semiautomatic method Automatic methodGene Locus tag EC number Gene Locus tag EC numberaccA Eab7 2059 6412 accA Eab7 2059 6412accB Eab7 0870 6412 accB Eab7 0870 64114accC Eab7 0871 6412 accC Eab7 0871 6412accD Eab7 2060 6412 accD Eab7 2060 64114fabD Eab7 1760 23139 fabD Eab7 1760 23139fabH1 Eab7 1911 231180 fabH1 Eab7 1911 231180fabF Eab7 1910 231179 fabF Eab7 1910 231179fabG Eab7 1795 111100 fabG Eab7 1795 111100fabZ Eab7 2463 42159 fabI Eab7 1885 13110fabI Eab7 1885 13110fabK Eab7 0377 1319mdash Eab7 2235 114192
(KEGG SEED and ENZYME) and the tool eQuilibrator [32]were used to analyze the reversibility of the reactions Thusthe thermodynamic constraints were respected
(iii) Addition of gene data and reaction location theArtemis tool was used to identify the genes of the reactions(enzymes) from their locations in the genome assessed usingthe draft generated
(iv) Assessment of Gene-Protein-Reaction (GPR) associ-ations in this step the function of each gene is indicatedGPRassociations were identified using databases of the organismand specific literature
(v) Definition of constraints some constraints weredefined in the model in this manual curation step includingstoichiometric and thermodynamic constraints (through thereversibility and irreversibility of fluxes)
24 Metabolic Model Design The metabolic model designedand refined following themanual curation stepwas convertedinto a mathematical representation termed a stoichiometricmatrix which encouraged the development of a wide varietyof computational tools to analyze network properties
The constraints of capacity which are the upper and lowerlimits defining the maximum and minimum fluxes allowedfor the reactions were added in this step The inputs of thestoichiometricmatrix are the coefficients of themetabolites inthe reactions with negative values for consumed metabolites(substrates) and positive values when themetabolites are pro-duced or secreted (products) (Additional File 1 see Supple-mentaryMaterial available online at httpdxdoiorg10115520167863706)
25Metabolic PathwayValidation Thecomputationalmodelsought to examine the metabolic capabilities and to evaluatethe systemproperties theymayperformunder the constraintsimposed on the cellThus the final step in the reconstructionprocess consisted of assessing evaluating and validating thefatty acid biosynthesis pathway of E antarcticum B7 Thevalidation of that metabolic model was performed using sim-ulation and flux analysis The fatty acid biosynthesis pathway
is well described in the literature because it is a highlyconserved process among organisms which facilitated itscomplete definition Thus most gaps had already been filledduring the manual curation process
3 Results and Discussion
The expectation to understand the relationship between thegenome and the physiology of a particular organismwas a keyincentive for reconstructing metabolic networks Protocoladaptations using semiautomatic and automatic methods arenecessary to reconstruct themetabolic networks of organismswith few reported data on theirmetabolic capabilities includ-ing E antarcticum B7
31 Pathway Reconstruction Using the Semiautomatic MethodThe draft of the metabolic network of E antarcticum B7was retrieved from the KEGG database [26] The KEGGMetabolic Pathway tool was used to highlight the fattyacid biosynthesis pathway from the resulting draft of themetabolic networkThe genes annotated and identified usingKEGG and their respective enzymes are shown in Figure 1(a)in green and the others are listed in white boxes Table 1shows the genes locus tags and enzymes identified using theKEGG Metabolic Pathway tool A total of 11 locus tags asso-ciated with their respective genes and Enzyme Commission(EC) numbers were identified only the locus tag Eab7 2235has no added gene associated with it
32 Pathway Reconstruction Using the Automatic MethodThe RASTSEED tool does not provide graphic display ofthe metabolic map draft as KEGG for this purpose it usesa standard table to list the 247 metabolic pathways that com-pose the network regardless of whether they were identifiedin the genome of the microorganism RAST identified 25reactions 40 compounds and 20ECnumbers in the fatty acidbiosynthesis pathway ofE antarcticumB7 Table 1 outlines thegenes locus tags and enzymes identified at this step Anno-tated genes identified by RAST and their respective enzymes
4 BioMed Research International
FAS2
FASN
FAS1 Fas
FASN
Fas
FabY
Fas
FASN
FAS1 FabD
6412
Acetyl-[acp]
Acetyl-CoA
Citrate cycle
Malonyl-CoA
ACP
Pyruvate metabolism
120573-Alanine metabolism
Malonyl-[acp]
FabH
FabF
FabB
Acetoacetyl- [acp]
FASN
FAS2
Fas
FabG
(R)-3-Hydroxy-butyryl-[acp]
FASN
FAS1
Fas FabZ
FabA
But-2-enoyl-[acp]
FASN
FAS1
Fas
FabI
FabK
FabL
FabV
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
FabA
FabI
FabK
FabL
FabV
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
FabA
FabI
FabK
FabL
FabV
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
FabA
FabI
FabK
FabL
FabV
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
FabA
FabI
FabK
FabL
FabV
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
FabA
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
FabA
FabI
FabK
FabL
FabV
(R)-3-Hydroxy-hexanoyl-[acp]
trans-Hex-2-enoyl-[acp]
3-Oxooctanoyl-[acp]
(R)-3-Hydroxy-octanoyl-[acp]
trans-Oct-2-enoyl-[acp]
3-Oxodecanoyl-[acp]
(R)-3-Hydroxy-decanoyl-[acp]
trans-Dec-2-enoyl-[acp]
3-Oxododecanoyl-[acp]
(R)-3-Hydroxy-dodecanoyl-[acp]
trans-Dodec-2-enoyl-[acp]
3-Oxo-tetradecanoyl-[acp]
(R)-3-Hydroxy-tetradecanoyl-[acp]
trans-Tetradec-2-enoyl-[acp]
FabI
FabK
FabL
FabV
3-Oxohexadecanoyl-[acp]
(R)-3-Hydroxy-hexadecanoyl-[acp]
trans-Hexadec-2-enoyl-[acp]
FabF
FabB
FabG
FabZ
FabI
FabK
Fas
Fas
Fas
Fas
3-Oxooctadecanoyl-[acp]
(R)-3-Hydroxy-octadecanoyl-[acp]
trans-Octadec-2-enoyl-[acp]
Butyryl-[acp] Hexanoyl-[acp] Octanoyl-[acp]
31221
Octanoic
Lipoic acid metabolism
53314
cis-Dec-3-enoyl-[acp]
n-7 unsatured acyl-[acp]
Decanoyl-[acp]
31221
Decanoic acid
Dodecanoyl-[acp]
31221
Dodecanoic acid
Tetradecanoyl-[acp]
31214 FASN
Tetradecanoic acid
Hexadeca-noyl-[acp]
31214 FASN
Hexadecanoic acid
6213
FAS1 Fas
Fatty acid degradation
Hexadecanoyl-CoA
6213
31214
Hexadecanoic acid
31214
Octadecanoic acid
6213 Octadecanoyl-[acp]
Octadecenoyl-[acp]
31214
Octadecenoic acid
Glycerolipid metabolism
Glycerophospholipid metabolism
Fatty acid elongation
3-Oxohexanoyl-[acp]
(a)
Fas
FAS2
FASN
FAS1 FabH
FASN
Fas
FabD
6412
Acetyl-[acp]
Acetyl-CoA
Citrate cycle
Malonyl-CoA
Pyruvate metabolism
120573-Alanine metabolism
Malonyl-[acp]
FabF
FabB
Acetoacetyl- [acp]
FASN
FAS2
Fas
FabG
(R)-3-Hydroxy-butyryl-[acp]
FASN
FAS1
Fas FabZ
42158
But-2-enoyl-[acp]
FASN
FAS1
Fas
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
42158
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
FabA
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
42161
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
42161
FAS2
FASN
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
42161
3-Oxohexanoyl-[acp]
(R)-3-Hydroxy-hexanoyl-[acp]
trans-Hex-2-enoyl-[acp]
3-Oxooctanoyl-[acp]
(R)-3-Hydroxy-octanoyl-[acp]
trans-Oct-2-enoyl-[acp]
3-Oxodecanoyl-[acp]
(R)-3-Hydroxy-decanoyl-[acp]
trans-Dec-2-enoyl-[acp]
3-Oxododecanoyl-[acp]
(R)-3-Hydroxy-dodecanoyl-[acp]
trans-Dodec-2-enoyl-[acp]
3-Oxo-tetradecanoyl- [acp]
(R)-3-Hydroxy-tetradecanoyl-
[acp]
trans-Tetradec-
3-Oxohexadecanoyl-[acp]
(R)-3-Hydroxy-hexadecanoyl-[acp]
trans-Hexadec-2-enoyl-[acp]
FabF
FabB
FabG
FabZ
FabI
FabK
Fas
Fas
Fas
Fas
3-Oxooctadecanoyl-[acp]
(R)-3-Hydroxy-octadecanoyl-[acp]
trans-Octadec-2-enoyl-[acp]
Butyryl-[acp] Hexanoyl-[acp] Octanoyl-[acp]
FatB
Octanoic acid
Decanoyl-[acp]
FatB
Decanoic acid
Dodecanoyl-[acp]
31221
Dodecanoic acid
Tetradecanoyl-[acp]
31214 FASN
Tetradecanoic acid
Hexadecanoyl-[acp]
31214
FASN
Hexadecanoic acid
Fas
114192
Hexadecanoic acid
31214
Octadecanoic acid
114192
Octadecanoyl-[acp]
Octadecenoyl-[acp]
31214
Octadecenoic acid
Fas
BCCA
FASN
FAS1
63414
Carboxybiotin-carboxyl-carrier protein
Biotin-carboxyl-carrier protein
FabH
FabI
FabK
FabL
FabI
FabK
FabL
42158
FabI
FabK
FabL
FabI
FabK
FabL
FabI
FabK
FabL
2-enoyl-[acp]
FabI
FabK
FabL
FabI
FabK
FabL
Fas
FatB FatA FASN
Fatty acid metabolism
31214
ACP
(b)
Figure 1 Drafts of the fatty acid biosynthesis pathway of E antarcticumB7 bacteriaThe drafts were designed using the followingmethods (a)semiautomatic method generated by the KEGG database and (b) automatic method generated by online tool RAST Colored boxes indicatethe possibility of the presence of enzymes in the pathway
BioMed Research International 5
Table 2 Relationships between components of the E antarcticum B7 fatty acid biosynthesis pathway following curation The signalsrArr andhArr indicate irreversible and reversible reactions respectively
Gene Locus tag EC number Enzyme Reaction Fold change (log2FC)
accA Eab7 2059 6412 Acetyl-CoA carboxylase carboxyltransferase alpha subunit
ATP + acetyl-CoA + HCO3minusrArr ADP +
orthophosphate + malonyl-CoA 04562
accB Eab7 0870 6412 Acetyl-CoA carboxylasebiotin-carboxyl carrier protein
ATP + acetyl-CoA + HCO3minusrArr ADP +
orthophosphate + malonyl-CoA minus005773
accC Eab7 0871 641264114
Acetyl-CoA carboxylase biotincarboxylase subunit
ATP + acetyl-CoA + HCO3minusrArr ADP +
orthophosphate + malonyl-CoA minus05623
accD Eab7 2060 6412 Acetyl-CoA carboxylasecarboxyl transferase beta subunit
ATP + acetyl-CoA + HCO3minusrArr ADP +
orthophosphate + malonyl-CoA minus02811
fabD Eab7 1760 23139 ACP S-malonyl transferase Malonyl-CoA + ACPhArr CoA +malonyl-(acp) 09448
fabH1 Eab7 1911 231180 3-Oxoacyl-ACP synthase III Acetyl-CoA + malonyl-(acp)hArracetoacetyl-(acp) + CoA + CO
2
08512
fabF Eab7 1910 231179 3-Oxoacyl-ACP synthase II Acetyl-(acp) + malonyl-(acp)rArracetoacetyl-(acp) + CO
2+ ACP 06942
fabG Eab7 1895 111100 3-Oxoacyl-ACP reductase Acetoacetyl-(acp) + NADPH + H+ hArr(R)-3-hydroxybutanoyl-(acp) + NADP+ 08523
fabZ Eab7 2463 42159 3-Hydroxyacyl-ACP dehydratase (R)-3-hydroxybutanoyl-(acp)hArrbut-2-enoyl-(acp) + H
2O 012902
fabI Eab7 1885 131913110 Enoyl-ACP reductase I But-2-enoyl-(acp) + NADH + H+ hArr
butyryl-(acp) + NAD+ 02969
mdash Eab7 2235 114192 Acyl-ACP desaturaseHexadecanoyl-(acp) + acceptor reduced+ O2rArr hexadecenoyl-(acp) + acceptor +
2H2O
13768
are colored in purple and the others are listed in white boxes(Figure 1(b))
The analysis of both fatty acid biosynthetic pathway draftsshows that the draft generated using KEGG apparently hasthe most complete flux except for enzyme 63414 whichis exclusively present in the draft resulting from the RASTtool The draft generated using RAST has a gap in which theenzymes FabA and FabB are not included in the pathwayelongation process The flux for the production of hexade-cenoic acid is also absent from the pathway generated usingRAST
Artemis software was used to confirm the presence ofall genes selected through the automatic and semiautomaticmethodsThe genes accABCD fabD fabH1 fabF fabG fabZand fabI and the locus tag Eab7 2235 were described in thegenome of E antarcticum B7 except fabK gene which wasdetected only by the automatic method
The KGML file produced by KEGG was submitted to thesoftware KEGGtranslator [33] to be converted into a SBML(System Biology Markup Language) file [34] This file wasconverted into an Excel spreadsheet using MATLAB func-tions The files in SBML format and the Excel spreadsheetsare the most used formats in metabolic reconstructions Thereactions and metabolites of the preliminary network gener-ated using KEGG could be visualized in the spreadsheet
The data generated using the RASTSEED tool were ana-lyzed and added to the first step of the process supplementingthe data collected using KEGGThe files generated with bothplatforms were used to manage the manual curation data
The larger number of genes identified using KEGG (12)compared to those found using RASTSEED (9) may beexplained because the former uses orthology (KEGGOrthol-ogy (KO)) through protein homology to identify the so-calledmetabolite genes [35] in a genome which facilitates findinggene-protein-reaction (GPR) associations
33 Manual Curation of the Metabolic Pathway of De NovoFatty Acid Synthesis The reactions of the fatty acid synthesispathway were annotated and refined The metabolites wereorganized into two compartments (cytoplasm and extracellu-lar compartment) based on the location of the enzymes asso-ciated with each pathway The cofactors and the reversibilityof the reactions were compiled from the data published in theliterature and online tools (ENZYME and BRENDA)The ECnumberwas noted and the genes were identified A summaryof those results is shown in Table 2Thermodynamic analysisof the reactions revealed that malonyl-CoA synthesis fromacetyl-CoA (AcCoA) is an irreversible process similar tothe process regarding the fabF gene the Eab7 2235 locustag and the extracellular metabolites the other processes arereversible
It is very important to assess the quality of the annotatedgenome submitted to the online tools during curation Theliterature categorically states that the quality of the recon-structed network directly depends on the annotated genomeof the organism The rule is to use the latest updated versionof the annotated genome [36ndash38]
6 BioMed Research International
297
295
296294
184
288300
64
32432533299185
44257
7018
26186182177
52286283280278276270268
81921920
10321
165 170 173 178 181 187 27160
54758697108119130142
162 167 168 175 176 183
154
66
147
157
156
153
158
155
67
163
259 61 60 59 58 536263262
161 166 169 174136 125 114
28188180179172171159 164
51 23 22
Figure 2 Layout of the fatty acid biosynthesis pathway generated using Cytoscape Green vertices fewer connections Yellow vertices regularnumber of connections Red vertices large number of connections This network is a free model scale
34 Curated Metabolic Network The previous step addedconstraints regarding the stoichiometry (chemical balance)thermodynamics (reversibility of reactions) and physiology(cofactors used) of the study model The result was a systemof equations that describes the cell metabolism according tothe metabolites of interest The mathematical representationof this model essentially consisted of describing the perfor-mance of the fatty acid biosynthesis pathway using a stoichio-metric matrix This data structure consists of 54 metabolitesand 59 reactions resulting in a 54 times 59 stoichiometric matrix(Additional File 1) Additional File 2 shows the list of bio-chemical reactions identified in the curated model
The reconstructed pathway model was converted intoSBML in the used MATLAB toolbox The SBML file wasvalidated using the tool SBML validator and was then sub-mitted to the tool Cytoscape which generated the networkof Figure 2 The gene-protein-reaction (GPR) representationtherein describes the degree of connectivity of each enzyme inthe pathway Vertices with few connections are in green thevertices with regular numbers of connections are in yellowand the vertices with large numbers of connections are in redThe network connectivity obeys a scale-free model [39]
35 Flux Balance Analysis (FBA) The FBA was coded inMATLAB implementing a constrained linear program usingthe GLPK (GNU Linear Programming Kit) linear optimiza-tion library [8] All fluxes were calculated in percentage of
the input flux of AcCoA (reaction 39) which was fixed to100 Hexadecenoic acid is the key metabolic product thusthe respective flux (reaction 41) was set as the maximiza-tion target for FBA To improve convergence upper andlower bounds were [0 100] for irreversible reactions and[minus100 100] for reversible reactionThe final optimized fluxesare shown inAdditional File 3The targetmaximum reactionR41 in Additional File 3 was 769 which may be read as 769moles of hexadecenoic acid produced for every 100 moles ofAcCoA consumed per unit time per unit cell mass
Figure 3(a) shows the generated flux plot which showsthe variation occurring between the response fluxes with themajority approximately 37 showing positive values smallerthan 20 while 15 are above that range
36 Influence of Transcriptome in FBA Calculation Log base2-fold change values (log
2FC) obtained in vitro by DallrsquoAgnol
and colleagues [22] were used to evaluate the influence ofdifferential expression in the FBA calculation These valueswere obtained by comparison of RPKM (reads per kilobaseper million reads sequenced) generated in the transcriptomeof the bacterium at 0∘C and 37∘C The log
2FC of genes that
composes the fatty acid biosynthesis pathway is shown inTable 2
As presented by DallrsquoAgnol and colleagues [22] theaerobic energetic metabolism of E antarcticum B7 at 0∘C isrepressed and a fraction of the acetyl-CoA is probably used as
BioMed Research International 7
Reaction
FBA
minus40
minus20
0
20
40
60
80
100
Flux
val
ue (
)
lsquolsquoR1rsquorsquo
lsquolsquoR4rsquorsquo
lsquolsquoR7rsquorsquo
lsquolsquoR10
rsquorsquolsquolsquoR
13rsquorsquo
lsquolsquoR16
rsquorsquolsquolsquoR
19rsquorsquo
lsquolsquoR22
rsquorsquolsquolsquoR
25rsquorsquo
lsquolsquoR28
rsquorsquolsquolsquoR
31rsquorsquo
lsquolsquoR34
rsquorsquolsquolsquoR
37rsquorsquo
lsquolsquoR40
rsquorsquolsquolsquoR
43rsquorsquo
lsquolsquoR46
rsquorsquolsquolsquoR
49rsquorsquo
lsquolsquoR52
rsquorsquolsquolsquoR
55rsquorsquo
lsquolsquoR58
rsquorsquo
(a)
Transcriptome
lsquolsquoR1rsquorsquo
lsquolsquoR4rsquorsquo
lsquolsquoR7rsquorsquo
lsquolsquoR10
rsquorsquolsquolsquoR
13rsquorsquo
lsquolsquoR16
rsquorsquolsquolsquoR
19rsquorsquo
lsquolsquoR22
rsquorsquolsquolsquoR
25rsquorsquo
lsquolsquoR28
rsquorsquolsquolsquoR
31rsquorsquo
lsquolsquoR34
rsquorsquolsquolsquoR
37rsquorsquo
lsquolsquoR40
rsquorsquolsquolsquoR
43rsquorsquo
lsquolsquoR46
rsquorsquolsquolsquoR
49rsquorsquo
lsquolsquoR52
rsquorsquolsquolsquoR
55rsquorsquo
lsquolsquoR58
rsquorsquo
Reaction
02468
10121416
Flux
val
ue (
)
(b)
Figure 3 Flux graph generated using the software MATLAB which uses methods based on FBA constraints The vertical axis represents thefluxes calculated from the input of 100 moles of acetyl-CoA (AcCoA) The horizontal axis represents the reactions participating in the fattyacid biosynthesis pathwayThe blue bars determine the output percentage for each pathway reaction (a) FBA for maximizing the productionof hexadecenoic acid (b) FBA calculated with the log
2FC values obtained from the transcriptome
a substrate to synthesize short-chain fatty acids in coldThe synthesis begins with the conversion of acetyl-CoA tomalonyl-CoA catalyzed by the multimeric enzyme encodedby the genes accABCD In FBA analysis only half of the inputof acetyl-CoA (100mM) is converted tomalonyl-CoA whichbinds to the acyl carrier protein (ACP) at 0∘C (Figure 3(b)) Asstated earlier the other half of acetyl-CoA is probably used forenergy generation
The remaining route is cyclical being the reactionscatalyzed by enzymes encoded by fabF fabG fabZ and fabIgenes At each round two carbons are added to the growingchain of fatty acid In these reactions the flow of metabolitesremains unchanged until the fatty acid molecule reachesa size of eight carbons (octadecanoyl-ACP in reaction 38)where the percentage of the flow amount decreases (Fig-ure 3(b)) These results are consistent with the previouslypublished data which affirm that bacteria decrease their fattyacid chains to survive in coldThese short fatty acidmoleculeswill be converted into membrane phospholipids in order tomaintain the fluidity of this biological barrier
37 Elementary Flux Modes The calculation of elementarymodes was performed in MATLAB using the Metatooltoolbox [40] (modo elementarm code in the SupplementalInformation) A total of 13 elementary fluxmodes were foundfor the fatty acid biosynthesis pathway of E antarcticum B7(Additional File 4) Of these only 4 elementary modes (25 8 and 11) have a positive nonzero coefficient for reaction41 which indicates that the target product hexadecenoicacid may only be generated by one of these four elementarymodes The routes identified in EM2 and EM5 begin at thesecond reaction (R2) which is catalyzed by the enzyme FabDThevalue of R2 for both elementarymodes indicates a consid-erable production of malonyl-CoA The remaining reactionsof EM2 and EM5 are presented in Additional File 4 In EM8
and EM11 the routes begin from R1The value 1 for this reac-tion indicates a lower activity reflecting in a lower productionof hexadecenoic acid Regarding reaction 41 the values ofelementary modes 2 and 5 in this case 1 are higher thanthe values of 8 and 11 (0142857 each) indicating that bothelementary modes 2 and 5 produce 1mol hexadecenoic acidwhen they are active while elementary modes 8 and 11produce 0142857 moles
4 Conclusions
The first metabolic pathway of E antarcticum B7 recon-structed following the steps defined in this work suggeststhat the protocol used is a suitable tool for further metabolicreconstruction studies Almost all the first steps of the processwere automated however manual curation was as usuallaborious because it required an intensive search for availabledata
The metabolic pathway of fatty acid biosynthesis wasrepresentative and consistent under the limits and boundaryconditions setThe FBA and elementary mode methods wereused to examine the hexadecenoic acid production potentialof the reconstructed pathway The application of constraint-based modeling revealed being very useful to assess networkoperation plasticity even if the intracellular kinetics areunknownThe in silico analysis performed using FBA enableda quantitative assessment of hexadecenoic acid productionpotential
Finally a key issue involves deciding when to stop theprocess and to consider the reconstruction finalized at leasttemporarilyThis decision is usually based on the reconstruc-tion purpose The most complete metabolic model currentlyavailable is the E coli model which has been researchedand refined for over 10 years [41ndash45] Other studies con-stantly updating their models are Homo sapiens with three
8 BioMed Research International
reconstructions [46ndash48] and S cerevisiae with more than adozen reconstructions including two in 2013 [49 50] Theprotocol reported in the present studymay be used to compileseveral data pieces available in the literature aimed at propos-ing possible metabolic pathways thereby enabling deeperresearch of the metabolism under study
Competing Interests
The authors declare that they have no competing interests
Acknowledgments
This study was supported by the Conselho Nacional de Des-envolvimento Cientıfico e Tecnologico (CNPq) and Coor-denacao de Aperfeicoamento de Pessoal de Nıvel Superior(CAPES)
References
[1] JW Lee DNa JM Park J Lee S Choi and S Y Lee ldquoSystemsmetabolic engineering of microorganisms for natural and non-natural chemicalsrdquo Nature Chemical Biology vol 8 no 6 pp536ndash546 2012
[2] J Becker and C Wittmann ldquoSystems and synthetic metabolicengineering for amino acid productionmdashthe heartbeat of indus-trial strain developmentrdquoCurrent Opinion in Biotechnology vol23 no 5 pp 718ndash726 2012
[3] C-H ChouW-C Chang C-M Chiu C-C Huang andH-DHuang ldquoFMM a web server for metabolic pathway reconstruc-tion and comparative analysisrdquo Nucleic Acids Research vol 37no 2 pp W129ndashW134 2009
[4] P Shannon A Markiel O Ozier et al ldquoCytoscape a softwareEnvironment for integratedmodels of biomolecular interactionnetworksrdquo Genome Research vol 13 no 11 pp 2498ndash25042003
[5] A Funahashi Y Matsuoka A Jouraku M Morohashi NKikuchi and H Kitano ldquoCellDesigner 35 a versatile modelingtool for biochemical networksrdquo Proceedings of the IEEE vol 96no 8 pp 1254ndash1265 2008
[6] F T Bergmann and HM Sauro ldquoSBWmdashamodular frameworkfor systems biologyrdquo in Proceedings of the 37th Winter Simula-tion Conference (WSC rsquo06) pp 1637ndash1645 December 2006
[7] S Hoops S Sahle R Gauges et al ldquoCOPASImdasha COmplexPAthway SImulatorrdquo Bioinformatics vol 22 no 24 pp 3067ndash3074 2006
[8] J Schellenberger R Que R M T Fleming et al ldquoQuantitativeprediction of cellular metabolism with constraint-based mod-els the COBRA Toolbox v20rdquo Nature Protocols vol 6 no 9pp 1290ndash1307 2011
[9] M Terzer N D Maynard M W Covert and J StellingldquoGenome-scale metabolic networksrdquo Wiley InterdisciplinaryReviews Systems Biology andMedicine vol 1 no 3 pp 285ndash2972009
[10] DMcCloskey B Oslash Palsson andAM Feist ldquoBasic and applieduses of genome-scale metabolic network reconstructions ofEscherichia colirdquo Molecular Systems Biology vol 9 article 6612013
[11] M Fondi I Maida E Perrin et al ldquoGenome-scale metabolicreconstruction and constraint-basedmodelling of the Antarctic
bacterium Pseudoalteromonas haloplanktis TAC125rdquo Environ-mental Microbiology vol 17 no 3 pp 751ndash766 2015
[12] DMcCloskey B Oslash Palsson andAM Feist ldquoBasic and applieduses of genomeminusscale metabolic network reconstructions ofEscherichia colirdquo Molecular Systems Biology vol 9 article 6612013
[13] A M Feist M J Herrgard I Thiele J L Reed and B OslashPalsson ldquoReconstruction of biochemical networks in microor-ganismsrdquoNature ReviewsMicrobiology vol 7 no 2 pp 129ndash1432009
[14] J L Reed I Famili IThiele and B Oslash Palsson ldquoTowardsmulti-dimensional genome annotationrdquoNature Reviews Genetics vol7 no 2 pp 130ndash141 2006
[15] I Thiele and B Oslash Palsson ldquoA protocol for generating ahigh-quality genome-scale metabolic reconstructionrdquo NatureProtocols vol 5 no 1 pp 93ndash121 2010
[16] J S Edwards R U Ibarra and B O Palsson ldquoIn silico pre-dictions of Escherichia colimetabolic capabilities are consistentwith experimental datardquoNature Biotechnology vol 19 no 2 pp125ndash130 2001
[17] R Mahadevan B Oslash Palsson and D R Lovley ldquoIn situ toin silico and back elucidating the physiology and ecology ofGeobacter spp using genome-scale modellingrdquo Nature ReviewsMicrobiology vol 9 no 1 pp 39ndash50 2011
[18] J D Trawick andCH Schilling ldquoUse of constraint-basedmod-eling for the prediction and validation of antimicrobial targetsrdquoBiochemical Pharmacology vol 71 no 7 pp 1026ndash1035 2006
[19] R Y Morita ldquoPsychrophilic bacteriardquo Bacteriological Reviewsvol 39 no 2 pp 144ndash167 1975
[20] R Mulvaney N J Abram R C A Hindmarsh et al ldquoRecentAntarctic Peninsula warming relative to Holocene climate andice-shelf historyrdquo Nature vol 489 no 7414 pp 141ndash144 2012
[21] A R Carneiro R T J Ramos H DallrsquoAgnol et al ldquoGenomesequence of Exiguobacterium antarcticum B7 isolated from abiofilm inGinger Lake KingGeorge Island Antarcticardquo Journalof Bacteriology vol 194 no 23 pp 6689ndash6690 2012
[22] HDallrsquoAgnol R A Barauna P de Sa et al ldquoOmics profiles usedto evaluate the gene expression of Exiguobacterium antarcticumB7 during cold adaptationrdquo BMCGenomics vol 15 no 1 article986 2014
[23] A Casanueva M Tuffin C Cary and D A Cowan ldquoMolecularadaptations to psychrophily the impact of lsquoomicrsquo technologiesrdquoTrends in Microbiology vol 18 no 8 pp 374ndash381 2010
[24] G E Schujman L Paoletti A D Grossman and D de Men-doza ldquoFapR a bacterial transcription factor involved in globalregulation ofmembrane lipid biosynthesisrdquoDevelopmental Cellvol 4 no 5 pp 663ndash672 2003
[25] E Pitkanen J Rousu and E Ukkonen ldquoComputational meth-ods for metabolic reconstructionrdquo Current Opinion in Biotech-nology vol 21 no 1 pp 70ndash77 2010
[26] H Ogata S Goto K Sato W Fujibuchi H Bono and MKanehisa ldquoKEGG kyoto encyclopedia of genes and genomesrdquoNucleic Acids Research vol 27 no 1 pp 29ndash34 1999
[27] T Carver S R Harris M Berriman J Parkhill and J AMcQuillan ldquoArtemis an integrated platform for visualizationand analysis of high-throughput sequence-based experimentaldatardquo Bioinformatics vol 28 no 4 pp 464ndash469 2012
[28] Y Moriya M Itoh S Okuda A C Yoshizawa and M Kane-hisa ldquoKAAS an automatic genome annotation and pathwayreconstruction serverrdquoNucleic Acids Research vol 35 no 2 ppW182ndashW185 2007
BioMed Research International 9
[29] R K Aziz D Bartels A Best et al ldquoThe RAST Server rapidannotations using subsystems technologyrdquo BMCGenomics vol9 article 75 2008
[30] A Bairoch ldquoThe ENZYME database in 2000rdquo Nucleic AcidsResearch vol 28 no 1 pp 304ndash305 2000
[31] R Overbeek R Olson G D Pusch et al ldquoThe SEED andthe rapid annotation of microbial genomes using SubsystemsTechnology (RAST)rdquo Nucleic Acids Research vol 42 no 1 ppD206ndashD214 2014
[32] A Flamholz E Noor A Bar-Even and R Milo ldquoEQuilibra-tormdashthe biochemical thermodynamics calculatorrdquo NucleicAcids Research vol 40 no 1 pp D770ndashD775 2012
[33] C Wrzodek A Drager and A Zell ldquoKEGGtranslator visualiz-ing and converting the KEGG PATHWAY database to variousformatsrdquo Bioinformatics vol 27 no 16 pp 2314ndash2315 2011
[34] M Hucka A Finney H M Sauro et al ldquoThe systems biologymarkup language (SBML) a medium for representation andexchange of biochemical network modelsrdquo Bioinformatics vol19 no 4 pp 524ndash531 2003
[35] M Ashburner C A Ball J A Blake et al ldquoGene ontology toolfor the unification of biologyrdquoNature Genetics vol 25 no 1 pp25ndash29 2000
[36] J J Hamilton and J L Reed ldquoSoftware platforms to facilitatereconstructing genome-scale metabolic networksrdquo Environ-mental Microbiology vol 16 no 1 pp 49ndash59 2014
[37] G J E Baart and D E Martens ldquoGenome-scale metabolicmodels reconstruction and analysisrdquo Methods in MolecularBiology vol 799 pp 107ndash126 2012
[38] V Lacroix L Cottret P Thebault and M-F Sagot ldquoAn intro-duction to metabolic networks and their structural analysisrdquoIEEEACM Transactions on Computational Biology and Bioin-formatics vol 5 no 4 pp 594ndash617 2008
[39] A-L Barabasi and Z N Oltvai ldquoNetwork biology understand-ing the cellrsquos functional organizationrdquo Nature Reviews Geneticsvol 5 no 2 pp 101ndash113 2004
[40] T Pfeiffer I Sanchez-Valdenebro J C Nuno F Montero andS Schuster ldquoMETATOOL for studying metabolic networksrdquoBioinformatics vol 15 no 3 pp 251ndash257 1999
[41] E Almaas B Kovacs T Vicsek Z NOltvai andA-L BarabasildquoGlobal organization of metabolic fluxes in the bacteriumEscherichia colirdquo Nature vol 427 no 6977 pp 839ndash843 2004
[42] J S Edwards and B Oslash Palsson ldquoThe Escherichia coli MG1655in silico metabolic genotype its definition characteristics andcapabilitiesrdquo Proceedings of the National Academy of Sciences ofthe United States of America vol 97 no 10 pp 5528ndash5533 2000
[43] I M Keseler J Collado-Vides S Gama-Castro et al ldquoEcoCyca comprehensive database resource for Escherichia colirdquoNucleicAcids Research vol 33 pp D334ndashD337 2005
[44] J D Orth T M Conrad J Na et al ldquoA comprehensivegenome-scale reconstruction of Escherichia coli metabolism-2011rdquoMolecular Systems Biology vol 7 article 535 2011
[45] A M Feist and B Oslash Palsson ldquoThe growing scope of appli-cations of genome-scale metabolic reconstructions usingEscherichia colirdquo Nature Biotechnology vol 26 no 6 pp 659ndash667 2008
[46] N C Duarte S A Becker N Jamshidi et al ldquoGlobal recon-struction of the human metabolic network based on genomicand bibliomic datardquo Proceedings of the National Academy ofSciences of the United States of America vol 104 no 6 pp 1777ndash1782 2007
[47] T Y Kim S B Sohn Y B KimW J Kim and S Y Lee ldquoRecentadvances in reconstruction and applications of genome-scalemetabolic modelsrdquo Current Opinion in Biotechnology vol 23no 4 pp 617ndash623 2012
[48] I Thiele N Swainston R M Fleming et al ldquoA community-driven global reconstruction of human metabolismrdquo NatureBiotechnology vol 31 no 5 pp 419ndash425 2013
[49] B D Heavner K Smallbone N D Price and L P WalkerldquoVersion 6 of the consensus yeast metabolic network refinesbiochemical coverage and improves model performancerdquoDatabase vol 2013 Article ID bat059 2013
[50] T Osterlund I Nookaew S Bordel and J Nielsen ldquoMap-ping condition-dependent regulation of metabolism in yeastthrough genome-scale modelingrdquo BMC Systems Biology vol 7article 36 2013
Submit your manuscripts athttpwwwhindawicom
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Anatomy Research International
PeptidesInternational Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
International Journal of
Volume 2014
Zoology
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Molecular Biology International
GenomicsInternational Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
BioinformaticsAdvances in
Marine BiologyJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Signal TransductionJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
BioMed Research International
Evolutionary BiologyInternational Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Biochemistry Research International
ArchaeaHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Genetics Research International
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Advances in
Virolog y
Hindawi Publishing Corporationhttpwwwhindawicom
Nucleic AcidsJournal of
Volume 2014
Stem CellsInternational
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Enzyme Research
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
Microbiology
4 BioMed Research International
FAS2
FASN
FAS1 Fas
FASN
Fas
FabY
Fas
FASN
FAS1 FabD
6412
Acetyl-[acp]
Acetyl-CoA
Citrate cycle
Malonyl-CoA
ACP
Pyruvate metabolism
120573-Alanine metabolism
Malonyl-[acp]
FabH
FabF
FabB
Acetoacetyl- [acp]
FASN
FAS2
Fas
FabG
(R)-3-Hydroxy-butyryl-[acp]
FASN
FAS1
Fas FabZ
FabA
But-2-enoyl-[acp]
FASN
FAS1
Fas
FabI
FabK
FabL
FabV
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
FabA
FabI
FabK
FabL
FabV
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
FabA
FabI
FabK
FabL
FabV
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
FabA
FabI
FabK
FabL
FabV
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
FabA
FabI
FabK
FabL
FabV
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
FabA
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
FabA
FabI
FabK
FabL
FabV
(R)-3-Hydroxy-hexanoyl-[acp]
trans-Hex-2-enoyl-[acp]
3-Oxooctanoyl-[acp]
(R)-3-Hydroxy-octanoyl-[acp]
trans-Oct-2-enoyl-[acp]
3-Oxodecanoyl-[acp]
(R)-3-Hydroxy-decanoyl-[acp]
trans-Dec-2-enoyl-[acp]
3-Oxododecanoyl-[acp]
(R)-3-Hydroxy-dodecanoyl-[acp]
trans-Dodec-2-enoyl-[acp]
3-Oxo-tetradecanoyl-[acp]
(R)-3-Hydroxy-tetradecanoyl-[acp]
trans-Tetradec-2-enoyl-[acp]
FabI
FabK
FabL
FabV
3-Oxohexadecanoyl-[acp]
(R)-3-Hydroxy-hexadecanoyl-[acp]
trans-Hexadec-2-enoyl-[acp]
FabF
FabB
FabG
FabZ
FabI
FabK
Fas
Fas
Fas
Fas
3-Oxooctadecanoyl-[acp]
(R)-3-Hydroxy-octadecanoyl-[acp]
trans-Octadec-2-enoyl-[acp]
Butyryl-[acp] Hexanoyl-[acp] Octanoyl-[acp]
31221
Octanoic
Lipoic acid metabolism
53314
cis-Dec-3-enoyl-[acp]
n-7 unsatured acyl-[acp]
Decanoyl-[acp]
31221
Decanoic acid
Dodecanoyl-[acp]
31221
Dodecanoic acid
Tetradecanoyl-[acp]
31214 FASN
Tetradecanoic acid
Hexadeca-noyl-[acp]
31214 FASN
Hexadecanoic acid
6213
FAS1 Fas
Fatty acid degradation
Hexadecanoyl-CoA
6213
31214
Hexadecanoic acid
31214
Octadecanoic acid
6213 Octadecanoyl-[acp]
Octadecenoyl-[acp]
31214
Octadecenoic acid
Glycerolipid metabolism
Glycerophospholipid metabolism
Fatty acid elongation
3-Oxohexanoyl-[acp]
(a)
Fas
FAS2
FASN
FAS1 FabH
FASN
Fas
FabD
6412
Acetyl-[acp]
Acetyl-CoA
Citrate cycle
Malonyl-CoA
Pyruvate metabolism
120573-Alanine metabolism
Malonyl-[acp]
FabF
FabB
Acetoacetyl- [acp]
FASN
FAS2
Fas
FabG
(R)-3-Hydroxy-butyryl-[acp]
FASN
FAS1
Fas FabZ
42158
But-2-enoyl-[acp]
FASN
FAS1
Fas
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
42158
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
FabA
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
42161
FAS2
FASN
Fas
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
42161
FAS2
FASN
FASN
FAS2
Fas
FASN
FAS1
Fas
FASN
FAS1
Fas
FabF
FabB
FabG
FabZ
42161
3-Oxohexanoyl-[acp]
(R)-3-Hydroxy-hexanoyl-[acp]
trans-Hex-2-enoyl-[acp]
3-Oxooctanoyl-[acp]
(R)-3-Hydroxy-octanoyl-[acp]
trans-Oct-2-enoyl-[acp]
3-Oxodecanoyl-[acp]
(R)-3-Hydroxy-decanoyl-[acp]
trans-Dec-2-enoyl-[acp]
3-Oxododecanoyl-[acp]
(R)-3-Hydroxy-dodecanoyl-[acp]
trans-Dodec-2-enoyl-[acp]
3-Oxo-tetradecanoyl- [acp]
(R)-3-Hydroxy-tetradecanoyl-
[acp]
trans-Tetradec-
3-Oxohexadecanoyl-[acp]
(R)-3-Hydroxy-hexadecanoyl-[acp]
trans-Hexadec-2-enoyl-[acp]
FabF
FabB
FabG
FabZ
FabI
FabK
Fas
Fas
Fas
Fas
3-Oxooctadecanoyl-[acp]
(R)-3-Hydroxy-octadecanoyl-[acp]
trans-Octadec-2-enoyl-[acp]
Butyryl-[acp] Hexanoyl-[acp] Octanoyl-[acp]
FatB
Octanoic acid
Decanoyl-[acp]
FatB
Decanoic acid
Dodecanoyl-[acp]
31221
Dodecanoic acid
Tetradecanoyl-[acp]
31214 FASN
Tetradecanoic acid
Hexadecanoyl-[acp]
31214
FASN
Hexadecanoic acid
Fas
114192
Hexadecanoic acid
31214
Octadecanoic acid
114192
Octadecanoyl-[acp]
Octadecenoyl-[acp]
31214
Octadecenoic acid
Fas
BCCA
FASN
FAS1
63414
Carboxybiotin-carboxyl-carrier protein
Biotin-carboxyl-carrier protein
FabH
FabI
FabK
FabL
FabI
FabK
FabL
42158
FabI
FabK
FabL
FabI
FabK
FabL
FabI
FabK
FabL
2-enoyl-[acp]
FabI
FabK
FabL
FabI
FabK
FabL
Fas
FatB FatA FASN
Fatty acid metabolism
31214
ACP
(b)
Figure 1 Drafts of the fatty acid biosynthesis pathway of E antarcticumB7 bacteriaThe drafts were designed using the followingmethods (a)semiautomatic method generated by the KEGG database and (b) automatic method generated by online tool RAST Colored boxes indicatethe possibility of the presence of enzymes in the pathway
BioMed Research International 5
Table 2 Relationships between components of the E antarcticum B7 fatty acid biosynthesis pathway following curation The signalsrArr andhArr indicate irreversible and reversible reactions respectively
Gene Locus tag EC number Enzyme Reaction Fold change (log2FC)
accA Eab7 2059 6412 Acetyl-CoA carboxylase carboxyltransferase alpha subunit
ATP + acetyl-CoA + HCO3minusrArr ADP +
orthophosphate + malonyl-CoA 04562
accB Eab7 0870 6412 Acetyl-CoA carboxylasebiotin-carboxyl carrier protein
ATP + acetyl-CoA + HCO3minusrArr ADP +
orthophosphate + malonyl-CoA minus005773
accC Eab7 0871 641264114
Acetyl-CoA carboxylase biotincarboxylase subunit
ATP + acetyl-CoA + HCO3minusrArr ADP +
orthophosphate + malonyl-CoA minus05623
accD Eab7 2060 6412 Acetyl-CoA carboxylasecarboxyl transferase beta subunit
ATP + acetyl-CoA + HCO3minusrArr ADP +
orthophosphate + malonyl-CoA minus02811
fabD Eab7 1760 23139 ACP S-malonyl transferase Malonyl-CoA + ACPhArr CoA +malonyl-(acp) 09448
fabH1 Eab7 1911 231180 3-Oxoacyl-ACP synthase III Acetyl-CoA + malonyl-(acp)hArracetoacetyl-(acp) + CoA + CO
2
08512
fabF Eab7 1910 231179 3-Oxoacyl-ACP synthase II Acetyl-(acp) + malonyl-(acp)rArracetoacetyl-(acp) + CO
2+ ACP 06942
fabG Eab7 1895 111100 3-Oxoacyl-ACP reductase Acetoacetyl-(acp) + NADPH + H+ hArr(R)-3-hydroxybutanoyl-(acp) + NADP+ 08523
fabZ Eab7 2463 42159 3-Hydroxyacyl-ACP dehydratase (R)-3-hydroxybutanoyl-(acp)hArrbut-2-enoyl-(acp) + H
2O 012902
fabI Eab7 1885 131913110 Enoyl-ACP reductase I But-2-enoyl-(acp) + NADH + H+ hArr
butyryl-(acp) + NAD+ 02969
mdash Eab7 2235 114192 Acyl-ACP desaturaseHexadecanoyl-(acp) + acceptor reduced+ O2rArr hexadecenoyl-(acp) + acceptor +
2H2O
13768
are colored in purple and the others are listed in white boxes(Figure 1(b))
The analysis of both fatty acid biosynthetic pathway draftsshows that the draft generated using KEGG apparently hasthe most complete flux except for enzyme 63414 whichis exclusively present in the draft resulting from the RASTtool The draft generated using RAST has a gap in which theenzymes FabA and FabB are not included in the pathwayelongation process The flux for the production of hexade-cenoic acid is also absent from the pathway generated usingRAST
Artemis software was used to confirm the presence ofall genes selected through the automatic and semiautomaticmethodsThe genes accABCD fabD fabH1 fabF fabG fabZand fabI and the locus tag Eab7 2235 were described in thegenome of E antarcticum B7 except fabK gene which wasdetected only by the automatic method
The KGML file produced by KEGG was submitted to thesoftware KEGGtranslator [33] to be converted into a SBML(System Biology Markup Language) file [34] This file wasconverted into an Excel spreadsheet using MATLAB func-tions The files in SBML format and the Excel spreadsheetsare the most used formats in metabolic reconstructions Thereactions and metabolites of the preliminary network gener-ated using KEGG could be visualized in the spreadsheet
The data generated using the RASTSEED tool were ana-lyzed and added to the first step of the process supplementingthe data collected using KEGGThe files generated with bothplatforms were used to manage the manual curation data
The larger number of genes identified using KEGG (12)compared to those found using RASTSEED (9) may beexplained because the former uses orthology (KEGGOrthol-ogy (KO)) through protein homology to identify the so-calledmetabolite genes [35] in a genome which facilitates findinggene-protein-reaction (GPR) associations
33 Manual Curation of the Metabolic Pathway of De NovoFatty Acid Synthesis The reactions of the fatty acid synthesispathway were annotated and refined The metabolites wereorganized into two compartments (cytoplasm and extracellu-lar compartment) based on the location of the enzymes asso-ciated with each pathway The cofactors and the reversibilityof the reactions were compiled from the data published in theliterature and online tools (ENZYME and BRENDA)The ECnumberwas noted and the genes were identified A summaryof those results is shown in Table 2Thermodynamic analysisof the reactions revealed that malonyl-CoA synthesis fromacetyl-CoA (AcCoA) is an irreversible process similar tothe process regarding the fabF gene the Eab7 2235 locustag and the extracellular metabolites the other processes arereversible
It is very important to assess the quality of the annotatedgenome submitted to the online tools during curation Theliterature categorically states that the quality of the recon-structed network directly depends on the annotated genomeof the organism The rule is to use the latest updated versionof the annotated genome [36ndash38]
6 BioMed Research International
297
295
296294
184
288300
64
32432533299185
44257
7018
26186182177
52286283280278276270268
81921920
10321
165 170 173 178 181 187 27160
54758697108119130142
162 167 168 175 176 183
154
66
147
157
156
153
158
155
67
163
259 61 60 59 58 536263262
161 166 169 174136 125 114
28188180179172171159 164
51 23 22
Figure 2 Layout of the fatty acid biosynthesis pathway generated using Cytoscape Green vertices fewer connections Yellow vertices regularnumber of connections Red vertices large number of connections This network is a free model scale
34 Curated Metabolic Network The previous step addedconstraints regarding the stoichiometry (chemical balance)thermodynamics (reversibility of reactions) and physiology(cofactors used) of the study model The result was a systemof equations that describes the cell metabolism according tothe metabolites of interest The mathematical representationof this model essentially consisted of describing the perfor-mance of the fatty acid biosynthesis pathway using a stoichio-metric matrix This data structure consists of 54 metabolitesand 59 reactions resulting in a 54 times 59 stoichiometric matrix(Additional File 1) Additional File 2 shows the list of bio-chemical reactions identified in the curated model
The reconstructed pathway model was converted intoSBML in the used MATLAB toolbox The SBML file wasvalidated using the tool SBML validator and was then sub-mitted to the tool Cytoscape which generated the networkof Figure 2 The gene-protein-reaction (GPR) representationtherein describes the degree of connectivity of each enzyme inthe pathway Vertices with few connections are in green thevertices with regular numbers of connections are in yellowand the vertices with large numbers of connections are in redThe network connectivity obeys a scale-free model [39]
35 Flux Balance Analysis (FBA) The FBA was coded inMATLAB implementing a constrained linear program usingthe GLPK (GNU Linear Programming Kit) linear optimiza-tion library [8] All fluxes were calculated in percentage of
the input flux of AcCoA (reaction 39) which was fixed to100 Hexadecenoic acid is the key metabolic product thusthe respective flux (reaction 41) was set as the maximiza-tion target for FBA To improve convergence upper andlower bounds were [0 100] for irreversible reactions and[minus100 100] for reversible reactionThe final optimized fluxesare shown inAdditional File 3The targetmaximum reactionR41 in Additional File 3 was 769 which may be read as 769moles of hexadecenoic acid produced for every 100 moles ofAcCoA consumed per unit time per unit cell mass
Figure 3(a) shows the generated flux plot which showsthe variation occurring between the response fluxes with themajority approximately 37 showing positive values smallerthan 20 while 15 are above that range
36 Influence of Transcriptome in FBA Calculation Log base2-fold change values (log
2FC) obtained in vitro by DallrsquoAgnol
and colleagues [22] were used to evaluate the influence ofdifferential expression in the FBA calculation These valueswere obtained by comparison of RPKM (reads per kilobaseper million reads sequenced) generated in the transcriptomeof the bacterium at 0∘C and 37∘C The log
2FC of genes that
composes the fatty acid biosynthesis pathway is shown inTable 2
As presented by DallrsquoAgnol and colleagues [22] theaerobic energetic metabolism of E antarcticum B7 at 0∘C isrepressed and a fraction of the acetyl-CoA is probably used as
BioMed Research International 7
Reaction
FBA
minus40
minus20
0
20
40
60
80
100
Flux
val
ue (
)
lsquolsquoR1rsquorsquo
lsquolsquoR4rsquorsquo
lsquolsquoR7rsquorsquo
lsquolsquoR10
rsquorsquolsquolsquoR
13rsquorsquo
lsquolsquoR16
rsquorsquolsquolsquoR
19rsquorsquo
lsquolsquoR22
rsquorsquolsquolsquoR
25rsquorsquo
lsquolsquoR28
rsquorsquolsquolsquoR
31rsquorsquo
lsquolsquoR34
rsquorsquolsquolsquoR
37rsquorsquo
lsquolsquoR40
rsquorsquolsquolsquoR
43rsquorsquo
lsquolsquoR46
rsquorsquolsquolsquoR
49rsquorsquo
lsquolsquoR52
rsquorsquolsquolsquoR
55rsquorsquo
lsquolsquoR58
rsquorsquo
(a)
Transcriptome
lsquolsquoR1rsquorsquo
lsquolsquoR4rsquorsquo
lsquolsquoR7rsquorsquo
lsquolsquoR10
rsquorsquolsquolsquoR
13rsquorsquo
lsquolsquoR16
rsquorsquolsquolsquoR
19rsquorsquo
lsquolsquoR22
rsquorsquolsquolsquoR
25rsquorsquo
lsquolsquoR28
rsquorsquolsquolsquoR
31rsquorsquo
lsquolsquoR34
rsquorsquolsquolsquoR
37rsquorsquo
lsquolsquoR40
rsquorsquolsquolsquoR
43rsquorsquo
lsquolsquoR46
rsquorsquolsquolsquoR
49rsquorsquo
lsquolsquoR52
rsquorsquolsquolsquoR
55rsquorsquo
lsquolsquoR58
rsquorsquo
Reaction
02468
10121416
Flux
val
ue (
)
(b)
Figure 3 Flux graph generated using the software MATLAB which uses methods based on FBA constraints The vertical axis represents thefluxes calculated from the input of 100 moles of acetyl-CoA (AcCoA) The horizontal axis represents the reactions participating in the fattyacid biosynthesis pathwayThe blue bars determine the output percentage for each pathway reaction (a) FBA for maximizing the productionof hexadecenoic acid (b) FBA calculated with the log
2FC values obtained from the transcriptome
a substrate to synthesize short-chain fatty acids in coldThe synthesis begins with the conversion of acetyl-CoA tomalonyl-CoA catalyzed by the multimeric enzyme encodedby the genes accABCD In FBA analysis only half of the inputof acetyl-CoA (100mM) is converted tomalonyl-CoA whichbinds to the acyl carrier protein (ACP) at 0∘C (Figure 3(b)) Asstated earlier the other half of acetyl-CoA is probably used forenergy generation
The remaining route is cyclical being the reactionscatalyzed by enzymes encoded by fabF fabG fabZ and fabIgenes At each round two carbons are added to the growingchain of fatty acid In these reactions the flow of metabolitesremains unchanged until the fatty acid molecule reachesa size of eight carbons (octadecanoyl-ACP in reaction 38)where the percentage of the flow amount decreases (Fig-ure 3(b)) These results are consistent with the previouslypublished data which affirm that bacteria decrease their fattyacid chains to survive in coldThese short fatty acidmoleculeswill be converted into membrane phospholipids in order tomaintain the fluidity of this biological barrier
37 Elementary Flux Modes The calculation of elementarymodes was performed in MATLAB using the Metatooltoolbox [40] (modo elementarm code in the SupplementalInformation) A total of 13 elementary fluxmodes were foundfor the fatty acid biosynthesis pathway of E antarcticum B7(Additional File 4) Of these only 4 elementary modes (25 8 and 11) have a positive nonzero coefficient for reaction41 which indicates that the target product hexadecenoicacid may only be generated by one of these four elementarymodes The routes identified in EM2 and EM5 begin at thesecond reaction (R2) which is catalyzed by the enzyme FabDThevalue of R2 for both elementarymodes indicates a consid-erable production of malonyl-CoA The remaining reactionsof EM2 and EM5 are presented in Additional File 4 In EM8
and EM11 the routes begin from R1The value 1 for this reac-tion indicates a lower activity reflecting in a lower productionof hexadecenoic acid Regarding reaction 41 the values ofelementary modes 2 and 5 in this case 1 are higher thanthe values of 8 and 11 (0142857 each) indicating that bothelementary modes 2 and 5 produce 1mol hexadecenoic acidwhen they are active while elementary modes 8 and 11produce 0142857 moles
4 Conclusions
The first metabolic pathway of E antarcticum B7 recon-structed following the steps defined in this work suggeststhat the protocol used is a suitable tool for further metabolicreconstruction studies Almost all the first steps of the processwere automated however manual curation was as usuallaborious because it required an intensive search for availabledata
The metabolic pathway of fatty acid biosynthesis wasrepresentative and consistent under the limits and boundaryconditions setThe FBA and elementary mode methods wereused to examine the hexadecenoic acid production potentialof the reconstructed pathway The application of constraint-based modeling revealed being very useful to assess networkoperation plasticity even if the intracellular kinetics areunknownThe in silico analysis performed using FBA enableda quantitative assessment of hexadecenoic acid productionpotential
Finally a key issue involves deciding when to stop theprocess and to consider the reconstruction finalized at leasttemporarilyThis decision is usually based on the reconstruc-tion purpose The most complete metabolic model currentlyavailable is the E coli model which has been researchedand refined for over 10 years [41ndash45] Other studies con-stantly updating their models are Homo sapiens with three
8 BioMed Research International
reconstructions [46ndash48] and S cerevisiae with more than adozen reconstructions including two in 2013 [49 50] Theprotocol reported in the present studymay be used to compileseveral data pieces available in the literature aimed at propos-ing possible metabolic pathways thereby enabling deeperresearch of the metabolism under study
Competing Interests
The authors declare that they have no competing interests
Acknowledgments
This study was supported by the Conselho Nacional de Des-envolvimento Cientıfico e Tecnologico (CNPq) and Coor-denacao de Aperfeicoamento de Pessoal de Nıvel Superior(CAPES)
References
[1] JW Lee DNa JM Park J Lee S Choi and S Y Lee ldquoSystemsmetabolic engineering of microorganisms for natural and non-natural chemicalsrdquo Nature Chemical Biology vol 8 no 6 pp536ndash546 2012
[2] J Becker and C Wittmann ldquoSystems and synthetic metabolicengineering for amino acid productionmdashthe heartbeat of indus-trial strain developmentrdquoCurrent Opinion in Biotechnology vol23 no 5 pp 718ndash726 2012
[3] C-H ChouW-C Chang C-M Chiu C-C Huang andH-DHuang ldquoFMM a web server for metabolic pathway reconstruc-tion and comparative analysisrdquo Nucleic Acids Research vol 37no 2 pp W129ndashW134 2009
[4] P Shannon A Markiel O Ozier et al ldquoCytoscape a softwareEnvironment for integratedmodels of biomolecular interactionnetworksrdquo Genome Research vol 13 no 11 pp 2498ndash25042003
[5] A Funahashi Y Matsuoka A Jouraku M Morohashi NKikuchi and H Kitano ldquoCellDesigner 35 a versatile modelingtool for biochemical networksrdquo Proceedings of the IEEE vol 96no 8 pp 1254ndash1265 2008
[6] F T Bergmann and HM Sauro ldquoSBWmdashamodular frameworkfor systems biologyrdquo in Proceedings of the 37th Winter Simula-tion Conference (WSC rsquo06) pp 1637ndash1645 December 2006
[7] S Hoops S Sahle R Gauges et al ldquoCOPASImdasha COmplexPAthway SImulatorrdquo Bioinformatics vol 22 no 24 pp 3067ndash3074 2006
[8] J Schellenberger R Que R M T Fleming et al ldquoQuantitativeprediction of cellular metabolism with constraint-based mod-els the COBRA Toolbox v20rdquo Nature Protocols vol 6 no 9pp 1290ndash1307 2011
[9] M Terzer N D Maynard M W Covert and J StellingldquoGenome-scale metabolic networksrdquo Wiley InterdisciplinaryReviews Systems Biology andMedicine vol 1 no 3 pp 285ndash2972009
[10] DMcCloskey B Oslash Palsson andAM Feist ldquoBasic and applieduses of genome-scale metabolic network reconstructions ofEscherichia colirdquo Molecular Systems Biology vol 9 article 6612013
[11] M Fondi I Maida E Perrin et al ldquoGenome-scale metabolicreconstruction and constraint-basedmodelling of the Antarctic
bacterium Pseudoalteromonas haloplanktis TAC125rdquo Environ-mental Microbiology vol 17 no 3 pp 751ndash766 2015
[12] DMcCloskey B Oslash Palsson andAM Feist ldquoBasic and applieduses of genomeminusscale metabolic network reconstructions ofEscherichia colirdquo Molecular Systems Biology vol 9 article 6612013
[13] A M Feist M J Herrgard I Thiele J L Reed and B OslashPalsson ldquoReconstruction of biochemical networks in microor-ganismsrdquoNature ReviewsMicrobiology vol 7 no 2 pp 129ndash1432009
[14] J L Reed I Famili IThiele and B Oslash Palsson ldquoTowardsmulti-dimensional genome annotationrdquoNature Reviews Genetics vol7 no 2 pp 130ndash141 2006
[15] I Thiele and B Oslash Palsson ldquoA protocol for generating ahigh-quality genome-scale metabolic reconstructionrdquo NatureProtocols vol 5 no 1 pp 93ndash121 2010
[16] J S Edwards R U Ibarra and B O Palsson ldquoIn silico pre-dictions of Escherichia colimetabolic capabilities are consistentwith experimental datardquoNature Biotechnology vol 19 no 2 pp125ndash130 2001
[17] R Mahadevan B Oslash Palsson and D R Lovley ldquoIn situ toin silico and back elucidating the physiology and ecology ofGeobacter spp using genome-scale modellingrdquo Nature ReviewsMicrobiology vol 9 no 1 pp 39ndash50 2011
[18] J D Trawick andCH Schilling ldquoUse of constraint-basedmod-eling for the prediction and validation of antimicrobial targetsrdquoBiochemical Pharmacology vol 71 no 7 pp 1026ndash1035 2006
[19] R Y Morita ldquoPsychrophilic bacteriardquo Bacteriological Reviewsvol 39 no 2 pp 144ndash167 1975
[20] R Mulvaney N J Abram R C A Hindmarsh et al ldquoRecentAntarctic Peninsula warming relative to Holocene climate andice-shelf historyrdquo Nature vol 489 no 7414 pp 141ndash144 2012
[21] A R Carneiro R T J Ramos H DallrsquoAgnol et al ldquoGenomesequence of Exiguobacterium antarcticum B7 isolated from abiofilm inGinger Lake KingGeorge Island Antarcticardquo Journalof Bacteriology vol 194 no 23 pp 6689ndash6690 2012
[22] HDallrsquoAgnol R A Barauna P de Sa et al ldquoOmics profiles usedto evaluate the gene expression of Exiguobacterium antarcticumB7 during cold adaptationrdquo BMCGenomics vol 15 no 1 article986 2014
[23] A Casanueva M Tuffin C Cary and D A Cowan ldquoMolecularadaptations to psychrophily the impact of lsquoomicrsquo technologiesrdquoTrends in Microbiology vol 18 no 8 pp 374ndash381 2010
[24] G E Schujman L Paoletti A D Grossman and D de Men-doza ldquoFapR a bacterial transcription factor involved in globalregulation ofmembrane lipid biosynthesisrdquoDevelopmental Cellvol 4 no 5 pp 663ndash672 2003
[25] E Pitkanen J Rousu and E Ukkonen ldquoComputational meth-ods for metabolic reconstructionrdquo Current Opinion in Biotech-nology vol 21 no 1 pp 70ndash77 2010
[26] H Ogata S Goto K Sato W Fujibuchi H Bono and MKanehisa ldquoKEGG kyoto encyclopedia of genes and genomesrdquoNucleic Acids Research vol 27 no 1 pp 29ndash34 1999
[27] T Carver S R Harris M Berriman J Parkhill and J AMcQuillan ldquoArtemis an integrated platform for visualizationand analysis of high-throughput sequence-based experimentaldatardquo Bioinformatics vol 28 no 4 pp 464ndash469 2012
[28] Y Moriya M Itoh S Okuda A C Yoshizawa and M Kane-hisa ldquoKAAS an automatic genome annotation and pathwayreconstruction serverrdquoNucleic Acids Research vol 35 no 2 ppW182ndashW185 2007
BioMed Research International 9
[29] R K Aziz D Bartels A Best et al ldquoThe RAST Server rapidannotations using subsystems technologyrdquo BMCGenomics vol9 article 75 2008
[30] A Bairoch ldquoThe ENZYME database in 2000rdquo Nucleic AcidsResearch vol 28 no 1 pp 304ndash305 2000
[31] R Overbeek R Olson G D Pusch et al ldquoThe SEED andthe rapid annotation of microbial genomes using SubsystemsTechnology (RAST)rdquo Nucleic Acids Research vol 42 no 1 ppD206ndashD214 2014
[32] A Flamholz E Noor A Bar-Even and R Milo ldquoEQuilibra-tormdashthe biochemical thermodynamics calculatorrdquo NucleicAcids Research vol 40 no 1 pp D770ndashD775 2012
[33] C Wrzodek A Drager and A Zell ldquoKEGGtranslator visualiz-ing and converting the KEGG PATHWAY database to variousformatsrdquo Bioinformatics vol 27 no 16 pp 2314ndash2315 2011
[34] M Hucka A Finney H M Sauro et al ldquoThe systems biologymarkup language (SBML) a medium for representation andexchange of biochemical network modelsrdquo Bioinformatics vol19 no 4 pp 524ndash531 2003
[35] M Ashburner C A Ball J A Blake et al ldquoGene ontology toolfor the unification of biologyrdquoNature Genetics vol 25 no 1 pp25ndash29 2000
[36] J J Hamilton and J L Reed ldquoSoftware platforms to facilitatereconstructing genome-scale metabolic networksrdquo Environ-mental Microbiology vol 16 no 1 pp 49ndash59 2014
[37] G J E Baart and D E Martens ldquoGenome-scale metabolicmodels reconstruction and analysisrdquo Methods in MolecularBiology vol 799 pp 107ndash126 2012
[38] V Lacroix L Cottret P Thebault and M-F Sagot ldquoAn intro-duction to metabolic networks and their structural analysisrdquoIEEEACM Transactions on Computational Biology and Bioin-formatics vol 5 no 4 pp 594ndash617 2008
[39] A-L Barabasi and Z N Oltvai ldquoNetwork biology understand-ing the cellrsquos functional organizationrdquo Nature Reviews Geneticsvol 5 no 2 pp 101ndash113 2004
[40] T Pfeiffer I Sanchez-Valdenebro J C Nuno F Montero andS Schuster ldquoMETATOOL for studying metabolic networksrdquoBioinformatics vol 15 no 3 pp 251ndash257 1999
[41] E Almaas B Kovacs T Vicsek Z NOltvai andA-L BarabasildquoGlobal organization of metabolic fluxes in the bacteriumEscherichia colirdquo Nature vol 427 no 6977 pp 839ndash843 2004
[42] J S Edwards and B Oslash Palsson ldquoThe Escherichia coli MG1655in silico metabolic genotype its definition characteristics andcapabilitiesrdquo Proceedings of the National Academy of Sciences ofthe United States of America vol 97 no 10 pp 5528ndash5533 2000
[43] I M Keseler J Collado-Vides S Gama-Castro et al ldquoEcoCyca comprehensive database resource for Escherichia colirdquoNucleicAcids Research vol 33 pp D334ndashD337 2005
[44] J D Orth T M Conrad J Na et al ldquoA comprehensivegenome-scale reconstruction of Escherichia coli metabolism-2011rdquoMolecular Systems Biology vol 7 article 535 2011
[45] A M Feist and B Oslash Palsson ldquoThe growing scope of appli-cations of genome-scale metabolic reconstructions usingEscherichia colirdquo Nature Biotechnology vol 26 no 6 pp 659ndash667 2008
[46] N C Duarte S A Becker N Jamshidi et al ldquoGlobal recon-struction of the human metabolic network based on genomicand bibliomic datardquo Proceedings of the National Academy ofSciences of the United States of America vol 104 no 6 pp 1777ndash1782 2007
[47] T Y Kim S B Sohn Y B KimW J Kim and S Y Lee ldquoRecentadvances in reconstruction and applications of genome-scalemetabolic modelsrdquo Current Opinion in Biotechnology vol 23no 4 pp 617ndash623 2012
[48] I Thiele N Swainston R M Fleming et al ldquoA community-driven global reconstruction of human metabolismrdquo NatureBiotechnology vol 31 no 5 pp 419ndash425 2013
[49] B D Heavner K Smallbone N D Price and L P WalkerldquoVersion 6 of the consensus yeast metabolic network refinesbiochemical coverage and improves model performancerdquoDatabase vol 2013 Article ID bat059 2013
[50] T Osterlund I Nookaew S Bordel and J Nielsen ldquoMap-ping condition-dependent regulation of metabolism in yeastthrough genome-scale modelingrdquo BMC Systems Biology vol 7article 36 2013
Submit your manuscripts athttpwwwhindawicom
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Anatomy Research International
PeptidesInternational Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
International Journal of
Volume 2014
Zoology
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Molecular Biology International
GenomicsInternational Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
BioinformaticsAdvances in
Marine BiologyJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Signal TransductionJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
BioMed Research International
Evolutionary BiologyInternational Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Biochemistry Research International
ArchaeaHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Genetics Research International
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Advances in
Virolog y
Hindawi Publishing Corporationhttpwwwhindawicom
Nucleic AcidsJournal of
Volume 2014
Stem CellsInternational
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Enzyme Research
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
Microbiology
BioMed Research International 5
Table 2 Relationships between components of the E antarcticum B7 fatty acid biosynthesis pathway following curation The signalsrArr andhArr indicate irreversible and reversible reactions respectively
Gene Locus tag EC number Enzyme Reaction Fold change (log2FC)
accA Eab7 2059 6412 Acetyl-CoA carboxylase carboxyltransferase alpha subunit
ATP + acetyl-CoA + HCO3minusrArr ADP +
orthophosphate + malonyl-CoA 04562
accB Eab7 0870 6412 Acetyl-CoA carboxylasebiotin-carboxyl carrier protein
ATP + acetyl-CoA + HCO3minusrArr ADP +
orthophosphate + malonyl-CoA minus005773
accC Eab7 0871 641264114
Acetyl-CoA carboxylase biotincarboxylase subunit
ATP + acetyl-CoA + HCO3minusrArr ADP +
orthophosphate + malonyl-CoA minus05623
accD Eab7 2060 6412 Acetyl-CoA carboxylasecarboxyl transferase beta subunit
ATP + acetyl-CoA + HCO3minusrArr ADP +
orthophosphate + malonyl-CoA minus02811
fabD Eab7 1760 23139 ACP S-malonyl transferase Malonyl-CoA + ACPhArr CoA +malonyl-(acp) 09448
fabH1 Eab7 1911 231180 3-Oxoacyl-ACP synthase III Acetyl-CoA + malonyl-(acp)hArracetoacetyl-(acp) + CoA + CO
2
08512
fabF Eab7 1910 231179 3-Oxoacyl-ACP synthase II Acetyl-(acp) + malonyl-(acp)rArracetoacetyl-(acp) + CO
2+ ACP 06942
fabG Eab7 1895 111100 3-Oxoacyl-ACP reductase Acetoacetyl-(acp) + NADPH + H+ hArr(R)-3-hydroxybutanoyl-(acp) + NADP+ 08523
fabZ Eab7 2463 42159 3-Hydroxyacyl-ACP dehydratase (R)-3-hydroxybutanoyl-(acp)hArrbut-2-enoyl-(acp) + H
2O 012902
fabI Eab7 1885 131913110 Enoyl-ACP reductase I But-2-enoyl-(acp) + NADH + H+ hArr
butyryl-(acp) + NAD+ 02969
mdash Eab7 2235 114192 Acyl-ACP desaturaseHexadecanoyl-(acp) + acceptor reduced+ O2rArr hexadecenoyl-(acp) + acceptor +
2H2O
13768
are colored in purple and the others are listed in white boxes(Figure 1(b))
The analysis of both fatty acid biosynthetic pathway draftsshows that the draft generated using KEGG apparently hasthe most complete flux except for enzyme 63414 whichis exclusively present in the draft resulting from the RASTtool The draft generated using RAST has a gap in which theenzymes FabA and FabB are not included in the pathwayelongation process The flux for the production of hexade-cenoic acid is also absent from the pathway generated usingRAST
Artemis software was used to confirm the presence ofall genes selected through the automatic and semiautomaticmethodsThe genes accABCD fabD fabH1 fabF fabG fabZand fabI and the locus tag Eab7 2235 were described in thegenome of E antarcticum B7 except fabK gene which wasdetected only by the automatic method
The KGML file produced by KEGG was submitted to thesoftware KEGGtranslator [33] to be converted into a SBML(System Biology Markup Language) file [34] This file wasconverted into an Excel spreadsheet using MATLAB func-tions The files in SBML format and the Excel spreadsheetsare the most used formats in metabolic reconstructions Thereactions and metabolites of the preliminary network gener-ated using KEGG could be visualized in the spreadsheet
The data generated using the RASTSEED tool were ana-lyzed and added to the first step of the process supplementingthe data collected using KEGGThe files generated with bothplatforms were used to manage the manual curation data
The larger number of genes identified using KEGG (12)compared to those found using RASTSEED (9) may beexplained because the former uses orthology (KEGGOrthol-ogy (KO)) through protein homology to identify the so-calledmetabolite genes [35] in a genome which facilitates findinggene-protein-reaction (GPR) associations
33 Manual Curation of the Metabolic Pathway of De NovoFatty Acid Synthesis The reactions of the fatty acid synthesispathway were annotated and refined The metabolites wereorganized into two compartments (cytoplasm and extracellu-lar compartment) based on the location of the enzymes asso-ciated with each pathway The cofactors and the reversibilityof the reactions were compiled from the data published in theliterature and online tools (ENZYME and BRENDA)The ECnumberwas noted and the genes were identified A summaryof those results is shown in Table 2Thermodynamic analysisof the reactions revealed that malonyl-CoA synthesis fromacetyl-CoA (AcCoA) is an irreversible process similar tothe process regarding the fabF gene the Eab7 2235 locustag and the extracellular metabolites the other processes arereversible
It is very important to assess the quality of the annotatedgenome submitted to the online tools during curation Theliterature categorically states that the quality of the recon-structed network directly depends on the annotated genomeof the organism The rule is to use the latest updated versionof the annotated genome [36ndash38]
6 BioMed Research International
297
295
296294
184
288300
64
32432533299185
44257
7018
26186182177
52286283280278276270268
81921920
10321
165 170 173 178 181 187 27160
54758697108119130142
162 167 168 175 176 183
154
66
147
157
156
153
158
155
67
163
259 61 60 59 58 536263262
161 166 169 174136 125 114
28188180179172171159 164
51 23 22
Figure 2 Layout of the fatty acid biosynthesis pathway generated using Cytoscape Green vertices fewer connections Yellow vertices regularnumber of connections Red vertices large number of connections This network is a free model scale
34 Curated Metabolic Network The previous step addedconstraints regarding the stoichiometry (chemical balance)thermodynamics (reversibility of reactions) and physiology(cofactors used) of the study model The result was a systemof equations that describes the cell metabolism according tothe metabolites of interest The mathematical representationof this model essentially consisted of describing the perfor-mance of the fatty acid biosynthesis pathway using a stoichio-metric matrix This data structure consists of 54 metabolitesand 59 reactions resulting in a 54 times 59 stoichiometric matrix(Additional File 1) Additional File 2 shows the list of bio-chemical reactions identified in the curated model
The reconstructed pathway model was converted intoSBML in the used MATLAB toolbox The SBML file wasvalidated using the tool SBML validator and was then sub-mitted to the tool Cytoscape which generated the networkof Figure 2 The gene-protein-reaction (GPR) representationtherein describes the degree of connectivity of each enzyme inthe pathway Vertices with few connections are in green thevertices with regular numbers of connections are in yellowand the vertices with large numbers of connections are in redThe network connectivity obeys a scale-free model [39]
35 Flux Balance Analysis (FBA) The FBA was coded inMATLAB implementing a constrained linear program usingthe GLPK (GNU Linear Programming Kit) linear optimiza-tion library [8] All fluxes were calculated in percentage of
the input flux of AcCoA (reaction 39) which was fixed to100 Hexadecenoic acid is the key metabolic product thusthe respective flux (reaction 41) was set as the maximiza-tion target for FBA To improve convergence upper andlower bounds were [0 100] for irreversible reactions and[minus100 100] for reversible reactionThe final optimized fluxesare shown inAdditional File 3The targetmaximum reactionR41 in Additional File 3 was 769 which may be read as 769moles of hexadecenoic acid produced for every 100 moles ofAcCoA consumed per unit time per unit cell mass
Figure 3(a) shows the generated flux plot which showsthe variation occurring between the response fluxes with themajority approximately 37 showing positive values smallerthan 20 while 15 are above that range
36 Influence of Transcriptome in FBA Calculation Log base2-fold change values (log
2FC) obtained in vitro by DallrsquoAgnol
and colleagues [22] were used to evaluate the influence ofdifferential expression in the FBA calculation These valueswere obtained by comparison of RPKM (reads per kilobaseper million reads sequenced) generated in the transcriptomeof the bacterium at 0∘C and 37∘C The log
2FC of genes that
composes the fatty acid biosynthesis pathway is shown inTable 2
As presented by DallrsquoAgnol and colleagues [22] theaerobic energetic metabolism of E antarcticum B7 at 0∘C isrepressed and a fraction of the acetyl-CoA is probably used as
BioMed Research International 7
Reaction
FBA
minus40
minus20
0
20
40
60
80
100
Flux
val
ue (
)
lsquolsquoR1rsquorsquo
lsquolsquoR4rsquorsquo
lsquolsquoR7rsquorsquo
lsquolsquoR10
rsquorsquolsquolsquoR
13rsquorsquo
lsquolsquoR16
rsquorsquolsquolsquoR
19rsquorsquo
lsquolsquoR22
rsquorsquolsquolsquoR
25rsquorsquo
lsquolsquoR28
rsquorsquolsquolsquoR
31rsquorsquo
lsquolsquoR34
rsquorsquolsquolsquoR
37rsquorsquo
lsquolsquoR40
rsquorsquolsquolsquoR
43rsquorsquo
lsquolsquoR46
rsquorsquolsquolsquoR
49rsquorsquo
lsquolsquoR52
rsquorsquolsquolsquoR
55rsquorsquo
lsquolsquoR58
rsquorsquo
(a)
Transcriptome
lsquolsquoR1rsquorsquo
lsquolsquoR4rsquorsquo
lsquolsquoR7rsquorsquo
lsquolsquoR10
rsquorsquolsquolsquoR
13rsquorsquo
lsquolsquoR16
rsquorsquolsquolsquoR
19rsquorsquo
lsquolsquoR22
rsquorsquolsquolsquoR
25rsquorsquo
lsquolsquoR28
rsquorsquolsquolsquoR
31rsquorsquo
lsquolsquoR34
rsquorsquolsquolsquoR
37rsquorsquo
lsquolsquoR40
rsquorsquolsquolsquoR
43rsquorsquo
lsquolsquoR46
rsquorsquolsquolsquoR
49rsquorsquo
lsquolsquoR52
rsquorsquolsquolsquoR
55rsquorsquo
lsquolsquoR58
rsquorsquo
Reaction
02468
10121416
Flux
val
ue (
)
(b)
Figure 3 Flux graph generated using the software MATLAB which uses methods based on FBA constraints The vertical axis represents thefluxes calculated from the input of 100 moles of acetyl-CoA (AcCoA) The horizontal axis represents the reactions participating in the fattyacid biosynthesis pathwayThe blue bars determine the output percentage for each pathway reaction (a) FBA for maximizing the productionof hexadecenoic acid (b) FBA calculated with the log
2FC values obtained from the transcriptome
a substrate to synthesize short-chain fatty acids in coldThe synthesis begins with the conversion of acetyl-CoA tomalonyl-CoA catalyzed by the multimeric enzyme encodedby the genes accABCD In FBA analysis only half of the inputof acetyl-CoA (100mM) is converted tomalonyl-CoA whichbinds to the acyl carrier protein (ACP) at 0∘C (Figure 3(b)) Asstated earlier the other half of acetyl-CoA is probably used forenergy generation
The remaining route is cyclical being the reactionscatalyzed by enzymes encoded by fabF fabG fabZ and fabIgenes At each round two carbons are added to the growingchain of fatty acid In these reactions the flow of metabolitesremains unchanged until the fatty acid molecule reachesa size of eight carbons (octadecanoyl-ACP in reaction 38)where the percentage of the flow amount decreases (Fig-ure 3(b)) These results are consistent with the previouslypublished data which affirm that bacteria decrease their fattyacid chains to survive in coldThese short fatty acidmoleculeswill be converted into membrane phospholipids in order tomaintain the fluidity of this biological barrier
37 Elementary Flux Modes The calculation of elementarymodes was performed in MATLAB using the Metatooltoolbox [40] (modo elementarm code in the SupplementalInformation) A total of 13 elementary fluxmodes were foundfor the fatty acid biosynthesis pathway of E antarcticum B7(Additional File 4) Of these only 4 elementary modes (25 8 and 11) have a positive nonzero coefficient for reaction41 which indicates that the target product hexadecenoicacid may only be generated by one of these four elementarymodes The routes identified in EM2 and EM5 begin at thesecond reaction (R2) which is catalyzed by the enzyme FabDThevalue of R2 for both elementarymodes indicates a consid-erable production of malonyl-CoA The remaining reactionsof EM2 and EM5 are presented in Additional File 4 In EM8
and EM11 the routes begin from R1The value 1 for this reac-tion indicates a lower activity reflecting in a lower productionof hexadecenoic acid Regarding reaction 41 the values ofelementary modes 2 and 5 in this case 1 are higher thanthe values of 8 and 11 (0142857 each) indicating that bothelementary modes 2 and 5 produce 1mol hexadecenoic acidwhen they are active while elementary modes 8 and 11produce 0142857 moles
4 Conclusions
The first metabolic pathway of E antarcticum B7 recon-structed following the steps defined in this work suggeststhat the protocol used is a suitable tool for further metabolicreconstruction studies Almost all the first steps of the processwere automated however manual curation was as usuallaborious because it required an intensive search for availabledata
The metabolic pathway of fatty acid biosynthesis wasrepresentative and consistent under the limits and boundaryconditions setThe FBA and elementary mode methods wereused to examine the hexadecenoic acid production potentialof the reconstructed pathway The application of constraint-based modeling revealed being very useful to assess networkoperation plasticity even if the intracellular kinetics areunknownThe in silico analysis performed using FBA enableda quantitative assessment of hexadecenoic acid productionpotential
Finally a key issue involves deciding when to stop theprocess and to consider the reconstruction finalized at leasttemporarilyThis decision is usually based on the reconstruc-tion purpose The most complete metabolic model currentlyavailable is the E coli model which has been researchedand refined for over 10 years [41ndash45] Other studies con-stantly updating their models are Homo sapiens with three
8 BioMed Research International
reconstructions [46ndash48] and S cerevisiae with more than adozen reconstructions including two in 2013 [49 50] Theprotocol reported in the present studymay be used to compileseveral data pieces available in the literature aimed at propos-ing possible metabolic pathways thereby enabling deeperresearch of the metabolism under study
Competing Interests
The authors declare that they have no competing interests
Acknowledgments
This study was supported by the Conselho Nacional de Des-envolvimento Cientıfico e Tecnologico (CNPq) and Coor-denacao de Aperfeicoamento de Pessoal de Nıvel Superior(CAPES)
References
[1] JW Lee DNa JM Park J Lee S Choi and S Y Lee ldquoSystemsmetabolic engineering of microorganisms for natural and non-natural chemicalsrdquo Nature Chemical Biology vol 8 no 6 pp536ndash546 2012
[2] J Becker and C Wittmann ldquoSystems and synthetic metabolicengineering for amino acid productionmdashthe heartbeat of indus-trial strain developmentrdquoCurrent Opinion in Biotechnology vol23 no 5 pp 718ndash726 2012
[3] C-H ChouW-C Chang C-M Chiu C-C Huang andH-DHuang ldquoFMM a web server for metabolic pathway reconstruc-tion and comparative analysisrdquo Nucleic Acids Research vol 37no 2 pp W129ndashW134 2009
[4] P Shannon A Markiel O Ozier et al ldquoCytoscape a softwareEnvironment for integratedmodels of biomolecular interactionnetworksrdquo Genome Research vol 13 no 11 pp 2498ndash25042003
[5] A Funahashi Y Matsuoka A Jouraku M Morohashi NKikuchi and H Kitano ldquoCellDesigner 35 a versatile modelingtool for biochemical networksrdquo Proceedings of the IEEE vol 96no 8 pp 1254ndash1265 2008
[6] F T Bergmann and HM Sauro ldquoSBWmdashamodular frameworkfor systems biologyrdquo in Proceedings of the 37th Winter Simula-tion Conference (WSC rsquo06) pp 1637ndash1645 December 2006
[7] S Hoops S Sahle R Gauges et al ldquoCOPASImdasha COmplexPAthway SImulatorrdquo Bioinformatics vol 22 no 24 pp 3067ndash3074 2006
[8] J Schellenberger R Que R M T Fleming et al ldquoQuantitativeprediction of cellular metabolism with constraint-based mod-els the COBRA Toolbox v20rdquo Nature Protocols vol 6 no 9pp 1290ndash1307 2011
[9] M Terzer N D Maynard M W Covert and J StellingldquoGenome-scale metabolic networksrdquo Wiley InterdisciplinaryReviews Systems Biology andMedicine vol 1 no 3 pp 285ndash2972009
[10] DMcCloskey B Oslash Palsson andAM Feist ldquoBasic and applieduses of genome-scale metabolic network reconstructions ofEscherichia colirdquo Molecular Systems Biology vol 9 article 6612013
[11] M Fondi I Maida E Perrin et al ldquoGenome-scale metabolicreconstruction and constraint-basedmodelling of the Antarctic
bacterium Pseudoalteromonas haloplanktis TAC125rdquo Environ-mental Microbiology vol 17 no 3 pp 751ndash766 2015
[12] DMcCloskey B Oslash Palsson andAM Feist ldquoBasic and applieduses of genomeminusscale metabolic network reconstructions ofEscherichia colirdquo Molecular Systems Biology vol 9 article 6612013
[13] A M Feist M J Herrgard I Thiele J L Reed and B OslashPalsson ldquoReconstruction of biochemical networks in microor-ganismsrdquoNature ReviewsMicrobiology vol 7 no 2 pp 129ndash1432009
[14] J L Reed I Famili IThiele and B Oslash Palsson ldquoTowardsmulti-dimensional genome annotationrdquoNature Reviews Genetics vol7 no 2 pp 130ndash141 2006
[15] I Thiele and B Oslash Palsson ldquoA protocol for generating ahigh-quality genome-scale metabolic reconstructionrdquo NatureProtocols vol 5 no 1 pp 93ndash121 2010
[16] J S Edwards R U Ibarra and B O Palsson ldquoIn silico pre-dictions of Escherichia colimetabolic capabilities are consistentwith experimental datardquoNature Biotechnology vol 19 no 2 pp125ndash130 2001
[17] R Mahadevan B Oslash Palsson and D R Lovley ldquoIn situ toin silico and back elucidating the physiology and ecology ofGeobacter spp using genome-scale modellingrdquo Nature ReviewsMicrobiology vol 9 no 1 pp 39ndash50 2011
[18] J D Trawick andCH Schilling ldquoUse of constraint-basedmod-eling for the prediction and validation of antimicrobial targetsrdquoBiochemical Pharmacology vol 71 no 7 pp 1026ndash1035 2006
[19] R Y Morita ldquoPsychrophilic bacteriardquo Bacteriological Reviewsvol 39 no 2 pp 144ndash167 1975
[20] R Mulvaney N J Abram R C A Hindmarsh et al ldquoRecentAntarctic Peninsula warming relative to Holocene climate andice-shelf historyrdquo Nature vol 489 no 7414 pp 141ndash144 2012
[21] A R Carneiro R T J Ramos H DallrsquoAgnol et al ldquoGenomesequence of Exiguobacterium antarcticum B7 isolated from abiofilm inGinger Lake KingGeorge Island Antarcticardquo Journalof Bacteriology vol 194 no 23 pp 6689ndash6690 2012
[22] HDallrsquoAgnol R A Barauna P de Sa et al ldquoOmics profiles usedto evaluate the gene expression of Exiguobacterium antarcticumB7 during cold adaptationrdquo BMCGenomics vol 15 no 1 article986 2014
[23] A Casanueva M Tuffin C Cary and D A Cowan ldquoMolecularadaptations to psychrophily the impact of lsquoomicrsquo technologiesrdquoTrends in Microbiology vol 18 no 8 pp 374ndash381 2010
[24] G E Schujman L Paoletti A D Grossman and D de Men-doza ldquoFapR a bacterial transcription factor involved in globalregulation ofmembrane lipid biosynthesisrdquoDevelopmental Cellvol 4 no 5 pp 663ndash672 2003
[25] E Pitkanen J Rousu and E Ukkonen ldquoComputational meth-ods for metabolic reconstructionrdquo Current Opinion in Biotech-nology vol 21 no 1 pp 70ndash77 2010
[26] H Ogata S Goto K Sato W Fujibuchi H Bono and MKanehisa ldquoKEGG kyoto encyclopedia of genes and genomesrdquoNucleic Acids Research vol 27 no 1 pp 29ndash34 1999
[27] T Carver S R Harris M Berriman J Parkhill and J AMcQuillan ldquoArtemis an integrated platform for visualizationand analysis of high-throughput sequence-based experimentaldatardquo Bioinformatics vol 28 no 4 pp 464ndash469 2012
[28] Y Moriya M Itoh S Okuda A C Yoshizawa and M Kane-hisa ldquoKAAS an automatic genome annotation and pathwayreconstruction serverrdquoNucleic Acids Research vol 35 no 2 ppW182ndashW185 2007
BioMed Research International 9
[29] R K Aziz D Bartels A Best et al ldquoThe RAST Server rapidannotations using subsystems technologyrdquo BMCGenomics vol9 article 75 2008
[30] A Bairoch ldquoThe ENZYME database in 2000rdquo Nucleic AcidsResearch vol 28 no 1 pp 304ndash305 2000
[31] R Overbeek R Olson G D Pusch et al ldquoThe SEED andthe rapid annotation of microbial genomes using SubsystemsTechnology (RAST)rdquo Nucleic Acids Research vol 42 no 1 ppD206ndashD214 2014
[32] A Flamholz E Noor A Bar-Even and R Milo ldquoEQuilibra-tormdashthe biochemical thermodynamics calculatorrdquo NucleicAcids Research vol 40 no 1 pp D770ndashD775 2012
[33] C Wrzodek A Drager and A Zell ldquoKEGGtranslator visualiz-ing and converting the KEGG PATHWAY database to variousformatsrdquo Bioinformatics vol 27 no 16 pp 2314ndash2315 2011
[34] M Hucka A Finney H M Sauro et al ldquoThe systems biologymarkup language (SBML) a medium for representation andexchange of biochemical network modelsrdquo Bioinformatics vol19 no 4 pp 524ndash531 2003
[35] M Ashburner C A Ball J A Blake et al ldquoGene ontology toolfor the unification of biologyrdquoNature Genetics vol 25 no 1 pp25ndash29 2000
[36] J J Hamilton and J L Reed ldquoSoftware platforms to facilitatereconstructing genome-scale metabolic networksrdquo Environ-mental Microbiology vol 16 no 1 pp 49ndash59 2014
[37] G J E Baart and D E Martens ldquoGenome-scale metabolicmodels reconstruction and analysisrdquo Methods in MolecularBiology vol 799 pp 107ndash126 2012
[38] V Lacroix L Cottret P Thebault and M-F Sagot ldquoAn intro-duction to metabolic networks and their structural analysisrdquoIEEEACM Transactions on Computational Biology and Bioin-formatics vol 5 no 4 pp 594ndash617 2008
[39] A-L Barabasi and Z N Oltvai ldquoNetwork biology understand-ing the cellrsquos functional organizationrdquo Nature Reviews Geneticsvol 5 no 2 pp 101ndash113 2004
[40] T Pfeiffer I Sanchez-Valdenebro J C Nuno F Montero andS Schuster ldquoMETATOOL for studying metabolic networksrdquoBioinformatics vol 15 no 3 pp 251ndash257 1999
[41] E Almaas B Kovacs T Vicsek Z NOltvai andA-L BarabasildquoGlobal organization of metabolic fluxes in the bacteriumEscherichia colirdquo Nature vol 427 no 6977 pp 839ndash843 2004
[42] J S Edwards and B Oslash Palsson ldquoThe Escherichia coli MG1655in silico metabolic genotype its definition characteristics andcapabilitiesrdquo Proceedings of the National Academy of Sciences ofthe United States of America vol 97 no 10 pp 5528ndash5533 2000
[43] I M Keseler J Collado-Vides S Gama-Castro et al ldquoEcoCyca comprehensive database resource for Escherichia colirdquoNucleicAcids Research vol 33 pp D334ndashD337 2005
[44] J D Orth T M Conrad J Na et al ldquoA comprehensivegenome-scale reconstruction of Escherichia coli metabolism-2011rdquoMolecular Systems Biology vol 7 article 535 2011
[45] A M Feist and B Oslash Palsson ldquoThe growing scope of appli-cations of genome-scale metabolic reconstructions usingEscherichia colirdquo Nature Biotechnology vol 26 no 6 pp 659ndash667 2008
[46] N C Duarte S A Becker N Jamshidi et al ldquoGlobal recon-struction of the human metabolic network based on genomicand bibliomic datardquo Proceedings of the National Academy ofSciences of the United States of America vol 104 no 6 pp 1777ndash1782 2007
[47] T Y Kim S B Sohn Y B KimW J Kim and S Y Lee ldquoRecentadvances in reconstruction and applications of genome-scalemetabolic modelsrdquo Current Opinion in Biotechnology vol 23no 4 pp 617ndash623 2012
[48] I Thiele N Swainston R M Fleming et al ldquoA community-driven global reconstruction of human metabolismrdquo NatureBiotechnology vol 31 no 5 pp 419ndash425 2013
[49] B D Heavner K Smallbone N D Price and L P WalkerldquoVersion 6 of the consensus yeast metabolic network refinesbiochemical coverage and improves model performancerdquoDatabase vol 2013 Article ID bat059 2013
[50] T Osterlund I Nookaew S Bordel and J Nielsen ldquoMap-ping condition-dependent regulation of metabolism in yeastthrough genome-scale modelingrdquo BMC Systems Biology vol 7article 36 2013
Submit your manuscripts athttpwwwhindawicom
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Anatomy Research International
PeptidesInternational Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
International Journal of
Volume 2014
Zoology
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Molecular Biology International
GenomicsInternational Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
BioinformaticsAdvances in
Marine BiologyJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Signal TransductionJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
BioMed Research International
Evolutionary BiologyInternational Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Biochemistry Research International
ArchaeaHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Genetics Research International
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Advances in
Virolog y
Hindawi Publishing Corporationhttpwwwhindawicom
Nucleic AcidsJournal of
Volume 2014
Stem CellsInternational
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Enzyme Research
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
Microbiology
6 BioMed Research International
297
295
296294
184
288300
64
32432533299185
44257
7018
26186182177
52286283280278276270268
81921920
10321
165 170 173 178 181 187 27160
54758697108119130142
162 167 168 175 176 183
154
66
147
157
156
153
158
155
67
163
259 61 60 59 58 536263262
161 166 169 174136 125 114
28188180179172171159 164
51 23 22
Figure 2 Layout of the fatty acid biosynthesis pathway generated using Cytoscape Green vertices fewer connections Yellow vertices regularnumber of connections Red vertices large number of connections This network is a free model scale
34 Curated Metabolic Network The previous step addedconstraints regarding the stoichiometry (chemical balance)thermodynamics (reversibility of reactions) and physiology(cofactors used) of the study model The result was a systemof equations that describes the cell metabolism according tothe metabolites of interest The mathematical representationof this model essentially consisted of describing the perfor-mance of the fatty acid biosynthesis pathway using a stoichio-metric matrix This data structure consists of 54 metabolitesand 59 reactions resulting in a 54 times 59 stoichiometric matrix(Additional File 1) Additional File 2 shows the list of bio-chemical reactions identified in the curated model
The reconstructed pathway model was converted intoSBML in the used MATLAB toolbox The SBML file wasvalidated using the tool SBML validator and was then sub-mitted to the tool Cytoscape which generated the networkof Figure 2 The gene-protein-reaction (GPR) representationtherein describes the degree of connectivity of each enzyme inthe pathway Vertices with few connections are in green thevertices with regular numbers of connections are in yellowand the vertices with large numbers of connections are in redThe network connectivity obeys a scale-free model [39]
35 Flux Balance Analysis (FBA) The FBA was coded inMATLAB implementing a constrained linear program usingthe GLPK (GNU Linear Programming Kit) linear optimiza-tion library [8] All fluxes were calculated in percentage of
the input flux of AcCoA (reaction 39) which was fixed to100 Hexadecenoic acid is the key metabolic product thusthe respective flux (reaction 41) was set as the maximiza-tion target for FBA To improve convergence upper andlower bounds were [0 100] for irreversible reactions and[minus100 100] for reversible reactionThe final optimized fluxesare shown inAdditional File 3The targetmaximum reactionR41 in Additional File 3 was 769 which may be read as 769moles of hexadecenoic acid produced for every 100 moles ofAcCoA consumed per unit time per unit cell mass
Figure 3(a) shows the generated flux plot which showsthe variation occurring between the response fluxes with themajority approximately 37 showing positive values smallerthan 20 while 15 are above that range
36 Influence of Transcriptome in FBA Calculation Log base2-fold change values (log
2FC) obtained in vitro by DallrsquoAgnol
and colleagues [22] were used to evaluate the influence ofdifferential expression in the FBA calculation These valueswere obtained by comparison of RPKM (reads per kilobaseper million reads sequenced) generated in the transcriptomeof the bacterium at 0∘C and 37∘C The log
2FC of genes that
composes the fatty acid biosynthesis pathway is shown inTable 2
As presented by DallrsquoAgnol and colleagues [22] theaerobic energetic metabolism of E antarcticum B7 at 0∘C isrepressed and a fraction of the acetyl-CoA is probably used as
BioMed Research International 7
Reaction
FBA
minus40
minus20
0
20
40
60
80
100
Flux
val
ue (
)
lsquolsquoR1rsquorsquo
lsquolsquoR4rsquorsquo
lsquolsquoR7rsquorsquo
lsquolsquoR10
rsquorsquolsquolsquoR
13rsquorsquo
lsquolsquoR16
rsquorsquolsquolsquoR
19rsquorsquo
lsquolsquoR22
rsquorsquolsquolsquoR
25rsquorsquo
lsquolsquoR28
rsquorsquolsquolsquoR
31rsquorsquo
lsquolsquoR34
rsquorsquolsquolsquoR
37rsquorsquo
lsquolsquoR40
rsquorsquolsquolsquoR
43rsquorsquo
lsquolsquoR46
rsquorsquolsquolsquoR
49rsquorsquo
lsquolsquoR52
rsquorsquolsquolsquoR
55rsquorsquo
lsquolsquoR58
rsquorsquo
(a)
Transcriptome
lsquolsquoR1rsquorsquo
lsquolsquoR4rsquorsquo
lsquolsquoR7rsquorsquo
lsquolsquoR10
rsquorsquolsquolsquoR
13rsquorsquo
lsquolsquoR16
rsquorsquolsquolsquoR
19rsquorsquo
lsquolsquoR22
rsquorsquolsquolsquoR
25rsquorsquo
lsquolsquoR28
rsquorsquolsquolsquoR
31rsquorsquo
lsquolsquoR34
rsquorsquolsquolsquoR
37rsquorsquo
lsquolsquoR40
rsquorsquolsquolsquoR
43rsquorsquo
lsquolsquoR46
rsquorsquolsquolsquoR
49rsquorsquo
lsquolsquoR52
rsquorsquolsquolsquoR
55rsquorsquo
lsquolsquoR58
rsquorsquo
Reaction
02468
10121416
Flux
val
ue (
)
(b)
Figure 3 Flux graph generated using the software MATLAB which uses methods based on FBA constraints The vertical axis represents thefluxes calculated from the input of 100 moles of acetyl-CoA (AcCoA) The horizontal axis represents the reactions participating in the fattyacid biosynthesis pathwayThe blue bars determine the output percentage for each pathway reaction (a) FBA for maximizing the productionof hexadecenoic acid (b) FBA calculated with the log
2FC values obtained from the transcriptome
a substrate to synthesize short-chain fatty acids in coldThe synthesis begins with the conversion of acetyl-CoA tomalonyl-CoA catalyzed by the multimeric enzyme encodedby the genes accABCD In FBA analysis only half of the inputof acetyl-CoA (100mM) is converted tomalonyl-CoA whichbinds to the acyl carrier protein (ACP) at 0∘C (Figure 3(b)) Asstated earlier the other half of acetyl-CoA is probably used forenergy generation
The remaining route is cyclical being the reactionscatalyzed by enzymes encoded by fabF fabG fabZ and fabIgenes At each round two carbons are added to the growingchain of fatty acid In these reactions the flow of metabolitesremains unchanged until the fatty acid molecule reachesa size of eight carbons (octadecanoyl-ACP in reaction 38)where the percentage of the flow amount decreases (Fig-ure 3(b)) These results are consistent with the previouslypublished data which affirm that bacteria decrease their fattyacid chains to survive in coldThese short fatty acidmoleculeswill be converted into membrane phospholipids in order tomaintain the fluidity of this biological barrier
37 Elementary Flux Modes The calculation of elementarymodes was performed in MATLAB using the Metatooltoolbox [40] (modo elementarm code in the SupplementalInformation) A total of 13 elementary fluxmodes were foundfor the fatty acid biosynthesis pathway of E antarcticum B7(Additional File 4) Of these only 4 elementary modes (25 8 and 11) have a positive nonzero coefficient for reaction41 which indicates that the target product hexadecenoicacid may only be generated by one of these four elementarymodes The routes identified in EM2 and EM5 begin at thesecond reaction (R2) which is catalyzed by the enzyme FabDThevalue of R2 for both elementarymodes indicates a consid-erable production of malonyl-CoA The remaining reactionsof EM2 and EM5 are presented in Additional File 4 In EM8
and EM11 the routes begin from R1The value 1 for this reac-tion indicates a lower activity reflecting in a lower productionof hexadecenoic acid Regarding reaction 41 the values ofelementary modes 2 and 5 in this case 1 are higher thanthe values of 8 and 11 (0142857 each) indicating that bothelementary modes 2 and 5 produce 1mol hexadecenoic acidwhen they are active while elementary modes 8 and 11produce 0142857 moles
4 Conclusions
The first metabolic pathway of E antarcticum B7 recon-structed following the steps defined in this work suggeststhat the protocol used is a suitable tool for further metabolicreconstruction studies Almost all the first steps of the processwere automated however manual curation was as usuallaborious because it required an intensive search for availabledata
The metabolic pathway of fatty acid biosynthesis wasrepresentative and consistent under the limits and boundaryconditions setThe FBA and elementary mode methods wereused to examine the hexadecenoic acid production potentialof the reconstructed pathway The application of constraint-based modeling revealed being very useful to assess networkoperation plasticity even if the intracellular kinetics areunknownThe in silico analysis performed using FBA enableda quantitative assessment of hexadecenoic acid productionpotential
Finally a key issue involves deciding when to stop theprocess and to consider the reconstruction finalized at leasttemporarilyThis decision is usually based on the reconstruc-tion purpose The most complete metabolic model currentlyavailable is the E coli model which has been researchedand refined for over 10 years [41ndash45] Other studies con-stantly updating their models are Homo sapiens with three
8 BioMed Research International
reconstructions [46ndash48] and S cerevisiae with more than adozen reconstructions including two in 2013 [49 50] Theprotocol reported in the present studymay be used to compileseveral data pieces available in the literature aimed at propos-ing possible metabolic pathways thereby enabling deeperresearch of the metabolism under study
Competing Interests
The authors declare that they have no competing interests
Acknowledgments
This study was supported by the Conselho Nacional de Des-envolvimento Cientıfico e Tecnologico (CNPq) and Coor-denacao de Aperfeicoamento de Pessoal de Nıvel Superior(CAPES)
References
[1] JW Lee DNa JM Park J Lee S Choi and S Y Lee ldquoSystemsmetabolic engineering of microorganisms for natural and non-natural chemicalsrdquo Nature Chemical Biology vol 8 no 6 pp536ndash546 2012
[2] J Becker and C Wittmann ldquoSystems and synthetic metabolicengineering for amino acid productionmdashthe heartbeat of indus-trial strain developmentrdquoCurrent Opinion in Biotechnology vol23 no 5 pp 718ndash726 2012
[3] C-H ChouW-C Chang C-M Chiu C-C Huang andH-DHuang ldquoFMM a web server for metabolic pathway reconstruc-tion and comparative analysisrdquo Nucleic Acids Research vol 37no 2 pp W129ndashW134 2009
[4] P Shannon A Markiel O Ozier et al ldquoCytoscape a softwareEnvironment for integratedmodels of biomolecular interactionnetworksrdquo Genome Research vol 13 no 11 pp 2498ndash25042003
[5] A Funahashi Y Matsuoka A Jouraku M Morohashi NKikuchi and H Kitano ldquoCellDesigner 35 a versatile modelingtool for biochemical networksrdquo Proceedings of the IEEE vol 96no 8 pp 1254ndash1265 2008
[6] F T Bergmann and HM Sauro ldquoSBWmdashamodular frameworkfor systems biologyrdquo in Proceedings of the 37th Winter Simula-tion Conference (WSC rsquo06) pp 1637ndash1645 December 2006
[7] S Hoops S Sahle R Gauges et al ldquoCOPASImdasha COmplexPAthway SImulatorrdquo Bioinformatics vol 22 no 24 pp 3067ndash3074 2006
[8] J Schellenberger R Que R M T Fleming et al ldquoQuantitativeprediction of cellular metabolism with constraint-based mod-els the COBRA Toolbox v20rdquo Nature Protocols vol 6 no 9pp 1290ndash1307 2011
[9] M Terzer N D Maynard M W Covert and J StellingldquoGenome-scale metabolic networksrdquo Wiley InterdisciplinaryReviews Systems Biology andMedicine vol 1 no 3 pp 285ndash2972009
[10] DMcCloskey B Oslash Palsson andAM Feist ldquoBasic and applieduses of genome-scale metabolic network reconstructions ofEscherichia colirdquo Molecular Systems Biology vol 9 article 6612013
[11] M Fondi I Maida E Perrin et al ldquoGenome-scale metabolicreconstruction and constraint-basedmodelling of the Antarctic
bacterium Pseudoalteromonas haloplanktis TAC125rdquo Environ-mental Microbiology vol 17 no 3 pp 751ndash766 2015
[12] DMcCloskey B Oslash Palsson andAM Feist ldquoBasic and applieduses of genomeminusscale metabolic network reconstructions ofEscherichia colirdquo Molecular Systems Biology vol 9 article 6612013
[13] A M Feist M J Herrgard I Thiele J L Reed and B OslashPalsson ldquoReconstruction of biochemical networks in microor-ganismsrdquoNature ReviewsMicrobiology vol 7 no 2 pp 129ndash1432009
[14] J L Reed I Famili IThiele and B Oslash Palsson ldquoTowardsmulti-dimensional genome annotationrdquoNature Reviews Genetics vol7 no 2 pp 130ndash141 2006
[15] I Thiele and B Oslash Palsson ldquoA protocol for generating ahigh-quality genome-scale metabolic reconstructionrdquo NatureProtocols vol 5 no 1 pp 93ndash121 2010
[16] J S Edwards R U Ibarra and B O Palsson ldquoIn silico pre-dictions of Escherichia colimetabolic capabilities are consistentwith experimental datardquoNature Biotechnology vol 19 no 2 pp125ndash130 2001
[17] R Mahadevan B Oslash Palsson and D R Lovley ldquoIn situ toin silico and back elucidating the physiology and ecology ofGeobacter spp using genome-scale modellingrdquo Nature ReviewsMicrobiology vol 9 no 1 pp 39ndash50 2011
[18] J D Trawick andCH Schilling ldquoUse of constraint-basedmod-eling for the prediction and validation of antimicrobial targetsrdquoBiochemical Pharmacology vol 71 no 7 pp 1026ndash1035 2006
[19] R Y Morita ldquoPsychrophilic bacteriardquo Bacteriological Reviewsvol 39 no 2 pp 144ndash167 1975
[20] R Mulvaney N J Abram R C A Hindmarsh et al ldquoRecentAntarctic Peninsula warming relative to Holocene climate andice-shelf historyrdquo Nature vol 489 no 7414 pp 141ndash144 2012
[21] A R Carneiro R T J Ramos H DallrsquoAgnol et al ldquoGenomesequence of Exiguobacterium antarcticum B7 isolated from abiofilm inGinger Lake KingGeorge Island Antarcticardquo Journalof Bacteriology vol 194 no 23 pp 6689ndash6690 2012
[22] HDallrsquoAgnol R A Barauna P de Sa et al ldquoOmics profiles usedto evaluate the gene expression of Exiguobacterium antarcticumB7 during cold adaptationrdquo BMCGenomics vol 15 no 1 article986 2014
[23] A Casanueva M Tuffin C Cary and D A Cowan ldquoMolecularadaptations to psychrophily the impact of lsquoomicrsquo technologiesrdquoTrends in Microbiology vol 18 no 8 pp 374ndash381 2010
[24] G E Schujman L Paoletti A D Grossman and D de Men-doza ldquoFapR a bacterial transcription factor involved in globalregulation ofmembrane lipid biosynthesisrdquoDevelopmental Cellvol 4 no 5 pp 663ndash672 2003
[25] E Pitkanen J Rousu and E Ukkonen ldquoComputational meth-ods for metabolic reconstructionrdquo Current Opinion in Biotech-nology vol 21 no 1 pp 70ndash77 2010
[26] H Ogata S Goto K Sato W Fujibuchi H Bono and MKanehisa ldquoKEGG kyoto encyclopedia of genes and genomesrdquoNucleic Acids Research vol 27 no 1 pp 29ndash34 1999
[27] T Carver S R Harris M Berriman J Parkhill and J AMcQuillan ldquoArtemis an integrated platform for visualizationand analysis of high-throughput sequence-based experimentaldatardquo Bioinformatics vol 28 no 4 pp 464ndash469 2012
[28] Y Moriya M Itoh S Okuda A C Yoshizawa and M Kane-hisa ldquoKAAS an automatic genome annotation and pathwayreconstruction serverrdquoNucleic Acids Research vol 35 no 2 ppW182ndashW185 2007
BioMed Research International 9
[29] R K Aziz D Bartels A Best et al ldquoThe RAST Server rapidannotations using subsystems technologyrdquo BMCGenomics vol9 article 75 2008
[30] A Bairoch ldquoThe ENZYME database in 2000rdquo Nucleic AcidsResearch vol 28 no 1 pp 304ndash305 2000
[31] R Overbeek R Olson G D Pusch et al ldquoThe SEED andthe rapid annotation of microbial genomes using SubsystemsTechnology (RAST)rdquo Nucleic Acids Research vol 42 no 1 ppD206ndashD214 2014
[32] A Flamholz E Noor A Bar-Even and R Milo ldquoEQuilibra-tormdashthe biochemical thermodynamics calculatorrdquo NucleicAcids Research vol 40 no 1 pp D770ndashD775 2012
[33] C Wrzodek A Drager and A Zell ldquoKEGGtranslator visualiz-ing and converting the KEGG PATHWAY database to variousformatsrdquo Bioinformatics vol 27 no 16 pp 2314ndash2315 2011
[34] M Hucka A Finney H M Sauro et al ldquoThe systems biologymarkup language (SBML) a medium for representation andexchange of biochemical network modelsrdquo Bioinformatics vol19 no 4 pp 524ndash531 2003
[35] M Ashburner C A Ball J A Blake et al ldquoGene ontology toolfor the unification of biologyrdquoNature Genetics vol 25 no 1 pp25ndash29 2000
[36] J J Hamilton and J L Reed ldquoSoftware platforms to facilitatereconstructing genome-scale metabolic networksrdquo Environ-mental Microbiology vol 16 no 1 pp 49ndash59 2014
[37] G J E Baart and D E Martens ldquoGenome-scale metabolicmodels reconstruction and analysisrdquo Methods in MolecularBiology vol 799 pp 107ndash126 2012
[38] V Lacroix L Cottret P Thebault and M-F Sagot ldquoAn intro-duction to metabolic networks and their structural analysisrdquoIEEEACM Transactions on Computational Biology and Bioin-formatics vol 5 no 4 pp 594ndash617 2008
[39] A-L Barabasi and Z N Oltvai ldquoNetwork biology understand-ing the cellrsquos functional organizationrdquo Nature Reviews Geneticsvol 5 no 2 pp 101ndash113 2004
[40] T Pfeiffer I Sanchez-Valdenebro J C Nuno F Montero andS Schuster ldquoMETATOOL for studying metabolic networksrdquoBioinformatics vol 15 no 3 pp 251ndash257 1999
[41] E Almaas B Kovacs T Vicsek Z NOltvai andA-L BarabasildquoGlobal organization of metabolic fluxes in the bacteriumEscherichia colirdquo Nature vol 427 no 6977 pp 839ndash843 2004
[42] J S Edwards and B Oslash Palsson ldquoThe Escherichia coli MG1655in silico metabolic genotype its definition characteristics andcapabilitiesrdquo Proceedings of the National Academy of Sciences ofthe United States of America vol 97 no 10 pp 5528ndash5533 2000
[43] I M Keseler J Collado-Vides S Gama-Castro et al ldquoEcoCyca comprehensive database resource for Escherichia colirdquoNucleicAcids Research vol 33 pp D334ndashD337 2005
[44] J D Orth T M Conrad J Na et al ldquoA comprehensivegenome-scale reconstruction of Escherichia coli metabolism-2011rdquoMolecular Systems Biology vol 7 article 535 2011
[45] A M Feist and B Oslash Palsson ldquoThe growing scope of appli-cations of genome-scale metabolic reconstructions usingEscherichia colirdquo Nature Biotechnology vol 26 no 6 pp 659ndash667 2008
[46] N C Duarte S A Becker N Jamshidi et al ldquoGlobal recon-struction of the human metabolic network based on genomicand bibliomic datardquo Proceedings of the National Academy ofSciences of the United States of America vol 104 no 6 pp 1777ndash1782 2007
[47] T Y Kim S B Sohn Y B KimW J Kim and S Y Lee ldquoRecentadvances in reconstruction and applications of genome-scalemetabolic modelsrdquo Current Opinion in Biotechnology vol 23no 4 pp 617ndash623 2012
[48] I Thiele N Swainston R M Fleming et al ldquoA community-driven global reconstruction of human metabolismrdquo NatureBiotechnology vol 31 no 5 pp 419ndash425 2013
[49] B D Heavner K Smallbone N D Price and L P WalkerldquoVersion 6 of the consensus yeast metabolic network refinesbiochemical coverage and improves model performancerdquoDatabase vol 2013 Article ID bat059 2013
[50] T Osterlund I Nookaew S Bordel and J Nielsen ldquoMap-ping condition-dependent regulation of metabolism in yeastthrough genome-scale modelingrdquo BMC Systems Biology vol 7article 36 2013
Submit your manuscripts athttpwwwhindawicom
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Anatomy Research International
PeptidesInternational Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
International Journal of
Volume 2014
Zoology
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Molecular Biology International
GenomicsInternational Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
BioinformaticsAdvances in
Marine BiologyJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Signal TransductionJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
BioMed Research International
Evolutionary BiologyInternational Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Biochemistry Research International
ArchaeaHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Genetics Research International
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Advances in
Virolog y
Hindawi Publishing Corporationhttpwwwhindawicom
Nucleic AcidsJournal of
Volume 2014
Stem CellsInternational
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Enzyme Research
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
Microbiology
BioMed Research International 7
Reaction
FBA
minus40
minus20
0
20
40
60
80
100
Flux
val
ue (
)
lsquolsquoR1rsquorsquo
lsquolsquoR4rsquorsquo
lsquolsquoR7rsquorsquo
lsquolsquoR10
rsquorsquolsquolsquoR
13rsquorsquo
lsquolsquoR16
rsquorsquolsquolsquoR
19rsquorsquo
lsquolsquoR22
rsquorsquolsquolsquoR
25rsquorsquo
lsquolsquoR28
rsquorsquolsquolsquoR
31rsquorsquo
lsquolsquoR34
rsquorsquolsquolsquoR
37rsquorsquo
lsquolsquoR40
rsquorsquolsquolsquoR
43rsquorsquo
lsquolsquoR46
rsquorsquolsquolsquoR
49rsquorsquo
lsquolsquoR52
rsquorsquolsquolsquoR
55rsquorsquo
lsquolsquoR58
rsquorsquo
(a)
Transcriptome
lsquolsquoR1rsquorsquo
lsquolsquoR4rsquorsquo
lsquolsquoR7rsquorsquo
lsquolsquoR10
rsquorsquolsquolsquoR
13rsquorsquo
lsquolsquoR16
rsquorsquolsquolsquoR
19rsquorsquo
lsquolsquoR22
rsquorsquolsquolsquoR
25rsquorsquo
lsquolsquoR28
rsquorsquolsquolsquoR
31rsquorsquo
lsquolsquoR34
rsquorsquolsquolsquoR
37rsquorsquo
lsquolsquoR40
rsquorsquolsquolsquoR
43rsquorsquo
lsquolsquoR46
rsquorsquolsquolsquoR
49rsquorsquo
lsquolsquoR52
rsquorsquolsquolsquoR
55rsquorsquo
lsquolsquoR58
rsquorsquo
Reaction
02468
10121416
Flux
val
ue (
)
(b)
Figure 3 Flux graph generated using the software MATLAB which uses methods based on FBA constraints The vertical axis represents thefluxes calculated from the input of 100 moles of acetyl-CoA (AcCoA) The horizontal axis represents the reactions participating in the fattyacid biosynthesis pathwayThe blue bars determine the output percentage for each pathway reaction (a) FBA for maximizing the productionof hexadecenoic acid (b) FBA calculated with the log
2FC values obtained from the transcriptome
a substrate to synthesize short-chain fatty acids in coldThe synthesis begins with the conversion of acetyl-CoA tomalonyl-CoA catalyzed by the multimeric enzyme encodedby the genes accABCD In FBA analysis only half of the inputof acetyl-CoA (100mM) is converted tomalonyl-CoA whichbinds to the acyl carrier protein (ACP) at 0∘C (Figure 3(b)) Asstated earlier the other half of acetyl-CoA is probably used forenergy generation
The remaining route is cyclical being the reactionscatalyzed by enzymes encoded by fabF fabG fabZ and fabIgenes At each round two carbons are added to the growingchain of fatty acid In these reactions the flow of metabolitesremains unchanged until the fatty acid molecule reachesa size of eight carbons (octadecanoyl-ACP in reaction 38)where the percentage of the flow amount decreases (Fig-ure 3(b)) These results are consistent with the previouslypublished data which affirm that bacteria decrease their fattyacid chains to survive in coldThese short fatty acidmoleculeswill be converted into membrane phospholipids in order tomaintain the fluidity of this biological barrier
37 Elementary Flux Modes The calculation of elementarymodes was performed in MATLAB using the Metatooltoolbox [40] (modo elementarm code in the SupplementalInformation) A total of 13 elementary fluxmodes were foundfor the fatty acid biosynthesis pathway of E antarcticum B7(Additional File 4) Of these only 4 elementary modes (25 8 and 11) have a positive nonzero coefficient for reaction41 which indicates that the target product hexadecenoicacid may only be generated by one of these four elementarymodes The routes identified in EM2 and EM5 begin at thesecond reaction (R2) which is catalyzed by the enzyme FabDThevalue of R2 for both elementarymodes indicates a consid-erable production of malonyl-CoA The remaining reactionsof EM2 and EM5 are presented in Additional File 4 In EM8
and EM11 the routes begin from R1The value 1 for this reac-tion indicates a lower activity reflecting in a lower productionof hexadecenoic acid Regarding reaction 41 the values ofelementary modes 2 and 5 in this case 1 are higher thanthe values of 8 and 11 (0142857 each) indicating that bothelementary modes 2 and 5 produce 1mol hexadecenoic acidwhen they are active while elementary modes 8 and 11produce 0142857 moles
4 Conclusions
The first metabolic pathway of E antarcticum B7 recon-structed following the steps defined in this work suggeststhat the protocol used is a suitable tool for further metabolicreconstruction studies Almost all the first steps of the processwere automated however manual curation was as usuallaborious because it required an intensive search for availabledata
The metabolic pathway of fatty acid biosynthesis wasrepresentative and consistent under the limits and boundaryconditions setThe FBA and elementary mode methods wereused to examine the hexadecenoic acid production potentialof the reconstructed pathway The application of constraint-based modeling revealed being very useful to assess networkoperation plasticity even if the intracellular kinetics areunknownThe in silico analysis performed using FBA enableda quantitative assessment of hexadecenoic acid productionpotential
Finally a key issue involves deciding when to stop theprocess and to consider the reconstruction finalized at leasttemporarilyThis decision is usually based on the reconstruc-tion purpose The most complete metabolic model currentlyavailable is the E coli model which has been researchedand refined for over 10 years [41ndash45] Other studies con-stantly updating their models are Homo sapiens with three
8 BioMed Research International
reconstructions [46ndash48] and S cerevisiae with more than adozen reconstructions including two in 2013 [49 50] Theprotocol reported in the present studymay be used to compileseveral data pieces available in the literature aimed at propos-ing possible metabolic pathways thereby enabling deeperresearch of the metabolism under study
Competing Interests
The authors declare that they have no competing interests
Acknowledgments
This study was supported by the Conselho Nacional de Des-envolvimento Cientıfico e Tecnologico (CNPq) and Coor-denacao de Aperfeicoamento de Pessoal de Nıvel Superior(CAPES)
References
[1] JW Lee DNa JM Park J Lee S Choi and S Y Lee ldquoSystemsmetabolic engineering of microorganisms for natural and non-natural chemicalsrdquo Nature Chemical Biology vol 8 no 6 pp536ndash546 2012
[2] J Becker and C Wittmann ldquoSystems and synthetic metabolicengineering for amino acid productionmdashthe heartbeat of indus-trial strain developmentrdquoCurrent Opinion in Biotechnology vol23 no 5 pp 718ndash726 2012
[3] C-H ChouW-C Chang C-M Chiu C-C Huang andH-DHuang ldquoFMM a web server for metabolic pathway reconstruc-tion and comparative analysisrdquo Nucleic Acids Research vol 37no 2 pp W129ndashW134 2009
[4] P Shannon A Markiel O Ozier et al ldquoCytoscape a softwareEnvironment for integratedmodels of biomolecular interactionnetworksrdquo Genome Research vol 13 no 11 pp 2498ndash25042003
[5] A Funahashi Y Matsuoka A Jouraku M Morohashi NKikuchi and H Kitano ldquoCellDesigner 35 a versatile modelingtool for biochemical networksrdquo Proceedings of the IEEE vol 96no 8 pp 1254ndash1265 2008
[6] F T Bergmann and HM Sauro ldquoSBWmdashamodular frameworkfor systems biologyrdquo in Proceedings of the 37th Winter Simula-tion Conference (WSC rsquo06) pp 1637ndash1645 December 2006
[7] S Hoops S Sahle R Gauges et al ldquoCOPASImdasha COmplexPAthway SImulatorrdquo Bioinformatics vol 22 no 24 pp 3067ndash3074 2006
[8] J Schellenberger R Que R M T Fleming et al ldquoQuantitativeprediction of cellular metabolism with constraint-based mod-els the COBRA Toolbox v20rdquo Nature Protocols vol 6 no 9pp 1290ndash1307 2011
[9] M Terzer N D Maynard M W Covert and J StellingldquoGenome-scale metabolic networksrdquo Wiley InterdisciplinaryReviews Systems Biology andMedicine vol 1 no 3 pp 285ndash2972009
[10] DMcCloskey B Oslash Palsson andAM Feist ldquoBasic and applieduses of genome-scale metabolic network reconstructions ofEscherichia colirdquo Molecular Systems Biology vol 9 article 6612013
[11] M Fondi I Maida E Perrin et al ldquoGenome-scale metabolicreconstruction and constraint-basedmodelling of the Antarctic
bacterium Pseudoalteromonas haloplanktis TAC125rdquo Environ-mental Microbiology vol 17 no 3 pp 751ndash766 2015
[12] DMcCloskey B Oslash Palsson andAM Feist ldquoBasic and applieduses of genomeminusscale metabolic network reconstructions ofEscherichia colirdquo Molecular Systems Biology vol 9 article 6612013
[13] A M Feist M J Herrgard I Thiele J L Reed and B OslashPalsson ldquoReconstruction of biochemical networks in microor-ganismsrdquoNature ReviewsMicrobiology vol 7 no 2 pp 129ndash1432009
[14] J L Reed I Famili IThiele and B Oslash Palsson ldquoTowardsmulti-dimensional genome annotationrdquoNature Reviews Genetics vol7 no 2 pp 130ndash141 2006
[15] I Thiele and B Oslash Palsson ldquoA protocol for generating ahigh-quality genome-scale metabolic reconstructionrdquo NatureProtocols vol 5 no 1 pp 93ndash121 2010
[16] J S Edwards R U Ibarra and B O Palsson ldquoIn silico pre-dictions of Escherichia colimetabolic capabilities are consistentwith experimental datardquoNature Biotechnology vol 19 no 2 pp125ndash130 2001
[17] R Mahadevan B Oslash Palsson and D R Lovley ldquoIn situ toin silico and back elucidating the physiology and ecology ofGeobacter spp using genome-scale modellingrdquo Nature ReviewsMicrobiology vol 9 no 1 pp 39ndash50 2011
[18] J D Trawick andCH Schilling ldquoUse of constraint-basedmod-eling for the prediction and validation of antimicrobial targetsrdquoBiochemical Pharmacology vol 71 no 7 pp 1026ndash1035 2006
[19] R Y Morita ldquoPsychrophilic bacteriardquo Bacteriological Reviewsvol 39 no 2 pp 144ndash167 1975
[20] R Mulvaney N J Abram R C A Hindmarsh et al ldquoRecentAntarctic Peninsula warming relative to Holocene climate andice-shelf historyrdquo Nature vol 489 no 7414 pp 141ndash144 2012
[21] A R Carneiro R T J Ramos H DallrsquoAgnol et al ldquoGenomesequence of Exiguobacterium antarcticum B7 isolated from abiofilm inGinger Lake KingGeorge Island Antarcticardquo Journalof Bacteriology vol 194 no 23 pp 6689ndash6690 2012
[22] HDallrsquoAgnol R A Barauna P de Sa et al ldquoOmics profiles usedto evaluate the gene expression of Exiguobacterium antarcticumB7 during cold adaptationrdquo BMCGenomics vol 15 no 1 article986 2014
[23] A Casanueva M Tuffin C Cary and D A Cowan ldquoMolecularadaptations to psychrophily the impact of lsquoomicrsquo technologiesrdquoTrends in Microbiology vol 18 no 8 pp 374ndash381 2010
[24] G E Schujman L Paoletti A D Grossman and D de Men-doza ldquoFapR a bacterial transcription factor involved in globalregulation ofmembrane lipid biosynthesisrdquoDevelopmental Cellvol 4 no 5 pp 663ndash672 2003
[25] E Pitkanen J Rousu and E Ukkonen ldquoComputational meth-ods for metabolic reconstructionrdquo Current Opinion in Biotech-nology vol 21 no 1 pp 70ndash77 2010
[26] H Ogata S Goto K Sato W Fujibuchi H Bono and MKanehisa ldquoKEGG kyoto encyclopedia of genes and genomesrdquoNucleic Acids Research vol 27 no 1 pp 29ndash34 1999
[27] T Carver S R Harris M Berriman J Parkhill and J AMcQuillan ldquoArtemis an integrated platform for visualizationand analysis of high-throughput sequence-based experimentaldatardquo Bioinformatics vol 28 no 4 pp 464ndash469 2012
[28] Y Moriya M Itoh S Okuda A C Yoshizawa and M Kane-hisa ldquoKAAS an automatic genome annotation and pathwayreconstruction serverrdquoNucleic Acids Research vol 35 no 2 ppW182ndashW185 2007
BioMed Research International 9
[29] R K Aziz D Bartels A Best et al ldquoThe RAST Server rapidannotations using subsystems technologyrdquo BMCGenomics vol9 article 75 2008
[30] A Bairoch ldquoThe ENZYME database in 2000rdquo Nucleic AcidsResearch vol 28 no 1 pp 304ndash305 2000
[31] R Overbeek R Olson G D Pusch et al ldquoThe SEED andthe rapid annotation of microbial genomes using SubsystemsTechnology (RAST)rdquo Nucleic Acids Research vol 42 no 1 ppD206ndashD214 2014
[32] A Flamholz E Noor A Bar-Even and R Milo ldquoEQuilibra-tormdashthe biochemical thermodynamics calculatorrdquo NucleicAcids Research vol 40 no 1 pp D770ndashD775 2012
[33] C Wrzodek A Drager and A Zell ldquoKEGGtranslator visualiz-ing and converting the KEGG PATHWAY database to variousformatsrdquo Bioinformatics vol 27 no 16 pp 2314ndash2315 2011
[34] M Hucka A Finney H M Sauro et al ldquoThe systems biologymarkup language (SBML) a medium for representation andexchange of biochemical network modelsrdquo Bioinformatics vol19 no 4 pp 524ndash531 2003
[35] M Ashburner C A Ball J A Blake et al ldquoGene ontology toolfor the unification of biologyrdquoNature Genetics vol 25 no 1 pp25ndash29 2000
[36] J J Hamilton and J L Reed ldquoSoftware platforms to facilitatereconstructing genome-scale metabolic networksrdquo Environ-mental Microbiology vol 16 no 1 pp 49ndash59 2014
[37] G J E Baart and D E Martens ldquoGenome-scale metabolicmodels reconstruction and analysisrdquo Methods in MolecularBiology vol 799 pp 107ndash126 2012
[38] V Lacroix L Cottret P Thebault and M-F Sagot ldquoAn intro-duction to metabolic networks and their structural analysisrdquoIEEEACM Transactions on Computational Biology and Bioin-formatics vol 5 no 4 pp 594ndash617 2008
[39] A-L Barabasi and Z N Oltvai ldquoNetwork biology understand-ing the cellrsquos functional organizationrdquo Nature Reviews Geneticsvol 5 no 2 pp 101ndash113 2004
[40] T Pfeiffer I Sanchez-Valdenebro J C Nuno F Montero andS Schuster ldquoMETATOOL for studying metabolic networksrdquoBioinformatics vol 15 no 3 pp 251ndash257 1999
[41] E Almaas B Kovacs T Vicsek Z NOltvai andA-L BarabasildquoGlobal organization of metabolic fluxes in the bacteriumEscherichia colirdquo Nature vol 427 no 6977 pp 839ndash843 2004
[42] J S Edwards and B Oslash Palsson ldquoThe Escherichia coli MG1655in silico metabolic genotype its definition characteristics andcapabilitiesrdquo Proceedings of the National Academy of Sciences ofthe United States of America vol 97 no 10 pp 5528ndash5533 2000
[43] I M Keseler J Collado-Vides S Gama-Castro et al ldquoEcoCyca comprehensive database resource for Escherichia colirdquoNucleicAcids Research vol 33 pp D334ndashD337 2005
[44] J D Orth T M Conrad J Na et al ldquoA comprehensivegenome-scale reconstruction of Escherichia coli metabolism-2011rdquoMolecular Systems Biology vol 7 article 535 2011
[45] A M Feist and B Oslash Palsson ldquoThe growing scope of appli-cations of genome-scale metabolic reconstructions usingEscherichia colirdquo Nature Biotechnology vol 26 no 6 pp 659ndash667 2008
[46] N C Duarte S A Becker N Jamshidi et al ldquoGlobal recon-struction of the human metabolic network based on genomicand bibliomic datardquo Proceedings of the National Academy ofSciences of the United States of America vol 104 no 6 pp 1777ndash1782 2007
[47] T Y Kim S B Sohn Y B KimW J Kim and S Y Lee ldquoRecentadvances in reconstruction and applications of genome-scalemetabolic modelsrdquo Current Opinion in Biotechnology vol 23no 4 pp 617ndash623 2012
[48] I Thiele N Swainston R M Fleming et al ldquoA community-driven global reconstruction of human metabolismrdquo NatureBiotechnology vol 31 no 5 pp 419ndash425 2013
[49] B D Heavner K Smallbone N D Price and L P WalkerldquoVersion 6 of the consensus yeast metabolic network refinesbiochemical coverage and improves model performancerdquoDatabase vol 2013 Article ID bat059 2013
[50] T Osterlund I Nookaew S Bordel and J Nielsen ldquoMap-ping condition-dependent regulation of metabolism in yeastthrough genome-scale modelingrdquo BMC Systems Biology vol 7article 36 2013
Submit your manuscripts athttpwwwhindawicom
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Anatomy Research International
PeptidesInternational Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
International Journal of
Volume 2014
Zoology
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Molecular Biology International
GenomicsInternational Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
BioinformaticsAdvances in
Marine BiologyJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Signal TransductionJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
BioMed Research International
Evolutionary BiologyInternational Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Biochemistry Research International
ArchaeaHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Genetics Research International
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Advances in
Virolog y
Hindawi Publishing Corporationhttpwwwhindawicom
Nucleic AcidsJournal of
Volume 2014
Stem CellsInternational
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Enzyme Research
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
Microbiology
8 BioMed Research International
reconstructions [46ndash48] and S cerevisiae with more than adozen reconstructions including two in 2013 [49 50] Theprotocol reported in the present studymay be used to compileseveral data pieces available in the literature aimed at propos-ing possible metabolic pathways thereby enabling deeperresearch of the metabolism under study
Competing Interests
The authors declare that they have no competing interests
Acknowledgments
This study was supported by the Conselho Nacional de Des-envolvimento Cientıfico e Tecnologico (CNPq) and Coor-denacao de Aperfeicoamento de Pessoal de Nıvel Superior(CAPES)
References
[1] JW Lee DNa JM Park J Lee S Choi and S Y Lee ldquoSystemsmetabolic engineering of microorganisms for natural and non-natural chemicalsrdquo Nature Chemical Biology vol 8 no 6 pp536ndash546 2012
[2] J Becker and C Wittmann ldquoSystems and synthetic metabolicengineering for amino acid productionmdashthe heartbeat of indus-trial strain developmentrdquoCurrent Opinion in Biotechnology vol23 no 5 pp 718ndash726 2012
[3] C-H ChouW-C Chang C-M Chiu C-C Huang andH-DHuang ldquoFMM a web server for metabolic pathway reconstruc-tion and comparative analysisrdquo Nucleic Acids Research vol 37no 2 pp W129ndashW134 2009
[4] P Shannon A Markiel O Ozier et al ldquoCytoscape a softwareEnvironment for integratedmodels of biomolecular interactionnetworksrdquo Genome Research vol 13 no 11 pp 2498ndash25042003
[5] A Funahashi Y Matsuoka A Jouraku M Morohashi NKikuchi and H Kitano ldquoCellDesigner 35 a versatile modelingtool for biochemical networksrdquo Proceedings of the IEEE vol 96no 8 pp 1254ndash1265 2008
[6] F T Bergmann and HM Sauro ldquoSBWmdashamodular frameworkfor systems biologyrdquo in Proceedings of the 37th Winter Simula-tion Conference (WSC rsquo06) pp 1637ndash1645 December 2006
[7] S Hoops S Sahle R Gauges et al ldquoCOPASImdasha COmplexPAthway SImulatorrdquo Bioinformatics vol 22 no 24 pp 3067ndash3074 2006
[8] J Schellenberger R Que R M T Fleming et al ldquoQuantitativeprediction of cellular metabolism with constraint-based mod-els the COBRA Toolbox v20rdquo Nature Protocols vol 6 no 9pp 1290ndash1307 2011
[9] M Terzer N D Maynard M W Covert and J StellingldquoGenome-scale metabolic networksrdquo Wiley InterdisciplinaryReviews Systems Biology andMedicine vol 1 no 3 pp 285ndash2972009
[10] DMcCloskey B Oslash Palsson andAM Feist ldquoBasic and applieduses of genome-scale metabolic network reconstructions ofEscherichia colirdquo Molecular Systems Biology vol 9 article 6612013
[11] M Fondi I Maida E Perrin et al ldquoGenome-scale metabolicreconstruction and constraint-basedmodelling of the Antarctic
bacterium Pseudoalteromonas haloplanktis TAC125rdquo Environ-mental Microbiology vol 17 no 3 pp 751ndash766 2015
[12] DMcCloskey B Oslash Palsson andAM Feist ldquoBasic and applieduses of genomeminusscale metabolic network reconstructions ofEscherichia colirdquo Molecular Systems Biology vol 9 article 6612013
[13] A M Feist M J Herrgard I Thiele J L Reed and B OslashPalsson ldquoReconstruction of biochemical networks in microor-ganismsrdquoNature ReviewsMicrobiology vol 7 no 2 pp 129ndash1432009
[14] J L Reed I Famili IThiele and B Oslash Palsson ldquoTowardsmulti-dimensional genome annotationrdquoNature Reviews Genetics vol7 no 2 pp 130ndash141 2006
[15] I Thiele and B Oslash Palsson ldquoA protocol for generating ahigh-quality genome-scale metabolic reconstructionrdquo NatureProtocols vol 5 no 1 pp 93ndash121 2010
[16] J S Edwards R U Ibarra and B O Palsson ldquoIn silico pre-dictions of Escherichia colimetabolic capabilities are consistentwith experimental datardquoNature Biotechnology vol 19 no 2 pp125ndash130 2001
[17] R Mahadevan B Oslash Palsson and D R Lovley ldquoIn situ toin silico and back elucidating the physiology and ecology ofGeobacter spp using genome-scale modellingrdquo Nature ReviewsMicrobiology vol 9 no 1 pp 39ndash50 2011
[18] J D Trawick andCH Schilling ldquoUse of constraint-basedmod-eling for the prediction and validation of antimicrobial targetsrdquoBiochemical Pharmacology vol 71 no 7 pp 1026ndash1035 2006
[19] R Y Morita ldquoPsychrophilic bacteriardquo Bacteriological Reviewsvol 39 no 2 pp 144ndash167 1975
[20] R Mulvaney N J Abram R C A Hindmarsh et al ldquoRecentAntarctic Peninsula warming relative to Holocene climate andice-shelf historyrdquo Nature vol 489 no 7414 pp 141ndash144 2012
[21] A R Carneiro R T J Ramos H DallrsquoAgnol et al ldquoGenomesequence of Exiguobacterium antarcticum B7 isolated from abiofilm inGinger Lake KingGeorge Island Antarcticardquo Journalof Bacteriology vol 194 no 23 pp 6689ndash6690 2012
[22] HDallrsquoAgnol R A Barauna P de Sa et al ldquoOmics profiles usedto evaluate the gene expression of Exiguobacterium antarcticumB7 during cold adaptationrdquo BMCGenomics vol 15 no 1 article986 2014
[23] A Casanueva M Tuffin C Cary and D A Cowan ldquoMolecularadaptations to psychrophily the impact of lsquoomicrsquo technologiesrdquoTrends in Microbiology vol 18 no 8 pp 374ndash381 2010
[24] G E Schujman L Paoletti A D Grossman and D de Men-doza ldquoFapR a bacterial transcription factor involved in globalregulation ofmembrane lipid biosynthesisrdquoDevelopmental Cellvol 4 no 5 pp 663ndash672 2003
[25] E Pitkanen J Rousu and E Ukkonen ldquoComputational meth-ods for metabolic reconstructionrdquo Current Opinion in Biotech-nology vol 21 no 1 pp 70ndash77 2010
[26] H Ogata S Goto K Sato W Fujibuchi H Bono and MKanehisa ldquoKEGG kyoto encyclopedia of genes and genomesrdquoNucleic Acids Research vol 27 no 1 pp 29ndash34 1999
[27] T Carver S R Harris M Berriman J Parkhill and J AMcQuillan ldquoArtemis an integrated platform for visualizationand analysis of high-throughput sequence-based experimentaldatardquo Bioinformatics vol 28 no 4 pp 464ndash469 2012
[28] Y Moriya M Itoh S Okuda A C Yoshizawa and M Kane-hisa ldquoKAAS an automatic genome annotation and pathwayreconstruction serverrdquoNucleic Acids Research vol 35 no 2 ppW182ndashW185 2007
BioMed Research International 9
[29] R K Aziz D Bartels A Best et al ldquoThe RAST Server rapidannotations using subsystems technologyrdquo BMCGenomics vol9 article 75 2008
[30] A Bairoch ldquoThe ENZYME database in 2000rdquo Nucleic AcidsResearch vol 28 no 1 pp 304ndash305 2000
[31] R Overbeek R Olson G D Pusch et al ldquoThe SEED andthe rapid annotation of microbial genomes using SubsystemsTechnology (RAST)rdquo Nucleic Acids Research vol 42 no 1 ppD206ndashD214 2014
[32] A Flamholz E Noor A Bar-Even and R Milo ldquoEQuilibra-tormdashthe biochemical thermodynamics calculatorrdquo NucleicAcids Research vol 40 no 1 pp D770ndashD775 2012
[33] C Wrzodek A Drager and A Zell ldquoKEGGtranslator visualiz-ing and converting the KEGG PATHWAY database to variousformatsrdquo Bioinformatics vol 27 no 16 pp 2314ndash2315 2011
[34] M Hucka A Finney H M Sauro et al ldquoThe systems biologymarkup language (SBML) a medium for representation andexchange of biochemical network modelsrdquo Bioinformatics vol19 no 4 pp 524ndash531 2003
[35] M Ashburner C A Ball J A Blake et al ldquoGene ontology toolfor the unification of biologyrdquoNature Genetics vol 25 no 1 pp25ndash29 2000
[36] J J Hamilton and J L Reed ldquoSoftware platforms to facilitatereconstructing genome-scale metabolic networksrdquo Environ-mental Microbiology vol 16 no 1 pp 49ndash59 2014
[37] G J E Baart and D E Martens ldquoGenome-scale metabolicmodels reconstruction and analysisrdquo Methods in MolecularBiology vol 799 pp 107ndash126 2012
[38] V Lacroix L Cottret P Thebault and M-F Sagot ldquoAn intro-duction to metabolic networks and their structural analysisrdquoIEEEACM Transactions on Computational Biology and Bioin-formatics vol 5 no 4 pp 594ndash617 2008
[39] A-L Barabasi and Z N Oltvai ldquoNetwork biology understand-ing the cellrsquos functional organizationrdquo Nature Reviews Geneticsvol 5 no 2 pp 101ndash113 2004
[40] T Pfeiffer I Sanchez-Valdenebro J C Nuno F Montero andS Schuster ldquoMETATOOL for studying metabolic networksrdquoBioinformatics vol 15 no 3 pp 251ndash257 1999
[41] E Almaas B Kovacs T Vicsek Z NOltvai andA-L BarabasildquoGlobal organization of metabolic fluxes in the bacteriumEscherichia colirdquo Nature vol 427 no 6977 pp 839ndash843 2004
[42] J S Edwards and B Oslash Palsson ldquoThe Escherichia coli MG1655in silico metabolic genotype its definition characteristics andcapabilitiesrdquo Proceedings of the National Academy of Sciences ofthe United States of America vol 97 no 10 pp 5528ndash5533 2000
[43] I M Keseler J Collado-Vides S Gama-Castro et al ldquoEcoCyca comprehensive database resource for Escherichia colirdquoNucleicAcids Research vol 33 pp D334ndashD337 2005
[44] J D Orth T M Conrad J Na et al ldquoA comprehensivegenome-scale reconstruction of Escherichia coli metabolism-2011rdquoMolecular Systems Biology vol 7 article 535 2011
[45] A M Feist and B Oslash Palsson ldquoThe growing scope of appli-cations of genome-scale metabolic reconstructions usingEscherichia colirdquo Nature Biotechnology vol 26 no 6 pp 659ndash667 2008
[46] N C Duarte S A Becker N Jamshidi et al ldquoGlobal recon-struction of the human metabolic network based on genomicand bibliomic datardquo Proceedings of the National Academy ofSciences of the United States of America vol 104 no 6 pp 1777ndash1782 2007
[47] T Y Kim S B Sohn Y B KimW J Kim and S Y Lee ldquoRecentadvances in reconstruction and applications of genome-scalemetabolic modelsrdquo Current Opinion in Biotechnology vol 23no 4 pp 617ndash623 2012
[48] I Thiele N Swainston R M Fleming et al ldquoA community-driven global reconstruction of human metabolismrdquo NatureBiotechnology vol 31 no 5 pp 419ndash425 2013
[49] B D Heavner K Smallbone N D Price and L P WalkerldquoVersion 6 of the consensus yeast metabolic network refinesbiochemical coverage and improves model performancerdquoDatabase vol 2013 Article ID bat059 2013
[50] T Osterlund I Nookaew S Bordel and J Nielsen ldquoMap-ping condition-dependent regulation of metabolism in yeastthrough genome-scale modelingrdquo BMC Systems Biology vol 7article 36 2013
Submit your manuscripts athttpwwwhindawicom
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Anatomy Research International
PeptidesInternational Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
International Journal of
Volume 2014
Zoology
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Molecular Biology International
GenomicsInternational Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
BioinformaticsAdvances in
Marine BiologyJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Signal TransductionJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
BioMed Research International
Evolutionary BiologyInternational Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Biochemistry Research International
ArchaeaHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Genetics Research International
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Advances in
Virolog y
Hindawi Publishing Corporationhttpwwwhindawicom
Nucleic AcidsJournal of
Volume 2014
Stem CellsInternational
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Enzyme Research
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
Microbiology
BioMed Research International 9
[29] R K Aziz D Bartels A Best et al ldquoThe RAST Server rapidannotations using subsystems technologyrdquo BMCGenomics vol9 article 75 2008
[30] A Bairoch ldquoThe ENZYME database in 2000rdquo Nucleic AcidsResearch vol 28 no 1 pp 304ndash305 2000
[31] R Overbeek R Olson G D Pusch et al ldquoThe SEED andthe rapid annotation of microbial genomes using SubsystemsTechnology (RAST)rdquo Nucleic Acids Research vol 42 no 1 ppD206ndashD214 2014
[32] A Flamholz E Noor A Bar-Even and R Milo ldquoEQuilibra-tormdashthe biochemical thermodynamics calculatorrdquo NucleicAcids Research vol 40 no 1 pp D770ndashD775 2012
[33] C Wrzodek A Drager and A Zell ldquoKEGGtranslator visualiz-ing and converting the KEGG PATHWAY database to variousformatsrdquo Bioinformatics vol 27 no 16 pp 2314ndash2315 2011
[34] M Hucka A Finney H M Sauro et al ldquoThe systems biologymarkup language (SBML) a medium for representation andexchange of biochemical network modelsrdquo Bioinformatics vol19 no 4 pp 524ndash531 2003
[35] M Ashburner C A Ball J A Blake et al ldquoGene ontology toolfor the unification of biologyrdquoNature Genetics vol 25 no 1 pp25ndash29 2000
[36] J J Hamilton and J L Reed ldquoSoftware platforms to facilitatereconstructing genome-scale metabolic networksrdquo Environ-mental Microbiology vol 16 no 1 pp 49ndash59 2014
[37] G J E Baart and D E Martens ldquoGenome-scale metabolicmodels reconstruction and analysisrdquo Methods in MolecularBiology vol 799 pp 107ndash126 2012
[38] V Lacroix L Cottret P Thebault and M-F Sagot ldquoAn intro-duction to metabolic networks and their structural analysisrdquoIEEEACM Transactions on Computational Biology and Bioin-formatics vol 5 no 4 pp 594ndash617 2008
[39] A-L Barabasi and Z N Oltvai ldquoNetwork biology understand-ing the cellrsquos functional organizationrdquo Nature Reviews Geneticsvol 5 no 2 pp 101ndash113 2004
[40] T Pfeiffer I Sanchez-Valdenebro J C Nuno F Montero andS Schuster ldquoMETATOOL for studying metabolic networksrdquoBioinformatics vol 15 no 3 pp 251ndash257 1999
[41] E Almaas B Kovacs T Vicsek Z NOltvai andA-L BarabasildquoGlobal organization of metabolic fluxes in the bacteriumEscherichia colirdquo Nature vol 427 no 6977 pp 839ndash843 2004
[42] J S Edwards and B Oslash Palsson ldquoThe Escherichia coli MG1655in silico metabolic genotype its definition characteristics andcapabilitiesrdquo Proceedings of the National Academy of Sciences ofthe United States of America vol 97 no 10 pp 5528ndash5533 2000
[43] I M Keseler J Collado-Vides S Gama-Castro et al ldquoEcoCyca comprehensive database resource for Escherichia colirdquoNucleicAcids Research vol 33 pp D334ndashD337 2005
[44] J D Orth T M Conrad J Na et al ldquoA comprehensivegenome-scale reconstruction of Escherichia coli metabolism-2011rdquoMolecular Systems Biology vol 7 article 535 2011
[45] A M Feist and B Oslash Palsson ldquoThe growing scope of appli-cations of genome-scale metabolic reconstructions usingEscherichia colirdquo Nature Biotechnology vol 26 no 6 pp 659ndash667 2008
[46] N C Duarte S A Becker N Jamshidi et al ldquoGlobal recon-struction of the human metabolic network based on genomicand bibliomic datardquo Proceedings of the National Academy ofSciences of the United States of America vol 104 no 6 pp 1777ndash1782 2007
[47] T Y Kim S B Sohn Y B KimW J Kim and S Y Lee ldquoRecentadvances in reconstruction and applications of genome-scalemetabolic modelsrdquo Current Opinion in Biotechnology vol 23no 4 pp 617ndash623 2012
[48] I Thiele N Swainston R M Fleming et al ldquoA community-driven global reconstruction of human metabolismrdquo NatureBiotechnology vol 31 no 5 pp 419ndash425 2013
[49] B D Heavner K Smallbone N D Price and L P WalkerldquoVersion 6 of the consensus yeast metabolic network refinesbiochemical coverage and improves model performancerdquoDatabase vol 2013 Article ID bat059 2013
[50] T Osterlund I Nookaew S Bordel and J Nielsen ldquoMap-ping condition-dependent regulation of metabolism in yeastthrough genome-scale modelingrdquo BMC Systems Biology vol 7article 36 2013
Submit your manuscripts athttpwwwhindawicom
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Anatomy Research International
PeptidesInternational Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
International Journal of
Volume 2014
Zoology
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Molecular Biology International
GenomicsInternational Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
BioinformaticsAdvances in
Marine BiologyJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Signal TransductionJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
BioMed Research International
Evolutionary BiologyInternational Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Biochemistry Research International
ArchaeaHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Genetics Research International
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Advances in
Virolog y
Hindawi Publishing Corporationhttpwwwhindawicom
Nucleic AcidsJournal of
Volume 2014
Stem CellsInternational
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Enzyme Research
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
Microbiology
Submit your manuscripts athttpwwwhindawicom
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Anatomy Research International
PeptidesInternational Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
International Journal of
Volume 2014
Zoology
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Molecular Biology International
GenomicsInternational Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
BioinformaticsAdvances in
Marine BiologyJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Signal TransductionJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
BioMed Research International
Evolutionary BiologyInternational Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Biochemistry Research International
ArchaeaHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Genetics Research International
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Advances in
Virolog y
Hindawi Publishing Corporationhttpwwwhindawicom
Nucleic AcidsJournal of
Volume 2014
Stem CellsInternational
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Enzyme Research
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
Microbiology