ACTA BIOLOGICA SLOVENICA LJUBLJANA 2011 Vol. 54, Št. 2: 15-30 ЛВ Molecular modelling of FtsZ proteins based on their homology in Escherichia coli and Mycobacterium tuberculosis as the key stage of rational design of new antituberculous compounds Molekularno modeliranje proteinov FtsZ na osnovi njihove homologije v Escherichia coli in Mycobacterium tuberculosis kot ključna stopnja racionalnega oblikovanja novih protituberkuloznih komponent Oleh Demchuk a, Pavel Karpov a, Peter Raspor b, Yaroslav Blume a* aDepartment of Genomics and Molecular Biotechnology, Institute of Food Biotechnology and Genomics, Natl. Academy of Sciences of Ukraine, Osipovskogo str, 2a, Kyiv, 04123, Ukraine; bBiotechnical Faculty, University of Ljubljana, Jamnikarjeva, 101, 1000 Ljubljana, Slovenia Correspondence: yablume@univ.kiev.ua Abstract: The analysis of the quality of X-ray structures from Mycobacterium tuberculosis FtsZ proteins, which are deposited in the ProteinDataBank, gave a possibility to select a 2Q1Y (Chain A) structure as a template for future in silico research. Also several spatial models of FtsZ protein from Escherichia coli were reconstructed with on-line servers »SWISS-MODEL Workspace« and I-TASSER, than the most appropriate structure was selected. Basing on complex bioinformatic study, the model, which was rebuilt by SwissModel server from 2Q1Y (chain A) template, was supposed as the most significant. Keywords: FtsZ, Escherichia coli, Mycobacterium tuberculosis, 3D-structure modelling and verification, in silico Izvleček: Analiza struktur proteinov FtsZ z X žarki iz Mycobacterium tuberculosis deponiranih v »ProteinDataBank« je dala možnost izbora strukture 2Q1Y (veriga A) kot matrice za nadaljno in silico raziskavo. Nekaj prostorskih modelov proteina FtsZ iz bakterije Escherichia coli je bilo rekonstruiranih na on-line serverju SwissModel in I-Tasser, kateremu je sledil izbor najprimernejše strukture. Na osnovi celovite bio-informacijske preverbe kaže, da je model narejen na platformi serverja SwissModel iz matrice 2Q1Y (veriga A) najbolj sprejemljiv za uporabo. Ključne besede: FtsZ, Escherichia coli, Mycobacterium tuberculosis, 3D-struk-turno modeliranje in preverjanje, in silico Introduction Tuberculosis is the leading cause of death in the world from a single infectious disease, claiming over three million lives each year (Huang et al. 2006). Furthermore, poor patient compliance and inadequate control programs have lead to the emergence of multidrug-resistant strains of M. tuberculosis (Raviglione 2000). Bacterial resistance to three or more 'second-line' antibiotics is classified as extremely drug-resistant tuberculosis. Therefore, there is an urgent need for the development of new anti-tuberculosis drugs with novel mechanism of action(s), which are active against drug-resistant as well as drug-sensitiveM tuberculosis strains. FtsZ (Filamentous temperature-sensitive protein Z) is an essential cell division protein has been shown to be a bacterial homolog of the mammalian cytoskeleton protein tubulin (Kumar et al. 2010). Accordingly, FtsZ protein is a very promising target for new antimicrobial drug development, and especially compounds effective against drug-resistant M. tuberculosis strains (Kumar et al. 2011). It is the perspective target for such numerous and diverse groups of low molecular weight compounds as benzimidazoles (Ohashi et al. 1999, Kumar et al. 2011), naphthalenesulfonates (Yu and Margolin 1998), azithromycins (Margalit et al. 2004), ethyl carbamates (White et al. 2000, 2002), diterpenoid phenols (Jaiswal et al. 2007) etc. At the same time, modern rational design of new compounds with antibacterial activity is impossible without stage of virtual screening, with application of accurate three-dimensional models of target FtsZ-proteins. The last is very important due to FtsZ - tubulin structure and function similarity, places particularly high demands on quality of 3D-models used for in silico molecular docking and virtual screening. At present, Worldwide ProteinDataBank (wwPDB - www.pdb.org: RCSB PDB (USA)/ PDBe (Europe)/PDBj (Japan)) (Berman et al. 2003) contain the row of bacterial FtsZ X-ray structures of different resolution: 2R6R (1.70 A) (Oliva et al. 2007) and 2R75 (1.40 A) (Läppchen et al. 2008) fromAquifex aeolicus; 2VAM (2.50 A) (Oliva et al. 2007), 2VXY (1.70 A) (Haydon et al. 2008), 2RHH (2.00A), 2RHJ (1.76 A), 2RHL (2.45 A) and 2RHO (2.45 A) (Raymond et al. 2009) from Bacillus subtilis; 1FSZ (2.80 A) (Löwe and Amos 1998), 1W5B (2.20 A), 1W5A (2.40 A), 1W58 (2.50A) and 1W59 (2.70 A) (Oliva et al. 2004) from Methanococcus jannaschii; 1RLU (2.08 A), 1RQ2 (1.86 A), 1RQ7 (2.60 A) (Leung et al. 2004), 2Q1X (2.35 A) and 2Q1Y (2.30 A) (Res-picio et al. database record) fromM. tuberculosis; 1OFU (2.10 A) (Cordell et al., 2003) and 2VAW (2.90 A) (Oliva et al. 2007) from Pseudomonas aeruginosa; 1W5F (2.00 A) (Oliva et al. 2004) from Thermotoga maritima. Most of these species belong to different phyla and subkingdoms of the Bacteria kingdom, and one, M. jannaschii, to the phylum Euryarchaeota (subkingdom Archaea). At the same time, there are a number of defects in all deposited in the Protein Data Bank structures. (Höltje et al. 2008) Most of deposited in PDB X-ray structures of bacterial FtsZs characterized by the loss of N- and C-terminal fragments (typically a few tens of residues), presence of gaps in protein globule, as well as absence of certain heavy atoms of side chains of amino acid residues. Unfortunately, until now, there are no more or less complete X-ray structures of E. coli FtsZ protein, model organism also plays an important role in modern biological engineering and industrial microbiology. Currently only the 1F47 (PDB) structure have the last 17 amino acid residues (Lys367-Asp383), forming a short unstructured region, ends with a two-helix turn at the C-terminal end (Mosyak et al. 2000). However, we are also interested in complete structure of this protein, due to the fact that commercial analytical kits for in vitro binding experiments are more available for E. coli FtsZ protein analysis than analytical kits for M. tuberculosis. In vitro and in silico modelling of interaction with low-molecular compounds of both, E. coli and M. tuberculosis FtsZ proteins, such as benzimidazole derivatives, allow us much accurate binding-site identification and analysis. Based on FtsZs structural homology, these experimentally confirmed binding site (or sites), can be extrapolated from E. coli protein to the structure of mycobacterial homolog. This allow us more accurate prediction of binding sites of such new and promising anti-TB compounds as benzimidazoles. Thus, the purpose of the research was in silico modelling of three-dimensional structure ofE. coli FtsZ protein, and qualitative reconstruction of M. tuberculosis FtsZ protein model based on comprehensive analysis of X-ray structures deposited in the Protein Data Bank. Methods Analysis of Protein Data Bank structures of FtsZ M. tuberculosis Complete amino acid sequence ofM tuberculosis FtsZ (P64170) (Cole et al. 1998, Fleischmann et al. 2002) was downloaded from UniProt (http:// www.uniprot.org/) database (The UniProt Consortium 2008). Multiple alignments of amino acid sequences ofM tuberculosis FtsZ PDB-structures and P64170 were realized in ClustalX 2.0.5 with a set of BLOSSUM matrices (http://www.clustal. org, Larkin et al. 2007). The PDB-structures of FtsZM. tuberculosis protein were analyzed using »DeepView - Swiss-PdbViewer 4.0.3« (Guex and Peitsch 1997; http://www.expasy.org/spdbv/). In the absence of heavy atoms in the side-chains the program generated a warning notice about the type and location of structural defects. Lack of amino acid residues was detected by using Ac-celrys Discovery Studio Visualizer 3.0 (Accelrys Software Inc. - http://accelrys.com/). Reconstruction of 3D model of E. coli FtsZ protein Complete amino acid sequence of E. coli O157:H7 FtsZ (P0A9A8) (Perna et al. 2001) was downloaded from UniProt (http://www.uniprot. org/) database (The UniProt Consortium 2008). A three-dimensional structural modelling was carried out on the I-TASSER server (Roy et al. 2010; http://zhanglab.ccmb.med.umich.edu/I-TASSER) and with »SwissModel Automatic Modelling Mode« of »SWISS-MODEL Workspace« server (http://swissmodel.expasy.org/) (Arnold et al. 2006). Both servers running in automatic mode of PDB structure (template) selection. As a result we generated 5 models of three-dimensional structures with I-TASSER, and and one more model with »SWISS-MODEL Workspace«. Additionally, with »SWISS-MODEL Workspace« server we constructed another model based on 2Q1Y (Chain A) template PDB structure which was specified in the manual mode. Root mean square deviations (RMSD) of the fitted 3-D structures were calculated using »molecule align« tool of PyMol 1.4 package (www.pymol.org). Estimation of protein model quality The 3-D structures quality was assessed by processing the models on the MolProbity server. (Chen et al. 2010, http://molprobity.biochem. duke.edu/) This study was performed to estimate the statistics of all-atom contacts (i.e. »all atoms Clashscore«) and protein geometry: defined percentages of poor rotamers, Ramachandran outliers, Ramachandran favoured, residues with bad bonds, residues with bad angles and defined Cß deviations >0.25Ä and MolProbity score. We used the Protein Structure and Model Assessment Tools available at »SWISS-MODEL Workspace« server (http://swissmodel.expasy.org) to assess the quality of the 3-D models based on Raw-score and Z-score of QMEAN6 (Composite scoring function for model quality estimation) (Benkert et al. 2009) and global model quality estimation based on »DFire energy« (all-atom distance-dependent statistical potential) (Zhou and Zhou 2002). Results Selection and quality checking of X-ray Protein Data Bank structures ofM. tuberculosis FtsZ Scanning the Worldwide Protein Data Bank (wwPDB) we revealed several crystal structures of the M. tuberculosis cell division protein FtsZ, determined at 1.86 to 2,60 Ä by X-ray method. The following PDB structures have been studied: 1RLU (Chains: A, B) and 2Q1Y (Chains: A, B) X-rays of FtsZ-GTPgammaS (5'-guanosine diphosphate monothiophosphate) complexes (Chains: A, B), 1RQ2 (Chains: A, B) and 2Q1X (Chains: A, B) X-rays of FtsZ-citrate complexes (Chains: A, B) and 1RQ7 X-ray of FtsZ-GDP complex (Chains: A, B). Using pairwise sequence alignment of polypeptide chains A and B from X-ray PDB structures with complete sequences from UniProt we tested them on presence of gaps (meaning occurrence of defects in structures). PDB-structures were verified on the presence of such artifacts as deficiency of heavy atoms (carbon, oxygen and nitrogen) in side chains of individual amino acid residues (using DeepView-Swiss-PdbViewer 4.0.3 software package). All gaps in polypeptide chains and residues with defective side chains were checked and represented in the Figure 1 and Table 1. As a result, of the ten available X-ray structures (considering chains A and B) the only one chain ofM. tuberculosis FtsZ protein was selected as a most complete experimentally proved structure and the base of subsequent work on its detailed 3-D reconstruction and in silico analysis. Such structure was a chain A of X-ray FtsZ-GTP-gamma-S complex from 2Q1Y (2.30 A, R-value=0.174, R-free=0.210) (D01:10.2210/pdb2q1y/pdb). It has no just first 7 (N-end) and the last 66 (C-end) amino acid residues, but, as opposed to the same A-chain of 1RLU (2.08 A, R-value=0.182, R-free= 0.224) (Leung et al., 2004), has a complete atomic composition of all available amino acids. The 2Q1Y chain A were analyzed with MolProbity server, and the values of evaluation functions (see Table 2), demonstrate its high quality for further in silico experiments and modelling its interactions with low molecular weight compounds. Table 1: Features and identified defects in Protein Data Bank X-rey structures of Mycobacterium tuberculosis FtsZ. Tabela 1: Lastnosti in okvare struktur PDB v proteinu FTsZ iz bakterije Mycobacterium tuberculosis. Defective regions of molecules PDB ,, , , Resolution, „ . - Strucrure Method a Chain N-terminal Tubulin/FtsZ family, GTPase С-terminal tail tile domain 1RLU 2.35 A Diffraction A aa:M1-Y7 ha: E29, K33, R64, L66, R181 aa:D313 - R379; ha: K236, E252 B aa:M1-H5 aa:R60-G70,G170-A173; aa:D313 - R379; ha: N6, L8, E73, K77, K120, R140 ha: K236, D301 1RQ2 Diffraction 186 A A aa:M1-Y7 aa:R64-A69; ha: K33, M177 B aa:M1-Y7 aa:R60-G70, D171-A173; ha: K33, Q45, E73 aa:D313 - R379; ha; K236 aa:V314-R379 A aa:M1-Y7 аа R64-A69; ha: K33, L48 aa:D313 -R379; ha: K236, Q255 1RQ7 Diffraction 260 A aa:R60-G70, D171-A173; B aa:M1-H5 ha: K33, E73, K120, R140, S141, E153 aa:V314-R379; ha: K236, D313 A aa:M1-Y7 aa:T63-G70 aa:D313 - R379 2Q1X Diffraction 2.35 A B aa:M1-H5 aa:R60-A71, R140-N142,Q168-A173 aa:D313 - R379 2Q1Y Diffraction 2.30 A A aa:M1-Y7 B aa:M1-H5 aa:R60-G70, G170-A173 aa:D313 - R379 aa:D313 - R379 aa - lack of respective amino acid residues in the X-rey structure; ha - absence of heavy atoms (carbon, oxygen or nitrogen) in the side chains of respective amino acid residues. CCCCCCrfJCCCrt;- ЕнЕнЕнЕнЕнЕнЕнЕнЕнЕнЕн->>>>>>>>>>>■ frifrjfrjfrjfrifrjfrifrifrifrifri- SääSI äägl ää: QQQQQQQQQQQ-ССССССССССС-CDCDCDCDCDCDCDCDCDCDCD-tóostótótótótótótótóec;- ыыыыыыыыымы-ыырдыыыыыыыы- ыиыыыыыыныы- QQQQQQQQQQQ-QQQQQQQQQQQ- а os i ČDOOUCDĆDČDČDĆDČDČD->>>>>>>>>>> ымыыыыыыыыы-.нсцсмсмсмрчсмсмсмсмсмоч- r-nriQQQQQQQQQ-\ < < < $ < < ' Ьцрчрчрчрцрчрчрчрцсцрч- ыыыыыыыыыыы->>>>>>>>>>>■ CDCDCDCDOCDCDCDCDCDO-уцуцудудуцуцуцудуцуцуд- CDCDCDCDCDCDCDCDCDCDCD- аосиоооооооо- ЫЫЫЫЫНЫЫЫЫЫ-Hl—IMI—II—II—IMI—II—II—II—I - 11111111111 SSSZSSSSZZZ->>>>>>>>>>> ÜCDÜCDOCDCDOCDÜCD-CDCDCDCDCDCDOOCDCDCD-CDCDÜOOCDCDOCDOO-CDCDOCDCDCDCDCDCDCDCD- HHHHHHHHHHH- CDCDCDCDCDCDCDCDCDCDCD- Ш1ШШ1: HHHHHHHHHHM- CCCCCCCCCCC- ► S ü CüDDDDDDDDD^ ДНННВЕн^ЕнНЯН эиоооооооио 4->*>1>Н>Ч>Ч>-1>Н>Ч*>Ч ,s|E|S|E|2|S|S|S|s|S| SNNNNIslNNNNN — BIWWWWWWWWW ОЕчЕнЕнЕнЕнЕнЕнЕнНЕн r^fafafrifrifrifafrifrifafri S «i Vi Vi Vi Vi"i СмЕ>Рсчсмг~-г-ХХ>|^ H йооюанннн акййсйсййаааа raHHHHHHN(NNCM Eh Eh Eh Eh Eh Eh Eh Eh Eh Eh Eh * Eh h Eh Eh Eh H Eh Eh Eh Eh Eh M M H н4 1-4 1-4 h4 1-4 1-4 1-4 h4 H4 1-4 h4 Q Q Q Q Q Q Q Q Q Q Q Eh Eh Eh Eh Eh Eh Eh Eh Eh Eh Eh H h ih CD CD CD CD О CD CD О О CD CD О a a a a О a a О a a > > > > > > > > > > > О CD CD CD о CD CD CD CD CD CD s S S S s S S S S s S * - 1-4 1-4 h4 1-4 1-4 h4 1-4 1-4 1-4 ь4 1-4 1-4 h-4 1-4 - 1-4 1-4 1-4 h4 .-4 1-4 > > > > > > > > > > > fri Ы fri fri fri fri fri fri fr] fri fr] * Q Q Q Q Q Q Q Q Q Q Q < < < < С < С <: < CO CO CO CO CO CO CO CO co co CO ce ed ce ed ed ed cd ed Cd fri tu tM ti £M fri ti ti fri ti fri < < С < < < < < < .< С p Q Q Q P p Q Q P Q Q S S S : S S S S ГО h4 ь4 ь4 h4 1-4 1-4 h4 h4 1-4 1-4 ь4 CO CO co co CO CO co co CO CO CO > > > > > > > > > > > * »Siiiiiit а а i о о i ► §§§ ОЭ J i-4 J ^ J J J «H Cd Cd Cd а а а SSS PL, p-l P-l „ Q CD CD O CD i-4 ,-4 i-4 >-4 J J J J Cd ti, CÙ ti. Q Q Q O S S S S CL| CL| p-l P-l O _ _ ,-4 i-4 .-4 i-4 - J J J J ■ Cd Cd Cd Cd - Q Q Q Q - S S S S - CM CM di PM - >>>>>>>>>>> ЕнЕнЕнЕнЕнЕнЕнЕнЕнЕнЕн-QDQQQDQQQQQ-OOOOUOUOUOO-WWWMWCOWWWWW-ЫЫЫЫЫЫЫЫЫЫЫ-cdcdcdcdcdcdcdcdcdcdcd- rt] - О О CD SSS см Ы Ы Ы с < с н а а о ► SSS CDCDCDCDCDCDCDCD4 SSSSSSSS-i иининниш ОООООООСИ г ss г г г i я я ................cd cd cd - i-iCDCDCDCDCDCDCDCDCDCDCD-ыыыыыыыыыыы- РчЬРчРчРчРчРчРчРчРчЕч-WWWWWCOWWWWM-рчрчрчрчрчьчрчрчрчрч&ч" CLiCUDjCLiCLiCUD-iCLiD-iCi-ieM-cdtfcdcdcdtfcdcdcdcdcd-ННННННЕнНННН- ....i:!... CDCDCDCDCDCDCDCDCDCDCD->>>>>>>>>>> EHEHEHE-IEHHHEHEHEHH- PPPJPPPJJPP- OCDCDCDOOCDOOCDO-,-4PPPPPPPi-4Pi-4- cdcdcdcdcdcdcdcdcdcdcd-wwwwwnwmwww- См Он < < О О CD CD H Eh CD CD EH EH CD CD О О CD CD Ы Ы CD CD см см < < CD О CD CD H EH CD CD Eh Eh CD CD О О CD CD Ы Ы CD CD см см < С О о CD О : < e i CD CD EH EH CD CD CD О CD CD Ы Ы CD CD CM CM CM CM CM -<<<;<;<;- CD CD CD CD CD - CD CD CD CD CD - EH EH EH En En - CD CD CD CD CD ■ EH EH EH EH EH - CD CD CD CD CD - CD CD CD CD CD - О О CD CD CD ■ Ы Ы Ы Ы Ы - CD CD CD CD CD ■ Я D D HjH>HjH>H>J>H xxxxxxfa NNNNNNNN (ОЮИИСЙИИСО ЕнЕнЕнЕнЕнЕнНЕн fr|fr|fr|fr|fr|fr|fafr| S^ST^äST, СЧСМГ^-Г-ХХ>1>Н аооанннн ййссйааою i—Ii—Ii—Ii—ICMCMCMCM CDCDCDCDCDCDCDCDCDCDCD-.-4l-4.-4l-4l-4.-4.-4l-4l-4l-4.-4-сососососососососососо- QQQQQQQQQQQ- QQQQQQQQQQQ->>>>>>>>>>>- ЕнЕнЕнЕнЕнЕнЕнЕнЕнЕнЕн-CDCDCDCDCDCDCDCDCDCDCD-РчСиРчРчСцРчРчЕчЬчРчЬч- QQQQQQQQQQQ-! СМСМСМСМСМСМСМСМСМСМСМ4 sqctlrtjrtjrtjisijrijisqrtjsiiri!-) QQQQQQQQQQQ-ì ООООООООООО) >>>>>>>>>>>i wwcowcowtowwcocoi rfieCrfrfsirfrfrrfrtJrfjrtiH ЫЫЫЫЫЫЫИЫЫИ-1 SSSSSSSSSSS-ì ы ы Рч Рч 1-4 1-4 CD CD н4 J Q Q CO C/5 CD О U CD С с Ы Ы fri fri P 1-4 CD CD P P Q Q CO co CD CD CD CD < <3 ы ы fri Рч P 1-4 CD CD 1-4 P Q Q CO CO CD CD CD CD < С и ы ы и ы ■ Рч fri Рч Рч Рч - уЦ уД н4 - CD CD CD CD CD ■ 1-4 1-4 1-4 1-4 1-4 - Q Q Q Q Q - CO CO CO CO CO - CD CD CD CD CD ■ CD CD CD CD CD - <3 rtj rtl <1 <1 - CO CO CO :> h4 н4 h4 > > > CD CD CD о а а с < < CD CD CD Ы Ы Ы SSS co co co < С С И ы ы h4 1-4 1-4 1-4 i-4 СМ СМ СМ со со со SSS со со s s ь4 1-4 > > О о а о О CD ы ы s s с л 00 С < ы ы н4 ь4 h4 А СМ СМ со со s s со со s s н4 1-4 > > с; с; а а с < о о ы ы s s оо оп < С и ы - 1-4 н4 н4 СМ СМ со со s s СО СО СО СО -i S S S S i 1-4 1-4 н4 1-4 -l > > > > -I CD CD CD CD -i си о о а -i Itg < sC >< H CD CD CD CD ■) Ы Ы Ы Ы -i S S S S -i со со со со -i <<<<■! ы ы ы ы -Ì А н4 УА h4 H у! А -i СМ СМ СМ СМ 4 СО СО СО СО -Ì S S S S 4 rt]rt!rt!i > Д Q < < h frj Q Q > > S S CD CD S S h-4 1-4 С с EH EH CD CD < С CD О CO co S S CD CD fcd ÜKj > > Q Q < С frj fr4 Q Q : S S CD CD S S ь4 1-4 < < EH EH CD CD С С CD CD CO CO S S CD CD > > Q Q < < fri Рч Q Q CD CD CD CD CD - s s s s s - J J J J J - <; <; < <3 - EH EH EH EH EH ■ CD CD CD CD CD ■ <3 > > > > S S SSS- i-4i-4i-4i-4i-4i-4i-4i-4i-4i-4i-4-CDCDCDCDÜCDCDCDCDCDCD-CMCMCMCMCMCMCMCMCMCMCM- o M D D D aj eh H eh Л U O ü 3 й S4 ^ • N N N S СО со со — EH EH EH О pei fri fri 5 У Л ifl D D N D D D O D D D H H H H EH EH H О О О О О О О >1 >н >1 ^ ^ ^ «iVi^iVi M N NN SI N N со со со со со ю со EH EH EH EH EH EH EH fri fri Рч fri fri fa Ьч СМ Г- Г- X X >Н >Н О О О н н и и cd cd cd о си O» о И Н И СМ CSI N см i i a i a i i i i jfrlfrlfrlfrlfrlpLlfrlfrlfrl _ -3CDCDCDCDCDCDCDCDCD HHHHHHHHHHH >>>>>>>>>>> EHEHEHEHEHEHEHEHEHEHEH cdcdcdcdcdcdcdcdcdcdcd >>>>>>>>>>> Wfr3fr]fr]fr]fr]frqf£|fr]fr!frq QQQQQQQQQQQ F > tu ruktu nt >н Л ^ a о 4 & ■ n a X - . ^ Ć>' (N - ^ f7 ■ & c^ ; N & j о a ( ^ Й 1 n e ue se gt n T ot r 4 ' <>H л^^оаоанннн Oigcdcdcdcdcdaaoa COrHi—Ii—Ii—Ii—Ii—ICMCMCMCM ^ « Z X d tn F a 's Ć> o и а § ES 98% ProteinGeometry Cß deviations > 0.25 A_0_Goal: 0 MolProbity score 1.20 99th percentile* Residues with bad bonds:_0.00%_Goal: 0% _Residues with bad angles:_0.00%_Goal: < 0.1% Clashscore is the number of serious steric overlaps (> 0.4 A) per 1000 atoms. * 100th percentile is the best among structures of comparable resolution; 0th percentile is the worst. л MolProbity score is defined as the following: 0.42574*log(1+clashscore) + 0.32996*log(1+max(0,pctRot0ut-1)) + 0.24979*log(1+max(0,100-pctRamaFavored-2)) + 0.5 Table 3: Top 10 template X-Ray structures selected by I-TASSER server for homology modelling of Escherichia coli FtsZ protein Tabela 3: Zgornjih 10 matric struktur z X žarki uporabljenih na I-TASSER-ju za modeliranje homolognosti FtsZ proteina bakterije Escherichia coli. Rank PDB Hit Ident1 Ident2 Cov. Norm. Z-score 1 2VAW (Chain A) 0.67 0.55 0.82 3.46 2 2VAW (Chain A) 0.67 0.55 0.82 5.88 3 1FSZ (Chain A) 0.44 0.40 0.86 6.38 4 2VAM (Chain A) 0.54 0.43 0.79 4.68 5 2R6R (Chain A) 0.46 0.39 0.84 6.20 6 2VAW (Chain A) 0.67 0.55 0.82 0.00 7 2VAW (Chain A) 0.67 0.55 0.82 13.50 8 1FSZ (Chain A) 0.48 0.26 0.53 5.77 9 1W5F (Chain A) 0.46 0.38 0.81 6.87 10 2RHL (Chain A) 0.52 0.43 0.82 5.00 (a) Rank of templates represents the top ten threading templates used by I-TASSER. (b) Ident1 is the percentage sequence identity of the templates in the threading aligned region with the query sequence. (c) Ident2 is the percentage sequence identity of the whole template chains with query sequence. (d) Cov. represents the coverage of the threading alignment and is equal to the number of aligned residues divided by the length of query protein. (e) Norm. Z-score is the normalized Z-score of the threading alignments. Alignment with a Normalized Z-score >1 mean a good alignment and vice versa. H tfc < £ a D F О p £ =з c 3 8 oo гз У 'S 00 гз T3 . о - e с £ с f u с о 0+ щ sn « sd S s О I tu "o a « °s • —Си pnl o ger g .§ .2 £ .2 ^ ^ ^ 3 'S a .2 ^ 3 a ais. >,43 P и et om щ в g. ster a« .tost a ^ с ü a ° "S sn < 8 .3 -g "O сл C -- laeu V £ o.e. — £ etn ^ "S s ад и S 3 sie. ue Vi s & s сд O O "rt O £ T В crte и "S •S ^ n. o 2 -3.2 3 C U cd o tj H =3 » rt o СД "O i.e 0 a hie. 1 * s s d.ie. a.. h с ~ s 3 ° int.entent tm fe of. pe m.e и ™ ^ у en ote p. eo nin £ a. ГЗ s. СЛ « p © £ § .а g o -e -o .S tican С сл .s o s i n c a o in ami e ut » g 3 & s s s es is niti o d I u О 3 < o .s a £ s ÖIO ^r au^o LO > O LO го au lo ^ л L) <£> LO EH О г- > О CO J o '-o ^ EH O^LO H M O CU и ^ O —tó O LO LO CM « О Ю co Q О мэ S О О со см H О OD ^ Ü О "-Đ о EH со 1-н о gj со со о СО СП О со сп о со сп о > СО СП О ti СО CTi СМ ^ со со о СО !-Н 00 О со О ст. ОГ-О Q О г- o — Л О оэ o P О со о m ^ со со о co С0 4>0 EH со ^ о Ü со ^ о H со со о СО cri о со cri о ион С ОГ-О Й о СП (О P О СП co ОСП см д о h д г- «ч д сп m сп со н депо Ен д сп см o — Й Д OW 00 О Д СП о CM > Д СП о EH Д СП со P д сп со E Д СГ\ о Д СП гН д сп со д со о Q О СП со О г- см < о ^ 'З1 ^ О со О спи со P ОСП со о — У, Д LO h О СП со 00 oć, д г- со O О СП о со Д COO < о lo о рц Д СО гН EH СО СО о |< д г- со H СО СП о Ш ООО СО СП о н о о о СО СП о Q О LO см СО «Ti о JU^o o со сп о >н О LO СМ kd со lo гн Q о ^ CM ОСМ со Ol о г- ^ ОГ-О Ы О со LO О со го ^ О Г- CO О г- ^ < О г- ^ ^ о lo о P О со со СО ^ H ОСОгН CL| о ^ p О со сС о г- ^ И О г- см О Ю О о -dumo О^о-Н ю s о г- lo ft о^н со > О г- ^ СО О СП <н со m н ч1 у, о lo H д г- гн СОг--^ < Д СП О > О Г- ^ s Д СП со Р-1 О г- Д СП -чГ ^ог-^г O < Д СП О OIO Д СП о Ы О LO LO CM Д СП п< OU ^ Д СП го eh о lo ^ д сп о ^ иш^ Д О^ i—1 QJ о Г-- 'З1 p иню rf О со ^ H s о г- "чГ o О со см о о *£> lo CQ О 'З1 со Д о ю -vf < со ^ о OIO ю > со со со о — OU ^ LO o со г- о со СО Г- гН со се; о "ч1 ^ а со со о q о lo lo со г- о S О LO ^ со ^ о > О LO < со ^ о i ! « 1 « JS . -ä О ' ^ ^ i § СУ , -о j § "В ! -ТЗ ^ О Щ in ai •Я : m & ■ s S 2 i & 4 о •Я g •a e H И n o ti a О •S rt -о - § ea S> -о ' 5 3 -3 ё iS e ^ S P & h ° & л ne C j ■ Sj < i* a 2 in m =й w N i i S И ■s ■o < 'Š I >o o aS . r ™ -o W «aS -я .2 > N & p N ^ o 2 2« > .S^l .а "^gO o СД СЛ rt o "" " ^ Š t >CJ T3 o & n ЯЁ 2 ! E 4 a л Evaluation of all functions demonstrates a significantly higher quality of models, built with SWISS-MODEL server (Table 4). Evaluation of all-atom contacts, namely the number of serious steric overlaps of all atoms, allows us to determine model based on PDB matrix structure 1OFU (Chain B) as the best among all E. coli FtsZ protein models. At the same time, evaluation indexes of protein geometry, listed in Table 4, characterize another SWISS-MODEL structure, based on PDB template 2Q1Y (Chain A), as the most accurate. According to all-atom statistical potential DFire index, among I-TASSER models, the undisputed leader are the Model 1 and an outsider - Model 4. At the same time, among SWISS-MODEL reconstructed models, structure, based on 2Q1Y (Chain A), demonstrate substantially superior quality in comparison with model based on Pseudomonas FtsZ structure. Evaluation indices »Raw score« and »Z-score« of QMEAN6, demonstrate undeniable superiority of E. coli FtsZ model based on 2Q1Y (Chain A) template. Discussion Choice of FtsZ M. tuberculosis model As shown in Figure 1, polypeptide chains of all FtsZ structures deposited in the database are lacking the first 5-7 a.a. residues of N-terminal region and 65-66 residues of C-terminal region (in fact, the last C-terminal region is absent at all). Also, eight chains, which are representing all five of the studied PDB-structures of FtsZ M. tuberculosis (in 1RLU and 2Q1Y - only Chains B), contain 5-10 residues gap within the loop, which corresponds to the loop between sheet S3 and helix H3 from 1FSZ - crystal structure of FtsZ protein from M. jannaschii (Oliva et al. 2004). B-chains of each structure are characterized by the absence of 3-6 residues in the loop that extend to helix H7. Additionally, the chain B of 2Q1X has a short gap (3 a.a. residues) within the a-helix that corresponds to helix H5 of 1FSZ. Relative to the artifacts of residues, only 2Q1X and 2Q1Y crystals have all heavy atoms in all amino acid residues of structure (Table 1). Both of this two structures of FtsZ M. tuberculosis are characterized by complete composition of amino acid side chains, whereas the X-Rey structure obtained by Leung et al. 2004 (PDB: 1RQ7) are suffer from lack of some heavy atoms in at least three residues in each chain. Thus, Chain A from the 2Q1Y structure, which does not possess the seven initial (N-end) and sixty six C-end amino acid residues but in contrast to the Chain A from 1RLU has full atom composition of all available amino acid residues, in our opinion is a good three-dimensional model of FtsZ M. tuberculosis. This assumption was confirmed by the scores of all MolProbity quality evaluating functions for this structure (model) (Table 2). Construction and verification of FtsZ E. coli models The »SWISS-MODEL Workspace« is a web-based integrated service dedicated to protein structure homology modelling. It assists in building protein homology models at different levels of complexity. Successful model building requires at least one experimentally determined 3D structure (template) that shows significant amino acid sequence similarity with the target sequence. Building a homology model comprises four main steps: identification of structural template(s), alignment of target sequence and template structure(s), model building, and model quality evaluation. These steps can be repeated until a satisfying modelling result is achieved. Each of the four steps requires specialized software and access to up-to-date protein sequence and structure databases. In full automatic mode server performs all four steps itself and gives the complete three-dimensional model of the target protein with specified matrix structure. There is also a possibility to specify the necessary structure as the matrix manualy (Arnold et al. 2006). Thus, in full automatic modelling of FtsZ E. coli server used Chain B from 1OFU structure as a matrix. It should be noted that selection of this object as a matrix was based on the origin of these two bacterial species. Both of them are belonging to the class y-proteobacteria from phylum Proteobacteria. Although orders of these species are different, an affiliation to a common class is important enough to explain/suppose a higher level of similarity between sequence of FtsZ from E. coli and P. aeruginosa (Vaughan et al. 2004, Demchuk and Blume 2005), than with sequences from other bacterial species from different phylums for which crystal structures are also presented in PDB database. However, such relatively close relationship between E. coli and P. aeruginosa was not sufficient to build a complete model of FtsZ E. coli. Incomplete N-terminal end of this model starting with the Asn24, which indicates the absence of first beta-sheet and significant part of the next alpha-helix in the model structure (Fig. 2a), although in the structure of the matrix this elements are present (Ala11-Gly23 in FtsZ from E.coli and Ala12-Gly24 in FtsZ from P. aeruginosa). This artifact is surprising because of the fact that primary sequence of missing part of FtsZ E. coli is completely similar to that area in the template structure - FtsZ protein from P. aeruginosa. We can only assume some failure in the algorithm of »SWISS-MODEL Workspace« server taking note the fact that the using a chain 2Q1Y_A as specific template allowed server to generate FtsZ E. coli model, which contains the mentioned area in its structure (Fig. 2b) and that is why it was stretched from the Ala11 to Ile315 inclusive. Ca-atom RMSD of obtained FtsZ E. coli models, whics were built on different matrices (Fig. 2c), is 0.862 Ä. This indicates a significant structural similarity of the obtained models (Gu and Bourne 2009). And it can become an additional demonstration of the structure conservatism of this class of proteins (Erickson, 1998) and significant modelling accuracy while taking proteins from bacteria phylums class as a template. Minor differences observed in the turn of main chain (Fig. 2c) are confined to the unstructured elements between ß-sheets and alpha-helices and fully capable of leveling through the lability of the secondary structure elements in the time (Nyporko and Blume 2001). In parallel with the modelling on »SWISS-MODEL Workspace« server we have performed modelling of FtsZ E. coli on I-TASSER server, an Internet service for protein structure and function predictions. It built 3-D models based on multiple-threading alignments by LOMETS and iterative TASSER assembly simulations; function insights are then derived by matching the predicted models with protein function databases. I-TASSER (as 'Zhang-Server') was ranked as the No 1 server for protein structure prediction in recent CASP7, CASP8 and CASP9 experiments (http://predic-tioncenter.org/). It was also ranked as the best for function prediction in CASP9. CASP (or Critical Assessment of Techniques for Protein Structure Prediction) is a community-wide experiment for testing the state-of-the-art of protein structure predictions which takes place every two years since 1994. The experiment (often referred as a competition) is strictly blind because the structures of testing proteins are unknown to the predictors (Roy et al. 2010, http://zhanglab.ccmb.med.umich. edu/I-TASSER/about.html). As a result, server generated five potential full length models of FtsZ proteins from E. coli (Fig. 3) with different values of C-score (see Table 4). C-scores a confidence score for estimating the quality of predicted models by I-TASSER. It is calculated based on the significance of threading template alignments and the convergence parameters of the structure assembly simulations. C-score is typically in the range of [-5,2], where a C-score of higher value signifies a model with a high confidence and vice-versa (Roy et al. 2010). As it is shown in the Table 4 the best value of this parameter belongs to the model 1. I-TASSER server estimated accuracy for this model: 0.56±0.15 TM-score and 9.6±4.6 Ä RMSD. TM-score and RMSD are known standards of structural similarity between two structures which are usually used as measure of model accuracy when the native structure is known. In case where the native structure is not known, it becomes necessary to predict the quality of the modelling prediction, i.e. what is the distance between the predicted model and the native structures. TM-score is a recently proposed scale for measuring the structural similarity between two structures (Zhang and Skolnick 2004). The purpose of proposing TM-score is to solve the problem of RMSD which is sensitive to the local error. Because RMSD is an average distance of all residue pairs in two structures, a local error (e.g. a misorientation of the tail) may cause a significant RMSD value although the global topology may be correct. In TM-score, however, the small distance is weighted stronger than the big distance which makes the score insensitive to the local modelling error. A TM-score >0.5 indicates a model of correct topology and a TM-score <0.17 means a random similarity. This cutoff does not depend on the protein length (Roy et al. 2010). Thus, the quality of prediction of model 1 is quite satisfactory for these scores. On Fig. 3 we presented a predicted secondary structures and properties of FtsZ protein from E. coli models generated by I-TASSER. It should be noted that the distribution of secondary structure elements (shown as H for alpha helixes, S for beta sheets and C for coils) matches with that for crystal structures of FtsZ (see figure 2 in Oliva et al., 2007). Confidence score values for the proposed structure also indicate a rather high quality of all predicted models. However, it should be noted that the value of this estimated parameter are at a slightly lower level for the C-tail region (which is not represented in any of the crystal structures of FtsZ proteins) in comparison with the globular N- and C-domains. We can speculate the possibility of predicted by I-TASSER server unstructured conformation of short N-terminal region in the structure of FtsZ E. coli because analogs are observed in structures 2VAW and 1FSZ. But predicted secondary structure of long and not structured C-terminal region between B10-sheet and final a-helix, provided in the FtsZ E. coli (Mosyak et al. 2000), does not inspire confidence in us. Low values of Confidence score and some crystal structures of FtsZ testified against coiled coil structure of this C-terminal region provided by I-TASSER server. Thus, in contrast to several crystal FtsZ structures that are starting with ß-sheet B1 and breaking immediately after the ß-sheet B10 (1OFU, 1RQ7, 1RLU, 1RQ2, 2Q1X and 2Q1Y), FtsZ structures from 2R6R and 2R75 as well as 2VAP and 1FSZ have two additional ß-sheets B11 and B12. 1W5X structure has also a short a-helix between B10 and B11, which is terminated element of C-end region in structures 1W5F, 2RHH and 2RHJ. All together with a variety of conformational representations of this C-terminal region in five models provided by I-TASSER server (Fig. 4) can not regard any of these models as the most likely and suitable for further bioinformatics researches. For the final determination of the best potential model of FtsZ E. coli among all constructed we have verified all seven candidates with the online server MolProbity and »Protein Structure & Model Assessment Tools« of »SWISS-MODEL Workspace«. MolProbity allows to evaluate the quality both of all atoms contact and so the protein geometry of any three-dimensional biopolymer molecule. Evaluation of all-atom contacts, specifically the number of serious steric overlaps of all atoms, allows to choose model, built basing on template 1OFU_B, as a favourable. In this case second model from »SWISS-MODEL Workspace« loses much less for this parameter in compare to models of I-TASSER. On the other hand, estimation of all parameters of protein geometry (Table 4), allow us to consider the »SWISS-MODEL Workspace« model built based on template structure 2Q1Y (chain A), as the best one. However, on such parameters as »poor rotamers« and »Ramachandran favoured« it's not much better than the second model from this server and all models from I-TASSER. A significant benefit of both models from »SWISS-MODEL Workspace« over models from I-TASSER is observed in rates of »Ramachandran outliers«, »Cß deviations«, »MolProbity score«. Significant differences are also observed in »Residues with bad angles« parameter. While two models from »SWISS-MODEL Workspace« have no residues with bad angles, in structures from I-TASSER models the quantity of these residues reaches 4-6 %. It should be noted that residues with bad bonds observed only in the case of model 2 among all seven analyzed models. Benefits of models built on the I-TASSER in all-atom statistical potential »DFire energy« due to the bigger size of this models that are completely full unlike to the incomplete models from SWISS-MODEL server. Therefore, this parameter from »Protein structure & model assessment tools« we used to compare structures built only on the same server. Finally, for evaluating quality of three-dimensional structures, it was implemented QMEAN6 score (http://swissmodel.expasy.org/ qmean/cgi/index.cgi). QMEAN6 is a reliability score for the whole model which can be used in order to compare and rank alternative models of the same target. The quality estimate ranges between 0 and 1 with higher values for better models. Additionally, the pseudo energies of the four contributing statistical potential terms are provided as well as the percentage agreement between predicted and measured features from the sequence and model, respectively. The comparison of the differences of the terms among the models may help to understand the reason for the differences in the estimated model quality. In addition to the »Raw scores«, »Z-scores« of the QMEAN composite score as well as all terms are provided relating the quality estimates to scores obtained for high-resolution reference structures solved experimentally by X-ray crystallography (Benkert et al. 2011). The QMEAN »Z-score« represents an measure of the absolute quality of a model by providing an estimate of the »degree of nativeness« of the structural features observed in a model and by describing the likelihood that a given model is of comparable quality to experimental structures. Models of low quality are expected to have strongly negative QMEAN Z-scores (i.e. the model's QMEAN score is several standard deviations lower than expected for experimental structures of similar size). Evaluation QMEAN6 scores, especially »Raw score« and »Z-score«, puts everything in its place, clearly demonstrating the undeniable advantage of E. coli FtsZ protein model based on template 2Q1Y (chain A) over all other constructed models. Among the variety of crystallographic structures of FtsZ proteins from tuberculous pathogen M. tuberculosis, deposited in »Protein Data Bank«, we have chosen chain 2Q1Y (chain A) as a reasonable model for further bioinformatics References researches. It lacks first seven and last 66 residues, but it is characterized by full atom composition of all available residues and excellent estimated values of parameters from MolProbity. Also, among the seven potential FtsZ E. coli models, built by I-TASSER server and »SWISS-MODEL Workspace«, we have selected a model from the last server, based on matrix 2Q1Y (chain A). This model prevailed for qualitative evaluation parameters not only full atom models of I-TASSER, but model, built on a matrix 1OFU (chain B), which represents FtsZ structure from P. aeruginosa - bacterial species systematically much closer to E. coli (the same class of Y-proteobacteria), than M. tuberculosis. Thus, were have successfuly reconstructed 3-D models of the FtsZ proteins of E. coli and M. tuberculosis is of sufficient quality for further in silico studies such as molecular docking, molecular dymamics simulations and computational drug discovery. Acknowledgments This work was supported by Scientific and Technologic Centre of Ukraine (STCU) grant #4939. Ukrainian participants were also supported by The Ukrainian National Grid (UNG)- http:// grid.nas.gov.ua/. Arnold, K., Bordoli, L., Kopp, J., Schwede, T., 2006. The »SWISS-MODEL Workspace«: A web-based environment for protein structure homology modelling. Bioinformatics, 22, 195-201. Benkert, P., Biasini, M., Schwede, T., 2011. Toward the estimation of the absolute quality of individual protein structure models. Bioinformatics, 27, (3), 343-350. Benkert, P., Schwede, T., Tosatto, S.C.E., 2009. QMEANclust: Estimation of protein model quality bycombining a composite scoring function with structural density information. BMC Struct. Biol., 9 (35), doi:10.1186/1472-6807-9-35. Berman, H.M., Henrick, K., Nakamura, H., 2003. Announcing the worldwide Protein Data Bank. Nat. Struct. Biol., 10 (12), p. 980. Chen, V.B., Arendall, W.B. III, Headd, J.J., Keedy, D.A., Immormino, R.M., Kapral, G.J., Murray, L.W., Richardson, J.S., Richardson, D.C., 2010. MolProbity: all-atom structure validation for macromolecular crystallography. Acta Crystallogr. D: Biol. Crystallogr., D66 (1), 12-21. Cole, S.T., Brosch, R., Parkhill, J., Garnier, T., Churcher, C.M., Harris, D.E., Gordon, S.V., Eiglmeier, K., Gas, S., Barry, C.E. III, Tekaia, F., Badcock, K., Basham, D., Brown, D., Chillingworth, T., Connor, R., Davies, R.M., Devlin, K., Barrell, B.G., 1998. Deciphering the biology of M. tuberculosis from the complete genome sequence, Nature, 393, 537-544. Cordell, S.C., Robinson, E.J.H., Löwe, J., 2003. Crystal structure of the SOS cell division inhibitor SulA and in complex with FtsZ. Proc. Natl. Acad. Sci. USA, 100 (13), 7889-7894. Demchuk, O.N., Blume, Ya.B., 2005. Phylogenetic tree of bacterial and eucaryotic FtsZ-proteins created according to the homology of their primary sequences, Cytol. Genetics, 39 (4), 3-12. Erickson, H.P., 1998. Atomic structures of tubulin and FtsZ, Trends Cell Biol., 8, 133-137. Fleischmann, R.D., Alland, D., Eisen, J.A., Carpenter, L., White, O., Peterson, J.D., DeBoy, R.T., Dodson, R.J., Gwinn, M.L., Haft, D.H., Hickey, E.K., Kolonay, J.F., Nelson, W.C., Umayam, L.A., Ermolaeva, M.D., Salzberg, S.L., Delcher, A., Utterback, T.R., Fraser, C.M., 2002. Whole-genome comparison ofM. tuberculosis clinical and laboratory strains. J. Bacteriol., 184, 5479-5490. Gu, J., Bourne, P.E., 2009. Structural Bioinformatics, 2nd ed. John Wiley and Sons, New Jersey, 1035 pp. Guex, N., Peitsch, M.C., 1997. SWISS-MODEL and the Swiss-PdbViewer: An environment for comparative protein modelling. Electrophoresis, 18 (15), 2714-2723. Haydon, D.J., Stokes, N.R., Ure, R., Galbraith, G., Bennett, J.M., Brown, D.R., Baker, P.J., Barynin, V.V., Rice, D.W., Sedelnikova, S.E., Heal, J.R., Sheridan, J.M., Aiwale, S.T., Chauhan, P.K., Srivastava, A., Taneja, A., Collins, I., Errington, J., Czaplewski, L.G., 2008. An inhibitor of FtsZ with potent and selective anti-staphylococcal activity. Science, 321, 1673-1675. Höltje, H.-D., Sippl, W., Rognan, D., Folkers, G., 2008. Molecular modeling. Basic Principles and Applications, 3rd Ed.. Wiley-VCH, p. 320, Huang, Q., Kirikae, F., Kirikae, T., Pepe, A., Slayden, R.A., Tonge, P. J., Ojima, I., 2006. Targeting FtsZ for anti-tuberculosis drug discovery: non-cytotoxic taxanes as novel anti-tuberculosis agents. J. Med. Chem., 49 (2), 463-466. Jaiswal, R., Beuria, T.K., Mohan, R., Mahajan, S.K., Panda, D., 2007. Totarol inhibits bacterial cytokinesis by perturbing the assembly dynamics of FtsZ. Biochem., 46 (14), 4211-4220. Kumar, K., Awasthi, D., Berger, W.T., Tonge, P. J., Slayden, R.A., Ojima, I., 2010. Discovery of anti-TB agents that target the cell-division protein FtsZ. Future Med Chem., 2 (8), 1305-1323. Kumar, K., Awasthi, D., Lee, S-Y., Zanardi, I., Ruzsicska, B., Knudson, S., Tonge, P.J., Slayden, R.A., Ojima, I., 2011.Novel trisubstituted benzimidazoles, targeting Mtb FtsZ, as a new class of antitubercular agents. J. Med. Chem., 54, 374-381. Läppchen, T., Pinas, V.A., Hartog, A.F., Koomen, G.J., Schaffner-Barbero, C., Andreu, J.M., Tram-baiolo, D., Löwe, J., Juhem, A., Popov, A.V., den Blaauwen, T., 2008. Probing FtsZ and tubulin with C8-substituted GTP analogs reveals differences in their nucleotide binding sites. Chem. Biol., 15, 189-199. Larkin, M.A., Blackshields, G., Brown, N.P., Chenna, R., McGettigan, P.A., McWilliam, H., Valentin, F., Wallace, I.M., Wilm, A., Lopez, R., Thompson, J.D., Gibson, T. J., Higgins, D.G., 2007. Clustal W and Clustal X version 2.0. Bioinformatics, 23, 2947-2948. Leung, A.K., Lucile White, E., Ross, L.J., Reynolds, R.C., DeVito, J.A., Borhani, D.W., 2004. Structure of M. tuberculosis FtsZ reveals unexpected, G protein-like conformational switches. J. Mol. Biol., 342 (3), 953-970. Li, Y., Zhang, Y., 2009. REMO: A new protocol to refine full atomic protein models from C-alpha traces by optimizing hydrogen-bonding networks. Proteins, 76, 665-676. Löwe, J., Amos, L.A., 1998. Crystal structure of the bacterial cell-division protein FtsZ. Nature, 391, 203-206. Margalit, D.N., Romberg, L., Mets, R.B., Hebert, A.M., Mitchison, T. J., Kirschner, M.W., Chaudhuri, D.R., 2004. Targeting cell division: small-molecule inhibitors of FtsZ GTPase perturb cytokinetic ring assembly and induce bacterial lethality. PNAS., 101, 11821-11826. Mosyak, L., Zhang, Y., Glasfeld, E., Haney, S., Stahl, M., Seehra, J., Somers, W.S., 2000. The bacterial cell-division protein ZipA and its interaction with an FtsZ fragment revealed by X-ray crystallography. EMBO J., 19 (13), 3179-3191. Nyporko, A.Yu., Blume, Ya.B., 2001. Comparative analysis of the tubulin secondary structure. Bio-polym. Cell, 17, (1), 61-69, in Russian. Ohashi, Y., Chijiiwa, Y., Suzuki, K., Takahashi, K., Nanamiya, H., Sato, T., Hosoya, Y., Ochi, K., Kawamura, F., 1999. The lethal effect of a benzamide derivative, 3-methoxybenzamide, can be suppressed by mutations within a cell division gene, ftsZ, in Bacillus subtilis. J. Bacteriol., 181 (4), 1348-1351. Oliva, M.A., Cordell, S.C., Löwe, J., 2004. Structural insights into FtsZ protofilament formation, Nat. Struct. Mol. Biol., 11 (12), 1243-1250. Oliva, M.A., Trambaiolo, D., Löwe, J., 2007. Structural insights into the conformational variability of FtsZ. J. Mol. Biol., 373 (5), 1229-1242. Perna, N.T., Plunkett 3rd, G., Burland, V., Mau, B., Glasner, J.D., Rose, D.J., Mayhew, G.F., Evans, P.S., Gregor, J., Kirkpatrick, H.A., Pósfai, G., Hackett, J., Klink, S., Boutin, A., Shao, Y., Miller, L., Grotbeck, E.J., Davis, N.W., Lim, A., Dimalanta, E.T., Potamousis, K.D., Apodaca, J., Anantha-raman, T.S., Lin, J., Yen, G., Schwartz, D.C., Welch, R.A., Blattner, F.R., 2001. Genome sequence of enterohaemorrhagic Escherichia coli O157:H7, Nature, 409 (6819), 529-533. Raymond, A., Lovell, S., Lorimer, D., Walchli, J., Mixon, M., Wallace, E., Thompkins, K., Archer, K., Burgin, A., Stewart, L., 2009. Combined protein construct and synthetic gene engineering for heterologous protein expression and crystallization using Gene Composer, BMC Biotechnol., 9 (37), doi: 10.1186/1472-6750-9-37. Raviglione, M.C., 2000. Issues facing TB control (7). Multiple drug-resistant tuberculosis, Scott. Med. J., 45, (5), 52-55. Respicio, L., Nair, P.A., Huang, Q., Anil, B., Tracz, S., Truglio, J.J., Kisker, C., Raleigh, D.P., Ojima, I., Knudson, D.L., Identification of FtsZ polymerization regulatory elements using aM. tuberculosis FtsZ temperature sensitive mutant. To be published, DOI:10.2210/pdb2q1y/pdb Roy, A., Kucukural, A., Zhang, Y. 2010. I-TASSER: a unified platform for automated protein structure and function prediction. Nat. Protoc., 5 (4), 725-738. Roy, A., Zhang, Y., 2011 COFACTOR: protein-ligand binding site predictions by global structure similarity match and local geometry refinement, - in print. The UniProt Consortium, 2008. The Universal Protein Resource (UniProt). Nucl Acids Res., 36, 190-195. Vaughan, S., Wickstead, B., Gull, K., Addinall, S.G., 2004. Molecular evolution of FtsZ protein sequences encoded within the genomes of archaea, bacteria, and eukaryote. J. Mol. Evol., 58 (1), 19-39. White, E.L., Ross, L.J., Reynolds, R.C., Seitz, L.E., Moore, G.D., Borhani, D.W., 2000. Slow polymerization ofM. tuberculosis FtsZ. J. Bacteriol., 182 (14), 4028-4034. White, E.L., Suling, W. J., Ross, L.J., Seitz, L.E., Reynolds, R.C., 2002. 2-alkoxycarbonylaminopyri-dines: inhibitors of M. tuberculosis FtsZ. J. Antimicrob Chemother., 50 (1), 111-114. Wu, S., Zhang, Y., 2007. LOMETS: a local meta-threading-server for protein structure prediction. Nucleic Acids Res., 35, 3375-3382. Yu, X.-C., Margolin, W., 1998. Inhibition of assembly of bacterial cell division protein FtsZ by the hydrophobic dye 5,5*-Bis-(8-anilino-1-aphthalenesulfonate). J. Biol. Chem., 273 (17), 10216-10222. Zhang, J., Zhang, Y., 2011. High-resolution protein structure refinement using fragment guided molecular dynamics simulations. Structure, in press. Zhang, Y., Kihara, D., Skolnick, J., 2002. Local energy landscape flattening: parallel hyperbolic Monte Carlo sampling of protein folding. Proteins, 48, 192-201. Zhang, Y., Skolnick, J., 2004. Scoring function for automated assessment of protein structure template quality. Proteins, 57 (4), 702-710. Zhou, H., Zhou, Y., 2002. Distance-scaled, finite ideal-gas reference state improves structure-derived potentials of mean force for structure selection and stability prediction. Protein Sci., 11, 2714-2726.