JBC Ideal method for primary cell transfection

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Originally published In Press as doi:10.1074/jbc.M703290200 on June 26, 2007

J. Biol. Chem., Vol. 282, Issue 37, 27334-27342, September 14, 2007
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Supplemental Data
Right arrow All Versions of this Article:
282/37/27334    most recent
M703290200v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Reddy, M. C. M.
Right arrow Articles by Sacchettini, J. C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Reddy, M. C. M.
Right arrow Articles by Sacchettini, J. C.
Social Bookmarking
Add to CiteULike   Add to Complore   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati  
What's this?

High Resolution Crystal Structures of Mycobacterium tuberculosis Adenosine Kinase

INSIGHTS INTO THE MECHANISM AND SPECIFICITY OF THIS NOVEL PROKARYOTIC ENZYME*Formula

Manchi C. M. Reddy1, Satheesh K. Palaninathan1, Nishant D. Shetty, Joshua L. Owen, Misty D. Watson, and James C. Sacchettini2

From the Department of Biochemistry & Biophysics, Texas A&M University, College Station, Texas 77843

Received for publication, April 19, 2007 , and in revised form, May 31, 2007.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS AND DISCUSSION
 REFERENCES
 
Adenosine kinase (ADK) catalyzes the phosphorylation of adenosine (Ado) to adenosine monophosphate (AMP). It is part of the purine salvage pathway that has been identified only in eukaryotes, with the single exception of Mycobacterium spp. Whereas it is not clear if Mycobacterium tuberculosis (Mtb) ADK is essential, it has been shown that the enzyme can selectively phosphorylate nucleoside analogs to produce products toxic to the cell. We have determined the crystal structure of Mtb ADK unliganded as well as ligand (Ado) bound at 1.5- and 1.9-Å resolution, respectively. The structure of the binary complexes with the inhibitor 2-fluoroadenosine (F-Ado) bound and with the adenosine 5'-(beta,{gamma}-methylene)triphosphate (AMP-PCP) (non-hydrolyzable ATP analog) bound were also solved at 1.9-Å resolution. These four structures indicate that Mtb ADK is a dimer formed by an extended beta sheet. The active site of the unliganded ADK is in an open conformation, and upon Ado binding a lid domain of the protein undergoes a large conformation change to close the active site. In the closed conformation, the lid forms direct interactions with the substrate and residues of the active site. Interestingly, AMP-PCP binding alone was not sufficient to produce the closed state of the enzyme. The binding mode of F-Ado was characterized to illustrate the role of additional non-bonding interactions in Mtb ADK compared with human ADK.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
EXPERIMENTAL PROCEDURES
 RESULTS AND DISCUSSION
 REFERENCES
 
Multidrug resistance in pathogenic microorganisms is increasing at an alarming rate, threatening human health around the world. Despite the World Health Organization's declaration over 10 years ago that tuberculosis was a global health emergency, the tuberculosis problem has worsened (1). Therefore, new antimycobacterial agents are needed to treat Mtb strains resistant to existing drugs and to shorten the treatment course to improve patient compliance (2, 3).

Nucleoside metabolic pathways are a good source of new targets for anti-bacterials, as the enzymes/pathways involved are often different from their human counterparts. One of the promising pathways for the development of novel nucleoside analogs with antitubercular activity is the purine salvage pathway. Mycobacteria preferentially use adenosine kinase (ADK)3 to directly phosphorylate Ado, even though they have purine nucleoside phosphorylase (Rv3307) and adenosine deaminase (Rv3313c) (4, 5). Therefore, ADK may represent a good bioactivator/prodrug target because of its inherent capability to phosphorylate therapeutically useful purine nucleoside analogs and convert them into active nucleosides (6).

ADK belongs to the phosphofructokinase B (PfkB) family of carbohydrate kinases, which includes ribokinase (RK), inosine-guanosine kinase, fructokinase, and 1-phosphofructokinase (7-9). However, Mtb ADK has a very low sequence similarity (less than 20%) with eukaryotic ADKs. Based on the amino acid sequence, Mtb ADK is closely related to RK from Escherichia coli (25% identity) (10). In E. coli, RK catalyzes the phosphorylation of ribose to ribose 5-phosphate, part of the pentose phosphate pathway. Both Mtb ADK and E. coli RK are dimers, whereas all other known ADKs, including human ADK, are monomers (11-13). Mtb ADK is encoded by the gene Rv2202c. It has recently been identified and cloned, and its gene product has been characterized biochemically (14).

Mtb ADK phosphorylates adenosine (Km = 0.8 µM) to adenosine monophosphate using ATP (14). Studies have shown that Ado analogs, like 2-methyladenosine, act as proinhibitors in Mtb, i.e. they require activation via phosphorylation before inhibiting other enzymes involved in DNA synthesis (15). The minimum inhibitory concentration of 2-methyladenosine on the cell growth of Mtb was determined to be 3.1 µg/ml (11 µM) (16). Indeed, the antitubercular activity of 2-methyladenosine is directly dependent on ADK activity (15, 16). It has been shown that ADK activity in 2-methyladoenosine- and 2-fluoroadenosine-resistant Mycobacterium smegmatis strains (SRI101 and SRI301) was significantly lower (<10 nmol/mg/min for SRI101 and <110 nmol/mg/min for SRI301) as compared with wild-type M. smegmatis mc2155 (3400 nmol/mg/min), suggesting that ADK is responsible for the activation of these nucleoside analogs to toxic metabolites (17). 2-Methyladenosine has also been shown to inhibit the growth of Mtb in infected macrophages in a hypoxic shift-down model (18). This suggests that nucleoside analogs phosphorylated by ADK may show inhibitory efficiency against Mtb in the persistent state (16).

2-Methyladenosine was shown to possess fairly selective activity against Mtb, signifying the differences in the substrate preferences between mycobacterial and human purine metabolic enzymes. Human ADK shows a Km of 960 µM for 2-methyladenosine, whereas Mtb ADK shows a Km of 79 µM (14). Besides 2-methyladenosine, several other nucleoside analogs with modifications to the adenine ring have been analyzed and found to possess promising antitubercular activity (19). For example, the fluorinated analogs, such as 2-fluoro-3-deaza-adenosine, 3-fluoro-3-deaza-adenosine, and 2,3-difluoro-3-deaza-adenosine, not only exhibited antitubercular activity but also proved to be better substrates for Mtb ADK than human ADK (20-23). The substrate-specific activities of 2-fluoro-3-deaza-adenosine, 3-fluoro-3-deaza-adenosine, and 2,3-difluoro-3-deaza-adenosine for Mtb ADK were 63, 16, and 81 nmol/mg/min, respectively; whereas for human ADK they were <0.5, 1, and 6 nmol/mg/min (23), indicating that differences exist between the substrate binding sites of these enzymes.

To provide further insight into the structure, function, and inhibition of Mtb ADK, we have solved the crystal structure of Mtb ADK to 1.5-Å resolution using multiwavelength anomalous dispersion (MAD). We also report the structural basis for the binding of adenosine, ATP analog (AMP-PCP), and 2-fluoroadenosine to ADK through three binary complex structures.


   EXPERIMENTAL PROCEDURES
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
RESULTS AND DISCUSSION
 REFERENCES
 
The substrates and their analogs (adenosine, 2-fluoroadenosine, and AMP-PCP) were obtained from Sigma. A Hi-Trap nickel chelating column from GE Healthcare was used for protein purification. Crystal screen solutions from Hampton Research and Emerald Biosciences were used for obtaining ADK crystals.

Cloning, Protein Expression, and Purification—A 972-bp DNA fragment containing the ADK (Rv2202c accession no. Q10391) gene was amplified by PCR with Mtb H37Rv genomic DNA as a template. The following oligonucleotides were used: 5'-AGATGAAGCATATGACGATCGCGGTAACCGGTTC-3' and 5'-AGATGACTCGAGTTAGGCCAGCACGGCGACGATCTC-3'. The amplified DNA fragment was digested with NdeI and XhoI restriction enzymes and subcloned into the corresponding restriction sites in the pET28b vector containing an N-terminal His tag (Novagen). The DNA was subcloned using NdeI and EcoRI restriction sites into the pET28b vector (Novagen) to generate a recombinant vector containing a 5' sequence encoding a 20-amino acid N-terminal His tag and a tobacco etch virus cleavage site. After ligation, the plasmid was transformed into E. coli BL21(DE3) for expression of apo-protein and E. coli B834(DE3) cells for producing selenomethionine-incorporated protein. For protein expression, cell cultures were grown in LB media at 37 °C. The cells were induced with 0.75 mM isopropyl 1-thio-beta-D-galactopyranoside when the cell density reached A600 ~ 0.6-1.0. Cell cultures were incubated for about 18 h at 20 °C before harvest. Selenomethionylated (SeMet) protein was prepared according to published methods (24).

The harvested cells were lysed using a French press, and the cell suspension was centrifuged at 15,000 x g for 1 h. Recombinant ADK protein was purified by a nickel affinity column, making use of the engineered His6 tag at the N terminus. For purification buffers, 20 mM Tris-HCl, pH 7.5, containing 500 mM NaCl, 10-500 mM imidazole, and 5% glycerol were used. The His tag was then cleaved by dialyzing the protein (in the presence of 20 mM Tris-HCl, pH 7.5, 50 mM NaCl, 5% glycerol, and 1 mM dithiothreitol) with tobacco etch virus protease. The tobacco etch virus and tag were then separated from ADK by passing the dialyzed sample through another nickel affinity chromatography column. Purified ADK was dialyzed a second time in the same composition buffer as before. Purified recombinant protein was concentrated to 20 mg/ml prior to crystallization. The protein was more than 95% pure as observed by SDS-PAGE.

Crystallization—Crystals were obtained by either hanging drop or sitting drop vapor diffusion methods. A protein concentration of 15-20 mg/ml was mixed with an equal volume (2 µl) of precipitant solution and incubated at 19 °C, forming a 4-5-µl drop. Apo-SeMet ADK was found to crystallize in 1.0 M potassium/sodium tartrate and 0.1 M MES, pH 6.0, by screening with Crystal Screen I and II (Hampton) and Wizard I and II (Emerald Biosciences). The crystals of the apo-ADK and ADK·AMP-PCP complex were obtained by mixing equal volumes (2 µl) of 18 mg/ml protein (preincubated for 2 h at room temperature with 5 mM AMP-PCP) with 20% PEG 8000, 100 mM sodium cacodylate, pH 6.5, and 200 mM magnesium acetate tetrahydrate. For the substrate-bound protein crystals, ADK was preincubated with 5 mM adenosine/2-fluoroadenosine for 1 h at room temperature prior to setting up crystal plates. Crystals were obtained in conditions containing 0.1 M HEPES-Na, pH 7.5, and 1.4 M trisodium citrate dihydrate.

Data Collection, Structure Determination, and Refinement—The initial native data set for Mtb apo-ADK was collected at 1.5-Å resolution at APS-23ID (Advanced Photon Source) in the C2 space group (Table 1). Molecular replacement of the apo-ADK data were attempted with a search model of 2-keto-3-deoxygluconate kinase from Thermus thermophilus (Protein Data Bank code 1v1s (25)) using various programs (PHASER, MOLREP, EPMR, and AMORE). However, no reasonable solution was obtained from the whole molecule or separate domains. Therefore, experimental phases from the SeMet incorporated apo-ADK crystal were obtained using the MAD phasing method (26) (supplementary materials Table 1). The SeMet ADK crystal diffracted up to 2.4-Å resolution at APS-23ID in the P4122 space group. Seven selenium sites were found using SHELXD with three different wavelength MAD data. SOLVE/RESOLVE (27) were used to refine the sites, calculate initial protein phases, and build the model. A final model was obtained after several cycles of manual model building using XTALVIEW (28). This 2.4-Å model was then used as a search model against 1.5-Å resolution data using PHASER (29) in the C2 space group. The solution for two molecules of the C2 asymmetric unit was refined using REFMAC (30) and built using XTALVIEW. Data sets of binary complexes of ADK·Ado, ADK·F-Ado, and ADK·AMP-PCP were collected on our home source using an R-axis IV imaging plate system and rotating anode x-ray generator equipped with osmic mirrors. All the ADK·ligand complex structures were solved with the molecular replacement method by PHASER using Molecule A of the 1.5-Å apo-ADK structure as a search model. Bias-minimized electron density maps were obtained using the Shake&wARP protocol (31). Clear electron density for Ado, 2-fluoroadenosine (F-Ado), and AMP-PCP (Fig. 1A and supplementary materials Fig. S1) was visible in the SNW map prior to any model building. Several cycles of manual model building and maximum likelihood refinement in REFMAC5 yielded the final model.


View this table:
[in this window]
[in a new window]

  		TABLE 1
Data collection and refinement statistics for apo and ligand bound ADK structures

Crystallographic statistics of Mtb apo-ADK and binary complex structures. Values in parentheses are for high resolution shells.

 
All the crystals were mounted in cryo-loops and were flash-cooled in a liquid N2 stream after brief soaks in N-paratone oil. The binary complex data sets were processed using CRYSTAL CLEAR (Molecular Structure Corporation, Inc., Woodlands, TX), and the apo-ADK data sets were processed using HKL2000 (32) (Table 1). Solvent molecules were identified using Xfit (28).

The final models of the apo-ADK, ADK·adenosine, ADK·2-fluoroadenosine, and ADK·AMP-PCP structures have excellent stereochemistry, as determined by Ramachandaran plot (calculated by PROCHECK (33, 34)), and the crystal data and final refinement statistics are given in Table 1. The figures were prepared using Spock (35), Raster3D (36), and Chimera (37). The atomic coordinates of apo and binary complexes have been deposited in the Protein Data Bank with the codes 2PKF (apo), 2PKM (Ado), 2PKK (F-Ado), and 2PKN (AMP-PCP).


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS AND DISCUSSION
REFERENCES
 
The x-ray crystal structure of Mtb apo-ADK with one molecule in the asymmetric unit was solved using the MAD method with crystals of selenomethionylated protein formed in the space group P4122 at 2.4-Å resolution. Apo-ADK also crystallized in the C2 space group with two molecules (A and B) in the asymmetric unit, and these crystals diffracted to 1.5-Å resolution. The comparison of both subunits of C2 apo-ADK with the SeMet ADK reveals that the overall structure of these protomer was not severely affected by their crystal packing; the r.m.s. deviation between these protomer (after omitting their crystal contact region ({alpha}2, {alpha}7-9)) is less than 0.2 Å. ADK co-crystals were produced after the incubation of protein with adenosine (Ado), AMP-PCP, or 2-fluoroadenosine (F-Ado) (Table 1). The ADK·AMP-PCP crystal was in the C2 space group and diffracted to 1.9-Å resolution; ADK·Ado and ADK·F-Ado crystals are in the P3121 space group and diffracted to 1.9-Å resolution. All of the ADK·ligand complex structures were solved with the molecular replacement method using the 1.5-Å resolution apo-ADK subunit as a search model.

Overall Structure—Like the E. coli RK and human ADK structures, each subunit of Mtb ADK contains a small lid domain (residues 9 to 48 and 99 to 120) and a large {alpha}/beta domain (residues 1 to 8, 49 to 98, and 121 to 323) (Fig. 1B). The small lid domain is formed by 5 beta strands (beta2, beta3, beta4, beta8, and beta9) and a short helix, {alpha}1. The large domain is an {alpha}/beta domain consisting of {alpha}(9-14)/beta (12-15) and {alpha}(3-8)/beta (1, 5, 6, 7, 10, 11). The beta1, beta5, beta6, beta10, and beta11 strands are all parallel, reside in approximately the same plane, and associate with {alpha}(3-8). beta Strands 12-15 are also roughly planar and associate with {alpha}(9-14). The large domain makes no significant interactions with the small domain in the apo-structure.


Figure 1
View larger version (36K):
[in this window]
[in a new window]

  		FIGURE 1.
A, electron density of (from left to right) Ado, F-Ado, and AMP-PCP complexed with Mtb ADK. Electron density of Shake&wARP (31) omit maps are contoured at the 1 {sigma} level. Ligand molecules were omitted from the model before the map calculation. The blob feature in XTALVIEW has been applied to limit the electron density display within 2.0 Å of the ligand and the final figure is rendered with Raster3D (36). B, structure of the Mtb ADK subunit. Ribbon representation of Mtb ADK subunit as observed in the apo-crystal structure. The secondary structure elements were calculated using DSSP (44); the helices are colored in golden yellow, beta strands in magenta, and loops in green. The lid domain (formed by beta2, beta3, beta4, beta8, beta9, and {alpha}1) is distinct from the large {alpha}/beta domain. The loops LA and LB are defined as large and small ATP binding loops, respectively, as per the nomenclature of E. coli RK (11).

 
In the crystal lattice of the apo-structure, each independent molecule of the asymmetric unit forms a tight, intact dimer with a twofold symmetry related subunit (Fig. 2). The majority of the dimer interface results from the interaction of the beta strands of the small lid domain that protrude from the large {alpha}/beta domain of each subunit. Four of the beta strands (beta2, beta4, beta8, and beta9) of the lid domain are more or less planar, whereas beta3 points away from this plane forming no intrasubunit interactions. However, the beta3' strand of the symmetry related subunit is anti-parallel with beta9, completing a 5 beta-stranded extended sheet (beta2, beta4, beta8, beta9, and beta3'). These beta strands pack close to each other, forming a flattened beta barrel at the center of the dimer. The dimer interface of the 5 beta-stranded sheet of each subunit is formed primarily via hydrophobic interactions (see supplementary materials Table 2 for distances) involving residues Met15, Phe21, Leu25, Leu30, Val33, Phe37, Leu43, Met44, Thr103, Ile113, and Ser115. Apart from the hydrophobic interactions, residues His13, Arg101', His13', and Arg101 of both subunits interact with each other through hydrogen bond interactions, forming a positive patch of electrostatic surface potential at the bottom of the beta barrel. The significance of this positive surface is not clear at this point; however, it is possible that this region might bind other molecules like inorganic phosphates. It has been previously shown that inorganic phosphates can act as activators of mammalian ADK (38). In both subunits of the apo-enzyme dimer, the substrate binding pocket is completely open and is freely accessible to solvent; hence, we refer to the apo-ADK structure as an "open dimer" (Fig. 2A).

The ADK·AMP-PCP Binary Complex Is Very Similar to the Apo-ADK Structure—The structure of the ADK·AMP-PCP complex reveals that each subunit of the dimer binds to a single AMP-PCP molecule. The ADK·AMP-PCP dimer is very similar to that of apo-ADK (r.m.s. deviation of the C{alpha}s of the dimers is 0.75 Å). This indicates that the binding of AMP-PCP has no significant effect on the relative orientation of the substrate binding pocket, lid domain, or dimer interface. AMP-PCP is located in a shallow cleft formed by the ends of beta strands beta12, beta13, beta14, and beta15, and {alpha} helices {alpha}11 and {alpha}12 on the enzyme surface. The small loop between beta13 and beta14 is named "small ATP binding loop," and the loop between beta15 and {alpha}11 is named "large ATP binding loop" (according to the E. coli RK nomenclature). The {gamma}-phosphate of the AMP-PCP is positioned at the end of this shallow binding pocket, ~5 Å away from the Ado binding pocket (Fig. 4A). The adenine and ribose rings of AMP-PCP are partially buried in a hydrophobic pocket formed by the side chains of residues Pro226, Val243, Glu246, Val255, Ala284, and Leu288. The N-1 and N-3 atoms of the base form water-mediated hydrogen bonds with the main chain nitrogen atom of Glu246 and main chain oxygen atom of Ser281, respectively. The Glu246 side chain is disordered in the electron density map; however, the backbone of Glu246 is positioned so that the side chain could interact with either N-1, N-6, or both of the base of AMP-PCP in one orientation. The ribose sugar is stacked with the side chain of Phe259 4.4 Å away. The O-3' atom of the ribose sugar makes a direct hydrogen bond (2.8 Å) with the main chain oxygen of Gly228, whereas O-2' makes a water-mediated interaction with the side chain oxygen of Ser281 (Fig. 4A). The phosphate groups of AMP-PCP form hydrogen bonds either directly with the protein or through intervening solvent molecules. Residues that interact with the phosphate groups are Thr223 (the side chain {gamma}-O is 3.0 Å from the O-2 of the {alpha}-phosphate), Gly225 (the main chain nitrogen is 2.7 Å from the O-2 of the {alpha}-phosphate), and Asn195 (the side chain {gamma}-N is 3.2 Å from the O-1 of the {alpha}-phosphate).


Figure 2
View larger version (53K):
[in this window]
[in a new window]

  		FIGURE 2.
The open and closed dimers of Mtb ADK. A, ribbon representation of the open dimer of Mtb apo-ADK shown with the electrostatic potential surface. The lid domain of each subunit exclusively forms the dimer interface. The beta sheets of the lid domain of both subunits interact with each other to form a flattened beta barrel at the center of the dimer. The arrows point to the Ado binding pocket that is situated beneath the lid domain. As a result of the orthogonal arrangement of the lid domain to the large {alpha}/beta domain, the Ado binding pocket is open and freely accessible in apo-ADK. B, ribbon representation of the closed dimer of Mtb ADK·Ado binary complex structure shown with the electrostatic potential surface. The relative arrangement of the lid to the large {alpha}/beta domain is quiet different compared with the native dimer. The Ado (shown as stick) is completely buried below the lid domain. C, comparison of the subunit of apo (yellow) and adenosine bound Mtb ADK (blue) binary structures. The {alpha}/beta domain of the apo structure is superimposed on the ADK·Ado binary complex. The beta3' strand from the other subunit that completes the 5 beta-stranded lid domain is drawn in red for the apo form and green for the Ado bound binary complex. The beta3 and beta3' strands extend from the lid domain beta sheets and contribute to the formation of the intact dimer in both apo- and Ado-bound ADK. The curved arrows indicate the movement of the lid for closure. When closed, the lid forms direct interactions with the Ado molecule. Ser36' of the beta3' strand (shown in magenta stick in the red and green strand) moves a distance of 17 Å, when the closed and apo forms are compared, to form a hydrogen bond with Ado. Phe116 of the beta9 strand (shown in "black" stick) moves 4 Å to stack with Ado in the closed form. Residues Phe102 (beta8) and Asp12 (beta2) of the lid domain also interact with the Ado (as detailed in supplementary materials Table 2) and are not shown here for clarity. The structures are superimposed with respect to the {alpha}/beta domain.

 
Crystal Structure of the ADK·Ado Binary Complex, the Closed Form—The structure of Mtb ADK complexed with its substrate Ado, reveals new insights into the role of dimerization in substrate binding. The structures of the individual domains of apo and ADK·Ado are very similar to each other (the r.m.s. deviation between C{alpha}s of the large domain of the Ado bound form and the apo form is 0.55 Å, r.m.s. deviation for the lid domain is 1.75 Å). However, the relative positioning of the lid to the base is very different in the two forms (Fig. 2C). In addition, the intact ADK.Ado dimer is quite different compared with that of the open form of apo-ADK, as shown in Fig. 2 (r.m.s. deviation of the C{alpha}s of the apo dimer and ADK.Ado dimer is 11.0 Å). Unlike the tight beta barrel formation of the open dimer, the 5 beta-stranded lids of the dimer move apart toward the large {alpha}/beta domain to form what we refer to as a "closed dimer" (Fig. 2B). The residues contributing to the intersubunit interactions of the beta barrel (His13, Met15, Phe21, Leu25, Leu26, Leu30, Val33, Leu35, Phe37, Leu42, Met44, Arg101, Thr103, Ile113, and Ser115 of both subunits) undergo significant displacement and several conformational changes to account for the formation of the closed dimer. In fact, the side chain atoms of residues Leu25, Leu26, Leu30, and His31 ({alpha}1) of the dimer are no longer ordered in the closed form. This is due to the large movement of the {alpha}1 helix in the closed form. For example, Leu43 (O) is 11.2 Å away from the His98' (NE) in the closed form, and they are 3.4 Å away from each other in the open dimer. The closed form loses four intermolecular hydrogen bonds and gains one new hydrogen bond compared with that of the open dimer (supplementary materials Table 2). These dramatic domain movements, rearrangement of residues, and formation of the closed dimer facilitate the direct interactions of the lid region with the substrate as described below (Fig. 2C).

Binding Interactions of Ado—In the ADK·Ado binary complex structure, a molecule of adenosine is buried in an active site under the 5 beta-stranded lid region of each protomer. The strongest stabilizing factors for the Ado in the binding pocket is the stacking of the side chain of Phe116 with the adenine ring moiety and the two hydrogen bond interactions of Asp12 with the O-3' and O-2' of the ribose ring. The side chain Phe102 stacks in between the C-2' atom of the ribose ring (CD2-C-2' distance is 3.6 Å) and the C-8 atom of Ado (at 3.2 Å distance). The N-7 atom of the base makes a hydrogen bond (2.7 Å) to {gamma}-O of Ser36' (beta3') of the symmetry related subunit of the dimer. Ser36' also makes weak hydrogen bond interactions with the N-6 atom of the Ado base (3.4 Å) and with the main chain oxygen of Gln172 (3.3 Å). Markedly, residues Asp12 (beta2), Ser36' (beta3'), Phe102 (beta8), and Phe116 (beta9) all belong to the lower part of the lid region. These residues are not close enough to make direct interactions with Ado in the open form. These residues also undergo large movements in response to Ado binding (Asp12 moves 2.4 Å, Ser36' moves 17 Å, Phe102 moves 5.7 Å, and Phe116 moves 4.2 Å, as detailed in Table 2) (Fig. 2C). Thus, it is evident that the loop closure is a critical requirement for the binding of the Ado substrate and the subsequent catalysis.


View this table:
[in this window]
[in a new window]

  		TABLE 2
Comparison of hydrogen bonds and close contacts involving adenosine (<3.6 Å) binding of human and Mtb ADK

Bold residues belong to the small lid domain of Mtb ADK.

 
Comparison with E. coli RK—The overall structure of Mtb ADK is similar to E. coli RK (Fig. 3A). The dimerization through the formation of the 10 beta-stranded beta barrel arrangement is comparable with the 8 beta-stranded beta clasp of the E. coli RK structure. Whereas notable lid movements were also observed in E. coli RK, compared with that of its native open form they were smaller in magnitude (39). As described in the Mtb apo-ADK structure discussion, the lid domain moves between 2.2 and 17.0 Å upon the binding of Ado. In contrast, the lid domain of E. coli RK undergoes a movement that is only about 1.0 Å after binding ribose. Fig. 3A shows the superimposition of the subunits of Mtb apo-ADK, apo-RK, and RK·ribose/AMP-PCP. It is evident from this superposition that the lid domain (particularly beta8 and beta9) is more orthogonal to the {alpha}/beta domain in the Mtb apo-ADK structure compared with that of E. coli apo-RK. The open and closed dimerization is very distinct in Mtb ADK. Superimposition of the Mtb ADK dimer with the apo or ternary complex dimer of E. coli RK yielded r.m.s. deviations of more than 7.0 Å, confirming that the relative arrangement of the lid and {alpha}/beta domain in Mtb ADK is quite different compared with that of E. coli RK. The axis defined by the central beta barrel of the apo-Mtb ADK is at a right angle to the domain and is located distinctly away from the substrate binding pockets, whereas in the E. coli apo-RK structure the lid domains were curved toward the {alpha}/beta domain. The E. coli apo-RK buried 7.4% of the surface area per subunit, and the substrate-bound RK complex structure buried 8.4%. Whereas Mtb apo-ADK buried 11.9% surface area per subunit, and the ADK·Ado binary complex buried 14.9%. As a result, the substrate binding pocket of Mtb apo-ADK is completely accessible, whereas it has limited accessibility in E. coli apo-RK.

Comparison of the Mtb ADK·Ado binary complex subunit with the ternary complex structure of E. coli RK (PDB code 1RKD) gives r.m.s. deviation of 1.3 Å for the C{alpha} carbons (198 C{alpha}s were compared). Similar comparison with the apo-RK structure gives a r.m.s. deviation of 1.5 Å. The most significant structural differences are found at two insertion regions and one deletion region of Mtb ADK compared with E. coli RK. The first insertion of 17 residues (residues 20-36) is observed in the beta3, {alpha}1 of the lid domain. Another difference between E. coli RK and Mtb ADK is evident at the C-terminal insertion in Mtb ADK (309-323 at {alpha}14). This C-terminal helix {alpha}14 makes a direct interaction with the ATP binding helix {alpha}12 and a large ATP binding loop (Arg306 forms a salt bridge with Glu289, OH of Tyr311 forms a hydrogen bond with the backbone oxygen of Gly242, and Tyr311 makes hydrophobic interaction with Val241) and accounts for the 12-residue deletion near the small ATP binding loop. The third difference is a 12-amino acid deletion of Mtb ADK (residues 196-208), compared with that of E. coli RK. In E. coli RK, Asp202 of this insertion makes a salt bridge (2.7 and 3.2 Å) with the Arg227 of the small ATP binding loop and may allosterically affect the ATP binding. Moreover, the backbone torsion of residue 226 of E. coli RK shows a maximum movement upon the binding of ATP (39).

A comparison of the apo and the ternary complex structure of E. coli RK revealed significant changes in the large and small ATP binding loops (11, 39) upon substrate binding. The large ATP binding loop of the E. coli RK structure has been shown to interact with the lid region upon the binding of the substrates and displayed significant conformational changes. It has been proposed that in E. coli RK, ribose is expected to bind the open form first and stabilize the closed conformation, whereas the lid region allosterically prepares the ATP binding pocket for the reaction (39). Mtb ADK does not seem to follow this induced fit mechanism proposed for E. coli RK, as there are no significant conformational changes evident in either of the ATP binding loops between the apo and binary Mtb ADK structures. Interestingly, AMP-PCP binds to ADK in the open form, proving that the closed form is not a prerequisite for ATP binding. Moreover, the ADK·Ado binary complex forms a closed dimer, thus limiting the accessibility of the ATP binding pocket.


Figure 3
View larger version (60K):
[in this window]
[in a new window]

  		FIGURE 3.
Comparison with E. coli RK and human ADK. A, overlay of the apo (green) and ternary complex (golden yellow) structures of E. coli RK on the Mtb apo-ADK subunit (cyan). The structurally conserved regions of the {alpha}/beta domain of the apo and ternary complex structures of E. coli RK (1RKA and 1RKD, respectively) are superimposed on Mtb ADK using the homology module of Insight II (Accelrys Inc.). The lid domain of apo-Mtb ADK is more open compared with that of E. coli apo-RK. Two insertions in Mtb ADK compared with E. coli RK are labeled as I1(Mtb) for the lid insertion and I2(Mtb) for the C-terminal insertion, whereas I(RK) points to the insertion in E. coli RK. The {alpha}/beta domain of apo E. coli is very similar to the {alpha}/beta domain of the E. coli ternary complex; it is not shown for clarity. B, overlay of the human ADK structure (yellow, PDB code 1BX4) on the Mtb apo-ADK subunit (cyan). The {alpha}2 helix of human ADK runs across its beta sheet of the lid domain and rules out the possibility of dimer formation in human ADK.

 
Comparison with Human ADK—Although the sequences of the Mtb and human ADK are less than 20% identical, their overall structures are similar (12). Fig. 3B shows the comparison of the human ADK·(Ado)2 complex with the Mtb apo-ADK. The large {alpha}/beta domains of both structures were similar to each other in the overlaid regions. The human ADK lid domain is composed of 4 beta strands and 2 {alpha} helices. Interestingly, the {alpha}2 helix of human ADK runs across its beta sheet of the lid domain and rules out the possibility of dimer formation (Fig. 3B). Mtb ADK and E. coli RK are dimers with only one helix in the lid domain.

ADK Active Site—The structure of the active ADK binary complexes indicates that the enzyme has a pre-formed anion hole. It has been shown that the formation of an anion hole immediately adjacent to the {gamma}-phosphate of the ATP is one of the structural requirements for the reaction mechanism (40, 41). Previous studies have also proposed that ATP binding might play an important role in inducing the local conformational changes near the {gamma}-phosphate to favor the anion hole formation (41). However, the conformation of the anion hole is identical in apo and ADK·Ado (see below) or ADK·AMP-PCP binary complex structures. Neither the AMP-PCP nor Ado binding seem to play any role in the formation of the anion hole.

Attempts to crystallize ADK with Ado and AMP-PCP or to soak AMP-PCP into crystals (in presence of MgCl2), and vice versa, were unsuccessful. Therefore, we have constructed a model of the ternary complex simply by superimposing the ADK·Ado and ADK·AMP-PCP binary complex structures. Fig. 4A shows the superposition of the binding pockets of ADK·AMP-PCP (non-hydrolyzable ATP analog) and ADK·Ado. Based on the constructed ternary complex model, the distance between the {gamma}-phosphorus of AMP-PCP and the O-5' atom of Ado is 5.6 Å. Comparable atoms in ternary complexes of phosphofructokinase are separated by 4.3-5.9 Å (42) and 5.6 Å in the E. coli RK (11). These distances are little less than ideal for an optimal phosphoryl transfer reaction. It has been suggested that these kinases will undergo minor structural rearrangement to force the ATP toward the substrate binding pocket (11, 42). Interestingly, two water molecules are conserved near the {gamma}-phosphate in both binary complex structures of Mtb ADK (labeled W1 and W2 in Fig. 4A). The active site conserved water molecule W-1 is well ordered in the anion hole, created by the combination of the main chain amide nitrogen of Val255 and Gly256. W-2 makes a hydrogen bond with the oxygen atom of both beta- and {gamma}-phosphates of AMP-PCP. Interestingly, W-2 is observed close to the proposed conserved magnesium binding motif, N195XXE198, for ADKs (43). In the Mtb ADK·AMP-PCP binary complex structure, the OE-2 atom of Glu198 is at 3.6 Å from the W-2 atom and ND-2 of Asn195 is at 4.4 Å. Although there is no strong electron density corresponding to magnesium ion binding in the Mtb ADK·AMP-PCP binary complex structure, it is possible that Mg2+ can either replace W-2 or bind between W-2 and Asn195 and Glu198 in the ternary complex structure.

In all known ADKs, an aspartic acid residue acts as a base to deprotonate the 5'-hydroxyl group of adenosine to activate it for nucleophilic attack on the {gamma}-phosphate of ATP (40). Mtb ADK uses Asp257 as the base to abstract a proton from the 5'-hydroxyl group of ribose. The negatively charged O-5' atom (at 2.6 Å from the OD-2 Asp257) could then make a water-mediated (W-1) nucleophilic attack on the {gamma}-phosphate of ATP. The resulting charge difference on the water can be nullified by an anion hole created by the main chain nitrogens of residues 255-257. This state may be further stabilized by the side chains of Gln172 and Asn252. Thus, as a consequence of water-mediated phosphorylation, the products AMP and ADP would be formed via an Sn2 reaction. The second water molecule, W-2, might also play a role in stabilizing the products. In contrast, based on proposed reaction mechanisms of all other ADK structures, the {gamma}-phosphorus should make direct contact with both the anion hole and the O-5' of the Ado. However, one cannot completely rule out the possibility of direct phosphorylation for Mtb ADK in which the localized conformational change of the binding pocket residues will push the ATP closer to the Ado so that the {gamma}-phosphate will replace W-1. For example, a blocked anion hole has been observed in the crystal structure of human ADK. It has been argued that the human enzyme needs to undergo local conformational changes to form an active anion hole and, in turn, an active monomer (12). In human and its close homolog apo Toxoplasma gondii ADK structure, the center of the oxyanion hole is blocked by the carbonyl oxygen of their -3rd residue (13) (comparable with the -3rd residue from Asp257 of Mtb ADK). Comparison of active and inactive anion holes of ADK is illustrated in supplementary materials Fig. S2. In the ternary complex structure of T. gondii, the -3rd residue moves away to form an unblocked anion hole. In contrast, the anion hole is well formed in Mtb ADK and is not blocked in either the apo- or substrate-bound structures and, therefore, does not need to undergo large conformational changes to perform the phosphorylation of Ado.


Figure 4
View larger version (29K):
[in this window]
[in a new window]

  		FIGURE 4.
ADK active site. A, ternary complex model of Mtb ADK was generated by superposing the binding pocket residues of the ADK·AMP-PCP (non-hydrolyzable ATP analog) binary complex structure onto the ADK·Ado binary complex structure. Stereo view of the key residues of the Mtb ADK active site of the ternary complex model is shown here. The water molecule W-1 hydrogen bonds with the O-5' atom of Ado, the {gamma}-phosphorus group of the AMP-PCP, and the oxyanion hole formed by residues 255-257 (shown in dotted lines). Although the distance between the{gamma}-phosphorus of AMP-PCP and the O-5' atom of Ado is (5.6 Å) not optimum for the phosphorylation, a water-mediated reaction is possible through W-1. B, superposition of the F-Ado binding pocket (yellow) of the Mtb ADK binary complex structure on the Mtb ADK·Ado complex structure (black). The interactions between the 2-fluoro atom and the protomer residues are shown in dotted lines.

 


Figure 5
View larger version (23K):
[in this window]
[in a new window]

  		FIGURE 5.
Comparison of the Ado binding pocket of human (A) and Mtb (B) ADK structures. None of the lid domain residues of human ADK is interacting with the Ado. In contrast, the lid domain residues Asp12, Phe102, Phe116, and Ser36 of Mtb ADK are directly involved in the Ado binding.

 
ADK·F-Ado Binary Complex Structure—Previous studies have shown that nucleoside analogs, like F-Ado and 2-methyladenosine, are preferred substrates to Mtb ADK compared with human ADK (15, 19). Recently, a series of halogenated 3-deaza-Ado analogs were shown to have ADK-dependent antitubercular activity with a minimum inhibitory concentration of 1.6 mg/liter (23). F-Ado readily yielded diffraction quality crystals in similar crystallization conditions as Ado-bound co-crystals. However, methyladenosine co-crystallization was unsuccessful. The overall structure of the ADK·F-Ado binary complex is very similar to that of the ADK·Ado complex structure (r.m.s. deviation for C{alpha}s is 0.20 Å). The superposition of Ado and F-Ado binding pockets in Mtb ADK were almost identical, suggesting that the binding of F-Ado does not alter the binding pocket relative to Ado (Fig. 4B). The significance of F-Ado lies in its complementarity to the ADK binding pocket. The 2-fluoro substitution forms seven additional non-bonding interactions with the binding pocket residues without altering the conformation of active site residues (Fig. 4B and Table 2). F-Ado also positions its O-5' atom in an identical position compared with that of the Ado complex structure for a phosphorylation reaction to produce the 2-fluoro-substituted product (2-fluoroadenosine monophosphate). Earlier studies have also shown that 2-methyladenosine (a close analog of F-Ado) is monophosphorylated intracellularly by ADK in Mtb (15). The Ado binding pocket of Mtb ADK is quiet different when compared with that of human ADK, as shown in Fig. 5. In a hypothetical model of F-Ado in the human structure, the 2-fluoro substitution can only make two additional interactions with the protomer (Table 2). This noticeable difference in the binding pocket of F-Ado between human and Mtb ADK was also reflected in the biochemical structure-activity relationship studies (19, 23). Thus, F-Ado can bind more specifically and strongly to Mtb ADK than it can to human ADK.


   FOOTNOTES
 
The atomic coordinates and structure factors (codes 2PKF, 2PKM, 2PKK, and 2PKN) have been deposited in the Protein Data Bank, Research Collaboratory for Structural Bioinformatics, Rutgers University, New Brunswick, NJ (http://www.rcsb.org/).

* This work was supported by Structural Genomics of Persistence Targets from Mycobacterium tuberculosis Grant PO1 AI 68135 from the National Institutes of Health, R. J. Wolfe-Welch Foundation Chair in Science Grants A-0015, and NSF-0521553. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Back

Formula The on-line version of this article (available at http://www.jbc.org) contains supplemental Tables S1-S2 and Figs. S1-S2. Back

1 Both authors contributed equally to this work. Back

2 To whom correspondence should be addressed. Tel.: 979-862-7636; Fax: 979-862-7638; E-mail: sacchett{at}tamu.edu.

3 The abbreviations used are: ADK, adenosine kinase; Mtb ADK, adenosine kinase from Mycobacterium tuberculosis; MAD, multiwavelength anomalous diffraction; r.m.s. deviation, root mean square deviation; AMP-PCP, adenosine 5'-(beta,{gamma}-methylene)triphosphate; F-Ado, 2-fluoroadenosine; RK, ribokinase; MES, 4-morpholineethanesulfonic acid; SeMet, selenomethionine; F-Ado, 2-fluoroadenosine. Back


   ACKNOWLEDGMENTS
 
We thank Arockiasamy Arulandu and the scientists of beamline 23-ID at the Advanced Photon Source (APS) of the Argonne National Laboratory (ANL) for help in data collection. We acknowledge Leslie Nicosia and Siaska Castro for comments on the manuscript and Dmitriy Verkhoturov for technical assistance. We thank Robert C. Reynolds (SRI) for providing 2-methyladenosine.



   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS AND DISCUSSION
 REFERENCES
 

  1. World Health Forum (1993) World Health Forum 14, 438[Medline] [Order article via Infotrieve]
  2. Basso, L. A., Zheng, R., Musser, J. M., Jacobs, W. R., Jr., and Blanchard, J. S. (1998) J. Infect. Dis. 178, 769-775[Medline] [Order article via Infotrieve]
  3. Duncan, K., and Sacchettini, J. C. (2000) in Molecular Genetics of Myco-bacteria (Hatfull, G. F., and Jacobs, W. R., Jr., eds) pp. 297-312, ASM Press, Washington, D. C.
  4. Basso, L. A., Santos, D. S., Shi, W., Furneaux, R. H., Tyler, P. C., Schramm, V. L., and Blanchard, J. S. (2001) Biochemistry 40, 8196-8203[CrossRef][Medline] [Order article via Infotrieve]
  5. Cole, S. T., Brosch, R., Parkhill, J., Garnier, T., Churcher, C., Harris, D., Gordon, S. V., Eiglmeier, K., Gas, S., Barry, C. E., 3rd, Tekaia, F., Badcock, K., Basham, D., Brown, D., Chillingworth, T., Connor, R., Davies, R., Devlin, K., Feltwell, T., Gentles, S., Hamlin, N., Holroyd, S., Hornsby, T., Jagels, K., Krogh, A., McLean, J., Moule, S., Murphy, L., Oliver, K., Osborne, J., Quail, M. A., Rajandream, M. A., Rogers, J., Rutter, S., Seeger, K., Skelton, J., Squares, R., Squares, S., Sulston, J. E., Taylor, K., Whitehead, S., and Barrell, B. G. (1998) Nature 393, 537-544[CrossRef][Medline] [Order article via Infotrieve]
  6. Miller, R. L., Adamczyk, D. L., Miller, W. H., Koszalka, G. W., Rideout, J. L., Beacham, L. M., 3rd, Chao, E. Y., Haggerty, J. J., Krenitsky, T. A., and Elion, G. B. (1979) J. Biol. Chem. 254, 2346-2352[Free Full Text]
  7. Wu, L. F., Reizer, A., Reizer, J., Cai, B., Tomich, J. M., and Saier, M. H., Jr. (1991) J. Bacteriol. 173, 3117-3127[Abstract/Free Full Text]
  8. Spychala, J., Datta, N. S., Takabayashi, K., Datta, M., Fox, I. H., Gribbin, T., and Mitchell, B. S. (1996) Proc. Natl. Acad. Sci. U. S. A. 93, 1232-1237[Abstract/Free Full Text]
  9. Bork, P., Sander, C., and Valencia, A. (1993) Protein Sci. 2, 31-40[Abstract]
  10. Sigrell, J. A., Cameron, A. D., Jones, T. A., and Mowbray, S. L. (1997) Protein Sci. 6, 2474-2476[Abstract]
  11. Sigrell, J. A., Cameron, A. D., Jones, T. A., and Mowbray, S. L. (1998) Structure 6, 183-193[Medline] [Order article via Infotrieve]
  12. Mathews, I. I., Erion, M. D., and Ealick, S. E. (1998) Biochemistry 37, 15607-15620[CrossRef][Medline] [Order article via Infotrieve]
  13. Cook, W. J., DeLucas, L. J., and Chattopadhyay, D. (2000) Protein Sci. 9, 704-712[Abstract]
  14. Long, M. C., Escuyer, V., and Parker, W. B. (2003) J. Bacteriol. 185, 6548-6555[Abstract/Free Full Text]
  15. Parker, W. B., Barrow, E. W., Allan, P. W., Shaddix, S. C., Long, M. C., Barrow, W. W., Bansal, N., and Maddry, J. A. (2004) Tuberculosis (Edinb.) 84, 327-336[CrossRef][Medline] [Order article via Infotrieve]
  16. Barrow, E. W., Westbrook, L., Bansal, N., Suling, W. J., Maddry, J. A., Parker, W. B., and Barrow, W. W. (2003) J. Antimicrob. Chemother. 52, 801-808[Abstract/Free Full Text]
  17. Chen, C. K., Barrow, E. W., Allan, P. W., Bansal, N., Maddry, J. A., Suling, W. J., Barrow, W. W., and Parker, W. B. (2002) Microbiology 148, 289-295[Abstract/Free Full Text]
  18. Wayne, L. G., and Hayes, L. G. (1996) Infect. Immunol. 64, 2062-2069[Abstract]
  19. Long, M. C., and Parker, W. B. (2006) Biochem. Pharmacol. 71, 1671-1682[CrossRef][Medline] [Order article via Infotrieve]
  20. Pittillo, R. F., Moncrief, C., Brockman, R. W., and Chambers, P. (1964) Antimicrob. Agents Chemother. 10, 474-484[Medline] [Order article via Infotrieve]
  21. Park, B. K., Kitteringham, N. R., and O'Neill, P. M. (2001) Annu. Rev. Pharmacol. Toxicol. 41, 443-470[CrossRef][Medline] [Order article via Infotrieve]
  22. Brockman, R. W., Cheng, Y. C., Schabel, F. M., Jr., and Montgomery, J. A. (1980) Cancer Res. 40, 3610-3615[Abstract/Free Full Text]
  23. Long, M. C., Allan, P. W., Luo, M. Z., Liu, M. C., Sartorelli, A. C., and Parker, W. B. (2007) J. Antimicrob. Chemother. 59, 118-121[Abstract/Free Full Text]
  24. Doublie, S. (1997) Methods Enzymol. 276, 523-530[Medline] [Order article via Infotrieve]
  25. Ohshima, N., Inagaki, E., Yasuike, K., Takio, K., and Tahirov, T. H. (2004) J. Mol. Biol. 340, 477-489[CrossRef][Medline] [Order article via Infotrieve]
  26. Ogata, C. M. (1998) Nat. Struct. Biol. 5, (suppl.) 638-640
  27. Terwilliger, T. C., and Berendzen, J. (1999) Acta Crystallogr. D Biol. Crystallogr. 55, 849-861[CrossRef][Medline] [Order article via Infotrieve]
  28. McRee, D. E. (1999) J. Struct. Biol. 125, 156-165[CrossRef][Medline] [Order article via Infotrieve]
  29. Storoni, L. C., McCoy, A. J., and Read, R. J. (2004) Acta Crystallogr. D Biol. Crystallogr. 60, 432-438[CrossRef][Medline] [Order article via Infotrieve]
  30. Collaborative Computational Project, Number 4 (1994) Acta Crystallogr. D Biol. Crystallogr. 50, 760-763[CrossRef][Medline] [Order article via Infotrieve]
  31. Kantardjieff, K. A., Hochtl, P., Segelke, B. W., Tao, F. M., and Rupp, B. (2002) Acta Crystallogr. D Biol. Crystallogr. 58, 735-743[CrossRef][Medline] [Order article via Infotrieve]
  32. Otwinowski, Z., and Minor, W. (1997) Methods Enzymol. 276, 307-326
  33. Laskowski, R. A., MacArthur, M. W., Moss, D. S., and Thornton, J. M. (1993) J. Appl. Crystallogr. 26, 283-291[CrossRef]
  34. Laskowski, R. A., Moss, D. S., and Thornton, J. M. (1993) J. Mol. Biol. 231, 1049-1067[CrossRef][Medline] [Order article via Infotrieve]
  35. Christopher, J. A. (1998) The Center for Macromolecular Design, Texas A&M University, College Station, TX
  36. Merritt, E. A., and Murphy, M. E. (1994) Acta Crystallogr. D Biol. Crystallogr. 50, 869-873[CrossRef][Medline] [Order article via Infotrieve]
  37. Pettersen, E. F., Goddard, T. D., Huang, C. C., Couch, G. S., Greenblatt, D. M., Meng, E. C., and Ferrin, T. E. (2004) J. Comput. Chem. 25, 1605-1612[CrossRef][Medline] [Order article via Infotrieve]
  38. Hao, W., and Gupta, R. S. (1996) Biochem. Mol. Biol. Int. 38, 889-899[Medline] [Order article via Infotrieve]
  39. Sigrell, J. A., Cameron, A. D., and Mowbray, S. L. (1999) J. Mol. Biol. 290, 1009-1018[CrossRef][Medline] [Order article via Infotrieve]
  40. Schumacher, M. A., Scott, D. M., Mathews, I. I., Ealick, S. E., Roos, D. S., Ullman, B., and Brennan, R. G. (2000) J. Mol. Biol. 296, 549-567[CrossRef][Medline] [Order article via Infotrieve]
  41. Zhang, Y., Dougherty, M., Downs, D. M., and Ealick, S. E. (2004) Structure 12, 1809-1821[Medline] [Order article via Infotrieve]
  42. Shirakihara, Y., and Evans, P. R. (1988) J. Mol. Biol. 204, 973-994[CrossRef][Medline] [Order article via Infotrieve]
  43. Maj, M. C., Singh, B., and Gupta, R. S. (2002) Biochemistry 41, 4059-4069[CrossRef][Medline] [Order article via Infotrieve]
  44. Kabsch, W., and Sander, C. (1983) Biopolymers 22, 2577-2637[CrossRef][Medline] [Order article via Infotrieve]

Add to CiteULike CiteULike   Add to Complore Complore   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us   Add to Digg Digg   Add to Reddit Reddit   Add to Technorati Technorati    What's this?



This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Supplemental Data
Right arrow All Versions of this Article:
282/37/27334    most recent
M703290200v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow