Chapter 4. The application of site-moiety map for characterizing protein-ligand binding
4.2 orthSiMMap methods
4.3.3 Inhibitors and inhibition assay
Following the SiMMap analysis, compounds were rescored using the rank-based consensus scoring (RCS54), which combines energy-based and anchor-based scoring functions. Since a compound simultaneously docked into apo and closed-form binding sites of orthSiMMaps was considered as a potentially useful hit, we selected common top-ranked compounds from the closed-form and apo-form orthSiMMap analysis for subsequent bioassay. After RCS ranks in both the Maybridge and NCI databases, 48 compounds that were available (either requested or purchased) were then subjected to MtSK and HpSK inhibitory assays (Fig. 4.4). Among those, 10 compounds had an IC50 value ≦ 100 μM for both HpSK and MtSK (Table 4.2), in which six (NSC45611, NSC162535, NSC45612, NSC45174, NSC45547, and NSC45609) demonstrated IC50 values of ≦10 μM (Table 4.3). In parallel, 65 existing kinase inhibitors were tested to evaluate their inhibitory effects against shikimate kinase. Of the two compounds
38
(AG538 and GW5074) that showed inhibitory effects, AG538 had a low IC50 value.
Enzymatic kinetic analysis showed that NSC45611, NSC162535, NSC45612, NSC45174, and AG538 were competitive inhibitors of ATP, in agreement with the docked poses (Table 4.2 and Table 4.3). Of these, NSC45611, NSC162535 and NSC45612 also competed with shikimate. Notably, NSC45611, NSC162535, NSC45612 and NSC45174 had low IC50 and αKi
values, showing potent inhibition. Figure 4.5 shows that three (NSC45611, NSC162535, and NSC45612) had lower values of IC50 (≤ 10 μM) and fit well over five hot spots (H1, V1, H3, V2, and E1). For those with IC50 ≥ 20 μM, these compounds lack of negatively charged groups to form electrostatic interactions with arginines (R57 and R136 in HpSK) on E1. On the other hand, kinase inhibitors AG538 and GW5074 did not occupy the shikimate site. Moieties with 1−3 rings were present at V1 and V2, yielding a number of vdW contacts. The binding groups of active inhibitors matched well with the identified moieties found from the consensus anchors. For example, the sulfonate groups of NSC162535, NSC45611, and NSC45612 were found to occupy H1. The moieties of NSC162535 (SO3- group), NSC45611 (CO2- group), and NSC45612 (CO2- group) occupied E1.
Table 4.2. Summary of 12 inhibitors with inhibition assay, compound structures, docked poses, and consensus anchors
Compound ID SK species
IC50
(μM)a Compound structure Docked pose
ATP site SKM site
H1 V1 H2 H3 V2 E1
NSC45611 Hp 4.8
Mt 1.5
NSC162535 Hp 4.9
Mt 1.6 NSC45612
Hp 6.1
NN NN
COO -OH -O3S OH
Mt 2.8
NSC45174 Hp 7.8
Mt 2.8
NSC45547 Hp 7.8
Mt 3.4
39
NSC45609 Hp 7.0
Mt 2.0
RH00037 Hp 23.8
O HN NH
O2N O O
Mt <100
RH00016 Hp 40.2
Mt <100
GK01385 Hp <100
Mt <100
SPB01099 Hp <100
Mt <100
AG538 Hp 2.3
Mt 0.4
GW5074 Hp 31.4
Mt 29.6
a The inhibition assay is done by Dr. Wen-Ching Wang of National Tsing Hua University51.
Table 4.3. Properties of potent inhibitors for HpSK and MtSK a Compound ID SK
species
Inhibition modeb αKi, ATP (μM)
αKi, SKM (μM)
ATP SKM
NSC45611 Hp ■ ■ 1.1 1.7
Mt ■ ■ 0.3 0.7
NSC162535 Hp ■ ■ 1.9 1.8
Mt ■ ■ 0.2 0.6
NSC45612 Hp ■ ■ 2.0 2.4
Mt ■ ■ 0.7 1.0
NSC45174 Hp ■ □ 1.7 12.8
Mt ■ □ 0.4 2.7
AG538 Hp ■ □ 3.1 5.4
Mt ■ □ 0.04 0.4
a The inhibition assay is done by Dr. Wen-Ching Wang of National Tsing Hua University51.
b■: Competitive inhibition; □: Non-competitive inhibition
40
Anchor Relative activity (%) of mutants
E1 R57A R132A R116A
2% 5% NA
H1 S12A S15A R116A
59% 1% NA
H2 S15A D31A D33A
1% 62% NA
-Figure 4.5. Characterization of shikimate kinase inhibitors by enzyme assay, orthSiMMaps, site-mutagenesis studies and analogues. (a–c) Structures of three inhibitors, NSC162535, NSC45611, and NSC45612. (d–f) The inhibitions of these compounds were analyzed by enzyme IC50 test on HpSK (filled) and MtSK. (g–i) The relationship between anchors and the docked mode of each inhibitor for HpSK. These compounds consistently include two negative charge moieties (SO3- or CO2-) that form hydrogen bonds with conserved interactions residues of anchors E1 and H1. (j) Comparison of relative activities of HpSK mutants. The conserved interacting residues for each anchor were mutated, respectively. R57, R132, R116, and F48 located in the shikimate site are critical for the enzymatic functions. (k) The potency of NSC162535 analogues. The substitution moieties of analogues are indicated in black. Those that lack the E1 moiety greatly lost the inhibitory effects (IC50 >100 μM). The inhibition assay is done by Dr. Wen-Ching Wang of National Tsing Hua University51.
41 4.3.4 Site-directed mutagenesis
A consensus anchor of orthologous targets, identified from the conserved binding pockets shared with conserved interacting residues and specific physico-chemical property, usually engages with specific functions in the enzymatic catalysis. We sought to investigate the roles of identified consensus anchor residues of the orthSiMMaps in catalysis. The site-directed mutagenesis study is done by Dr. Wen-Ching Wang of National Tsing Hua University51. We first investigated mutants of E1 residues (R57, R116, and R132) that contact with shikimate55. Enzymatic analysis revealed that these arginines had extremely low activity (Fig. 4.5j), suggesting the importance of these residues in catalysis. Indeed, R117 of MtSK that corresponds to R116 of HpSK has thus been suggested as a primary candidate to stabilize the transition-state intermediate56.
For the H3 (D33) and V2 (F48) residues, D33A completely lost the enzymatic activity while F48A exhibited hardly any detectable activity (1%). D33 and F48 are in direct contact with shikimate. More importantly, it should be noted that D33 forms a hydrogen bond to the 3-OH group of shikimate, which may increase the nucleophilicity of the O atom or accept the proton from the 3-OH group of shikimate, facilitating the catalysis. E114A, a LID residue whose side chain faces the solvent, retained 82% relative activity. On the other hand, the F48 side chain contacts with those from several residues nearby (V44, E53, F56, R57 and P117), which may form a stable platform to interact with the ligand for subsequent catalytic reaction.
We then evaluated residues from H1, H2, and V1 located at the nucleotide site. H1 residues are primarily from the Walker A motif (P loop; residues 11−16, GSGKSS) surrounding the phosphate groups of the nucleotides. Of the three mutants (S12A, S15A, and S16A), S12A and S16A remained >50% of the relative activity, while S15A had extremely low activity (1%).
The S15 side chain resides nearby the β-phosphate of ADP. Furthermore, the adjacent lysine (K14) corresponding to K15 of MtSK has been identified as a critical catalytic residue in MtSK since its side chain points toward the γ-phosphate56. The other H2 mutant D31A retained 62%
of the relative activity (62%), possibly due to its remote location to the phosphate group. For V1 that is just next to H1, several H1 residues are also shared by V1. Enzymatic analysis showed that M10A remained 38% relative activity. These results suggest that the conserved interacting residues from E1 (R57, R116 and R132), H1 (S15 and R116), H2 (S15 and D33),
42
H3 (D33), V1 (S15) and V2 (F48) contribute significantly to catalytic power and substrate binding.
4.3.5 Analogues assay and orthSiMMap
To validate the moiety preferences of consensus anchors, we identified four analogues (NSC45547, NSC45609, NSC37215, and NSC45208) of NSC162535 for inhibitory assays (Fig.
4.5k). NSC45547 and NSC45609 that occupy E1 (SO3- group) and H1 (SO3- and NO2 groups) retained good IC50 values (7.8 and 7.0 μM for HpSK; 3.4 and 2.0 μM for MtSK). Conversely, NSC37215 and NSC45208, that cannot anchor at E1, lost the inhibitory.
To evaluate the significance of pocket-moiety interaction preferences of consensus anchors in the orthSiMMaps, we performed clustering analysis on 27 inhibitory assay compounds. These compounds can be roughly clustered into three groups (Fig. 4.6). The potent inhibitors of group I (NSC162535, NSC45609, NSC45547, NSC45174, NSC45611, and NSC45612) match more than 5 consensus anchors (Fig. 4.5g-i, and Table 4.2). For Group II compounds (RH00037, RH00016, GK01385, and SPB01099), each compound matches four of six anchors; Group III are kinase inhibitors (AG538 and GW5074) and these compounds share anchors of ATP site.
For inactive compounds, there are fewer matched consensus anchors in the HpSK/MtSK (usually 4), particularly E1 is the least seen. While the inhibitors of group I and II agreed with anchors of ATP site and shikimate site, the kinetic assay showed competitive inhibitions for ATP and shikimate acid (Table 4.3). The kinase inhibitors of group III occupied the anchors of ATP site, and only showed the competitive inhibitions for ATP. Generally, the pocket environment of ATP is conserved for kinase family, and the inhibitors of group III also have the broadband inhibition for multiple kinases, such as the inhibition of AG538 observed on insulin-like growth factor-1 receptor (IGF-1R)57, IR, EGFR58, and Src kinases59.
While there are the same number of consensus anchors (E1, H1, H2, H3, V1 and V2), the spatial arrangement of these anchors were closer in the closed form (Fig. 4.2c and 4.2d).
Residues (D31 and D33) that contribute to H2 of the apo form were in closer proximity in the closed conformation, resulting in a reduced volume at this site. Likewise, the corresponding site at V2 surrounded by F48, G80, and G81 in HpSK had less space in the closed form, hindering the accommodation of large moieties carrying one or two rings at this site. The above evidences demonstrate that induced LID conformation of shikimate kinases was sensitive in the
structure-43 based drug discovery strategy.
NSC162535
ATP SKM ATP SKM
a
G11 S12 G13 K14 S15 M10 G11 S12 G13 K14 S15 S15 D31 D33 K14 D33 G80 D33 F48 G80 G81 R57 R132 G12 S13 G14 K15 P11 G12 S13 G14 K15 K15 S16 D32 K15 D34 G80 D34 G79 G80 R58 R136
H1 V1 H2 H3 V2 E1 H1 V1 H2 H3 V2 E1
G11 S12 G13 K14 S15 M10 G11 S12 G13 K14 S15 S15 D31 D33 K14 D33 G80 D33 F48 G80 G81 R57 R132 G12 S13 G14 K15 P11 G12 S13 G14 K15 K15 S16 D32 K15 D34 G80 D34 G79 G80 R58 R136
H1 V1 H2 H3 V2 E1 H1 V1 H2 H3 V2 E1
G11 S12 G13 K14 S15 M10 G11 S12 G13 K14 S15 S15 D31 D33 K14 D33 G80 D33 F48 G80 G81 R57 R132 G12 S13 G14 K15 P11 G12 S13 G14 K15 K15 S16 D32 K15 D34 G80 D34 G79 G80 R58 R136
H1 V1 H2 H3 V2 E1 H1 V1 H2 H3 V2 E1
Figure 4.6. Interaction profiles between selected anchor residues and 27 tested compounds. (a) The anchor profile of tested compounds on shikimate kinases. (b) Group I: the NCI compounds (orange). (c) Group II: the Maybridge compounds (yellow). (d) Group III: kinase inhibitors (cyan). The NCI compounds consistently occupy anchors E1 and V2 locating in both ATP and shikimate sites. Except for NSC45174, the NCI compounds are competitive inhibitors with both ATP and shikimate. For the Maybridge compounds, none form electrostatic interactions with R57 and R132 on the consensus anchor E1. The two kinase compounds are located at the ATP site, in good agreement with the kinetic results showing that they exhibited competitive inhibition with ATP and noncompetitive inhibition with shikimate.
4.3.6 Structural mechanism of the inhibitor binding for shikimate kinases
Superposition of various structures (apo HpSK, HpSK·shikimate·PO4, HpSK·S3P·ADP,
44
and E114A·162535) reveals a significant conformational change in the LID-containing segment after β4 of the CORE domain (residues 101 to 138; α5, LID and α6) (Fig. 4.7).
Furthermore, the SB region (residues 32−60) shows a small rotation among the different unliganded/liganded states, in accord with MtSK structures53.
Of the three conserved arginines (R57, R116, and R132), it is noted the Cα atom of R57 superimposes relatively well, while that of R132 has a small shift among various structures (Fig. 4.7a-4.7d). A shorter Cα-atom (R57-R132) distance (~0.6 Å) is noted for HpSK·S3P·ADP and HpSK·shikimate·PO4 as compared to the apo-form HpSK. On the other hand, there is a significant drift for R116 due to the distinct conformations of the LID loop (Fig. 4.7a-4.7d).
Our results suggest that these arginines contribute to the movement of the lid region and the shikimate-binding domain upon ligand binding. R116, when visible, makes a significant shift to form hydrogen-bonding interactions with various ligands in the binding pocket: (i) shikimate in HpSK·shikimate·PO4; (ii) β-phosphate of ADP in HpSK·S3P·ADP; and (iii) NSC162535 in E114A·162535. In the MtSK·shikimate·AMPPCP structure, a direct contact is also observed between R117 (corresponding to R116 in HpSK) and γ-phosphate of AMPPCP, an ATP analogue, which supports its catalytic role in the γ-phosphoryl transfer56.
To evaluate whether NSC162535 will come in contact with R116 in other forms, we have docked NSC162535 into the binding pockets of various HpSK structures (Fig. 4.7e-4.7i). In the apo form, HpSK with a flexible LID presents a wide-opening pocket, allowing entry of promising substrates (Figs. 4.7a, 4.7e and 4.7j). No close contacts are found between R116 and the docked NSC162535 in the binding pockets of the apo and HpSK·shikimate·PO4 forms (Fig.
4.7e and 4.7f). In the HpSK·S3P·ADP state, NSC162535 is docked into a site where the Nη1 and Nη2 of the guanidino group in R116 make no significant contacts. NSC162535, on the other hand, is docked into a comparable site in the E114A·162535 form, where it contacts directly with the Nη1 and Nη2 atoms of R116 just like that of the crystal structure. Thus, it is likely that R116 plays a crucial role during the course of a conformational cycle in conducting a catalytic event (Fig. 4.7i and 4.7j). Upon diffusion into the binding pocket, NSC162535 that carries two SO3- and a -N=N- groups may bind to the active site, interact with R57 and R132,
45
Figure 4.7. Probing the affinity pockets in HpSK. (a-d) The binding pockets of HpSK (a), HpSK·shikimate·SO4 (b), HpSK·S3P·ADP (c), and E114A·162535 (d) structures. The bound ligands, D33, F48, R57, R116, and R132 are drawn as sticks. The LID segments (residues 109−123) are drawn as the ribbon structures. (e−h) The docked NSC162535 models in the binding pockets of HpSK (e), HpSK·shikimate·SO4 (f), HpSK·S3P·ADP (g), and E114A·162535 (h) structures. Superposition of three residues (R57, R116, and R132), docked and bound NSC162535 among HpSK (blue), HpSK·SKM·SO4 (yellow), HpSK·S3P·ADP (cyan), and E114A·162535 (orange) structures. (i) Superimposed docked structures (e−h). The conformation of LID segment (residues 113−119, ribbon) having R116 (thick stick) demonstrates the greatest conformational changes induced by bound ligands. (j) Schematic
46
diagram of induced-fit conformational changes upon binding to ligands. The view of LID regions corresponding to apo HpSK, HpSK·S3P·ADP, and E114A·162535 is colored as blue, cyan, and orange, respectively. The crystallized structures studies are done by Dr. Wen-Ching Wang of National Tsing Hua University.
and then trigger a conformational change cycle. As a result, R116 along with R57 and R132 will trap the inhibitor, yielding an optimized anchor (E1). These results also suggest an unusually elastic LID region, which allows accommodating various ligands, as demonstrated here. In this section, the crystallized structures studies are done by Dr. Wen-Ching Wang of National Tsing Hua University.
4.3.7 Performance of the orthSiMMap method
We then evaluated the accuracy of the orthSiMMap approach. The orthSiMMap score (solid lines) significantly outperformed those (dashed lines) of energy-based scoring methods, which are often used in docking tools, on apo-form HpSK and MtSK (Fig. 4.8a). The average enrichments of 3.73 (HpSK), 1.59 (MtSK), and 2.74 (fusion of HpSK and MtSK) were obtained using energy-based scoring methods, as compared to 11.18 (HpSK), 35.51 (MtSK), and 93.69 (fusion of HpSK and MtSK) using the orthSiMMap scoring method. Additionally, the orthSiMMap scores exhibited a higher accuracy than that of the SiMMap score from a single target (HpSK or MtSK).
The orthSiMMap is able to reduce the deleterious effects of screening ligand structures that are rich in charged or polar atoms. Generally, energy-based scoring functions favor the selection of high-molecular-weight compounds yielding high vdW potentials, as well as polar compounds that produce hydrogen-bonding and/or electrostatic potentials54. The average molecular weights of the top ranked 100 compounds of the orthSiMMap and the energy-based scoring methods were 459.9 and 532.6, respectively; the average numbers of polar atoms were 11.3 (orthSiMMap method) and 14.1 (energy-based method) (Fig. 4.8b,c). The ranks of those 10 active compounds were much higher in the orthSiMMap scoring analysis than in the energy-based analysis. It should be noted that NSC162535 was ranked as 1 and 1821 using the apo-form orthSiMMap and energy-based scoring methods, respectively (Table 4.4).
47
0 1000 2000 3000 4000 5000 6000 Rank of compound
True hit (%)
Hp+Mt (orthSiMMap) Hp+Mt (Energy) HpSK (orthSiMMap) HpSK (Energy) MtSK (orthSiMMap) MtSK (Energy)
0 Number of polar atoms
Number of compounds
375-400400-425425-450450-475475-500500-5 25
525-550550-575575-600600-625625-650 Molecular weight
Number of compounds
orthSiMMap Energy
Figure 4.8. Performance of the orthSiMMap method on apo-form HpSK and MtSK. (a) The true-hit rates of energy-based and orthSiMMap scoring approaches. The orthSiMMap scores (solid line) of adaptive inhibitors significantly outperform energy-based scores (dashed line) using the top ranked 6000 compounds by combining the Maybridge and NCI databases. (b) Distribution of number of polar atoms, and (c) molecular weight of top 100 compounds from orthSiMMap scores and energy-based score.
Table 4.4. The ranks of active compounds using orthSiMMap, energy-bases, and combination scoring methods for apo and closed forms of HpSK and MtSK
Compound ID Compound structure Apo form Closed form
RCS a orthSiMMap Energy orthSiMMap Energy
NSC45611 48 435 515 827 96
NSC162535 1 1821 242 253 25
NSC45612 32 106 238 229 31
NSC45174 38 162 737 110 147
48
NSC45547 130 5999 308 1540 67
NSC45609 18 4017 5 17 3
RH00037 786 876 891 1549 371
RH00016 3765 6000 1219 5824 1837
GK01385 286 1774 730 49 199
SPB01099 117 5940 68 2871 19
a The rank is the rank combination of orthSiMMap and energy.
4.4 Summary
The largest obstacle by far in structure-based drug discovery is the relatively low hit rates in scoring methods due to the lack of adequate quantities of binding partners for a given target.
In other words, there is no adequate training set to establish the veracity or utility of an algorithm. Under these circumstances, the accuracy of a given individual scoring function is generally unknown and/or cannot be evaluated at a critical point. The current emphasis of the orthSiMMap scoring developed here thus provides a useful index to improve the screening accuracy for identification of adaptive inhibitors when the target proteins shared conserved binding sites. Through the employment of this developed method, we successfully found six new potent inhibitors (<8.0 μM) of HpSK and MtSK. Two of the 65 kinase inhibitors were also found to inhibit both HpSK and MtSK activity. The finding that NSC45611, NSC162535, and NSC45612 were competitive inhibitors of ATP and shikimate suggests that they belong to a novel class of shikimate kinase inhibitors. Based on the novel inhibitor - NSC162535, the inhibitor complex crystal structure, E114A·162535, was determined by Dr. Wang’s group of National Tsing Hua University. These results illustrate a robust orthSiMMap-based approach to identify selective kinase inhibitors.
49
Table 4.5. Some selected top-ranked compounds using orthSiMMap, energy-bases, and combination scoring methods for apo and closed forms of HpSK and MtSK
Compound ID Compound structure Apo form Closed form
RCS a Bioassay iSiMMap Energy iSiMMap Energy
NSC131133 2 2153 7 456 1 -b
NSC407257 3 91 17 317 2 -
NSC644745 430 1 3152 745 1085 -
NSC714539 431 3 1 9 64 Inactive
NSC524127 920 3090 2 189 175 -
NSC2460 313 2633 4 2073 49 Inactive
ZINC05823979 1321 578 8 1021 265 -
NSC83262
NO2 O2N
NH O -O
NH3+
11 28 21 1 5 -
NSC16220 728 37 10 2 137 Inactive
NSC82523 13 170 15 238 4 Inactive
NSC624285 39 199 22 8 8 -
NSC85597 OSO NO2 O
O O -O2N
O O
-O 28 68 31 476 6 -
a The rank is the rank combination of orthSiMMap and energy.
b The compound is not tested.
The developed orthSiMMap scoring method appears to outperform the energy-based method (Tables 4.4 and 4.5). Of six potent inhibitors, it was interesting to find that aside from NSC45609, the others have a higher rank in the apo-form than in the closed-form orthSiMMap scoring analysis. Additionally, the top-ranked inhibitors from the apo-form orthSiMMap scoring analysis often possess larger moieties (e.g. naphthalene or nitrobenzene) at both sides as opposed to those with a relatively small moiety (e.g. amide or aliphatic chain). The closed-form orthSiMMap scoring analysis has, nonetheless, yielded useful hits including NSC45609
50 and SPB01099.
P-loop kinase fold consists of functionally diverse kinase classes, such as shikimate kinase, NTPases and GTPases60. They frequently share conserved binding environments (e.g., P-loop and walker A/B motifs) for interacting with partners (e.g., small compounds and proteins). The molecules inhibit P-loop kinases that play a key role in various diseases, such as cancer, cardiovascular diseases, gastric diseases or infections. Although a number of inhibitors in clinical trials 61-63 or on the market (omeprazole and ciprofloxacin) inhibit the activity of P-loop kinases, few of them bind to the ATP-binding site64. Meanwhile, target proteins with dynamic induced-fit forms, like the P-loop SKs, represent a major limitation for the structure-based screening approach. The approach of orthSiMMap designing the competitive ATP inhibitors with specific substrate pocket presents a novel strategy of targeting P-loop kinases.
The developed orthSiMMap method is database independent. Comparable anchors were identified in compounds from the Maybridge and NCI databases. Each of the anchors also included analogous chemical moieties. Nonetheless, the derived proportion of these moieties was different because the Maybridge and NCI databases contain heterogeneous distribution of compounds. For example, the proportion of carboxyl, sulfonate, and phosphate was significantly higher in compounds from the NCI database than in those from the Maybridge database. On the other hand, the derived model was sensitive to binding-site properties, as illustrated by the difference between the apo- and closed-form models (Fig. 4.2). In summary, we anticipate that the orthSiMMap method can be useful in discovering new inhibitors, investigating the binding mechanisms, and guiding the lead optimization for orthologous targets. Additionally, crystal structures reveal the details of ligand binding in the induced-fit P-loop kinases and will be valuable in the development of novel P-P-loop kinase inhibitors.
51
Chapter 5 Conclusion
5.1 Summary
Briefly, the major contributions of this thesis can be summarized in the following:
(1) The concept of site-moiety map (SiMMap) was proposed for predicting protein-ligand binding modes and characterizing protein-ligand binding sites in structure-based drug design. SiMMap statistically infers the site-moiety map describing the relationship between the moiety preferences and physico-chemical properties of the binding site. Our experimental results showed that the site-moiety map is useful to reflect biological functions and identify active compounds from thousands of compounds. In addition, the site-moiety map can guide to assemble potential leads by optimal steric, hydrogen-bonding, and electronic moieties.
(2) Members of individual protein families often share a homologous fold and conserved structural features to interact with chemically similar ligands throughout evolution, despite low sequence identity. A structure-based site-moiety screening method, orthSiMMap, was developed to discover the inhibitors for a family of orthologous proteins. Here, we utilized the orthSiMMap to pharmacologically interrogate orthologous shikimate kinases (SKs) from Mycobacterium tuberculosis and Helicobacter pylori. The derived apo/closed core site-moiety maps and the anchor scores were used to identify six potent inhibitors (<8.0 μM). Site-directed mutagenesis (these studies done by Dr. W.C. Wang of National Tsing Hua University) and analogues studies revealed that critical conserved interacting residues contribute to a given pocket-moiety interaction spot. Crystal structures of HpSK·SO4,
(2) Members of individual protein families often share a homologous fold and conserved structural features to interact with chemically similar ligands throughout evolution, despite low sequence identity. A structure-based site-moiety screening method, orthSiMMap, was developed to discover the inhibitors for a family of orthologous proteins. Here, we utilized the orthSiMMap to pharmacologically interrogate orthologous shikimate kinases (SKs) from Mycobacterium tuberculosis and Helicobacter pylori. The derived apo/closed core site-moiety maps and the anchor scores were used to identify six potent inhibitors (<8.0 μM). Site-directed mutagenesis (these studies done by Dr. W.C. Wang of National Tsing Hua University) and analogues studies revealed that critical conserved interacting residues contribute to a given pocket-moiety interaction spot. Crystal structures of HpSK·SO4,