Skip to main content
NCBI home page
Protein Science : A Publication of the Protein Society logoLink to Protein Science : A Publication of the Protein Society
. 2005 Jul;14(7):1760–1771. doi: 10.1110/ps.051432405

Crystallographic studies on acyl ureas, a new class of glycogen phosphorylase inhibitors, as potential antidiabetic drugs

Nikos G Oikonomakos 1,2, Magda N Kosmopoulou 1, Evangelia D Chrysina 1, Demetres D Leonidas 1, Ioannis D Kostas 1, K Ulrich Wendt 3, Thomas Klabunde 3, Elisabeth Defossa 3
PMCID: PMC2253349  PMID: 15987904

Abstract

Acyl ureas were discovered as a novel class of inhibitors for glycogen phosphorylase, a molecular target to control hyperglycemia in type 2 diabetics. This series is exemplified by 6-{2,6-Dichloro- 4-[3-(2-chloro-benzoyl)-ureido]-phenoxy}-hexanoic acid, which inhibits human liver glycogen phosphorylase a with an IC50 of 2.0 μM. Here we analyze four crystal structures of acyl urea derivatives in complex with rabbit muscle glycogen phosphorylase b to elucidate the mechanism of inhibition of these inhibitors. The structures were determined and refined to 2.26Å resolution and demonstrate that the inhibitors bind at the allosteric activator site, where the physiological activator AMP binds. Acyl ureas induce conformational changes in the vicinity of the allosteric site. Our findings suggest that acyl ureas inhibit glycogen phosphorylase by direct inhibition of AMP binding and by indirect inhibition of substrate binding through stabilization of the T′ state.

Keywords: type 2 diabetes, glycogen phosphorylase, acyl ureas, inhibition, X-ray crystallography

Inhibition of glycogen phosphorylase (GP) has been proposed as a therapeutic strategy for the treatment of type 2 diabetes (Aiston et al. 2001, 2003; McCormack et al. 2001; Treadway et al. 2001) and several binding sites on the enzyme such as the catalytic, allosteric, inhibitor, and the new allosteric site (Fig. 1 ▶) have been identified as specific targets for inhibitor binding (Somsák et al. 2001, 2003; Treadway et al. 2001; Oikonomakos 2002; Kurukulasuriya et al. 2003). In some cases detailed X-ray crystallographic studies have revealed key interactions responsible for inhibitor potency and have elucidated the structural mechanisms of enzyme inhibition (for review, see Oikonomakos 2002).

Figure 1.

Figure 1.

Open in a new tab

A schematic diagram of the T-state rmGPb dimeric molecule, for residues 13 to 837, viewed down the molecular dyad, showing the positions for the catalytic, allosteric, the inhibitor, and the new allosteric binding sites. The catalytic site, which includes the essential cofactor PLP (not shown), is buried at the center of the subunit accessible to the bulk solvent through a 15-Å-long channel. Glucose (shown in black) binds at this site and promotes the less active T state through stabilization of the closed position of the 280s loop (black). The allosteric site, which binds the activator AMP, the inhibitors Glc-6-P, the Bayer compound W1807, and acyl urea compound 1 (black) is situated at the subunit–subunit interface some 30 Å from the catalytic site. The inhibitor site, which binds purine compounds such as caffeine and flavopiridol (dark gray) (Oikonomakos et al. 2000a), is situated at the entrance to the catalytic site tunnel, formed by two hydrophobic residues of Phe285 and Tyr613. The new allosteric or indole binding site, located inside the central cavity formed on association of the two subunits, binds CP320626 molecule (light gray) and is some 15 Å from the allosteric effector site, 33 Å from the catalytic site, and 37 Å from the inhibitor site (Oikonomakos et al. 2000b).

The allosteric site has lately attracted considerable interest. More specifically, the site has been shown to bind Bayer’s inhibitor 1,4-dihydropyridine-2,3-dicarboxylate (W1807) and several dihydropyridine diacid analogs (Zographos et al. 1997; Oikonomakos et al. 1999; Ogawa et al. 2003), which exhibited glucose lowering in a db/db mouse model. Rational inhibitor design efforts have led to synthesis of phenyl diacid analogs (Lu et al. 2003) and phenoxy-phthalates (Kristiansen et al. 2004), which inhibited both the basal and the glucagon-induced glucose production when tested in cultured primary hepatocytes.

Recently acyl ureas were reported as human liver glycogen phosphorylase a (hlGPa) inhibitors, which bind to the allosteric site of the enzyme (T. Klabunde, K.U. Wendt, D. Kadereit, V. Brachvogel, H.-J. Burger, A.W. Herling, N.G. Oikonomakos, M.N. Kosmopoulou, D. Schmoll, E. Sarubbi, et al., in prep.). Here we report on the detailed analysis of four crystal structures of acyl urea inhibitors (1–4) (Scheme 1 ▶) in complex with rabbit muscle glycogen phosphorylase (rmGPb). These data show that compounds 1–4 bind at the allosteric site of the enzyme, where they occupy a position similar to that of the allosteric activator AMP. Binding of 1–4 induces significant conformational changes in the vicinity of the site, and stabilizes the T′-state conformation.

Scheme 1.

Scheme 1.

Open in a new tab

Chemical structures of the acyl urea compounds 1–4, showing the numbering system used.

Results and Discussion

Compounds 1–4 were found to inhibit hlGPa (IC50 values of 0.65–2.48 μM), and rmGPb (IC50 values of 1.6–2.9 μM) with similar potencies (Table 1) as expected from the high sequence identity (79%) between the two isoforms (Rath et al. 1987; T. Klabunde, K.U. Wendt, D. Kadereit, V. Brachvogel, H.-J. Burger, A.W. Herling, N.G. Oikonomakos, M.N. Kosmopoulou, D. Schmoll, E. Sarubbi, et al., in prep.). In order to elucidate the structural basis of inhibition, we have determined the crystal structure of rmGPb in complex with 1–4. A summary of the data processing and refinement statistics for the rmGPb–1, rmGPb–2, rmGPb–3, and rmGPb–4 complex structures is given in Table 2. For all complexes, the 2Fo−Fc Fourier electron density maps indicated that 1–4 bound at the allosteric site. The allosteric site is situated where the C termini of the helices α2 (residues 47–78) and α8 (residues 289–314) come together. It is lined by strands of the central core of β4 (residues 153–160) and β11 (residues 237–247), and surrounded on the third side by the short β7 strand (residues 191–193) and the following loop to residue 197. The site is closed by the cap′ region (residues 36′ to 47′; the prime refers to residues from the symmetry-related subunit).

Table 1.

Inhibition studies of compounds 1–4

Compound IC50 (μM) (±SD) hlGPa IC50 (μM) (±SD) rmGPb
1 2.00 (±0.16) 1.9 (±0.1)
2 0.65 (±0.05) 1.6 (±0.1)
3 2.48 (±0.13) 2.9 (±0.1)
4 1.30 (±0.06) 2.8 (±0.4)
Open in a new tab

Table 2.

Summary of diffraction data and refinement statistics for rmGPb: 1–4 complexes

rmGPb–1 rmGPb–2 rmGPb–3 rmGPb–4
Space group P43212 P43212 P43212 P43212
No. of images (°) 90 (72°) 55 (55°) 50 (50°) 25 (37.5°)
Unit cell dimensions (Å) a = b = 128.7, c = 116.0 a = b = 128.6, c = 116.0 a = b = 128.7, c = 116.1 a = b = 128.3, c = 115.9
Resolution range (Å) 29.5–2.26 29.36–2.26 29.49–2.26 29.24–2.26
No. of observations 394,073 279,449 246,348 143,690
No. of unique reflections 44,163 43,984 44,282 43,769
Outermost shell (Å) 2.30–2.26 2.30–2.26 2.30–2.26 2.30–2.26
Redundancy (outermost shell) 5.5 (5.4) 4.1 (2.9) 4.3 (3.2) 2.9 (2.5)
<I/σ(I)> (outermost shell) 15.6 (5.9) 6.7 (1.7) 9.9 (2.0) 6.6 (1.9)
Completeness (outermost shell) (%) 95.7 (88.6) 95.4 (83.5) 95.9 (81.9) 94.9 (65.7)
Rm (outermost shell) 0.063 (0.248) 0.127 (0.505) 0.116 (0.551) 0.129 (0.392)
Refinement (resolution) (Å) 29.5–2.26 29.36–2.26 29.49–2.26 29.24–2.26
No. of reflections used (free) 41,890 (2248) 41,749 (2216) 42,041 (2241) 41,573 (2196)
Residues included 13–314, 324–837 13–314, 324–837 13–314, 324–837 13–314, 324–837
No. of protein atoms 6634 6634 6634 6634
No. of water molecules 240 237 236 250
No. of ligand atoms 15 (PLP), 23 (compound 1) 15 (PLP), 23 (compound 2) 15 (PLP), 24 (compound 3) 15 (PLP), 23 (compound 4)
Final R (Rfree ) (%) 19.3 (23.3) 19.1 (23.2) 19.7 (23.9) 18.9 (23.0)
RMSD in bond lengths (Å) 0.008 0.007 0.007 0.007
RMSD in bond angles (°) 1.36 1.31 1.32 1.31
RMSD in dihedrals (°) 25.3 25.3 25.3 25.2
RMSD in impropers (°) 0.72 0.71 0.70 0.69
Average B (Å 2) for residues 16–250, 261–314, 324–837 16–250, 261–314, 324–837 16–250, 261–314, 324–837 16–250, 261–314, 324–837
Overall 27.8 34.0 33.5 29.0
Cα, C, N, O 26.9 32.6 32.1 27.5
Side chain 30.3 35.4 34.9 30.4
Average B (Å 2) for heteroatoms 14.9 (PLP) 35.3 (1) 21.6 (PLP) 46.5 (2) 20.3 (PLP) 53.1 (3) 16.0 (PLP) 35.6 (4)
Average B (Å 2) for water molecules 33.9 36.9 36.9 33.5
Open in a new tab

Merging Rm is defined as Rm=∑i∑h|<Ih>− Iih| /∑i∑hIih, where<Ih>and Iih are the mean and ith measurements of intensity for reflection h, respectively. σ(I) is the standard deviation of I. The crystallogaphic R-factor is defined as R=∑ | |Fo | − |Fc | | /∑ |Fo |, where |Fo | and |Fc | are the observed and calculated structure factor amplitudes, respectively. Rfree is the corresponding R-value for a randomly chosen 5% of the reflections that were not included in the refinement.

Portions of the 2Fo−Fc electron density maps for molecules 1–4 are shown in Figure 2 ▶. The molecules could be fitted unambiguously at the allosteric site, since clear density was present for all atoms of the inhibitor except for the aliphatic parts of hexanoic, butyric, and pentanoic acids. We describe below the rmGPb : 1 interactions and briefly the rmGPb : 2–4 interactions at the allosteric site.

Figure 2.

Figure 2.

Open in a new tab

Stereo diagrams of the 2Fo−Fc electron density maps, contoured at 1σ, for the bound compounds 1 (A), 2 (B), 3 (C), and 4 (D) at the allosteric site. Electron density maps were calculated using the standard protocol as implements in X-PLOR 3.8 (Brünger 1992) before incorporating ligand coordinates.

Ligand–enzyme interactions of compound 1

Compound 1makes polar contacts to the protein, involving all of the inhibitor’s potential hydrogen-bonding groups except N2 as well as van der Waals contacts. In the complex structure, 1 makes a total of three hydrogen bonds and 73 van der Waals interactions (1 polar/polar, 45 polar/nonpolar, and 27 nonpolar/nonpolar interactions) (Tables 3, 4). There are 31 contacts to the symmetry-related subunit of which 10 are interactions between nonpolar atoms. In specific, N1 makes a direct contact to main-chain O of Val40′, O1 forms an indirect contact to Arg193 NH1 via a water molecule (Wat195) and to Thr240 OG1 and Asp227 OD1 via another water molecule (Wat214), and O2 makes a hydrogen bond to the main-chain N of Asp42′. The hydrogen- bonding interactions formed between the ligand and the protein are illustrated in Figure 3A ▶. Compound 1 exploits numerous van der Waals contacts that are dominated by the substantial interactions to Val40′, Val45′, Trp67, Tyr75, and Arg193. These comprise mainly CH/π electron interactions between the hydrogen atoms of the aliphatic carbons and the π electrons of the aromatic ring (Nishio et al. 1995) (Val40′ side-chain/chlorophenyl group, Val45′ side-chain/dichlorophenyl group), aromatic/aromatic interactions (chlorophenyl group/CD2, CE2, CE3, CZ2, CZ3, and CH2 of Trp67), and nonpolar/nonpolar interactions (dichlorophenyl group/aliphatic part of Gln72, aliphatic chain of aliphatic part of hexanoic acid/ Tyr75). The side chain of Arg193 stacks against the chlorophenyl ring making some 10 van der Waals contacts, and this stacking may involve interactions characteristic of amino–aromatic interactions (Burley and Petsko 1988).

Table 3.

Hydrogen–bond interactions between compounds 1–4 and residues at the allosteric site of rmGPb

rmGPb–1 complex rmGPb–2 complex rmGPb–3 complex rmGPb–4 complex
Inhibitor atom Protein atom Distance (Å) Protein atom Distance (Å) Protein atom Distance (Å) Protein atom Distance (Å)
N1 Val40′ O 3.0 Val40′ O 3.1 Val40′ O 3.1 Val40′ O 3.0
O1 Wat195 2.8 Wat186 3.0 Wat188 3.1 Wat186 3.2
O2 Asp42′ N 3.0 Asp42′ N 3.1 Asp42′ N 3.0 Asp42′ N 2.9
Open in a new tab

RmGPb–1: Wat195 (or Wat186 or Wat188) is hydrogen–bonded to Arg193 NH1 and to Thr240 OG1 and Asp227 OD1 through another water molecule Wat214.

Table 4.

Van der Waals interactions between compounds 1–4 and residues at the allosteric site of rmGP

rmGPb–1 rmGPb–2 rmGPb–3 rmGPb–4
Inhibitor atom Protein atom No. of contacts Protein atom No. of contacts Protein atom No. of contacts Protein atom No. of contacts
C4 Arg193 CD, NE, CZ, NH1, NH2; Leu39′ O; Val40′ O, CB 8 Arg193 CD, NE, CZ, NH1, NH2; Val40′ O, CB 7 Arg193 CD, NE, CZ, NH1, NH2; Val40′ O, CB 7 Arg193 CD, NE, CZ, NH1, NH2; Val40′ O,CB 7
C5 Arg193 CZ, NH1, NH2; Val40′ O, C 5 Arg193 CZ, NH1, NH2; Val40′ O 4 Arg193 CZ, NH1, NH2; Val40′ O, 4 Arg193 CZ, NH1, NH2; Val40′ O 4
C6 Arg193 NH1; Val40′ O; Wat195 3 Arg193 NH1; Val40′ O; Wat186 3 Arg193 NH1; Val40′ O; Wat186 3 Arg193 NH1; Val40′ O; Wat186 3
C1 Val40′ O, CGI 2 Val40′ O, CGI 2 Val40′ O, CGI 1 Val40′ O, CGI; Wat186 3
CL1 Trp67 CD1, CB, CG, C, O, NE1, CD2, CE2; Gln71 CG 9 Trp67 CD1, CB, CG, C, O, NE1, CD2, CE2; Ile68 N, CA; Gln71 CG 11 Trp67 CD1, CB, CG, C, O, NE1, CD2, CE2; Ile68 N, CA; Gln71 CB, CG 12 Trp67 CD1, CG, C, O, NE1, CD2, CE2; Ile68 N; Gln71 CG 9
C2 Trp67 CD2, CE2, CE3, CZ2, CZ3, CH2; Val40′ CB, CG1 8 Trp67 CD2, CE2, CE3, CZ2, CZ3, CH2; Val40′ CG1 7 Trp67 CD2, CE2, CE3, CZ2, CZ3, CH2; Val40′ CG1 7 Trp67 CD2, CE2, CE3, CZ2, CZ3, CH2; Val40′ CG1 7
C3 Lys191 CG; Arg193 NH1; Val40′ O, CB, CG1 5 Arg193 NHI; Val40′ CB, CG1 3 Arg193 NHI; Val40′ CB, CG1 4 Arg193 CD, CZ, NH1; Val40′ CB, CG1 5
N1 Ile68 CG1 1 Ile68 CGI 1 Ile68 CG1 1 Ile68 CG1 1
C7 Val40′ O; Wat195 2 Val40′ O; Wat186 2 Val40′ O; Wat188 2 Val40′ O; Wat186 2
O1 Gln71 CB 1 Gln71 CB 1
N2 Val45′ CG2 1 Val45′ CG2 1 Val45′ CG2 1 Val45′ CG2 1
C9 Val45′ CG2 1 Val45′ CG2 1 Val45′ CG2 1 Val45′ CG2 1
C10 Val45′ CG2 1 Asp42′ CB; Val45′ CG2 2 Asp42′ CB, CG; Val45′ CG2 3 Asp42′CB, CG, OD2; Val45′ CG2 4
C11 Gln72 CB 1 Asp42′ CG, OD2 2 Asp42′ CB, CG, OD2 3
CL2 Gln72 OE1; Asp42′ CG, OD2; Asn44′ CB, CG, ND2 6 Trp67 CZ3, CH2; Lys191 CB, CG; Arg193 CD 5 Trp67 CZ3, CH2; Lys191 CB, CG; Arg193 CD 5 Trp67 CZ3, CH2; Lys191 CB, CG; Arg193 CD 5
C12 Gln72 CA, CB 2 Gln72 CB 1 Gln72 CB, OE1 2 Gln72 CB, OE1 2
O3 Tyr75 CD2 1 Gln72 OE1 1 Gln72 CA, CB; Tyr75 CD2 3
C13 Gln72 CA, CB 2 Gln72 CA, CB 2 Gln72 CA, CB 2
CL3 Gln72 CA; Tyr75 CB, CG, CD1, CD2 5 Gln71 C, O, CB; Gln72 N 4
C14 Gln72 CA 1 Gln72 CA, CB 2 Gln72 CA, CB 2
C15 Asn44′ CB, CG 2 Gln72 CA 1
C8 Ile68 CG1; Val45′ CG2 2 Ile68 CG1; CG2; Val45′ CG2 3 Ile68 CG1, CG2; Val40′ O; Val45′ CG2 4 Ile68 CG1, CG2; Val40′ O; Asp42′ N; Val45′ CG2 5
O2 Ile68 CG1, CG2; Lys41′ CA, C, N; Asp42′ CA, CB; Val45′ CG2 8 Ile68 CG1, CG2; Lys41′ CA, C; Asp42′ CA, CB; Val45′ CG2 7 Ile68 CG1, CG2; Val40′ O; Lys41′ CA, C; Asp42′ CA, CB; Val45′ CG2 8 Ile68 CB, CG1, CG2, OD1; Val40′ O; Lys41′ CA, C; Asp42′ CA, CB; Val45′ CG2 10
C20 Gln72 CA; Tyr75 CB; CG, CD2 4
C16 Tyr75 CE2, CG, CD2 3
Total 73 70 79 79
Open in a new tab

Figure 3.

Figure 3.

Open in a new tab

(A) Interactions between compound 1 (BN2) (shown in orange) and rmGPb at the allosteric site. Residues from subunit 1 are shown in green, and residues from subunit 2 are shown in cyan. (B) Comparison between the rmGPb–1 complex (shown in green) and T-state rmGPb (shown in blue) in the vicinity of the allosteric site. (C) rmGPa in the R state viewed in a similar orientation to that of the T-state rmGPb; the position of 1 is superimposed. Residues from subunit 1 are shown in light gray, and residues from subunit 2 are shown in dark gray.

The inhibitor becomes buried upon complex formation with rmGPb. The solvent accessibilities of the free and bound molecules are 724Å2 and 168Å2, indicating that 556Å 2 (77%) of the surface of 1 becomes inaccessible once bound to the enzyme allosteric binding site. The inhibitor is associated with both subunits (surface areas of 360Å 2 and 196Å 2, respectively). While both polar and nonpolar groups of the inhibitor are buried, nonpolar groups contribute 455Å 2 (82%) to the surface that becomes inaccessible. On binding of 1, a total of 191Å 2 of solvent-accessible protein surface area becomes inaccessible.

Comparison with the native T-state structure

Superposition of the activation locus (residues 23–40, 41–47, 48–78, 94–103, 104–111, 118–125 from both subunits) of the structure of the native T-state rmGPb with the activation locus of structure of the rmGPb–1 complex gave an r.m.s. deviation of 0.22Å for Cα atoms, indicating that the two structures have very similar overall conformations within the limits of the 2.26Å resolution data. The major conformational changes on binding of 1 to rmGPb occur in the vicinity of the allosteric site. Shifts for main-chain atoms are observed for residues 47′ to 49′ (between 0.5 and 0.8Å ), and residues 193 to 196 (between 0.5 and 1.7Å ) that affect the subunit–subunit interface in the region between the cap′ and the loop between β7 (residues 191 to 193) and β8 (residues 198 to 209) strands. The binding of 1 to rmGPb is accompanied by local conformational changes in the vicinity of the allosteric site. The greatest changes include shifts of the side-chain atoms of residues 193 to 196 by −1.0 to 3.8Å , side-chain atoms of residues 40′ to 41′ by −0.4 to 3.7Å , and also shifts of the side-chain atoms of residues 47′ to 49′ of −0.6 to 1.9Å . Similar shifts were observed previously on binding of W1807 to the allosteric site of rmGPb (Zographos et al. 1997) and rmGPa (Oikonomakos et al. 1999) and appear important in stabilizing a modified T state, denoted T′, that is more tensed than the T state.

In the native T-state rmGPb structure, hydrogen bonds across the subunit interface in the vicinity of the allosteric site are thought to stabilize the dimer structure. Arg193 hydrogen bonds to the main-chain oxygens of residues Leu39′ (2.9Å ) and Val40′ (2.9 Å) and Glu195 hydrogen bonds to Ly41′ (2.9 Å ). These contacts are present in both T- and R-state enzymes. In the complex structure with 1, the aforementioned major shift in Arg193 and minor shifts in residues 39′–40′ are responsible for the disruption of the intersubunit hydrogen bonds between Arg193 and Leu39′ and Val40′. However, Glu195 (OE1) is still able to hydrogen-bond to Lys41′ (3.1Å ). The Tyr185′/Pro194 interaction, an important subunit/subunit contact and a major link between the cap′/α2 and the tower/tower′, is also retained in the 1 complex. A comparison of the two structures in the vicinity of the allosteric site is shown in Figure 3B ▶.

Comparison with R-state GP structures

Comparison of the rmGPb–1 complex with the R-state rmGPa (Barford et al. 1991) suggests that the inhibitor is likely to have lower affinity for the R-state conformation. Superposition of the activation locus (residues 23–40, 41–78, 94–111, 118–125 from both subunits, as defined by Sprang et al. 1991) of the structure of the R-state rmGPa (subunit A) with the activation locus of structure of the rmGPb–1 complex gave an r.m.s. deviation of 1.21Å for Cα atoms. The transformation (obtained with LSQKAB) that allows superposition of the rmGPb–1 complex structure to the R-state rmGPa involves a rotation of one subunit by −5.2Å so as to bring the two subunits closer together at the twofold axis of the dimer. Incorporation of 1 into the allosteric site of the R-state rmGPa would result in clashes with the side chains of residues Val40′, Asp42′, Gln72, Tyr75, and Arg193. Movements of these residues that would enable binding of 1, as seen in the rmGPb–1 complex, appear to be suppressed by the subunit–subunit contacts that promote the R state (Fig. 3C ▶). Similarly, superposition of the activation loci in the R-state rmGPb–AMP complex (Barford et al. 1991) and in the rmGPb–1 complex gave an r.m.s. deviation of 1.31Å for Cα atoms. The requirement for shifts for the residues 40′, 42′, 44′, 67, and 193 is also apparent.

A comparison of the positions of AMP (R-state) and 1 (T′-state) bound at the allosteric site is shown in Figure 4 ▶. The ribose partially overlaps with the central acyl urea moiety of 1 (ribose C2′/N2, O3′/O1, and O2′/N1 separations are, respectively, 1.0Å , 1.6 Å , and 1.9 Å ), and the adenine partially overlaps with the dichlorophenyl group (adenine atoms C2, C4, N3, and N4 make close contacts [0.5–1.4 Å ] to O3, C13, C12, and C14 atoms of the dichlorophenyl group, respectively). In the R-state AMP complex, there is a significant shift in the side chain of Tyr75 from its T-state position, which results in van der Waals contacts from the tyrosine to the adenine. The loop region 193–196 and residues 39′–40′ and 47′–49′ shift on binding AMP, but these shifts are in the opposite direction to those observed with 1 binding.

Figure 4.

Figure 4.

Open in a new tab

Comparison of the positions of the inhibitor compound 1 (BN2) and the activator AMP bound at the allosteric site of rmGPb, after superimposing the rmGPb–1 protein structure onto the R-state rmGPb–AMP complex.

Ligand–enzyme interaction of compounds 2–4

The complex structures with derivatives 2–4 provide some additional insight into potential pathways for inhibitor optimization. In derivative 2 a chlorine atom (CL2) is introduced in the para (C3) position of the benzoyl ring. Also, the CL3 is shifted from the meta (C13) to the ortho position (C14) of the phenyl ring to avoid the steric clash with Tyr75. The second chloro group at the phenyl ring is removed. The pattern of polar and nonpolar interactions between the inhibitor and enzyme is maintained when comparing the complex structures of derivatives 1 and 2 (Tables 3, 4; Fig. 5 ▶). We would have expected larger effects from these changes, particularly from removing the steric clash seen between CL3 and Tyr75 in the complex with 1. Anyhow, it is interesting to observe that in the absence of the CL2 group, the meta CL3 is positioned toward the phosphate-recognition region of the allosteric site.

Figure 5.

Figure 5.

Open in a new tab

Interactions of compound 2 (BN3) (A), compound 3 (BN4) (B), and compound 4 (BN5) (C) with rmGPb in the vicinity of the allosteric site. Color code as in Figure 3A ▶.

The positions and conformations of 3 and 4, bound at the allosteric site, are also similar with those observed for 1 and 2. As expected from complex 1 the methyl group (C20) of 3 is found in close contact with Tyr75, and it seems that this is not a favorable environment for this substituent. In compound 4 the phenoxy-pentanoic acid is placed in the meta position of the phenyl ring and could thereby form additional polar interactions of the carboxylic function with the adjacent polar residues, for example, Asn72. Still, this modification resulted neither in improved ordering of the pentanoic acid chain nor in improved affinity of derivative 4.

Overall, it is notable that, as in the case of compound 1, exactly the same conformational rearrangements are observed on binding of 2–4 to the allosteric site of rmGPb, and this indicates that binding of acyl ureas 1–4 to rmGPb stabilize similar conformations. The LSQKAB superposition of the structure of the native T-state rmGPb with the refined rmGPb–2, rmGPb–3, and rmGPb–4 complex structures over the activation loci gave r.m.s. deviations of 0.25 Å , 0.24 Å , and 0.27 Å for Cα atoms, respectively, indicating that the four structures have very similar overall conformations. Similarly, the LSQKAB superposition of the rmGPb–1 complex with rmGPb–2, rmGPb–3, and rmGPb–4 complex structures gave an r.m.s. deviation of 0.06 Å for Cα atoms.

Conclusions

The X-ray crystallographic study of four rmGPb complexes with acyl urea inhibitors showed that compounds 1–4 bind at the allosteric site, at the subunit–subunit interface of the dimer. All four derivatives form three hydrogen bonds with the main-chain carbonyl of Val40′, the main-chain NH of Asp42′ and an ordered water molecule (through the central acyl urea moiety), and make extensive nonpolar interactions with the enzyme that involve the side chains of Trp67, Asn72, Tyr75, Arg193, and Val40′, Lys41′, and Val45′ from the symmetry-related subunit. These interactions provide a rationale for their potency to inhibit hlGPa activity. In contrast, to previously known GP inhibitors that bind to the allosteric site, compounds 1–4 do not exploit the phosphate-recognition site, which comprises Arg309, Arg310, and a more distant arginine, Arg242. The phosphate-recognition site appears accessible by substituents positioned at the C13 or C14 atoms of the acyl urea scaffold (Scheme I ▶). This site recognizes a variety of phosphorylated compounds, such as AMP, ATP, Glc-6-P, and the nonphysiological phosphate- mimetic W1807.

The position of inhibitors 1–4 is distinct but partially overlapping with the position of AMP bound to the relaxed state of the enzyme (Fig. 4 ▶). The binding of acyl ureas and AMP is therefore mutually exclusive. The formation of a stable rmGPb–acyl urea complex requires conformational changes in the vicinity of the allosteric site, particularly in the backbone and side chain of residues 193 to 196 and in the cap′ region of the other subunit (39′–40′ and 47′–49′). The ligand-induced conformational changes are characteristic of the T′-state conformation. The key rearrangement is probably the backbone displacement of the loop 193–196 that allows for van der Waals interactions with the ligand similar to those observed with W1807 (Zographos et al. 1997; Oikonomakos et al. 1999). Overall, we suggest that acyl ureas inhibit the enzyme directly by preventing binding of the allosteric activator AMP and indirectly (allosterically) by stabilizing the T′ state.

The results with W1807 and several dihydropyridine diacid analogs (Zographos et al. 1997; Oikonomakos et al. 1999; Ogawa et al. 2003), phenyl diacid analogs (Lu et al. 2003), phenoxy-phthalates (Kristiansen et al. 2004), and acyl ureas (this work) show how nonphysiological compounds could be potent inhibitors of glycogenolysis. The structural results obtained with the acyl ureas can be further exploited by means of chemical optimization to yield new potent inhibitors. In fact, rational design and parallel synthesis, used to develop a series of acyl ureas, led to the discovery of hlGPa inhibitors (T. Klabunde, K.U. Wendt, D. Kadereit, V. Brachvogel, H.-J. Burger, A.W. Herling, N.G. Oikonomakos, M.N. Kosmopoulou, D. Schmoll, E. Sarubbi, et al., unpubl.), which when administered to anaesthetized Wistar rats caused a dose-dependent reduction of the glucagoninduced hyperglycemic peak.

Materials and methods

The synthesis of compounds 1–4 is described in Defossa et al. (2001). RmGPb was isolated, purified, recrystallized, and assayed as previously described (Oikonomakos et al. 2002). Kinetics studies were performed in the direction of glycogen synthesis with 10 μg/mL enzyme, constant concentrations of glycogen (1% w/v), Glc-1-P (4 mM), AMP (40 μM), 1% DMSO, and various concentrations of inhibitors 1–4 (1–20 μM). The inorganic phosphate released in the reaction was determined according to Saheki et al. (1985). HlGPa was expressed and purified according to the procedures described by Rath et al. (2000). The activity of recombinant hlGPa was monitored in the direction of glycogen synthesis. The standard assay mixture (100 μL final volume) contained 30 mM HEPES (pH 7.2), 60mMKCl, 1.5 mMEDTA, 1.5 mMMgCl2, 1 mg/mL glycogen, 1 mM Glc-1-P, 7 mU hlGPa, 1% DMSO, and the respective inhibitors at a concentration within the range 0–100 μM. The reaction was initiated by the addition of Glc-1-P and incubated for 40 min at 25°C. The inorganic phosphate released was assayed according to Drueckes et al. (1995).

RmGPb complexes with compounds 1–4 were cocrystallized in a medium consisting of 23–25 mg/mL enzyme, compounds 1–4 (at molar ratios of compound/enzyme varied from 2 to 10), 3 mM dithiothreitol, 10 mM Bes, 0.1 mM EDTA, and 0.02 % sodium azide (pH 6.7). Crystallographic data were collected from small single crystals on an image plate RAXIS IV using a Rigaku Ru-H3RHB belt drive rotating anode (λ =1.5418 Å ), operating at 60 kV, 100 mA or on an image plate at EMBL-Hamburg outstation (beamline X31, λ =1.24 Å ). Crystal orientation, integration of reflections, interframe scaling, partial reflection summation, data reduction, and post-refinement were all performed using DENZO and SCALEPACK (Otwinowski and Minor 1997).

Crystallographic refinement of the rmGPb complexes was performed with X-PLOR version 3.8 (Brünger 1992) using bulk solvent corrections. All data were included with no σ cutoff. The starting model was a refined structure of the native T-state rmGPb at 1.8 Å resolution (N.G. Oikonomakos, E.D. Chrysina, D.D Leonidas, and M.N. Kosmopoulou, unpubl.), with water molecules removed. Throughout the refinement, 5% of the data were flagged for calculation of Rfree. The Fourier maps calculated with SIGMAA (Read 1986) weighted (2mFo−DFc) and (Fo−Fc) coefficients indicated binding of compounds 1–4 at the allosteric site. Minimized conformers of compounds 1–4 generated using the program SYBYL (Tripos Associates Inc.) were fitted to the electron density map after small adjustments of the torsion angles. Map interpretation was performed using the program O (Jones et al. 1991). Several side chains of the enzyme model were adjusted; water molecules were added and retained only if they met stereochemical requirements by using WATERPICK. The final models were refined by the conventional positional and restrained individual B-factor refinement protocol as implemented in X-PLOR.

The structures were analyzed with the graphics program O (Jones et al. 1991). Hydrogen bonds and van der Waals interactions were calculated with the program CONTACT (CCP4 1994) applying a distance cutoff of 3.3 Å and 4.0 Å , respectively. The program calculates the angle O…H…N (where the hydrogen position is unambiguous) and the angle source…oxygen-bonded carbon atom. Suitable values are 120° and 90° RmGPb complex structures were superimposed using LSQKAB (Collaborative Computational Project No. 4 1994). Coordinate sets for comparison were: room temperature T-state rmGPb, R-state rmGPa (PDB code 1GPA), and R-state rmGPb-AMP complex (PDB code 7GPB). The schematic representation of the crystal structures presented in all figures were prepared with the programs MolScript (Kraulis 1991) and BobScript (Esnouf 1997) and rendered with Raster3D (Merritt and Bacon 1997). The coordinates of the new structures have been deposited with the RCSB Protein Data Bank (http://www.rcsb.org/pdb) with codes 1WUT (rmGPb–1 complex), 1WVX (rmGPb–2 complex), 1WV0 (rmGPb–3 complex), and 1WV1 (rmGPb–4 complex).

Acknowledgments

This work was supported by the Greek General Secretariat for Research and Technology (GSRT) through the programs PENED-204/2001 (MNK), ENTER-EP6/2001 (EDC), Aventis Pharma Deutschland GmbH, a company of the Sanofi-Aventis Group, and the EMBL-Hamburg outstation under FP6 “Structuring the European Research Area Programme” contract no. RII3/CT/2004/5060008. The assistance of V.T. Skamnaki in protein crystallization is also acknowledged. We also acknowledge the assistance of the staff at EMBL-Hamburg outstation for providing excellent facilities for X-ray data collection.

Abbreviations

  • GP, glycogen phosphorylase, 1,4-α-D-glucan:orthophosphate α-glucosyltransferase (EC 2.4.1.1)

  • rmGPb, rabbit muscle glycogen phosphorylase b

  • rmGPa, rabbit muscle glycogen phosphorylase a

  • hlGPa, human liver glycogen phosphorylase a

  • PLP, pyridoxal 5′-phosphate

  • glucose, α-D-glucose

  • Glc-1-P, α-D-glucose 1-phosphate

  • Glc-6-P, D-glucose 6-phosphate

  • W1807, (−)(S)-3-isopropyl 4-(2-chlorophenyl)-1,4-dihydro-1-ethyl-2-methyl-pyridine-3,5,6-tricarboxylate

  • CP320626, 5-chloro-1H-indole-2-carboxylic acid [1-(4-fluorobenzyl)- 2-(4-hydroxypiperidin-1-yl)-2-oxoethyl]amide

  • compound 1, 6-{2,6-dichloro-4-[3-(2-chloro-benzoyl)-ureido]-phenoxy}-hexanoic acid

  • compound 2, 4-{3-chloro-4-[3-(2,4-dichloro-benzoyl)-ureido]- phenoxy}-butyric acid

  • compound 3, 4-{4-[3-(2,4-dichloro-benzoyl)-ureido]-2,3-dimethyl-phenoxy}-butyric acid

  • compound 4, 5-{3-[3- (2,4-dichloro-benzoyl)-ureido]-2-methyl-phenoxy}-pentanoic acid

  • r.m.s. deviation, root-mean-square deviation.

Article and publication are at http://www.proteinscience.org/cgi/doi/10.1110/ps.051432405.

References

  1. Aiston, S., Hampson, L., Gómez-Foix, A.M., Guinovart, J.J., and Agius, L. 2001. Hepatic glycogen synthesis is highly sensitive to phosphorylase activity. J. Biol. Chem. 276 23858–23866. [DOI] [PubMed] [Google Scholar]
  2. Aiston, S., Coghlan, M.P., and Agius, L. 2003. Inactivation of phosphorylase is a major component of the mechanism by which insulin stimulates hepatic glycogen synthesis. Eur. J. Biochem. 270 2773–2781. [DOI] [PubMed] [Google Scholar]
  3. Barford, D., Hu, S.H., and Johnson, L.N. 1991. Structural mechanism for glycogen phosphorylase control by phosphorylation and AMP. J. Mol. Biol. 218 233–260. [DOI] [PubMed] [Google Scholar]
  4. Brünger, A.T. 1992. X-PLOR: Version 3.1; a system for protein crystallography and NMR. Yale University Press, New Haven, CT.
  5. Burley, S.K. and Petsko, G.A. 1988. Weakly polar interactions in proteins. Adv. Protein Chem. 39 125–189. [DOI] [PubMed] [Google Scholar]
  6. Collaborative Computational Project No. 4. 1994. The CCP4 suite: Programs for protein crystallography. Acta Crystallogr. D50 760–763. [DOI] [PubMed] [Google Scholar]
  7. Defossa, E., Klabunde, T., Burger, H.-J., Herling, A., and Baringhaus, K.-H. 2001. New acylphenylurea derivatives are hypoglycemic agents and glycogen phosphorylase inhibitors—useful for the treatment of hypertension and type II diabetes. Priority: 04 April 2001 (DE-116768), Filing date: 26 May 2001 as E06030.
  8. Drueckes, P., Schinzel, R., and Palm, D. 1995. Photometric microtiter assay of inorganic phosphate in the presence of acid-labile organic phosphates. Anal. Biochem. 230 173–177. [DOI] [PubMed] [Google Scholar]
  9. Esnouf, R.M. 1997. An extensively modified version of Molscript that includes greatly enhanced coloring capabilities. J. Mol. Graph. Model. 15 132–134. [DOI] [PubMed] [Google Scholar]
  10. Jones, T.A., Zou, J.Y., Cowan, S.W., and Kjeldgaard, M. 1991. Improved methods for the building of protein models in electron density maps and the location of errors in these models. Acta Crystallogr. A47 110–119. [DOI] [PubMed] [Google Scholar]
  11. Kraulis, P. 1991. MOLSCRIPT: A program to produce both detailed and schematic plots of protein structures. J. Appl. Crystallogr. 24 946–950. [Google Scholar]
  12. Kristiansen, M., Andersen, B., Iversen, L.F., and Westergaard, N. 2004. Identification, synthesis, and characterization of new glycogen phosphorylase inhibitors binding to the allosteric AMP site. J. Med. Chem. 47 3537–3545. [DOI] [PubMed] [Google Scholar]
  13. Kurukulasuriya, R., Link, J.T., Madar, D.J., Pei, Z., Richards, S.J., Rohde, J.J., Souers, A.J., and Szczepankiewicz, B.G. 2003. Potential drug targets and progress towards pharmacologic inhibition of hepatic glucose production. Curr. Med. Chem. 10 123–153. [DOI] [PubMed] [Google Scholar]
  14. Lu, Z., Bohn, J., Bergeron, R., Deng, Q., Ellsworth, K.P., Geissler, W.M., Harris, G., McCann, P.E., McKeever, B., Myers, R.W., et al. 2003. A new class of glycogen phosphorylase inhibitors. Bioorg. Med. Chem. Lett. 13 4125–4128. [DOI] [PubMed] [Google Scholar]
  15. McCormack, J.C., Westergaard, N., Kristiansen, M., Brand, C.L., and Lau, J. 2001. Pharmacological approaches to inhibit endogenous glucose production as a means of anti-diabetic therapy. Curr. Pharm. Des. 7 1451–1474. [DOI] [PubMed] [Google Scholar]
  16. Merritt, E.A. and Bacon, D.J. 1997. Raster3D: Photorealistic molecular graphics. Methods Enzymol. 277 505–524. [DOI] [PubMed] [Google Scholar]
  17. Nishio, M., Umezawa, Y., Hirota, M., and Takeuchi, Y. 1995. The CH/π interaction significance in molecular recognition. Tetrahedron 51 8665–8701. [Google Scholar]
  18. Ogawa, A.K., Willoughby, C.A., Bergeron, R., Ellsworth, K.P., Geissler, W.M., Myers, R.W., Yao, J., Harris, G., and Chapman, K.T. 2003. Glucose-lowering in a db/db mouse model by dihydropyridine diacid glycogen phosphorylase inhibitors. Bioorg. Med. Chem. Lett. 13 3405–3408. [DOI] [PubMed] [Google Scholar]
  19. Oikonomakos, N.G. 2002. Glycogen phosphorylase as a molecular target for type 2 diabetes therapy. Curr. Protein Pept. Sci. 3 561–586. [DOI] [PubMed] [Google Scholar]
  20. Oikonomakos, N.G., Tsitsanou, K.E., Zographos, S.E., Skamnaki, V.T., Goldmann, S., and Bischoff, H. 1999. Allosteric inhibition of glycogen phosphorylase a by the potential antidiabetic drug 3-isopropyl 4-(2-chlorophenyl)-1,4-dihydro-1-ethyl-2-methyl-pyridine-3,5,6-tricarboxylate. Protein Sci. 8 1930–1945. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Oikonomakos, N.G., Schnier, J.B., Zographos, S.E., Skamnaki, V.T., Tsitsanou, K.E., and Johnson, L.N. 2000a. Flavopiridol inhibits glycogen phosphorylase by binding at the inhibitor site. J. Biol. Chem. 275 34566–34573. [DOI] [PubMed] [Google Scholar]
  22. Oikonomakos, N.G., Skamnaki, V.T., Tsitsanou, K.E., Gavalas, N.G., and Johnson, L.N. 2000b. A new allosteric site in glycogen phosphorylase b as a target for drug interactions. Structure 8 575–584. [DOI] [PubMed] [Google Scholar]
  23. Oikonomakos, N.G., Kosmopoulou, M., Zographos, S.E., Leonidas, D.D., Chrysina, E.D., Somsák L., Nagy, V., Praly, J.-P., Docsa, T., Tóth, B., et al. 2002. The binding of N′-acetyl- and benzoyl-N′-β-D-glucopyranosyl ureas to glycogen phosphorylase b: Kinetic and crystallographic studies. Eur. J. Biochem. 269 1–13. [DOI] [PubMed] [Google Scholar]
  24. Otwinowski, Z. and Minor, W. 1997. Processing of X-ray diffraction data collected in oscillation mode. Methods Enzymol. 276 307–326. [DOI] [PubMed] [Google Scholar]
  25. Rath, V.L., Newgard, C.B., Sprang, S.R., Goldsmith, E.J., and Fletterick, R.J. 1987. Modeling the biochemical differences between rabbit muscle and human liver phosphorylase. Proteins Struct. Funct. Genet. 2 225–235. [DOI] [PubMed] [Google Scholar]
  26. Rath, V.L., Ammirati, M., LeMotte, P.K., Fennell, K.F., Mansour, M.N., Danley, D.E., Hynes, T.R., Schulte, G.K., Wasilko, D.J., and Pandit J. 2000. Activation of human liver glycogen phosphorylase by alteration of the secondary structure and packing of the catalytic core. Mol. Cell 1 139–148. [PubMed] [Google Scholar]
  27. Read, R.J. 1986. Improved Fourier coefficients for maps using phases from partial structures with errors. Acta Crystallogr. A42 140–149. [Google Scholar]
  28. Saheki, S., Takeda, A., and Shimazu, T. 1985. Assay of inorganic phosphate in the mild pH range, suitable for measurement of glycogen phosphorylase activity. Anal. Biochem. 148 277–281. [DOI] [PubMed] [Google Scholar]
  29. Somsák, L., Kovács, L., Tóth, M., Ösz, E., Szilágyi, L., Györgydeák, Z., Dinya, Z., Docsa, T., Tóth, B., and Gergely, P. 2001. Synthesis of and a comparative study on the inhibition of muscle and liver glycogen phosphorylases by epimeric pairs of D-gluco- and D-xylopyranosylidene-spiro- (thio)hydantoins and N-(D-glucopyranosyl) amides. J. Med. Chem. 44 2843–2848. [DOI] [PubMed] [Google Scholar]
  30. Somsák, L., Nagy, V., Hadady, Z., Docsa, T., and Gergely, P. 2003. Glucose analog inhibitors of glycogen phosphorylases as potential antidiabetic agents: Recent developments. Curr. Pharm. Des. 9 1177–1189. [DOI] [PubMed] [Google Scholar]
  31. Sprang, S.R., Withers, S.G., Goldsmith, E.J., Fletterick, R.J., and Madsen, N.B. 1991. Structural basis for activation of glycogen phosphorylase b by adenosine monophosphate. Science 254 1367–1371. [DOI] [PubMed] [Google Scholar]
  32. Treadway, J.L., Mendys, P., and Hoover, D.J. 2001. Glycogen phosphorylase inhibitors for treatment of type 2 diabetes mellitus. Expert. Opin. Investig. Drugs 10 439–454. [DOI] [PubMed] [Google Scholar]
  33. Zographos, S.E., Oikonomakos, N.G., Tsitsanou, K.E., Leonidas, D.D., Chrysina, E.D., Skamnaki, V.T., Bischoff, H., Goldmann, S., Watson, K.A., and Johnson, L.N. 1997. The structure of glycogen phosphorylase b with an alkyldihydropyridine-dicarboxylic acid compound, a novel and potent inhibitor. Structure 5 1413–1425. [DOI] [PubMed] [Google Scholar]

Articles from Protein Science : A Publication of the Protein Society are provided here courtesy of The Protein Society

RESOURCES