C7/C8-cyclitols and C7N-aminocyclitols find applications in the pharmaceutical sector as α-glucosidase inhibitors and in the agricultural sector as fungicides and insecticides. In this study, we identified C7/C8-cyclitols and C7N-aminocyclitols as potential inhibitors of Streptomyces coelicolor (Sco) GlgEI-V279S based on the docking scores. The protein and the ligand (targets 11, 12, and 13) were prepared, the states were generated at pH 7.0 ± 2.0, and the ligands were docked into the active sites of the receptor via Glide™. The synthetic route to these targets was similar to our previously reported route used to obtain 4-⍺-glucoside of valienamine (AGV), except the protecting group for target 12 was a p-bromobenzyl (PBB) ether to preserve the alkene upon deprotection. While compounds 11-13 did not inhibit Sco GlgEI-V279S at the concentrations evaluated, an X-ray crystal structure of the Sco GlgE1-V279S/13 complex was solved to a resolution of 2.73 Å. This structure allowed assessment differences and commonality with our previously reported inhibitors and was useful for identifying enzyme-compound interactions that may be important for future inhibitor development. The Asp 394 nucleophile formed a bidentate hydrogen bond interaction with the exocyclic oxygen atoms (C(3)-OH and C(7)-OH) similar to the observed interactions with the Sco GlgEI-V279S in a complex with AGV (PDB:7MGY). In addition, the data suggest replacing the cyclohexyl group with more isosteric and hydrogen bond-donating groups to increase binding interactions in the + 1 binding site.
C7/C8-cyclitols and C7N-aminocyclitols find applications in the pharmaceutical sector as α-glucosidase inhibitors and in the agricultural sector as fungicides and insecticides. In this study, we identified C7/C8-cyclitols and C7N-aminocyclitols as potential inhibitors of Streptomyces coelicolor (Sco) GlgEI-V279S based on the docking scores. The protein and the ligand (targets 11, 12, and 13) were prepared, the states were generated at pH 7.0 ± 2.0, and the ligands were docked into the active sites of the receptor via Glide™. The synthetic route to these targets was similar to our previously reported route used to obtain 4-⍺-glucoside of valienamine (AGV), except the protecting group for target 12 was a p-bromobenzyl (PBB) ether to preserve the alkene upon deprotection. While compounds 11-13 did not inhibit Sco GlgEI-V279S at the concentrations evaluated, an X-ray crystal structure of the Sco GlgE1-V279S/13 complex was solved to a resolution of 2.73 Å. This structure allowed assessment differences and commonality with our previously reported inhibitors and was useful for identifying enzyme-compound interactions that may be important for future inhibitor development. The Asp 394 nucleophile formed a bidentate hydrogen bond interaction with the exocyclic oxygen atoms (C(3)-OH and C(7)-OH) similar to the observed interactions with the Sco GlgEI-V279S in a complex with AGV (PDB:7MGY). In addition, the data suggest replacing the cyclohexyl group with more isosteric and hydrogen bond-donating groups to increase binding interactions in the + 1 binding site.
Cyclitols and aminocyclitols are an important class of compounds found as both natural products or produced by chemical means. Of these, C7-cyclitols, some C8-cyclitols, and C7N-aminocyclitols stand out for their ability to mimic carbohydrates and interfere with a variety of enzymatic processes. For example, derivatives of valienol (streptol) (Hsiao et al., 2019) or some inositols such as C7-cyclitols 1 and C8-cyclitols 2, Figure 1 each with good leaving groups at the C-1 position have found roles as mechanism-based inhibitors of the α-glucosidases (Chakladar et al., 2014; Adamson et al., 2016; Shamsi Kazem Abadi et al., 2017). The strained cyclopropyl in C8-cyclitol 2 is believed to enable the enzyme-catalyzed formation of cationic intermediates within the active sites of retaining α-galactosidases (Chakladar et al., 2014). Closely related C7N-aminocyclitols such as the amylostatins (3) (Fukuhara et al., 1982a;b;Sakairi and Kuzuhara, 1982) and adiposins (4) are also known inhibitors of α-glucosidases, Figure 1. On the other hand, validamycin A (5) has important applications in agriculture due to its insecticidal, fungicidal, and fungistatic activities. The antifungal activity of 5 is attributed to its ability to inhibit trehalase (Treh) in the fungi (Neyman et al., 2022; Ren et al., 2022). (+)-Validoxylamine-A (6) is a common core for validamycins A, C, D, E, and F and is also known inhibitor of Treh of various origins (Kameda et al., 1987; Ishikawa et al., 2005). Valienamine (7), an essential unit in many commercial glucosidase inhibitors such as 6 and acarbose, is a potent inhibitor itself. Other closely related aminosugars such as trehazolin (8) (Liebl et al., 2010; El Nemr and El Ashry, 2011) and kirkamide (9) (Sieber et al., 2020) share similar abilities to inhibit Treh, while the C6 epimer of valienamine, epi-valienamine is an inhibitor for various hexosaminidases (Scaffidi et al., 2007) and mannosidases (Ramstadius et al., 2009). The latter has potential as a therapeutic treatment for Gaucher disease (Lin et al., 2004).
FIGURE 1
Previously isolated/synthesized C7/C8-cyclitols and C7N-aminocyclitols inhibitors.
Previously isolated/synthesized C7/C8-cyclitols and C7N-aminocyclitols inhibitors.Although most of these covalent and non-covalent inhibitors are naturally occurring, various synthetic approaches have been adopted to chemically synthesize the C7/C8-cyclitols and C7N-aminocyclitols. Bennett et al. have synthesized 1 and 2, among other covalent inhibitors, using a linear synthetic route from 2,3,4,6-tetra-O-benzyl glucopyranose (Chakladar et al., 2014; Adamson et al., 2016; Shamsi Kazem Abadi et al., 2017). The leaving groups were introduced toward the end using a nucleophilic aromatic substitution (Beenakker et al., 2017). Compounds 3, 4, 5, and 6 and their derivatives have been synthesized by various groups (Knapp et al., 1992; Kapferer et al., 1999; Scaffidi et al., 2007) using similar approaches starting with glucose or mannose scaffolds (Ramstadius et al., 2009). Oligosaccharides such as adiposins (4) and acarbose (Ogawa et al., 1985; Shibata and Ogawa, 1989) have been obtained using common glycosylation methods (Mcauliffe et al., 1996). Birch reduction was widely used to deprotect the benzyl protecting groups as the conditions leave the C5–C6 double bond intact in the valienamine-derived series (Kapferer et al., 1999). Recently, our group has reported an optimized stereoselective route to obtain maltose-based carbasugar derivatives of valienamine and epi-valienamine starting from α-D-maltose (Si et al., 2021). In this work, we used a common intermediate (10 (Si et al., 2021) or 10′) to obtain the targets 11, 12, and 13 (Figure 2), which we hoped would be inhibitors of Streptomyces coelicolor (Sco) GlgEI V279S.
FIGURE 2
Proposed potential inhibitors 11, 12, and 13 from carbasugar intermediate 10.
Proposed potential inhibitors 11, 12, and 13 from carbasugar intermediate 10.Sco GlgEI is a homolog of Mycobacteria tuberculosis (Mtb) GlgE. In Mtb, GlgE is part of a biosynthetic pathway that generates α-glucans from trehalose and is a genetically validated anti-TB target (Kalscheuer et al., 2010; Miah et al., 2013). GlgE enzymes polymerize maltose-1-phosphate (M1P) to linear α-(1→4) glucans, Figure 3A. Some of this GlgE-derived glycogen is exported in Mtb to form a capsule correlated with increased virulence (Sambou et al., 2008; Koliwer-Brandl et al., 2016). Mechanistically, Mtb GlgE and Sco GlgEI are similar to glycoside hydrolases (GH), but have no hydrolase activity on α-glucan, and are present in over 10% of sequenced genomes from bacteria and archaea, see Figure 3A (Chandra et al., 2011). Increasingly, new connections between bacterial glycogen and trehalose metabolism are being discovered (Chandra et al., 2011). Recently, it was found that carbon flux is diverted from cell wall synthesis in a process called the trehalose-catalytic shift which diverts trehalose into the GlgE pathway during the early stage of drug-tolerant persister bacilli formation, suggesting a vital role for the pathway in both bacilli persistence and drug resistance (Lee et al., 2019). In these studies, we used Sco GlgEI-V279S, which possesses the same active site topology and active site residues as Mtb GlgE; however, Sco GlgEI-V279S is more stable and is highly amenable to structural studies (Veleti et al., 2014a). For some time, our lab has been exploring the development of compounds which can be used to inhibit Sco GlgEI-V279S and Mtb GlgE (Lindenberger et al., 2015; Thanna et al., 2015; Veleti et al., 2017; Veleti et al., 2017; Si et al., 2021).
FIGURE 3
(A)
Mtb GlgE Sco GlgEI catalyzing the conversion of M1P to α-glucan. (B) Structures of mechanism-based inactivators of glycoside hydrolases and proposed conformations of their ionized intermediates.
(A)
Mtb GlgE Sco GlgEI catalyzing the conversion of M1P to α-glucan. (B) Structures of mechanism-based inactivators of glycoside hydrolases and proposed conformations of their ionized intermediates.These current studies seek to identify additional reversible and covalent inhibitors which can inhibit Sco GlgEI-V279S. 2-Deoxy-2-fluoroglycosides have already been shown to covalently inhibit Sco GlgEI, and these types of deoxy glycoside inhibitors also work on several other classes of glycoside hydrolases (Vocadlo and Davies, 2008; Syson et al., 2014). Other emerging inhibitor classes include the cyclophellitol-based epoxides, aziridines, and cyclosulfates that have been shown to modify retaining α- or β-glucosidases (Gloster et al., 2007; Kallemeijn et al., 2012; Artola et al., 2017; Chen et al., 2021). Quinone methide–based glycosides also serve as activity-based inhibitors to this class of enzymes (Chauvigne-Hines et al., 2012). Our attention was drawn to the vinyl 1 and cyclopropyl 2 carbasugar inhibitors, as shown in Figure 3B, which were shown to inhibit α-retaining hydrolases (Adamson et al., 2016; Shamsi Kazem Abadi et al., 2017; Ren et al., 2018) based on our prior maltose-based carbasugar work (Si et al., 2021).Vinyl 1 and cyclopropyl 2 carbasugars have been suggested to ionize to their resonance stabilized carbocations followed by a reaction with the active site residues. The cyclopropyl cation is noteworthy as it can adopt the bisected conformation leading to an overall 1,4B conformer in the carbasugar, a conformation close to the 1S3 conformation predicted to exist immediately after a nucleophilic attack on the natural substrate for GH13 class enzymes (Figure 3B). If the vinyl carbasugar were to ionize, it would be predicted to adopt an E3 conformation for maximal orbital overlap with the π-system, as shown in Figure 3B (Shamsi Kazem Abadi et al., 2017). The E3 conformation is very close to 4H3 conformation, which is generally invoked as the transition state conformation for GH13 enzymes (Shamsi Kazem Abadi et al., 2017). Thus, in this work, we sought to investigate whether maltose-based homologs of cyclopropyl and vinyl carbasugars, compounds 11 and 12, respectively, could potentially inhibit Sco GlgEI-V279S. In addition, we report that the common starting material 10 can be used for entry into amylostatin GXG-like derivatives 13, which were found to form a complex with Sco GlgEI-V279S.
Results and discussion
Chemistry: Targets 11 and 12 were obtained via a 14-step synthetic route but using two different global protecting groups. While the benzyl group was convenient to synthesize 11, it proved to be challenging to deprotect benzyl using traditional methods without affecting the alkene group in 12. Therefore, we used a modifiable global protecting group p-bromobenzyl (PBB) which could later be converted into a labile leaving group via a Buchwald–Hartwig amination (Plante et al., 2000). We started the synthesis using commercially available D-(+)-maltose on a gram scale. Compound 14 was obtained using reported literature in two steps (Veleti et al., 2014b). After acetyl deprotection of 14 using Zemplén conditions, we installed benzyl and p-bromobenzyl protecting group to access previously reported thiomaltoside 15 (Si et al., 2021) and also the new PBB-protected 15′ compound, respectively (Scheme 1).
SCHEME 1
Using different protecting groups to make intermediates 15 and 15′.
Using different protecting groups to make intermediates 15 and 15′.The protected thiomaltosides 15 and 15′ were each subjected to N-bromosuccinimide in 9:1 acetone:water to obtain hemiacetals 16 and 16′. Compounds 16 and 16′ were subjected to Wittig olefination, Moffatt oxidation, and Grignard reactions to obtain the dienes 19A, 19B, 19A′, and 19B′ as intermediates (Scheme 2) (Pfitzner and Moffatt, 1963; Chiara et al., 2008). After obtaining dienes 19A′ and 19B′, we could conclude that 19A′ was the desired diene owing to the similarities in the NMR splitting pattern of the alkene protons with previously reported 19A in comparison to 19A′ and previously reported 19B in comparison to 19B′ (Si et al., 2021). The upfield shift observed for 19A′ (H-1′ = 5.26 ppm) relative to 19B′ (H-1′ = 5.41 ppm) was similar to the upfield shift of 19A (H-1′ = 5.30 ppm) relative to 19B (H-1′ = 5.44 ppm), which is indicative of the desired isomer.
SCHEME 2
Synthesis of intermediates 19A, 19B, 19A′, and 19B′. Yields are reported for new compounds 15′ to 19A′/19B′. Compounds 15–19A/19B are previously reported and have similar yields (Si et al., 2021).
Synthesis of intermediates 19A, 19B, 19A′, and 19B′. Yields are reported for new compounds 15′ to 19A′/19B′. Compounds 15–19A/19B are previously reported and have similar yields (Si et al., 2021).Intermediate 19A′ was subjected to a ring-closing metathesis to afford 10A′ in 80% yield, which was notably higher than the same reaction reported for the benzyl-protected analogue 19A which proceed in 25% yield (Scheme 3) (Si et al., 2021). Although the intermediates 19A and 19A′ both experience similar steric bulk from their protecting groups, the presence of an electron-withdrawing group in 19A′ appears to affect the yield significantly. After obtaining the desired pseudosugars 10 and 10′, we protected the tertiary alcohols with an acetyl group affording 20 and 20′ to set up a [1,3]-sigmatropic shift catalyzed by palladium (II) under reflux conditions to afford 21 and 21′ after Zemplén deacetylation (Lim et al., 2009; Shamsi Kazem Abadi et al., 2017). Compound 21 was subjected to Simmons–Smith conditions (Davies et al., 2007) to convert the C5–C6 alkene to the cyclopropyl adduct 22. Compound 21′ was treated with 1,3,5-trifluorobenzene under nucleophilic aromatic substitution conditions to afford compound 23 (Scheme 3) (Chakladar et al., 2014).
SCHEME 3
Grubbs ring-closing metathesis reaction to obtain alkenes 20 and 20′ and synthesis of key intermediates 22 and 23.
Grubbs ring-closing metathesis reaction to obtain alkenes 20 and 20′ and synthesis of key intermediates 22 and 23.Target 11 was obtained from intermediate 22 in two steps involving nucleophilic substitution using 1,3,5-trifluorobenzene and, ultimately, global deprotection using hydrogenolysis (Scheme 4). To obtain target 12, we converted the p-bromobenzyl groups of compound 23 into labile leaving groups via Buchwald–Hartwig amination using N-methylaniline and, finally, deprotection in a weakly acidic medium as per the reported literature (Scheme 4) (Plante et al., 2000).
SCHEME 4
Synthesis of targets 11 and 12.
Synthesis of targets 11 and 12.For target 13, intermediate 10 was converted to an intermediate allylic bromide 26 (α:β, 1:1) formed using phosphorous tribromide in 55% yield (Cumpstey et al., 2011). Compound 27 was obtained by subjecting the mixture of allylic bromide 26 to an excess of cyclohexylamine and triethylamine. The final compound 13 was quantitatively obtained by Birch reduction (Scheme 5) (Kapferer et al., 1999; Si et al., 2021).
SCHEME 5
Synthesis of target 13.
Synthesis of target 13.Computational studies: Since compounds 11–13 did not inhibit the enzyme at the concentrations tested, the Sco GlgEI-V279S/13 complex crystal structure was used as a basis for additional computational docking studies. The protein and the ligands 11–13 were prepared, states generated at pH 7.0 ± 2.0, and the ligands docked into the active sites of the receptor via Glide™ module of the Schrödinger suite (Supplementary Figures S1, S2) (Schrödinger, L.L.C. (2017); Kumar et al., 2020; Oyeneyin et al., 2021). The docked Sco GlgEI-V279S/13 complex is shown in Figure 4.
FIGURE 4
Compound 13 docked in the Sco GlgEI-V279S M1P binding site.
Compound 13 docked in the Sco GlgEI-V279S M1P binding site.Among the three compounds, the docking scores were similar, where the hexylamine derivative recorded the highest docking score (−12.363 kcal/mol) and XP GScore (−12.377 kcal/mol) (Supplementary Table S1).As noted, compound 13 lacks any apparent inhibition of Sco GlgEI-V279S at concentrations as high as 1 mM. To assess any difference or commonality with our previously tested inhibitors and identify any additional enzyme–compound interactions that may be important for future drug development, we analyzed the X-ray crystal structure of the Sco GlgEI-V279S/13 complex. In particular, it is important to assess the conformation of the carbocycle in the −1 site as the computational docking model of compound 13 differs from the carbocycle conformation in the previously published Sco GlgEI-V279S complex with a 4-α-glucoside of valienamine (AGV). The Sco GlgEI-V279S/13 complex structure was resolved to a resolution of 2.73 Å, and the 2Fo-Fc composite omit map illustrated the presence of compound 13 in the M1P binding site of Sco GlgEI-V279S (Figure 5). The difference density maps show strong density for the glucose moiety in the -2 subsite and the carbocycle in the −1 subsite. However, the cyclohexyl moiety of compound 13 exhibits only weak difference density suggesting a lack of sufficient binding interactions with the residues forming the Sco GlgEI-V279S +1 site. This is consistent with computational modeling.
FIGURE 5
Compound 13 (pink carbon atoms) bound within the active site of Sco GlgEI-V279S (cyan carbon atoms). The 2Fo-Fc map is contoured at 1.5σ.
Compound 13 (pink carbon atoms) bound within the active site of Sco GlgEI-V279S (cyan carbon atoms). The 2Fo-Fc map is contoured at 1.5σ.Inspection of the enzyme–compound interactions shows many similarities to previously determine Sco GlgEI structures. Specifically, the glucose moiety of compound 13 bound in the −2 subsite resembles the interactions observed in previously published Sco GlgEI-V279S/inhibitor complex structures (Syson et al., 2011; Lindenberger et al., 2015) (Figure 6A).
FIGURE 6
Comparison of interactions in the −1 sites of (A)
Sco GlgEI-V279S/13 (pink carbon atoms) and (B)
Sco GlgEI-V279S/AGV (gray carbon atoms) structures.
Comparison of interactions in the −1 sites of (A)
Sco GlgEI-V279S/13 (pink carbon atoms) and (B)
Sco GlgEI-V279S/AGV (gray carbon atoms) structures.Inspection of the −1 subsite illustrates that the carbocycle C(3)-OH and C(7)-OH functional groups are in an axial conformational and form a bidentate hydrogen-bonded interaction with the side chain oxygen atoms of the Asp 394 nucleophile (Figure 6A). A similar bidentate interaction was observed with the Sco GlgEI-V279S in a complex with AGV (Figure 6B) (PDB:7MGY) (Si et al., 2021). As AGV was used as the basis for designing compound 13, this consistency gives confidence to the molecular design but contrasts with the computational modeling. The computational model shows these hydroxyls taking an equatorial orientation, which is conformationally lower in energy but apparently lacks appropriate interactions with the Sco GlgEI −1 subsite to support binding. Further comparison of these complexes demonstrates that the polar interactions within the −1 subsite of the Sco GlgEI-V279S/AGV complex are shorter and likely stronger than those of the Sco GlgEI-V279S/13 complex. Specifically, the side chain of Asp 394 in the Sco GlgEI-V279S/AGV complex is well oriented toward the hydroxyl groups of C2 and C7, thereby facilitating the formation of shorter hydrogen-bonded interactions (Figure 6B). But in the Sco GlgEI-V279S/13 complex, the Asp 394 side chain has slightly rotated away from the hydroxyl groups of C2 and C7 (Figure 6A).Since both Sco GlgEI-V279S/AGV and Sco GlgEi-V279S/13 crystals were formed at pH 7.5 and due to the proximity of the secondary amine of compound 13 with the side chain of Glu 423, it is hypothesized that the nitrogen atom of compound 13 is protonated and harbors a positive charge as anticipated for the primary ammonium moiety in AGV. Therefore, the 3.2 Å distant interaction between Glu 423 and the amine of 13 should be considered an ionic interaction. In both cases this is meant to emulate the positively charged oxocarbenium intermediate formed during the GlgE reaction mechanism. Furthermore, Figure 6A illustrates that the Glu 423 (general acid/base) in the Sco GlgEI-V279S/13 complex lacks interactions with compound 13. That is, due to the secondary amine of compound 13 shifting away from Glu 423, thereby forming a longer ionic interaction between the side chain of Glu 423 and the amine linker. In addition, this movement facilitates the formation of a hydrogen-bonded interaction between the amine and the side chain of the Asn 395 (Figure 6A).Furthermore, the oxocarbenium ion formed during catalysis is expected to take a flat chair conformation in each transition state. Figure 6A illustrates that the carbocycle of compound 13 successfully resembles the flat chair conformation. The superimposed active sites of Sco GlgEI-V279S/13 and Sco GlgEI-V279S/AGV complexes further illustrate that the carbocycles of compounds 13 and AVG have the same conformation and the introduction of cyclohexane moiety does not affect the conformation of 13. Superposition of the Sco GlgEI-V279S/13 and Sco GlgEI-V279S/AGV complexes further highlights the shift of the amine linker in compound 13 away from the Glu 423 side chain, compared to the ammonium ion of the AGV in the Sco GlgEI-V279S/AVG complex (Figure 7).
FIGURE 7
Superimposed active sites of Sco GlgEI-V279S/13 (pink carbon atoms) and Sco GlgEI-V279S/AGV (gray carbon atoms) structures.
Superimposed active sites of Sco GlgEI-V279S/13 (pink carbon atoms) and Sco GlgEI-V279S/AGV (gray carbon atoms) structures.The superimposed structures of Sco GlgEI-V279S/13 and Sco GlgEI-V279S/AGV complexes further highlight a 22° χ1 dihedral angle (between Cα and Cβ atoms) difference between Asp 394 in the Sco GlgEI-V279S/13 and Sco GlgEI-V279S/AGV complexes (Figure 7). The rotated Asp 394 in Sco GlgEI-V279S/13 complex is further stabilized by forming a hydrogen-bonded interaction with the side chain NH of Gln 324 (Figure 6A).Finally, the modeling of the cyclohexyl moiety of 13 in the X-ray crystal structure indicates the possible formation of only three hydrophobic interactions with the side chains of residues Asp 480, Ile 481, and Phe 425. These interactions, in contrast to the docking results, suggest a little role in supporting compound 13 binding. Alternatively, the weak difference density in Figure 5 may suggest undesired rotation of the sigma bonds flanking the amine, which could afford relatively free rotation of this moiety within the enzyme active site, which would weaken binding affinity.
Conclusion
Based on literature (Si et al., 2021), a 750 μM concentration of AGV reduced Sco GlgEI-V279S activity by 55%. To attempt to improve on the modest inhibitory activity, we thought to couple a cyclohexane moiety to the amine moiety to examine the effect of the 3rd ring on interacting with the amino acid residues in the +1 binding site. To achieve that goal, we discovered that allylic bromide 26 could be coupled to amine nucleophiles to extend the pseudosaccharide core. This may represent a useful strategy for entry into other naturally occurring C7N-aminocyclitols such as the amylostatins. As expected, the Sco GlgEI-V279S/13 complex structure illustrates that the carbocycle of compound 13 resembles a flat chair confirmation and maintains the axial positions for the hydroxyls at positions 2 and 3. Unfortunately, the compound 13 carbocycle interactions with the Glu 423 (general acid/base) were weakened by addition of the cyclohexyl moiety due to steric hindrance. This also promotes bond rotation in the Asp 394 side chain, which weakens an otherwise strong bidentate hydrogen-bonded interaction between Asp 394 and hydroxyls 2 and 7 of compound 13. Unfortunately, these minor disruptions in binding lack compensating interactions between the +1 subsite and the cyclohexyl moiety of 13, which appears to form only transitory interactions with side chains of the +1 subside. However, it is still possible to use compound 13 as the lead molecule in future GlgE inhibitor design studies. Taken together, the structure of the Sco GlgEI-V279S/13 complex strongly supports changing the stereo centers at carbocycle positions 2 and 3 to allow for equatorial hydroxyl groups at these positions. However, this would eliminate one of the hydrogen bonds to Asp 394 and lower conformational energy of the molecule and the addition of three other hydrogen bonds with Arg 392 and Asp 480 as seen in GlgE structures with bound maltose or maltose analogs (Veleti et al., 2014a; Lindenberger et al., 2015; Thanna et al., 2015; Veleti et al., 2017; Veleti et al., 2017; Si et al., 2021). In addition, replacing the cyclohexyl moiety with isosteric groups possessing hydrogen bond donors may support binding specificity as well as stabilizing hydrophobic interactions with the side chains of nearby residues.
Experimental methods
General methods
All chemicals and solvents were purchased from Fisher Scientific, Acros Organics, Alfa Aesar, or Sigma-Aldrich. Solvents were dried by using a solvent purification system by passing through activated alumina and copper catalyst columns. All reactions were carried out at room temperature under a nitrogen atmosphere using a nitrogen balloon, unless mentioned otherwise. Reactions were monitored by TLC (silica gel, f254) under UV light or by charring (5% H2SO4-MeOH), and the purification was performed by column chromatography on a silica gel (230–400 mesh) using the solvent system specified. Solvents were used without purification for chromatography. 1H NMR was recorded on the Bruker Avance III 600 MHz spectrometer using CDCl3 and D2O as solvents with residual CDCl3 and D2O as references. 13C were recorded on the Bruker Avance III 600 MHz spectrometer using CHCl3 and HDO as internal references. High resolution mass spectrometry was carried out using the TOF MS-ES + instrument. Low resolution mass spectrometry was carried out on ESquire-LC-MS.
Ammonia was condensed into a solution of 27 (94.0 mg, 0.09 mmol) in tetrahydrofuran (3.0 ml) using a dry ice cooled cold finger apparatus. The solution was treated with sodium in small pieces, until a blue color in the solution persisted. After stirring for 10 min at −78°C , the mixture was treated with NH4Cl (70 mg), stirred at room temperature overnight and evaporated. The residue was extracted with methanol, filtered, and evaporated. The residue (25.0 mg) was redissolved in water and purified using reverse phase C18 column with a gradient of water:methanol to give 13 as a white solid: yield quantitative (32.0 mg). 1H NMR (600 MHz, MeOD) δ 5.85 (s, 1H, H-1), 5.32 (d, J = 3.0 Hz, 1H, H-1′), 4.21 (d, J = 4.0 Hz, 1H, H-3), 4.14 (s, 2H, H-4′, H-4), 4.10–4.05 (m, 1H, H-2), 3.99–3.95 (m, 1H, H-6a), 3.93 (s, 1H), 3.79 (d, J = 10.4 Hz, 1H, H-6b), 3.69 (dd, J = 16.2, 6.8 Hz, 1H, H-2′), 3.59– 3.54 (m, 1H), 3.54–3.50 (m, 1H, H-5′), 3.34 (d, J = 8.7 Hz, 1H, H-6a′), 3.11 (d, J = 7.3 Hz, 1H, H-6b;), 3.05 (s, 1H), 1.97 (d, J = 28.6 Hz, 3H, C-Hex), 1.77 (d, J = 34.4 Hz, 2H, C-Hex), 1.58 (d, J = 12.4 Hz, 1H, C-Hex), and 1.29–1.17 (m, 4H, C-Hex). 13C NMR (151 MHz, MeOD) δ 138.24 (C5), 119.43 (C6) 97.43 (C1′), 74.43 (C2), 72.81 (C3′), 72.76 (C2′), 71.12 (C3,C4′), 69.68 (C5′), 69.30 (C6′), 66.54 (C7), 61.85 (C1), 60.45, 55.24 (CN-Hex), 51.53 (C-Hex), 48.83 (C-Hex), 29.84 (C-Hex), 24.72 (C-Hex), and 24.18 (C-Hex). Mass spectrum (HRMS), m/z = 420.4788 (M + H)+; C19H33NO9 requires 420.4791.Computational Methodology: The Sco GlgEI crystal structure was prepared using the Protein Preparation Wizard incorporated in Maestro (v.11.8) and Prime modules (Schrödinger, L.L.C. (2017)) to assign bond orders and charges and to remove water molecules and all heteroatoms. Restrained minimization was carried out to converge heavy atoms to RMSD of 0.3 Å using the OPLS3e force field. The ligands were drawn using ChemDraw (v21), imported into the workspace and prepared using the Ligprep module incorporated in Maestro (v11.8) via the OPLS3e force field (Kumar et al., 2020). Epik was used to generate possible states at target pH 7.0 ± 2.0 for accurate tautomer and ionization. The prepared ligands were docked into the active sites of the receptor (x = 25.27, y = −32.26, z = 277.11 via Glide docking in Maestro (v11.8) (Oyeneyin et al., 2021).Protein purification: The protein purification was carried out according to the previously published method (Veleti et al., 2014b). In brief, the Sco GlgEI-V279S construct was used to transform T7 express E. coli cells. At 37°C, the largescale cultures were grown in LB media with a 264 mM concentration of carbenicillin. When O.D600 nm reached 0.6, the temperature was lowered to 16°C, and 1 mM Isopropyl β-D-1-thiogalactopyranoside was added to cultures. After 16 h of induction, cells were harvested and resuspended in the lysis buffer (20 mM Tris pH 7.5, 500 mM NaCl, 10% glycerol, 0.5 mM imidazole, and 0.3 mM tris(2-carboxyethyl) phosphine (TECP)), and the resulting cell suspension was incubated on ice with lysozyme and DNase I. After 30 mins, the cell suspension was sonicated and centrifuged for 45 min. The clarified lysate was applied to a 5 ml metal affinity cobalt column, which was pre-equilibrated with lysis buffer. To remove unbound protein, 25 column volumes of lysis buffer were passed through the column and unbound protein was washed away. Finally, Sco GlgEI-V279S was eluted isocratically with elution buffer (20 mM Tris pH 7.5, 500 mM NaCl, 150 mM imidazole, and 0.3 mM TCEP) and dialyzed against 20 mM Tris pH 7.5, 150 mM NaCl, and 0.3 mM TCEP to get rid of excess salt and imidazole.Inhibition studies: The inhibition studies on compound 13 were performed separately using a real-time assay and the EnzChek Phosphate Assay Kit. According to the literature, when the phosphate concentration of the reaction mixture is above 25 mM, GlgE catalyzes the reverse reaction of the production of M1P from glucans (Syson et al., 2011). In the real-time assay, the Sco GlgEI-V279S reverse reaction was coupled with α-glucosidase and performed according to the previously published method. In the presence of phosphate, Sco GlgEI-V279S catalyzes the hydrolysis of 2-chloro-4-nitrophenyl-D-maltotrioside (substrate) into M1P and 2-chloro-4-nitrophenyl glucose12,13. The 2-chloro-4-nitrophenyl glucose is further hydrolyzed into 2-chloro-4-nitrophenyl and glucose by α-glucosidase. The resulting 2-chloro-4-nitrophenyl exhibits an increased absorbance at 410 nm wavelength, and production of 2-chloro-4-nitrophenyl was measured continuously during the reaction. The reactions were carried out in a 96-well format, with a 50 µl reaction volume at 25°C for 40 min, and the velocity was measured at 410 nm wavelength using a Synergy H4 plate reader (Bio Tek). The reaction mixture consisted of 1.5 mM 2-chloro-4-nitrophenyl-D-maltotrioside, 20 mU α-glucosidase, 100 mM sodium phosphate, and 20 mM Tris pH 7.5. Finally, the reaction was initiated by adding 250 nM Sco GlgI-V279S to the reaction mixture. In the reaction, the compound 13 concentrations varied from 0 to 3000 µM. The positive and negative controls were lacking compound 13 and Sco GlgEI-V279S, respectively. The percent enzymatic activity was calculated by using the equation (V/V0) × 100, where V and V0 denote the rates of the inhibited and uninhibited enzyme (positive control), respectively.The percent inhibition of α-glucosidase was calculated at each compound 13 concentration. In the reaction, 2-chloro-4-nitrophenyl-α-D-glucopyranoside was used as the substrate, and the production of 2-chloro-4-nitrophenyl was monitored throughout the reaction. The reaction mixture comprises 1.5 mM 2-chloro-4-nitrophenyl-α-D-glucopyranoside, 20 mM Tris pH 7.5, and each concentration of compound 13. The reaction was initiated by adding 20 mU α-glucosidase. The compound 13 concentrations varied from 0 to 3000 μM. The percent enzymatic activity was calculated using the equation mentioned above. Finally, the total inhibition observed for Sco GlgEI-V279S and α-glucosidase was deducted using α-glucosidase inhibition at each concentration, and the actual GlgE inhibition was calculated. The results indicated that a 3000 μM concentration of compound 13 inhibits the activity of GlgE by 50%.The EnzChek phosphate detection assay was performed according to the previously published method. The reaction mixture consisted of 20 mM Tris pH 7.5, 150 mM NaCl, 500 nM MESG, 0.25 U PNP, 0.5 mM glycogen (as maltose acceptor), and 250 µM M1P (Veleti et al., 2014a; Veleti et al., 2014b). The reaction was initiated by adding the 50 nM Sco GlgEI-V279S to a final concentration of 50 nM. All the reactions were carried out in 96 well formats in a triplicate manner, and a Synergy H4 plate reader (Bio Tek) was used to measure the absorbance at 360 nm. The 100 and 1000 µM concentrations of compound 13 were tested against Sco GlgEI-V279S, and neither of those concentrations exhibited inhibition of the Sco GlgE1-V279S activity.Crystallization experiments: The Sco GlgEI-V279S/13 mixture consisted of 8 mg/ml Sco GlgEI-V279S and 10 mM concentration of compound 13. The crystallization experiments were performed using the hanging drop vapor diffusion method. Each drop consisted of 2 μl of Sco GlgEI V279S/compound mixture and 2 μl of the well solution of 0.2 mM sodium citrate pH 7 and 10% w/v PEG 3,350 (Lindenberger et al., 2015). The crystallization drops were equilibrated against 100 µl of well solution. To prepare for crystal cryoprotection, 4 μl of 50% w/v PEG 2000 was added to the drops containing the Sco GlgEI-V279S/ligand cocrystals and cryoprotected crystals were harvested and flash-cooled by immersing in liquid nitrogen.X-ray diffraction experiments: The LS-CAT beamline at Advanced Photon Source of Argonne National Labs, IL, was used to perform X-ray diffraction experiments. The collected data were indexed, integrated, and scaled by using HKL 2000 (Otwinowski and Minor, 1997). The molecular replacement used the previously published Sco GlgEI-V279S in complex with a maltose-C-phosphonate structure (PDB ID:4U31), using Phaser (Mccoy et al., 2007) in PHENIX (Adams et al., 2010). The refinements were carried out using PHENIX. The visualization and manual refinements were performed using COOT (Emsley and Cowtan, 2004). Finally, MolProbity (Williams et al., 2018) in PHENIX was used for structure validation.
Authors: Adrian Scaffidi; Keith A Stubbs; Rebecca J Dennis; Edward J Taylor; Gideon J Davies; David J Vocadlo; Robert V Stick Journal: Org Biomol Chem Date: 2007-08-15 Impact factor: 3.876
Authors: Christopher Adamson; Robert J Pengelly; Saeideh Shamsi Kazem Abadi; Saswati Chakladar; Jason Draper; Robert Britton; Tracey M Gloster; Andrew J Bennet Journal: Angew Chem Int Ed Engl Date: 2016-10-26 Impact factor: 15.336
Authors: Jared J Lindenberger; Sri Kumar Veleti; Brittney N Wilson; Steven J Sucheck; Donald R Ronning Journal: Sci Rep Date: 2015-08-06 Impact factor: 4.379
Authors: Airlie J McCoy; Ralf W Grosse-Kunstleve; Paul D Adams; Martyn D Winn; Laurent C Storoni; Randy J Read Journal: J Appl Crystallogr Date: 2007-07-13 Impact factor: 3.304
FIGURE 1
FIGURE 2
FIGURE 3
SCHEME 1
SCHEME 2
SCHEME 3
SCHEME 4
SCHEME 5
FIGURE 4
FIGURE 5
FIGURE 6
FIGURE 7