Glycoside Hydrolase Family 4

Glycoside Hydrolase Family 4
Clan	GH-x
Mechanism	retaining
Active site residues	not known
CAZy DB link

Substrate specificities

The majority of the enzymes are of bacterial origin, but they have recently been found in the archaeal taxa [1]. Unlike other families of glycosidases, some of the enzymes in Family 4 possess distinct substrate specificities from each other[1], [2]. This family contains α-glucosidases, α-galactosidases, α-glucuronidases, 6-phospho-α-glucosidases, and 6-phospho-β-glucosidases. Similar to GH1, some enzymes prefer phosphorylated substrates over non-phosphorylated substrates [2], [3]. Unique to GH4 is the presence of both α- and β-glycosidases. The ability to hydrolyze natural substrates of different anomeric configurations has not been discovered in other glycoside hydrolase families to-date. GH4 enzymes were the first glycosidases shown to demonstrate an absolute requirement for NAD⁺ and a divalent metal ion and in some instances reducing environments for catalytic activity [4], [5], [6], [7], [8], [9], [10], [11]. The cofactors are proposed to play key mechanistic roles in the atypical glycosidase mechanism. Unlike GH4, metal ions are required by some glycosidases for structural integrity [12], [13] or proper configuration of an active enzyme [14], [15], [16], [17]. Meanwhile, GH4 is the first family of glycoside hydrolases reported to require a NAD⁺ cofactor for catalytic activity [18], [5], [6], [7], [8], [10], [11], [19], [20], [21], [22], [23]. Subsequently, GH109 has more recently been shown to demonstrate the same use of a NAD⁺ cofactor, but this family does not require metal ions or reducing conditions for activity. The hydrolysis of thioglycosides with activated leaving groups has been documented in GH84 [24], GH1 [25], [26], [27], [28], [29], [30]. However, GH4 is currently the only family that has been shown to catalyze the hydrolysis of unactivated thioglycoside substrates [31]. The remarkable substrate specificities are all feasible due to the β-elimination mechanism discussed below.

Kinetics and Mechanism

GH4 enzymes have been shown to be retaining glycosidases [32], [33], [34], [35]. However, the enzymes do not utilize the double-displacement mechanism proposed by Koshland [36, 37, 37, 38, 39, 40, 41, 42, 43, 44, 45] and GlvA (6-phospho-α-glucosidase) from Bacillus subtilis [33] are two of the most thoroughly studied enzymes of this family, and both are proposed to utilize an oxidation-elimination-addition-reduction mechanism via anionic transition states. Exchange of the substrate C2 proton with solvent deuterium during the enzyme-catalyzed reaction was the first indication of a C2-H2 bond cleavage in the substrate, which is inconsistent with the double-displacement mechanism [32], [34]. This fact coupled with the proximity of the NAD⁺ cofactor to C3 of the glucose 6-phosphate product and the high structural similarity to lactate/malate dehydrogenases as garnered from the x-ray crystallographic data [9], [32], [34], [35] led to the following proposed mechanism [32], [34]: 1) C3 hydride abstraction via the reduction of NAD⁺ cofactor to NADH and consequent oxidation of the C3 hydroxyl group; 2) adjacent to the ketone functionality, the C2 proton is activated to deprotonation by a near-by catalytic base residue; 3) cleavage of the C1-O1 linkage occurs via an α? β-elimination mechanism, producing an α? β-unsaturated intermediate; 5) a 1,4-Michael-like addition of a water molecule at C1 takes place; and 6) reduction of the C3 carbonyl functionality by the "on-board" NADH generates the final hydrolysis product.

Primary kinetic isotope effects and the determination of Brønsted relationships revealed that both the C3 hydride and C2 proton abstractions are partially rate-limiting and that cleavage of the C1-O1 linkage is not rate-limiting [33], [34], [45]. This led to the proposed E1 _cb mechanism for both BglT and GlvA [33], [45].

Catalytic Residues

The proposed catalytic residues were mostly derived from x-ray crystallographic data and pH-dependent activity profiles obtained for BglT and GlvA [32], [33], [34], [45]. A Tyr residue that is conserved in all 6-phospho-α- and 6-phospho-β-glycosidases was found to be approximately 4 Å away from C2 of the reaction product, which can potentially act as a catalytic base [32], [33], [34], [35], [45]. The pH-dependent activity profiles were that of double ionization curves with pH_opt at approximately 8 [33], [45]. Two pK_a values of approximately 7 and 9 were determined [33], [45]. The pK_a value of approximately 7 was proposed to correspond to that of the Tyr catalytic base, since this residue would need to be deprotonated for enzyme activity. The position of the Tyr in the enzyme active site was used as a possible explanation for the relatively low pK_a compared to that of a free Tyr (pK_a = 10). In the case of BglT, the Tyr residue is located at a distance of 3.4 Å from the glycosidic oxygen, making it ideal to assume a second role in providing general acid catalysis to the departing oxygen as well as C2 deprotonation [32], [45]. The pK_a value of approximately 9 was proposed to correspond to the ionization of the conserved Arg residue(s) that were found to be within hydrogen bonding distance of the phosphate group of the substrate in these two 6-phospho-glycosidases [32], [35], [45]. Again, the normal pK_a of 12 for Arg could be lowered to 9, and this matches the expectation that protonated Arg residue(s) would be needed to form electrostatic interactions with the substrate phosphate moiety. Two other conserved residues, Cys and His, were shown to be responsible for chelating the metal ion in BglT and GlvA. In the AglA structure, an Asp residue occupies the same position as the catalytic Tyr base, so the Asp is proposed to take on this role in AglA and possibly other GH4 enzymes that prefer non-phosphorylated substrates.

Three-dimensional structures

Crystallographic data is available for a number of GH4 enzymes, including α- and β-members as well as those enzymes specific for phosphorylated and non-phosphorylated substrates. The x-ray crystal structures currently available for this family are those of the Thermotoga maritima α-glucosidase AglA (PDB 1OBB) [9], the T. maritima α-glucosidase Agu4B (PDB 1VJT) [46], the Bacillus subtilis 6-phospo-α-glucosidase GlvA (PDB 1U8X) [35], and two 6-phospho-β-glucosidases: one from Geobacillus stearothermophilus (PDB 1S6Y) [46] and other BglT from T. maritima (PDB 1UP6) [32], [34]. Each enzyme possesses the characteristic dinucleotide-binding Rossman fold [47]. The crystal structures of Agu4B (PDB 1VJT) and the 6-phospho-β-glucosidase (PDB 1SY6) from do not provide useful information about the active sites. No ligands are bound in the structure of the 6-phospho-β-glucosidase from G. stearothermophilus. In the Agu4B structure, only the NAD ⁺ cofactor is found at the enzyme active site, but no data for the nicotinamide ring is provided, possibly because it is not well-defined in the crystal structure. The AglA structure is that of an inactive enzyme with no divalent metal ion. The GH4 enzymes show some structural similarities to lactate/malate dehydrogenases as well as dehydratases [9], [32], [34], [35], [46].

Family Firsts

’’’First stereochemistry determination’’’: Thermotoga maritima BglT (6-phospho-β-glucosidase) via NMR and HPLC analysis of methyl-glycoside product [34]
’’’First catalytic residue identification’’’: Thermotoga maritima BglT (6-phospho-β-glucosidase) via x-ray crystal structure [34]
’’’First 3-D structural determination’’’: Thermotoga maritima AlgA (α-glucosidase) (inactive enzyme) [9]; Thermotoga maritima BglT (6-phospho-β-glucosidase) (active enzyme with all cofactors present) [32], [34]
’’’First GH4 enzyme shown to hydrolyze both α- and β-substrates’’’: Fusobacterium mortiferum MalH (6-phosph-α-glucosidase) [20]
’’’First β-glycosidase’’’: Escherichia coli CelF (6-phospho-β-glucosidase) [23]

References

All Medline abstracts: PubMed