Metalloglycomics: a new perspective upon competitive metal–carbohydrate binding using EPR spectroscopy

Ternary complexes formed between calcium, the oxochromium(v) ion and N-acetylneuraminic (sialic) acid (naH6) of the form, Ca(ii)–oxoCr(v)–naH6, have electronic structures and equilibrium distributions distinct from the binary oxoCr(v)–naH6 analogues, as illustrated by electron paramagnetic resonance (EPR) spectroscopy.

Sialic acids are a family of carbohydrates ubiquitous throughout the animal kingdom that are structurally defined as 9-membered carbon chains with an α-keto acid head group.1

The most abundant sialic acid in humans is N-acetylneuraminic acid (naH6; Scheme 1) which occurs predominantly as the terminal residue of protein- and lipid-based cell-surface glycoconjugates.1

Structural features of naH6 include a tert-2-hydroxycarboxylate ‘head’ group, which is deprotonated at physiological pH (pKa ∼ 2.6), and a glycerol ‘tail’.

The 2′-OH α-glycoside link from naH6 to the penultimate glycoconjugate residue orients the glycerol tail towards the extracellular milieu inviting interactions with biometals.

The interaction between Ca(ii) and naH6 has been investigated previously;2 the structure of a binary (1 : 1) complex determined from 1H and 13C NMR spectroscopy (Scheme 1) proposes Ca(ii) binds to naH6 (as the free acid, where β : α ∼ 92 : 8) via the carboxylate O1, OR, O7 and O8 groups.3

More recently, both the tert-2-hydroxycarboxylate ‘head’ group of naH6 (O1,O2) in addition to the glycerol ‘tail’ (O7,O8,O9) have been established as viable chelates of select transition metal ions using electron paramagnetic resonance (EPR) spectroscopy,4 NMR spectroscopy,5 or potentiometry;6 EPR spectroscopy in particular continues to emerge as a powerful tool in studying metal–bioligand speciation.4,7

Strong EPR signals from solutions of Cr(vi) and glutathione (GSH) in the presence of excess naH6 (Fig. 1; black) are attributed to a distribution of five-coordinate oxoCr(v)–naH6 complexes (d1) in which two naH6 units chelate the oxoCr(v) ion ([naH6] > [oxoCr(v)]) via combinations of the O1,O2, O7,O8 or O8,O9 groups.4

The multiplicity of the central xCr (x = 50, 52, 54; I = 0) signal (Fig. 1; RHS) conveys the number (n) of protons (I = ½; 2nI + 1) in the second coordination sphere that perturb the Cr(v) redox (dxy) orbital (e.g.; n = 0, bis-O1,O2 singlet; n = 4, bis-O7,O8 quintet; n = 6, bis-O8,O9 septet).4

There is a pH dependence of the distribution of oxoCr(v)–naH6 species whereby isomers of the O1,O2-coordinated species (giso = 1.9785; 1.9792) dominate under acidic conditions (Fig. 1a) and the series of O7,O8- and O8,O9-coordinated species (giso ∼ 1.9800) dominate at alkaline pH values (Fig. 1c, d).4

This work shows that in the presence of excess Ca(ii) ([Ca(ii)] : [naH6] = 10) there is significant modulation to the Cr(v) EPR signals (Fig. 1; purple) which suggests the formation of ternary Ca(ii)–oxoCr(v)–naH6 complexes that have electronic structures (as indicated by variations in giso values and 1H aiso values) distinct from the binary oxoCr(v)–naH6 analogues. Also, the pH-dependent equilibrium distribution of oxoCr(v)–naH6 species clearly differs in the presence of Ca(ii); the appearance of complex multiplets at lower pH values in the ternary system, relative to the binary system (Fig. 1b), is likely due to the Ca(ii)-induced lowering of the hydroxyl pKa values of naH6, thereby increasing the relative concentration of ternary Ca(ii)–oxoCr(v)–naH6(diolato) species.

These observations were investigated using EPR simulation procedures. The parent Ca(ii)–oxoCr(v)–naH6 signals (at pH values ≥ 4.49) simulated (correlation > 0.998) as comprising two singlets, two quartets, two septets (Table 1) and up to five additional high-field singlets (present in very low (<3.5%) concentrations at pH values ≥ 7.14).

Noteworthy differences between the binary oxoCr(v)–naH6 and ternary Ca(ii)–oxoCr(v)–naH6 systems are as follows.

First, the presence of excess Ca(ii) appears to mitigate against the formation of oxoCr(v)–naH6 species involving coordination via the O7,O8-diolato group, since (unlike the binary system) the simplest simulation model for the Ca(ii)–oxoCr(v)–naH6 system did not require the inclusion of a triplet, a quintet or a sextet to model the ternary analogues of [CrO(O1,O2-naH4)(O7,O8-naH3)]2− (n = 2), [CrO(O7,O8-naH3)2]3− (n = 4) or [CrO(O7,O8-naH3)(O8,O9-naH3)]3− (n = 5), respectively.

Second, an excellent fit between the experimental and simulated spectra for the Ca(ii)–oxoCr(v)–naH6 system was obtained with the inclusion of two isomers each of the singlet, quartet and septet (Table 1).

Individual EPR spectra for each species present in the Ca(ii)–oxoCr(v)–naH6 equilibrium solution at pH = 4.49 are shown in Fig. 2 (RHS) aligned with the spectra from the related binary oxoCr(v)–naH6 species (LHS) at pH = 4.37.

This spectral deconvolution highlights the marked differences in the relative concentrations of related species between the binary oxoCr(v)–naH6 and ternary Ca(ii)–oxoCr(v)–naH6 systems (where the absence of a species is represented by a dotted line).

The sum of the component spectra (purple) in both the binary oxoCr(v)–naH6 (pH = 4.37) and ternary Ca(ii)–oxoCr(v)–naH6 (pH = 4.49) systems closely maps onto the experimental (black) spectrum (Fig. 2; lower trace).

In the binary oxoCr(v)–naH6 system, while the presence of two singlets (Fig. 1a) is evident (ascribed to geometrical isomers of [CrO(O1,O2-naH4)] in which donor atoms are juxtaposed about the oxoCr(v) trigonal bipyramid),4,9 only one isomer is observed (as determined by the limits of EPR simulation) for the remaining linkage isomers.

Third, the spectral fit in the Ca(ii)–oxoCr(v)–naH6 system required the inclusion of non-uniform 1H aiso values (∼0.59 G, 0.80 G, 0.96 G) for second coordination sphere protons within a single species (Table 1).

This is distinct from the binary oxoCr(v)–naH6 system, where the conformational flexibility of the glycerol tail of naH6 confers magnetic equivalence upon these protons (1H aiso ∼ 0.75 G).

These differences suggest that the interaction between Ca(ii) and naH6, as modelled from NMR spectroscopic data (O1, OR, O7, O8),3 restricts the conformational flexibility of the glycerol tail, effectively acting as an ‘ionic lock’.

The ternary Ca(ii)–oxoCr(v)–naH6 species (two geometrical isomers each) are assigned (Table 1) as {Ca2[CrO(O1,O2-naH4)2]}3+ (CaI (a,b)), {Ca2[CrO(O1,O2-naH4)(O8,O9-naH3)]}2+ (CaIII (a,b); Scheme 2) and {Ca2[CrO(O8,O9-naH3)2]}+ (CaVI (a,b); Scheme 2).

The increase in the strength of the sets of donor atoms (bis-hydroxycarboxylate < hydroxycarboxylate-diolato < bis-diolato) correlates with the increase in the giso values for CaI (a,b), CaIII (a,b) and CaVI (a,b), respectively, which supports the species assignment, together with the signal multiplicities (Table 1).

The stoichiometry of the ternary Ca(ii)–oxoCr(v)–naH6 complexes (2 : 1 : 2) is based upon the known Ca(ii) : naH6 stoichiometry (1 : 1) determined from NMR experiments (even where [Ca(ii)] > [naH6]),2,3 and from studies of oxoCr(v)–bioligand systems that show the preferential formation of bis-chelate species (relative to mono-chelate species) where [ligand] : [Cr(v)] ≥ 2..54,9

The absence of O7,O8-coordinated oxoCr(v)–naH6 species in the ternary Ca(ii)–oxoCr(v)–naH6 system suggests that the O7 group of naH6 has a stronger affinity toward Ca(ii) than toward the oxoCr(v) ion.

Conversely, based upon the presence of O8,O9-coordinated Ca(ii)–oxoCr(v)–naH6 species, the O8 group of naH6 has an affinity toward the oxoCr(v) ion (and may be simultaneously binding Ca(ii)).

The magnetic inequivalence of the protons in the second coordination sphere of ternary Ca(ii)–oxoCr(v)–naH6 complexes (Fig. 2; Table 1) is also explained by the ‘ionic lock’ mechanism, since the {Ca(ii)–naH5}+ structure places H8, H9 and H9′ in unique magnetic environments.3

These environments evidently increase the extent of orbital overlap between the second coordination sphere protons and the dxy oxoCr(v) orbital, as measured by the increased 1H aiso values of the Ca(ii)–oxoCr(v)–naH6 complexes, relative to the binary oxoCr(v)–naH6 analogues.

Also, there are shifts in giso values for related species within the binary oxoCr(v)–naH6 and ternary Ca(ii)–oxoCr(v)–naH6 systems (Fig. 1).

The giso values of the species involving tert-2-hydroxycarboxylate coordination ([CrO(O1,O2-naH4)2] or {Ca2[CrO(O1,O2-naH4)2]}3+ in the binary or ternary system, respectively) move to higher field (lower giso) in the presence of Ca(ii), while the giso values of diolato coordinated species ([CrO(O8,O9-naH3)2]3− or {Ca2[CrO(O8,O9-naH3)2]}+ in the binary or ternary system, respectively) move to lower field (higher giso) in the presence of Ca(ii).

This latter trend is consistent with [CrO(glycerol(2−))2] (giso = 1.9800), where in the presence of Ca(ii), giso = 1.9803.

This suggests that in the case of oxoCr(v)–diolato complexes, the presence of Ca(ii) modifies the polarizability of the donor oxygen atom (increases the ‘hardness’), which correlates with an increase in the giso value.

No change in giso value is observed for well characterised tert-2-hydroxycarboxylate oxoCr(v) complexes ([CrO(hmba(2−))2] (hmba(2−) = 2-hydroxy-2-methylbutanoato), giso = 1.9784) in the presence of Ca(ii), which indicates a poor affinity of Ca(ii) towards an isolated hydroxycarboxylate group and suggests that the polyfunctionality of naH6 is an important determinant of Ca(ii) binding.

This is an elegant illustration of the use of EPR spectroscopy in studying competitive metal–carbohydrate binding (termed here ‘metalloglycomics’) that may have wide ranging implications in understanding transition metal–bioligand speciation in Ca(ii)-rich (and/or Mg(ii)) matrices that model the biological milieu.

The author acknowledges a Sesqui New Staff Support grant and access to the EPR spectroscopy facility at the University of Sydney.