Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

Vanadium salts mimic the actions of insulin in vitro (Shechter and Karlish, 1980; Shechter, 1990; Shechter et al., 1995). In diabetic rats, vanadium therapy induces normoglycemia and improves glucose homeostasis in insulin-deficient and -resistant diabetic rodents (Meyerovitch et al., 1987; Brichard and Henquin, 1995). Accumulating data suggest that vanadium acts through alternative noninsulin pathways, involving inhibition of protein phosphotyrosine phosphatases and activation of nonreceptor protein tyrosine kinases (Green, 1986; Fantus et al., 1989; Mooney et al., 1989; Strout et al., 1989; Shisheva et al., 1991; Venkatesan et al., 1991; Shisheva and Shechter, 1992, 1993; Elberg et al., 1994, 1997). Accordingly, vanadium salts are effective in diabetic rodents that do not respond at all to insulin (reviewed in Brichard and Henquin, 1995).

Initial clinical studies have already been performed with low doses of vanadium (2 mg/kg/day) to avoid toxicity. Although 10- to 20-fold lower than those used in most animal studies, even these doses of vanadium yielded several beneficial effects (Cohen et al., 1995; Goldfine et al., 1995; Halberstam et al., 1996). Any manipulation that elevates the insulinomimetic efficacy of vanadium without increasing its toxicity is of significant potential interest for the future care of diabetes in humans.

Various organically chelated vanadium compounds are more potent than free vanadium salts in facilitating insulin-like effects in vitro and in vivo (McNeill et al., 1992; Shechter et al., 1992; Sakurai et al., 1995; Li et al., 1996). Recently, we found that the L-isomer of Glu()monohydroxamate (HXM) is particularly active; in streptozocin (STZ)-rats, L-Glu()HXM-vanadium complex (2:1 stoichiometry) potentiated vanadium-induced normalization of circulating glucose levels by 4- to 5-fold. In rat adipocytes, in vitro, L-Glu()HXM itself, in the absence of vanadium, activated hexose transport and lipogenesis to 20 to 35% of maximal stimulation (Goldwaser et al., 1999). We thus consider L-Glu()HXM an optimal-synergizing vanadium chelator.

In this study, we wished to determine whether the vanadium-enhancing effect of L-Glu()HXM can be attributed to a specific feature or features of this ligand and examine other less active vanadium ligands for these characteristics. To this end, we developed three experimental systems. First, we determined the approximate formation constants of various ligands toward vanadium(+4) and vanadium(+5) at neutral pH. Second, vanadium-enriched adipocytes were exposed to different vanadium ligands to assess their potential lipogenic activity. Finally, we examined whether a favorable geometry (such as that of vanadium-L-Glu()HXM) may have physicochemical attributes that are reflected by simple spectrophotometric measurements.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials. D-[U-¹⁴C]Glucose was from New England Nuclear (Boston, MA). Collagenase type I (134 U/mg) was from Worthington Biochemicals (Freehold, NJ). Porcine insulin was from Eli Lilly Co. (Indianapolis, IN).

Cell Preparation and Bioassay for Lipogenesis. Rat adipocytes were prepared from fat pads of male Wistar rats (130 to 150 g) by collagenase digestion according to the method of Rodbell (1964). Viability of cell preparations, evaluated by Trypan blue exclusion, was more than 95% at least 3 h after digestion. Assay of lipogenesis (the incorporation of U-¹⁴C-labeled glucose into lipids) was performed essentially according to Moody et al. (1974). Briefly, freshly prepared rat adipocytes were suspended in KRBH/0.7% BSA buffer and divided into 50 plastic vials, each containing 0.5 ml of adipocyte suspension (about 1.5 × 10⁵ cells). These were incubated with 0.16 mM [U-¹⁴C]glucose at 37°C for 2 h under an atmosphere of 95% O₂/5% CO₂. Each assay was compoed of vials with and without 17 nM insulin and the various test compounds. Lipogenesis was terminated by adding toluene-based scintillation fluid, after which radioactivity was counted in the extracted lipids (Moody et al., 1974). Results are expressed as percentage of maximal insulin response. Only assays in which insulin activated lipogenesis 5- to 6-fold above basal level (~4000 cpm per 1.5 × 10⁵ cells/2 h; V_ins = 20,000-24,000 cpm per 1.5 × 10⁵ cells/2 h) were taken into consideration. Insulin activated lipogenesis in this assay at an ED₅₀ of 33 ± 0.03 pM. A concentration of 0.3 nM insulin already facilitated maximal (100%) response (Shisheva and Shechter, 1992). All assays were performed in duplicate or triplicate.

Saccharomyces cerevisiae. S. cerevisiae (wild type) was obtained from the Weizmann Institute biological services and grown overnight in yeast extract/peptone/dextrose media at 25°C. Before the experiment, cells were suspended and diluted in the same medium, brought to pH 7.4 with NaOH, divided into several vials, and supplemented with none or the indicated concentrations of NaVO₃(+5) or VOCl₂(+4) in quadruplicate for each measurement. The pH was readjusted to 7.4 (when needed). Cells were then incubated for 5 h at 37°C with moderate shaking. Growth was evaluated by measuring turbidity at 540 nm.

Preparation of Vanadium-Enriched Adipocytes. Male-Wistar rats (7 weeks old, 140 ± 10 g) were s.c. injected daily with NaVO₃ (0.1 mmol/kg at 10:00 AM) over a period of 4 days. The rats were sacrificed by decapitation 4 to 6 h after the last injection. Epididymal fat pads were removed to determine vanadium levels by atomic emission spectroscopy and prepare rat adipocytes for use in the lipogenic assay. Vanadium-enriched adipocytes prepared in this way had higher basal activity, attributed to the higher level of intracellular vanadium [~8400 ± 300 cpm per 1.5 × 10⁵ cells/2 h, whereas V_ins remained nonmodified (V_ins = 20,000-24,000 cpm per 1.5 × 10⁵ cells/2 h)]. After enrichment in vivo, the level of vanadium in the adipose tissue was 0.39 ± 0.03 µg/g (~7.65 µM), as determined by inductively coupled plasma atomic emission spectroscopy (ICP-AES model spectroscope, Kleeve, Germany). Using this procedure (sensitivity, 2.55 µM), the vanadium level was undetectable in the adipose tissue of control rats [although others have reported a control level of 0.3 µM (Sitpirija and Eiam-Ong, 1998)]. Thus in vivo enrichment elevated vanadium level about 25-fold.

Evaluating Relative Affinities of Ligands for Vanadium: Reversal of Vanadium-Evoked Inhibition of Acid Phosphatase by Vanadium Chelators. This assay evaluates the efficacy of vanadium chelators by measuring their ability to reverse vanadium-evoked inhibition of acid phosphatase at pH 7.2. The assay is performed in 4-ml plastic tubes (0.5 ml per tube) in 0.05 M HEPES buffer, pH 7.2, containing 1.0 M KCl, 0.2 mM p-nitrophenylphosphate (pNPP), either NaVO₃ or VOCl₂ (final concentration, 20 µM), increasing concentrations of the chelator to be tested, and wheat-germ acid phosphatase (50 µg per tube). pNPP hydrolysis was terminated after 10 min at 25°C by adding NaOH (50 µl of 5 M NaOH per tube), and the absorbency at 410 nm, corresponding to p-nitro-phenolate, was determined. Because several chelators enhance pNPP-hydrolysis by acid phosphatase, rate of hydrolysis was determined individually for each concentration of chelator in the absence (Vo) and the presence (Vi) of 20 µM NaVO₃ or VOCl₂. The plot of inhibited-extent of hydrolysis versus the chelator concentration is linear with a slope proportional to the affinity constant of the ligand for vanadium. RC₅₀ is defined here as the concentration of the chelator that half-maximally reverses vanadium-evoked inhibition of acid phosphatase. All values are presented as means ± S.D.

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

Evaluation of the Intact Adipose Cell System as a Possible Measure for Classifying Insulinomimetic-Synergizing Vanadium Ligands. Initially, we examined whether the intact adipose cell was an appropriate experimental system for classifying vanadium ligands in respect to their potentiating effect. Figure 1A shows that 10 µM concentrations of either VOCl₂(+4), bis-(picolinato)oxovanadium_, bis-(maltolato)oxovanadium_, vanadyl acetylacetonate, or VO(L-Glu()HXM)₂ activated lipogenesis to the extent of 4, 5, 16, 30, and 47% of maximal insulin response, respectively, indicating that several preformed vanadium(+4) complexes are significantly more potent than free vanadium(+4) in activating lipogenesis.

View larger version (21K): [in this window] [in a new window]   Fig. 1.   Lipogenic capacity of low concentration of vanadium(+4), with several organic chelators: comparison between preformed vanadium(+4) complexes and vanadium(+4)-ligand mixtures. A, freshly prepared rat adipocytes (3 × 105 cells/ml) were suspended in KRBH-buffer, pH 7.4, containing 0.7% BSA and preincubated for 10 min with 10 µM VOCl2 and 10 µM concentrations of the preformed organically chelated vanadium(+4) compounds. B, the indicated organic chelators (20 µM) were added first to the adipocytic cell suspension followed by addition of VOCl2 (10 µM) 10 s later. Cells were then supplemented with [U-14C]glucose and lipogenesis was performed for 2 h at 37°C. Radioactivity incorporated into extracted lipids was then determined. The reference maximal response (100%) was that obtained in the presence of 17 nM insulin

Optimal Vanadium Ligands, When Added Alone to Vanadium-Enriched Adipocytes, Activate Lipogenesis. The basal lipogenic activity of adipocytes enriched with vanadium, obtained by treating rats with vanadate (see Experimental Procedures), rose to 15 to 25% of maximal stimulation; maximal insulin response (Fig. 2A) and the dose-response curve to insulin (not shown) were the same in normal and vanadium-enriched adipocytes. This procedure yielded adipocytes that were enriched with vanadium by about 25-fold (see Experimental Procedures). In this system, we define an optimal ligand as one that, upon its addition to vanadium-enriched adipocytes by itself and at low concentrations (10 µM), is capable of permeating into the cell interior. There it associates with the enriched pool of vanadium, independent of whether vanadium is intracellularly bound or free, at (+4) or (+5) state of oxidation, and turns it into an insulinomimetic active species, which thus activates lipogenesis.

View larger version (19K): [in this window] [in a new window]   Fig. 2.   Capacity of vanadium ligands, in the absence of exogenous vanadium, to activate lipogenesis in vanadium-enriched and in control adipocytes. Male Wistar rats received daily s.c. administrations of NaVO3 for 4 days before the preparation of adipose cells (see under Experimental Procedures). A, the vanadium-enriched adipocytes were preincubated with 10 µM concentrations of the ligands listed for 10 min. Lipogenesis was then performed for 2 h. B and C, lipogenesis was performed with control adipocytes and 100 µM concentrations of the ligands.

Establishing Cell-Free Procedures for Evaluating Affinities of Ligands to Vanadium at Physiological pH. The complex aqueous chemistry of vanadium can complicate the interpretation of biological experiments. Gross differences in shape, geometry, and coordinating capacity of the vanadium atom may occur in aqueous media at varying pH values (Butler, 1990). Therefore, we searched for a simple cell-free experimental system for estimating approximate ligand affinities to vanadium(+4) and vanadium(+5) at physiological pH values. The reversal of vanadium-evoked inhibition of acid phosphatase from wheat-germ origin at pH 7.2 was finally found most suitable. This acid phosphatase displays a wide, bell-shaped pH profile (Hollander, 1971). Its K_m value with pNPP is not altered upon elevating the pH from 4.8 to 7.2, whereas the maximum turnover number is decreased by 6- to 7-fold only at the higher pH value (data not shown). Wheat-germ acid phosphatase was documented to be inhibited by vanadate(+5) at a K_i value of 6.7 µM (VanEtten et al., 1974), allowing us to use the low vanadium concentrations, which were effective under physiological conditions. In our assay system, vanadium(+5) is stable. The rate of vanadium(+5) oligomerization is minimal at concentrations lower than 50 µM (Crans, 1994), and reduction to vanadium(+4) does not occur because reductants are absent.

1

1

1

1

View larger version (20K): [in this window] [in a new window]   Fig. 3.   Reversal of vanadate(+5)-evoked inhibition of acid-phosphatase as a function of ligand concentration at pH 7.2. The assay mixture (0.5 ml) contained 50 mM HEPES, pH 7.2, 1.0 M KCl, 20 µM NaVO3, 0.2 mM pNPP, 100 µg/ml wheat-germ acid-phosphatase, and increasing concentrations of the tested vanadium ligands. Rates of hydrolysis are defined as the quantity of pNPP hydrolyzed during a 10-min incubation period at 25°C and pH 7.2. Rates were determined individually for each ligand concentration in the absence and presence of 20 µM NaVO3.

1

Vanadium(+4) Is Preserved as Such at Aqueous Neutral Conditions in the Presence of Weakly Interacting Ligands. Before studying further ligands affinities toward vanadium(+4) at physiological pH, we wished to determine the existence of vanadium(+4), at neutral pH, under the experimental conditions applied here. Previous studies claimed that free vanadium(+4) is stable only at acidic pH (pH <3.0), whereas in aqueous solution at neutral pH it is quickly hydrolyzed and/or slowly oxidized to vanadium(+5) (Macara, 1980). Millimolar concentrations of free vanadium(+4) are rather insoluble at neutral pH values. However, in citrate buffer, a weakly interacting ligand (Crans et al., 1989; and this study) in a solution of up to 10 mM can be maintained in a fully soluble form at pH 7.4. Figure 4, A and B, shows the electron spin resonance (ESR) spectrum of VOCl₂, pH <3.0, and in 0.1 M citrate buffer, pH = 7.4. At a neutral pH value, the ESR-spectrum of vanadium(+4) disappeared, but upon acidification at 30, 60, and 90 min after preparation, it reappeared (Fig. 4, C, D, and E). Thus, vanadium(+4) seems to remain in its (+4) oxidation state at aqueous physiological pH. The loss of the ESR signal does not reflect vanadium(+4) hydrolysis, or oxidation, as previously assumed, but rather dimerization, which tends to occur at pH values >3.0 (D. Crans, personal communication).

View larger version (14K): [in this window] [in a new window]   Fig. 4.   ESR spectra of vanadium(+4). Loss of the typical ESR spectrum of vanadium(+4) in aqueous buffer at pH 7.4 and its reappearance upon acidification at different time points. ESR spectrum of 10 mM vanadium dichloride at pH <3.0 (A). B, as in A, but in 0.1 M citrate buffer, pH 7.2. C, as in A, but in 0.1 M citrate-buffer, pH 7.2, maintained for 30 min at 25°C and then acidified with concentrated HCl to pH <3.0. D, as in B, but maintained at pH 7.2, 25°C for 60 min before acidification. E, as in B, but maintained in citrate buffer, pH 7.4, for 90 min before acidified and remonitored. A Bruker ESR-spectrometer (ER-200) was used. Recording conditions were as follows: modulation amplitude, 3.2G; microwave power, 24 mW; time constant, 640 ms; scan time, 200 sec; sweep width, 2000G; and center field, 3500G.

View larger version (17K): [in this window] [in a new window]   Fig. 5.   Growth response of S. cerevisiae in the presence of vanadate and vanadyl. Cells were grown overnight and resuspended in yeast extract/peptone/dextrose medium, pH 7.4. The suspension was divided into several vials and incubated for 5 h at 37°C in the presence or absence of NaVO3(+5) or VOCl2(+4). Cell growth was evaluated spectroscopically at 540 nm.

Ligand Affinities toward Vanadium(+4). Fig. 6 shows the reversal of VOCl₂(+4)-evoked inhibition of acid phosphatase by EDTA and other ligands at pH 7.2. Half-maximal values taken from Fig. 6 are 27 ± 3, 170 ± 14, and 480 ± 30 µM for EDTA, picolinate, and maltolate, respectively. RC₅₀ values for the amino acid hydroxamates were 6.0 ± 0.4 and 21 ± 3 mM for L-Glu()HXM and L-Ile()HXM, respectively. Reduced glutathione, citrate, L-glutamine, L-asparagine, and L-glutamic acid at 100 mM concentration did not reverse vanadium(+4)-evoked inhibition of acid-phosphatase (summarized in Table 2). Thus, under similar experimental conditions, EDTA is 3-fold more potent in reversing vanadium(+4)- than vanadium(+5)-evoked inhibition of acid phosphatase. Table 2 presents the ligands' affinities for vanadium(+4), taking EDTA as an internal standard and using K^{eff/pH 7.2} of 4.14 × 10⁴ M¹. Affinities of formations were 5.8 ± 0.3 and 2.07 ± 0.2 × 10³ M¹ for picolinate and maltolate, repectively, and 1.65 ± 0.2, 1.84 ± 0.3, and 0.47 ± 0.03 × 10² M¹ for L-Glu()HXM, L-Asp()HXM, and L-Ile()HXM, respectively. Thus, at pH 7.2, the affinity of L-Ile()HXM for vanadium(+4) is considerably weaker than for the vanadium(+5) anion (Tables 1 and 2). This, however, is not the case with either L-Glu()HXM or L-Asp()HXM, which have nearly the same binding affinities toward both vanadium(+4) and (+5) at physiological pH (Table 3).

View larger version (18K): [in this window] [in a new window]   Fig. 6.   Reversal of vanadium(+4) evoked inhibition of acid phosphatase as a function of ligand concentration at pH 7.2. The assay mixture (0.5 ml) contained 50 mM HEPES, pH 7.2, 1.0 M KCl, 20 µM VOCl2(+4), 0.2 mM pNPP, and increasing concentrations of the tested vanadium ligands. Rates of hydrolysis are defined as the quantity of pNPP hydrolyzed during a 10-min incubation period at 25° and pH 7.2. Rates were determined individually for each ligand concentration in the absence and presence of 20 µM VOCl2.

Spectroscopic Monitoring of Insulinomimetic-Potentiating Vanadium Ligands. Vanadium ions in aqueous media assume several topographies, depending on the type and number of ancillary ligands occupying the coordination sphere of the vanadium atom (reviewed in Butler, 1990). Here we determined whether these shape-differences are reflected by measurable spectroscopic changes. In Fig. 7, VOCl₂ (2 mM, +4) maintained at 0.1 M acetic acid, pH 2.7, was titrated with various ligands, and the absorbance at 300 nm was recorded. After preliminary studies, pH 2.7 to 3.0 was chosen for conducting these spectroscopic studies, because vanadium(+4) itself absorbed negligibly at this wavelength (_{300 nm} = 50 ± 3).

View larger version (23K): [in this window] [in a new window]   Fig. 7.   Spectroscopic titration of vanadium(+4) with ligands at 300 nm. A solution of VOCl2 (0.3 ml, 2 mM in 0.1 M acetic acid) was placed in a cuvette and the absorbance at 300 nm was monitored. This was sequentially titrated with 1 to 10 equivalents of each of the ligands and the absorbance was measured after each addition. Each equivalent was 3 µl from a 200 mM ligand solution (in 0.1 M acetic acid). None of the vanadium ligands presented in the figure absorb at 300 nm.

View larger version (20K): [in this window] [in a new window]   Fig. 8.   Spectroscopic titration of vanadium(+4) with maltolate, acetylacetonate, and picolinate at 324 nm. Spectroscopic titration was performed essentially as described in the legend to Fig. 7, except that a solution of 0.2 mM VOCl2 was titrated with the studied ligands, and absorbance was monitored at 324 nm. None of the individual ligands presented in the figure absorb at 324 nm.

View larger version (22K): [in this window] [in a new window]   Fig. 9.   Spectroscopic titration of VOCl2 with chemically modified derivatives of L-Glu()HXM: N-acetyl- L-Glu()HXM, but not L-Glu()HXM-COOCH3, associates with vanadium(+4). VOCl2 (2.0 mM) was titrated with L-Glu()HXM and its derivatives. The absorbance at 300 nm was monitored as described in the legend to Fig. 7. L-Glu()HXM-COOCH3 has some absorbance at 300 nm, which was subtracted to obtain net absorbance at this wavelength.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

Complexes of vanadium with L-Glu()HXM (Goldwaser et al., 1999) and with L-Asp()HXM (this study) are especially potent in activating glucose uptake and glucose metabolism in rat adipocytes. In STZ-rats, L-Glu()HXM·vanadium (2:1) is 4 to 5 times more powerful than free vanadium in reducing circulating glucose levels (Goldwaser et al., 1999). In general, for a ligand and vanadium to act synergistically, particularly in vivo, certain requirements must be fulfilled. Most crucial is the appropriate kinetics of complexation and decomplexation, which ensures delivery and adequate presentation of the metalo-oxide ion in an associated form to key sites of glucose metabolism (Orvig et al., 1995; Setyawati et al., 1998). Our notion of L-Glu()HXM being an optimal vanadium ligand is based on its high efficacy in STZ rats in vivo, and in several type II diabetic rodents (Goldwaser et al., 1999; I. Goldwaser, M. Fridkin, and Y. Shechter, manuscript in preparation).

Vanadium can readily fluctuate between vanadate(+5, anionic) and vanadyl(+4, cationic) according to the prevailing conditions (Shechter, 1990; Brichard and Henquin, 1995; Shechter et al., 1995). We therefore a priori envisioned that an optimal chelator should have singular bonds with vanadium at each of the two states of oxidation.

This study demonstrated that L-Glu()HXM and L-Asp()HXM at 10 µM concentrations activated lipogenesis in vanadium-enriched and nonenriched rat adipocytes (Fig. 2). The -amino acid hydroxamates (represented by L-Ile()HXM) were about 7 times less potent. Picolinate and maltolate, although their complexes with vanadium(+4) in STZ-rats were somewhat more active than free vanadium, did not stimulate vanadium-enriched adipocytes. Examining the affinities of formation of all these ligands toward vanadium(+5), we found that the amino acid HXMs have calculated K^{eff/pH 7.2} values significantly lower than picolinate and maltolate. We assumed that this narrow and rather low range of amino acid HXM ligand affinities, toward vanadium(+5) (1.3-1.9 × 10² M¹) coincide with the potentiating effect of vanadium upon complexation; thus, we searched further for an explanation for the lower synergistic potencies of the -amino acid hydroxamates.

With the recognition that vanadium(+4), although paramagnetically silent at pH 7.2, remains in its +4 state of oxidation (Fig. 4 and 5), we studied ligands' affinity for vanadyl(+4). We obtained K^{eff/pH 7.2} values of 1.65 and 1.85 × 10² M¹ for L-Glu()HXM and L-Asp()HXM, respectively, whereas the K^{eff/pH 7.2} for L-Ile()HXM was considerably lower (0.47 × 10² M¹; Table 2).

Comparison of ligands' affinities for vanadium(+4) and (+5) revealed higher affinities of maltolate (4.2-fold) and picolinate (1.7-fold) toward vanadium(+4) compared with vanadium(+5). L-Glu()HXM and L-Asp()HXM, however, showed nearly the same affinity toward vanadium(+4) and vanadium(+5) (affinity ratios of 0.9 and 1.3, respectively; Table 3). L-Ile()HXM is an exception: its affinity for vanadium(+5) exceeded by nearly 3-fold its affinity for vanadium(+4) (Table 3). We therefore concluded that comparable affinities of a ligand toward both states of vanadium oxidation are beneficial for optimal synergistic efficacy. Figure 2C suggests that nonmodified -amino, -carboxyl, and the L configuration are required for Glu()HXM entry into the adipose cell. The free -carboxyl moiety seems necessary for associating with the vanadyl(+4) cation but not with the vanadate(+5) anion, as also suggested, inferred from the spectroscopic measurements (Fig. 9).

Having established the K_d value for L-Glu()HXM·vanadate (2:1), we can now estimate the complex circulating levels shortly after i.p. administration to rats under the experimental conditions applied in our previous study. Considering that the amount of complex administered was 20 µmol/rat and assuming a 10-fold dilution, the complex level is estimated at 2 µmol/ml, of which about 20% would be complexed at physiological pH, using a K_d value of 6.25 ± 1.0 × 10³ M. This level drops within several hours after administration. Nevertheless, normoglycemia has been attained even at 24 h after administration (Fig. 5 in Goldwaser et al., 1999). Thus an initial complex-circulating level of about 0.4 µmol/ml is sufficient for maintaining long-lasting normoglycemic state.

Finally, simple spectroscopic titration of vanadium(+4) with 1, 2, 3, 5, and 10 equivalents of insulinomimetic-active ligands revealed large graded increases in the absorbency of vanadium(+4) at 300 nm with L-Glu()HXM, or L-Asp()HXM, but not with L-Ile()HXM, indicating the efficacy of optimal ligands in inducing intense `spectroscopic-signature' upon associating with vanadium(+4) at slightly acidic, aqueous media. Although these spectroscopic measurements could not be carried out at physiological pH, comparison of the patterns obtained for the different ligands at pH 2.7 to 3.0 may turn out to be useful, as a means of predicting properties of complexes at physiological pH. Thus ligands forming high ultraviolet absorbing complexes also `force' vanadium into an insulinomimetically-active form. The -amino acid HXMs cannot induce this desirable structural configuration. Maltolate and acetylacetonate also form intense ultraviolet absorbing complexes with vanadium(+4) (Fig. 8), but they differ from L-Glu()HXM in forming a much tighter complex with vanadium and in possessing higher affinity for vanadium(+4) than for vanadium(+5) (Tables 1 to 3). These ligands are, therefore, weaker, less effective vanadium donators.

In summary, we propose that optimal insulinomimetic synergizing ligands have to fulfill these three criteria: 1) they must complex with vanadium(+5) within the narrow range of K^{eff/pH 7.2} = 1.3 to 1.9 × 10² M¹ at physiological pH value; 2) they must have nearly equal binding affinities for vanadium(+4) and vanadium(+5) at this pH range; and 3) they must form, with vanadium(+4), an intense ultraviolet-absorbing complex upon titrating the metalo-oxide with 1 to 3 M equivalents of selected ligands at pH ~3.0. Assuming a 2:1 M stoichiometry-dimeric type of complex, each mole of an optimal ligand contributes with a ₃₂₄=1680 ± 100 upon association with vanadium(+4) (calculated from Fig. 8). This is an extensive, large signal, at a wavelength at which the individual components of the complex absorb negligibly or not at all.

Finally, a free -carboxyl moiety (e.g., as in L-Glu()HXM) seems necessary for ligand-vanadyl(+4) cation binding, whereas the -amino side chain is not (Table 3; Fig. 9). In newly designed second generation amino acid HXMs, substitutions can therefore be potentially introduced through the -amino moiety that augment the insulinomimetic effects of vanadium. This and other aspects raised here are currently being studied.