- Email: [email protected]
Composition and biological activity of chromium-pyridine carboxylate complexes
Composition and Biological Activity of Chromium-Pyridine Carboxylate Complexes G. W. Evans and D. J. Pouchnik Department of Chemistty,Bemidji State University,Bemiuji, Mnnesota
ABSTRACT Coordination complexes of chromium (0) and two pyridine carboxylate isomers, nicotinate (nicl and picolinate (pit) were synthesized and analyzed. Cr mono and dinicotinate complexes were formed with 1:l and 1:2 ratios of Cr3+ and nit at pH 7.5. Cr dinicotinate was the only complex formed from a 1:3 ratio of Cr3+ and nit. Mono, di, and tri picolinate complexes were formed with l:l, 1:2, and 1:3 ratios of Cr3+ and pit at pH 7.5. Cr is coordinated with nit through the carboxyl carbon while Cr is coordinated with pit through both the pyridine nitrogen and the carboxyl
carbon. Cr dinicotinate enhanced insulin activity in isolated adipose tissue. None of the other complexes were active in this assay system. In contrast, Cr tripicolinate, which is lipophilic, increased glucose uptake by skeletal muscle cultures but none of the other complexes were effective. In addition, dietary supplements of Cr tripicolinate increased rate of lean body mass development in humans and decreased hemoglobin glycation in aging rats. None of the other complexes was effective in these in vivo assays. The results of this investigation prove that the chemical properties of Cr nit and Cr pit complexes differ markedly. The chemical differences result in a vast difference in the biological action of the complexes.
INTRODUCTION In 1959, Schwan and Mertz [l] first presented evidence that an organic complex which contained chromium was essential to maintain maximal insulin function. Several years later, Mertz and his associates partially purified a chromium complex from yeast [2]. The complex was never completely characterized but preliminary evidence suggested that the complex contained nicotinate (pyridine3-carboxylate) and either glutathione or some amino acid. Experiments with the yeast preparation as well as synthetic complexes of chromium and nicotinate
Address for reprint requests correspondence: Dr. Gary W. Evans, Department Bemidji State University, Bemidji, MN 56601-2699. Journal of Inorganic Biochemishy, 49,177-187 (1993)
of Chemistry,
177 0 1993 Elsevier Science Publishing Co., Inc., 655 Avenue of the Americas, NY, NY 10010 0162-0134/93/$5.00
178
G. W Evans and D. J. Pouchnik
demonstrated that this type of complex would increase the oxidation of glucose in isolated fat pads and in yeast 13, 41. While examining the nature of transition metal complexes with picolinate (pyridine-2-carboxylate), we obtained evidence that chromium in yeast is coordinated with picolinate [51. Synthetic complexes of chromium picolinate added to yeast growth medium resulted in a marked increase in growth rate but did not increase the rate of glucose oxidation [Lillequist and Evans, manuscript submitted]. In addition, synthetic chromium picolinate supplements administered to both animals and humans have resulted in decreased cholesterol [6, 71, decreased serum glucose [8,91, increased rate of lean body mass development, and increased rate of body fat loss [9-121. Several studies prove that picolinate is a bidentate chelating ligand which coordinates with chromium through the pyridine nitrogen and the carboxyl oxygen [13, 141. As described above, several recent studies demonstrate that chromium picolinate affects metabolic parameters regulated by insulin. Since much less is known about the chemical nature of chromium nicotinate complexes or their action in vivo, we compared the chemical and biological characteristics of synthetic preparations of these two pyridine carboxylate isomers coordinated with chromium. METHODS The chromium picolinate complexes were prepared by dissolving 0.01 mole chromium chloride hexahydrate in 25 ml deionized water followed by 0.01 mole, 0.02 mole, or 0.03 mole picolinic acid. The solutions were stirred until a reddish color appeared, indicating complex formation. With the solution containing 0.03 mole picolinic acid, a red solid formed and precipitated from solution. When the other two solutions were adjusted to pH 7.5 with NaOH, a red solid precipitated from solution. The reactions listed above required approximately 1 hr at room temperature, while at 37°C the reactions were complete within 10 min. The solids recovered in the reactions were filtered, washed with deionized water, and air-dried prior to assay. The chromium nicotinate complexes were prepared by dissolving 0.01 mole chromium chloride hexahydrate in 25 ml deionized water followed by 0.01 mole, 0.02 mole, or 0.03 mole nicotinic acid. Unlike the reactions between chromium and picolinate, these solutions had to be heated (60°C) before the deep green of the solution began to turn a royal blue, indicating that a complex had formed. No precipitate formed with these mixtures until the solutions were adjusted to pH 7.5 with NaOH. The solids recovered in the reactions were filtered, washed with deionized water, and air-dried prior to assay. To insure that the solid complexes recovered and ultimately assayed were homogeneous, the solids were assayed by high performance liquid chromatography (HPLC). The solids from each reaction described above were first separately dissolved in deionized water after which the solution was filtered. A 20 microliter sample was then injected into the port of a Beckman 1lOB Solvent Delivery System with an Econosphere Cl8 5U column attached. Samples were eluted with methanol/water (50/50) at a flow rate of 0.8 ml/min. Sample elution was detected with a Beckman System Gold Analog Interface Module 406 set at 262 nm for the nicotinate complexes and 264 for the picolinate complexes.
CHROMIUM
COORDINATION
COMPLEXES
179
Mass spectrographic analysis was accomplished by means of fast atom bombardment, Each of the solid samples described above was dissolved or dispersed in 3-nitrobenxyl alcohol that contained 5 percent ethanol and subsequently applied to the probe rod of a Varian VGZAB3F mass spectrometer equipped with a Varian-MAT 55200 data system. An 8 keV xenon beam was used to generate the mass spectrum. Molecular mass of each of the recovered solid complexes was determined by the use of melting point depression of camphor. A 0.25 g sample of each of the solids described above was mixed separately with 5.0 g recrystallized camphor. The melting point of the solid mixture was determined with a Laboratory Devises Mel-Temp II apparatus. Molecular mass was calculated from the formula: AT = k, x m, where AT is melting point depression, k, = 4O.O’C/m, and m is molality. Ultraviolet (UV) analysis of each of the complexes was accomplished with a Perkin-Elmer Lambda 3B ultraviolet spectrophotometer. The biological activity of the chromium complexes was examined in two different in vitro systems and in two different in vivo systems. Glucose oxidation in isolated fat cells was assayed by the method described by Anderson et al. [151. Fat pads were removed from rats fed ad libitum a Torula yeast-based diet (Teklad, Madison, WI). The chromium content of this diet was less than 400 nanograms/g diet. The fat cells were isolated by use of collagenase after which the cells were washed and subsequently suspended in 10 ml of a Krebs Ringer Phosphate-albumin solution. To assay glucose oxidation, 0.05 ml of the fat cell preparation was added to 1.95 ml of KRP-albumin that contained 5 nM [U- CID-glucose (ICN, Costa Mesa, CA, Sp. Act. = 290 Ci/mol), 1.0 nM insulin, and 10 PM chromium complex or chromium chloride in sealed tubes. Radioactive carbon dioxide trapped in hyamine was determined in a Beckman LS 5000 scintillation counter. The biological activity of the chromium complexes was also assayed in cultured muscle cells. Culture media and all ingredients were purchased from Sigma Chemical Co., St. Louis, MO. Rat skeletal muscle myoblasts (American Type Culture Collection, Rockville, MD) were initially established for 48 hr in Rulbecco’s modtied Eagle’s medium supplemented with 10% fetal calf serum. Thereafter, the cells were cultured in serum-free treatment medium which was replaced daily. Each of the culture treatment medium contained Dulbecco’s modified Eagle’s medium, 1.0 nM insulin, 50 ng/ml basic fibroblast growth factor, and 50 ng/ml insulin-like growth factor II, and 1.0 PM chromium complex or ,l.O PM chromium chloride. After 72 hr, cells from each treatment medium were washed’and plated at 5.0 x lo4 cells/cm* in 0.5 ml medium that contained additions appropriate for the experiment, To assay the effect of chromium on glucose metabolism, 1.0 nM insulin and 5 nM [6-3H]D-glucose (ICN, Costa Mesa, CA, Sp. Act. = 35 Ci/mmol) was added to Dulbecco’s modified Eagle’s medium. An identical assay was carried out with insulin omitted from the medium. The cells were incubated at 37°C for 1 hr, and removed and rinsed with cold saline over a vacuum filter. The cells were then transferred to vials for radioactivity assay in a Beckman LS 5000 scintillation counter. The biological activity of the chromium complexes was measured in vivo in rats and humans. Weanling, male Long-Evans rats were housed individually in
180
G. W. Evans and D. J. Pouchnik
plastic cages and fed ad libitum a purified diet [16] to which either chromium chloride or one of the chromium complexes was added at a level of 1 pg Cr/g diet. Since we were merely attempting to determine the effect of chromium additions to a basal diet, we did not prepare chromium deficient diets for this group of experiments. The chromium content of the basal diet was 0.8 pg Cr/g diet. After 200 days, 50 PL blood was drawn from the tail of each rat between the hours of 1400 and 1500. Plasma glucose was assayed with the glucose oxidase kit obtained from Sigma Chemical Co. Glycated hemoglobin (HbA,,) was measured with the Glycotest Kit from Pierce Chemical Company (Rockford, IL). For the tests with human volunteers, 12 male and 12 female participants (age 25-36) in a weekly aerobics class were divided into two groups of 6 males and 6 females. Each of the males was given a coded bottle of tablets that contained either 3.3 mg chromium tripicolinate (400 micrograms Cr+3) mixed with 5 mg calcium phosphate or 3.1 mg chromium dinicotinate (400 micrograms Cr”) mixed with 5 mg calcium phosphate. The females were given a coded bottle of tablets that contained either 1.6 mg chromium tripicolinate (200 micrograms Cr+3) mixed with 5 mg calcium phosphate or 1.5 mg chromium dinicotinate (200 micrograms Cr+3) mixed with 5 mg calcium phosphate. The participants were instructed to take one capsule each day at breakfast. Lean body mass was determined prior to beginning the class and exercise regimen and again after 12 weeks. Lean body mass was calculated from total body resistivity measured as described by Segal et al. [171 with a four-terminal portable impedance analyzer (RJL Systems, Detroit, MI). Data are expressed as mean f S.E. Data were analyzed by analysis of variance and Duncan’s multiple range test.
RESULTS Plcollnate Complexes The precipitate formed with a 1:l ratio of chromium and picolinate (pick proved to be homogeneous when assayed with HPLC. A single, symmetrical peak was eluted after 2.92 min with this complex. The composition of this homogeneous complex was deduced from a combination of mass spectral analysis, melting point depression, and ultraviolet spectroscopy. Mass spectral analysis of this precipitate showed a large peak at 52 daltons (chromium ion) and a parent peak at 244 daltons which corresponds with 1 mole chromium coordinated with 1 mole picolinate and four moles’water. Melting point depression (AT = 7.66’0 yielded a molecular mass of 261 f 2 daltons which corresponds with 1 mole chromium, 1 mole picolinate, ,and 5 moles water. When 2.61 mg of this complex was dissolved in 1.0 liter deionized water, peak W absorbance was observed at 264 nm and the absorbance was 0.052 at pH -7.0. Peak absorbance of picolinate is found at 264 nm and the molar absorptivity of picolinate at 264 nm is 4020 L mol-’ cm-’ at pH 7.0 where the pyridine nitrogen would be unprotonated. At pH 3.0, where the pyridine nitrogen would be protonated, the molar absorptivity of picolinate at 264 nm is 7220 L mol-’ cm-‘. Thus, assuming that the molar absorptivity of picolinate coordinated with chromium would fall between 4020 L mol-’ cm-’ and 7220 L mol-’ cm-‘, the molecular formula of the precipitate
CHROMIUM
COORDINATION
COMPLEXES
181
formed with a 1:l ratio of chromium and picolinate was Cr pit (H,O),(OH),. (H,O). The solubility df this precipitated complex in pH 7.0 water was 11.1 PM. The precipitate formed with a 1:2 ratio of chromium and picolinate proved to be homogeneous when assayed with HPLC. A single, symmetrical peak was eluted after 3.59 min with this complex. The composition of this homogeneous complex was deduced from a combination of mass spectral analysis, melting point depression, and ultraviolet spectroscopy. Mass spectral analysis of this precipitate formed with a 1:2 ratio of chromium and showed a large peak at 52 daltons and a parent peak at 331 daltons which corresponds with 1 mole chromium coordinated with 2 moles picolinate and two moles water. Melting point depression (AT = 5.73”C) yielded a molecular mass of 349 f 3 daltons which corresponds with 1 mole chromium, 2 moles picolinate, and 3 moles water. When 3.49 mg of this complex was dissolved in 1.0 liter deionized water, peak UV absorbance was observed at 264 nm and the absorbance was 0.104 at pH 7.0. Thus, the molecular formula of the precipitate formed with a 1:2 ratio of chromium and picolinate was Cr(pic),(H,OXOH)(H,O). The solubility of this precipitated complex in pH 7.0 water was 14.1 PM. Neither the mono- nor the dipicolinate complex was active in the in vitro systems (Tables 1 and 2). The precipitate formed with a 1:3 ratio of chromium and picol,jnate also proved to be homogeneous when assayed with HPLC. A single, @metrical peak was eluted after 4.95 min with this complex. The composition of this homogeneous complex was deduced from a combination of mass spectral analysis, melting point depression, and ultraviolet spectroscopy. Mass spectral analysis of this precipitate showed a large peak at 52 daltons and a parent peak at 418 daltons which corresponds with 1 mole chromium coordinated with 3 moles picolinate. Melting point depression (AT = 4.78”C) yielded a molecular mass of 435 f 2 daltons which corresponds with 1 mole chromium, 3 moles picolinate, and 1 mole water. When 4.35 mg of this complex was dissolved in 1.0 liter deionized water, peak UV absorbance was observed at 264 nm and the absorbance was 0.155 at pH 7.0. Thus, the molecular formula of the precipitate formed with a 1:3 ratio of chromium and picolinate was Cr(pic),-H,O. The solubility of this crystalline complex in pH 7.0 water was 0.6 mM and 2.0 mM in chloroform. Molar absorptivity of this complex at 264 nm was 15546 L mol-’ cm-’ at both pH 3.0 and pH 7.0. As described previously, the molar absorptivity
TABLE 1. The Effect of Chromium Fyridine Carboxylates on Glucose Gxidation in Isolated Fat Cells Addition None Cr mono pit Cr di pit Cr tri pit Cr mono nit Cr di nit Cr chloride
Radioactivity Qm 219 f 34 225 f 45 229*37 231f49 218 f 54 623 f 147’ 243 f 29
Increase 1.02 1.05 1.06 1.0 2.84 1.11
Values are mean f S.E. %lue is significantlydifferent (P < 0.01) from other values in the column.
182 G. W. Evans and D. J. Pouchnik
TABLE 2. Uptake of Glucose Into Skeletal Muscle Cells Cultured With Various Forms of Chromium Addition None Cr mono pit Cr di pit Cr tri pit Cr mono nit Cr di nit Cr chloride PiC
Radioactivity Cpm 369 f 43 3% 408 629 354 367 382 375
f f f f + f f
54 52 6ga 56 45 53 67
Increase 1.07 1.10 1.70 1.0 1.0 1.03 1.01
Values are mean f SE. ‘Value is significantly different (P < 0.01) from other values in the column.
of picolinate at 264 nm is 4020 L mol-’ cm-’ at pH 7.0 but is 7220 L mol-’ cm-’ at pH 3.0. The molar absorptivity of the chromium tripicolinate complex was not altered by change in pH, suggesting that the pyridine nitrogen is involved in the coordination with chromium. The chromium tripicolinate complex was not active in the rat fat pad assay (Table 1). However, chromium tripicolinate added to rat skeletal muscle cultures produced a significant increase in the uptake of glucose (Table 2). Moreover, plasma glucose was significantly less (P < 0.01) in rats fed chromium tripicolinate and this chromium complex also prevented glycation of hemoglobin (Table 3). Chromium tripicolinate supplements resulted in a significant increase (P < 0.01) in lean body mass in both males and females (Table 4) participating in an aerobics class. Nicotinate Complexes The precipitate formed with a 1:l ratio of chromium and nicotinate (nit) proved to be homogeneous when assayed with HPLC. A single, symmetrical peak was eluted after 1.21 min with this complex. The composition of this homogeneous complex was deduced from a combination of mass spectral analysis, melting point depression, and ultraviolet spectroscopy. Mass spectral analysis of this precipitate showed a large peak at 52 daltons and a parent peak at 264 daltons which corresponds with 1 mole chromium coordinated with 1 mole nicotinate
TABLE 3. Plasma Glucose and Gylcated Hemoglobin in Aging Rats Fed Various Chromium Compounds Diet Supplement
Plasma Glucose mM
None
7.9 f 0.11
Cr tri pit Cr di nit Cr chloride
6.6 f 0.12a 7.7 f 0.12 7.8 f 0.11
HfJAI, (o/o) 5.46 3.31 5.19 5.39
f f f f
0.08 OLW 0.09 0.08
Values are mean f S.E. ‘Value is significantiy different (P < 0.01) from other values in the cohmul.
CHROMIUM
COORDINATION
COMPLEXES
183
TABLE 4. Lean Body Mass Development in Volunteers Given Chromium-Pyridine carboxylate supplements Supplement Females Initial resistance (Q) Initial LBM (kg) Final resistance (QI Final LBM (kg) Change in resistance (Q) Change in LBM (kg) Males Initial resistance (Q) Initial LBM (kg) Final resistance (Q) Final LBM (kg) Change in resistance (Q) Change in LBM (kg)
Cr pit
Cr nit
563 f 52 46 f 6 540*45 48 f 4 -24 f 5a + 1.8 f 0.3a
554 f 49 44*5 546 f 47 45 f 5 -7f2 +0.6 f 0.2
455 f 54 64*9 441 f 49 67 f 7 - 15 + 38 +2.1 f 0.3”
461 f 56 62 f 9 454 f 43 63 f 8 -5* 1 0.7 f 0.3
Values are mean f SE. aValue is significantly different (P < 0.01) from other values in the row.
and five moles water. Melting point depression (AT = 7.12”C) yielded a molecular mass of 281 f 2 daltons which corresponds with 1 mole chromium, 1 mole nicotinate, and 6 moles water. When 2.81 mg of this complex was dissolved in 1.0 liter deionized water, peak UV absorbance was observed at 262 nm and the absorbance was 0.031 at pH 7.0. Peak absorbance of nicotinate is found at 262 nm and the molar absorptivity of nicotinate at 262 nm is 3000 L mol-’ cm-’ at pH 6.0 where the pyridine nitrogen would be unprotonated. At pH 3.0, where the pyridine nitrogen would be protonated, the molar absorptivity of nicotinate at 262 nm is 5010 L mol-’ cm-‘. Thus, assuming that the molar absorptivity of nicotinate coordinated with chromium would fall between 3000 L mol-’ cm-i and 5010 L mol-’ cm-i, the molecular formula of the precipitate formed with a 1:l ratio of chromium and nicotinate was Cr nit (H,O),(OH),*(H,O). The solubility of this precipitated complex in pH 7.0 water was 10.9 PM. This complex showed no biological activity (Tables 1 and 2). The precipitates formed with either a 1:2 ratio of chromium and nicotinate or a 1:3 ratio of chromium and nicotinate proved to be homogeneous and identical when assayed with HPLC. A single, symmetrical peak was eluted after 2.32 min when the precipitate from either of these preparations was assayed. The composition of these two homogeneous complexes was deduced from a combination of mass spectral analysis, melting point depression, and ultraviolet spectroscopy. Mass spectral analysis of the precipitates from both of these preparations were identical and showed a large peak at 52 daltons with a parent peak at 369 daltons which corresponds with 1 mole chromium coordinated with 2 moles nicotinate and four moles water. Melting point depression (AT = 5.17”C) yielded a molecular mass of 387 + 2 daltons which corresponds with 1 mole chromium, 2 moles nicotinate, and 5 moles water. When 3.87 mg of this complex was dissolved in 1.0 liter deionized water, peak UV absorbance was observed at 262 nm and the absorbance was 0.060 at pH 6.0. From these observations, the
184
G. W Evans and D. J. Pouchnik
molecular formula of the precipitated complex formed with either a 1:2 or a 1:3 ratio of chromium and nicotinate was C&ric),(H,O),(OH)-H,O. The solubility of the dinicotinate complex in pH 7.0 water was 12.4 PM. The solubility of chromium dinicotinate in chloroform was 52 PM. When the dinicotinate complex was solubilized in 1.0 M HCl and subsequently adjusted to pH 3.0, the molar absorptivity of the complex at 262 nm was 10200 L mol-’ cm-‘. However, as described above, the molar absorptivity of the dinicotinate complex was 6000 L mall’ cm-’ at pH 6.0. These results indicate that the pyridine nitrogen of nicotinate is not involved in coordination of chromium since the molar absorptivity of nicotinate is 5000 L mol-’ cm-’ when the nitrogen is protonated and only 3000 L mol-’ cm-’ when the hydrogen ion dissociates. The chromium dinicotinate precipitate produced a significant increase in glucose oxidation when assayed in the fat pad assay system (Table 1). In contrast, this complex did not enhance glucose uptake into cultured skeletal muscle (Table 2). The precipitated chromium dinicotinate complexes were not effective in either of the in vivo assays (Tables 3 and 4). Chromium picolinate added to rat diets resulted in a decreased concentration of plasma glucose and a concomitant decrease in glycated hemoglobin but chromium nicotinate had no effect on these parameters. In addition, chromium nicotinate supplements given to humans enrolled in an aerobics class failed to produce a significant increase (P > 0.05) in development of lean body mass. DISC1JSSION
The results of our experiments prove that chromium coordination complexes with nicotinate or picolinate differ markedly in both chemical composition and biological action. Our results and those of Cooper et al. [4] provide substantial evidence that chromium is coordinated with nicotinate through the carboxyl oxygen. Green and Tong [18] showed that the molar absorptivity of nicotinate at 262 nm is approximately 5000 L mol-’ cm-’ when the pyridine nitrogen is protonated but decreases to 3000 L mol-’ cm-’ when the proton dissociates. As described above, when the hydrogen ion concentration was decreased in a solution that contained chromium dinicotinate, the molar absorptivity decreased. This observation proves that the pyridine nitrogen in the chromium nicotinate complex is not coordinated. The position of the carboxyl group in the nicotinate molecule obviously precludes formation of a five-membered ring which in turn apparently prevents complexation of chromium with three molecules nicotinate as found in the picolinate complex. Since nicotinate complexes with chromium are monodentate, the remaining coordination sites must be occupied by water or hydroxide unless molecules capable of replacing these are present in the medium. When coordinated with nicotinate and water, the complexes are insoluble at physiological pH. The insolubility at physiological pH arises from the dissociation of the pyridine nitrogens (pK = 4.8) and the conversion of the coordinated water to hydroxide (pK = 41. Once precipitated, chromium nicotinate complexes are extremely difficult to solubilize due to the presence of the coordinated hydroxides which form strong hydrogen bonds in the precipitate. This strong hydrogen bonding may account in part for the ineffectiveness of the chromium nicotinate complexes in the in vivo assays. Although ineffective in vivo, chromium dinicotinate potentiated glucose
CHROMIUM
COORDINATION
COMPLEXES
185
oxidation in isolated fat cells. Similar results have been obtained by other investigators [2-41 but the mechanism whereby chromium dinicotinate enhances glucose oxidation has not been elucidated. Mertz [2] has suggested that a ternary complex is formed with chromium, insulin, and insulin receptors. However, Cooper et al. [4] demonstrated that chromium dinicotinate enhances glucose oxidation in yeast where no insulin is present. Cooper et al. 141 have also presented evidence that the two pyridine nitrogen groups are somehow necessary for activity in the glucose oxidation assay system. The results of our experiments and those of other investigators indicate that picolinate forms bidentate complexes with chromium through the carboxyl oxygen and .the pyridine nitrogen [13, 141. We have recently confirmed this structure by use of x-ray crystallography. The bidentate coordination permits formation of a tripicolinate complex which is uncharged at all hydrogen ion concentrations. Chromium tripicolinate is slightly soluble in water (0.6 mM) but is about three times more soluble in chloroform (2 mM). In a previous series of experiments, we demonstrated that chromium picolinate added to synthetic liposomes increased membrane fluidity. In those experiments, we also discovered that chromium picolinate increased insulin internalization, glucose uptake, and leucine uptake into cultured skeletal muscle. Based on these observations, we have suggested that chromium picolinate potentiates insulin action by increasing membrane fluidity [ 191. In the study described in this paper, chromium picolinate supplements accelerated development of lean body mass in exercising volunteers while chromium nicotinate was ineffective. Several previous studies with both animals and humans prove that chromium picolinate promotes development of muscle with a concomitant loss of body fat. In two separate studies at this university, chromium picolinate supplements resulted in accelerated body fat loss and lean body mass development in males involved in weight-lifting programs [9]. Hasten et al. [lo] detected a significant increase in lean body mass in females and a significant increase in muscle circumference in both males and females given chromium picolinate supplements. Kaats et al. [ll] found significant body fat loss and development of lean body mass in sedentary individuals. During experiments with swine, Page et al. [12] discovered that chromium picolinate supplements in the diet resulted in decreased fat content and increased muscle in rib loins. These investigators also detected significantly lower cholesterol levels in swine supplemented with chromium picolinate [7]. In experiments with human volunteers, Press et al. [6] detected a significant decrease in total cholesterol and LDL-cholesterol during supplementation with chromium picolinate. At the same medical center, investigators detected a significant decrease in serum glucose and glycated hemoglobin in non-insulindependent diabetics [9]. In addition, a marked decrease in blood glucose was produced in native American volunteers taking chromium picolinate supplements 181. As described above, both serum glucose and glycated hemoglobin were significantly decreased in aging rats fed chromium picolinate from the time of weaning which confirms the in vivo effect of chromium picolinate on insulin regulated metabolic functions. Each of the metabolic parameters affected by chromium picolinate supplementation can be linked to the action of insulin. Insulin produces a wide variety of responses that ultimately affect carbohydrate and lipid and protein metabolism
186 G. W Evans and D. J. Pouchnik
[201. Previous studies with muscle cultures suggest that chromium picolinate may
affect the action of insulin through an effect on the rate of insulin internalization, which, by an unidentified mechanism, regulates the synthesis and/or insertion of insulin receptors into the plasma membrane [19]. Because of the lipophilic nature of chromium picolinate, the complex may affect insulin internalization by maintaining a level of membrane fluidity necessary ‘for efficient removal of the insulin-receptor complex from the surface of the plasma membrane. Alternatively, the picolinate complex with chromium may simply provide a delivery system enabling the chromium ion to traverse the,plasma membrane to reach the interior of the cell where it is utilized in some as yet unidentified biochemical pathway. Although the nicotinate and picolinate complexes enhanced insulin action through entirely different mechanisms, in each case chromium was complexed with a pyridine carboxylate. Chromium salts were not effective in either of the in vitro studies and the chromium salt failed to prevent elevated plasma glucose and protein glycation in aging rats. Chromium chloride added to the diets of swine had no effect on body fat or lean body mass [121 and several previous investigations demonstrate that ionic chromium has no effect on insulin action [2,3, 151. Although our results and those of other investigators demonstrate that chromium ion has no effect upon insulin action, positive results have been obtained following oral or intravenous administration of chromium salts [2, 211. Since chromium ion has no apparent function, these observations suggest that chromium can form biologically active complexes within mammalian cells under certain conditions. Several years ago researchers discovered that picolinate carboxylase, the enzyme involved in synthesis of picolinate, is much more active in diabetic animals [22]. This observation suggests that more picolinate is available for combination with chromium in humans and animals with impaired insulin sensitivity. In most, if not all, the successful experiments in which chromium salts were used, the subjects had some form of insulin resistance [2, 211. Chromium coordinates with picolinate readily at 37°C and though not extremely soluble in water is several times more soluble than chromium nicotinate. Chromium picolinate is also soluble in chloroform which suggests that this complex can pass through or function within biological membranes. The formation of chromium nicotinate requires heat, and once formed, these complexes are not extremely soluble in either water or lipid at physiological pH. As Mertz [2] has suggested, in order for a chromium dinicotinate complex to exist and be effective at physiological pH, the complex would have to contain groups that preclude chromium coordination with water. While exogenously synthesized chromium-pyridine carboxylate complexes influence insulin function, whether or not these types of chromium complexes are actually formed in mammalian cells remains to be determined.
REFERENCES 1. K. Schwarz and W. Mertz, Arch. Biochem. BiopIys. 85,292 (1959). 2. W. Mertz, Nutrition Revs. 33, 129 (1975). 3. E. W. Toepfer, W. Mertz, M. M. Polansb, E. E. Roginski, and W. R. Wolf, J. Agric. Food Chem. 25, 162 (1977).
CHROMIUM COORDINATION COMPLEXES
187
4. J. A. Cooper, B. F. Anderson, P. D. Buckley, and L. F. Blackwell, Inorganica Chimicu Actu. 91, 1 (1984). 5. P. Munnis, Bemidji State Uniuersity J Stuknt Res 7, 11 (1987). 6. R. I Press, J. Geller, and G. W. Evans, Western Journul of Medicine 152, 41 (1990). 7. T. G. Page, T. L. Ward, L. L. Southern, and D. L. Thompson, J. Animal Science Suppf. 2 69, 356 (1991). 8. G .W. Evans, Western Journal of Medicine 155,549 (1991). 9. G .W. Evans, Znt. J. Bios. and Med. Res. 11, 163(1989). 10. D. L. Hasten, E. P. Rome, and B. D. Franks, Am. College of Sports Med. Southeast Regional Chupt. Abstmcts 18, 49 (1991). 11. G. R. Kaats, J. A. Fisher, K. Blum, and J. A. Adelman, Am. Aging Assoc. Abstructs 21, 10 (1991). 12. T. G. Page, T. L. Ward, and L. L. Southern, J. Animal Science Suppl. Z 69, 356 (1991). 13. L. Campanella, E. Chiacchierini, G. De Angelis, and V. Petrone, Annali di Chimicu 67, 385 (1977). 14. S. Takata, E. Kyuno, and R. Tsuchiya, Bull Chem. Sot. Japan 41,3416 (1968). 15. R. A. Anderson, J. H. Brantner, and M. M. Polansky, J. A&c. Food Chem. 26,1219 (1978). 16. G. W. Evans and E. C. Johnson, J. Nutr. 111,68 (1981). 17. K. R. Segal, M. Van Loan, P. I. Fitzgerald, J. A. Hodgdon, and T. B. Van' Itallie, Am. J. Clin. Nutr. 47, 7 (1988). 18. R. W. Green and H. K. Tong, J. Am. Chem. Sot. 78,48% (1976). 19. G. W. Evans and T. D. Bowman, J. Inorgan. B&hem., 46, 243 (1992). 20. C. R. Kahn, Ann. Rev. Med. 36,429 (1985). 21. R. A. Anderson, Clin Physiol Biochem 4,31 (1986). 22. A. H. Mehler, E. G. McDaniel, and J. M. Hundley, J. Biol. Chem. 232, 323 (1958). Received March 5, 1992; accepted July 2, 1992
ABSTRACT Coordination complexes of chromium (0) and two pyridine carboxylate isomers, nicotinate (nicl and picolinate (pit) were synthesized and analyzed. Cr mono and dinicotinate complexes were formed with 1:l and 1:2 ratios of Cr3+ and nit at pH 7.5. Cr dinicotinate was the only complex formed from a 1:3 ratio of Cr3+ and nit. Mono, di, and tri picolinate complexes were formed with l:l, 1:2, and 1:3 ratios of Cr3+ and pit at pH 7.5. Cr is coordinated with nit through the carboxyl carbon while Cr is coordinated with pit through both the pyridine nitrogen and the carboxyl
carbon. Cr dinicotinate enhanced insulin activity in isolated adipose tissue. None of the other complexes were active in this assay system. In contrast, Cr tripicolinate, which is lipophilic, increased glucose uptake by skeletal muscle cultures but none of the other complexes were effective. In addition, dietary supplements of Cr tripicolinate increased rate of lean body mass development in humans and decreased hemoglobin glycation in aging rats. None of the other complexes was effective in these in vivo assays. The results of this investigation prove that the chemical properties of Cr nit and Cr pit complexes differ markedly. The chemical differences result in a vast difference in the biological action of the complexes.
INTRODUCTION In 1959, Schwan and Mertz [l] first presented evidence that an organic complex which contained chromium was essential to maintain maximal insulin function. Several years later, Mertz and his associates partially purified a chromium complex from yeast [2]. The complex was never completely characterized but preliminary evidence suggested that the complex contained nicotinate (pyridine3-carboxylate) and either glutathione or some amino acid. Experiments with the yeast preparation as well as synthetic complexes of chromium and nicotinate
Address for reprint requests correspondence: Dr. Gary W. Evans, Department Bemidji State University, Bemidji, MN 56601-2699. Journal of Inorganic Biochemishy, 49,177-187 (1993)
of Chemistry,
177 0 1993 Elsevier Science Publishing Co., Inc., 655 Avenue of the Americas, NY, NY 10010 0162-0134/93/$5.00
178
G. W Evans and D. J. Pouchnik
demonstrated that this type of complex would increase the oxidation of glucose in isolated fat pads and in yeast 13, 41. While examining the nature of transition metal complexes with picolinate (pyridine-2-carboxylate), we obtained evidence that chromium in yeast is coordinated with picolinate [51. Synthetic complexes of chromium picolinate added to yeast growth medium resulted in a marked increase in growth rate but did not increase the rate of glucose oxidation [Lillequist and Evans, manuscript submitted]. In addition, synthetic chromium picolinate supplements administered to both animals and humans have resulted in decreased cholesterol [6, 71, decreased serum glucose [8,91, increased rate of lean body mass development, and increased rate of body fat loss [9-121. Several studies prove that picolinate is a bidentate chelating ligand which coordinates with chromium through the pyridine nitrogen and the carboxyl oxygen [13, 141. As described above, several recent studies demonstrate that chromium picolinate affects metabolic parameters regulated by insulin. Since much less is known about the chemical nature of chromium nicotinate complexes or their action in vivo, we compared the chemical and biological characteristics of synthetic preparations of these two pyridine carboxylate isomers coordinated with chromium. METHODS The chromium picolinate complexes were prepared by dissolving 0.01 mole chromium chloride hexahydrate in 25 ml deionized water followed by 0.01 mole, 0.02 mole, or 0.03 mole picolinic acid. The solutions were stirred until a reddish color appeared, indicating complex formation. With the solution containing 0.03 mole picolinic acid, a red solid formed and precipitated from solution. When the other two solutions were adjusted to pH 7.5 with NaOH, a red solid precipitated from solution. The reactions listed above required approximately 1 hr at room temperature, while at 37°C the reactions were complete within 10 min. The solids recovered in the reactions were filtered, washed with deionized water, and air-dried prior to assay. The chromium nicotinate complexes were prepared by dissolving 0.01 mole chromium chloride hexahydrate in 25 ml deionized water followed by 0.01 mole, 0.02 mole, or 0.03 mole nicotinic acid. Unlike the reactions between chromium and picolinate, these solutions had to be heated (60°C) before the deep green of the solution began to turn a royal blue, indicating that a complex had formed. No precipitate formed with these mixtures until the solutions were adjusted to pH 7.5 with NaOH. The solids recovered in the reactions were filtered, washed with deionized water, and air-dried prior to assay. To insure that the solid complexes recovered and ultimately assayed were homogeneous, the solids were assayed by high performance liquid chromatography (HPLC). The solids from each reaction described above were first separately dissolved in deionized water after which the solution was filtered. A 20 microliter sample was then injected into the port of a Beckman 1lOB Solvent Delivery System with an Econosphere Cl8 5U column attached. Samples were eluted with methanol/water (50/50) at a flow rate of 0.8 ml/min. Sample elution was detected with a Beckman System Gold Analog Interface Module 406 set at 262 nm for the nicotinate complexes and 264 for the picolinate complexes.
CHROMIUM
COORDINATION
COMPLEXES
179
Mass spectrographic analysis was accomplished by means of fast atom bombardment, Each of the solid samples described above was dissolved or dispersed in 3-nitrobenxyl alcohol that contained 5 percent ethanol and subsequently applied to the probe rod of a Varian VGZAB3F mass spectrometer equipped with a Varian-MAT 55200 data system. An 8 keV xenon beam was used to generate the mass spectrum. Molecular mass of each of the recovered solid complexes was determined by the use of melting point depression of camphor. A 0.25 g sample of each of the solids described above was mixed separately with 5.0 g recrystallized camphor. The melting point of the solid mixture was determined with a Laboratory Devises Mel-Temp II apparatus. Molecular mass was calculated from the formula: AT = k, x m, where AT is melting point depression, k, = 4O.O’C/m, and m is molality. Ultraviolet (UV) analysis of each of the complexes was accomplished with a Perkin-Elmer Lambda 3B ultraviolet spectrophotometer. The biological activity of the chromium complexes was examined in two different in vitro systems and in two different in vivo systems. Glucose oxidation in isolated fat cells was assayed by the method described by Anderson et al. [151. Fat pads were removed from rats fed ad libitum a Torula yeast-based diet (Teklad, Madison, WI). The chromium content of this diet was less than 400 nanograms/g diet. The fat cells were isolated by use of collagenase after which the cells were washed and subsequently suspended in 10 ml of a Krebs Ringer Phosphate-albumin solution. To assay glucose oxidation, 0.05 ml of the fat cell preparation was added to 1.95 ml of KRP-albumin that contained 5 nM [U- CID-glucose (ICN, Costa Mesa, CA, Sp. Act. = 290 Ci/mol), 1.0 nM insulin, and 10 PM chromium complex or chromium chloride in sealed tubes. Radioactive carbon dioxide trapped in hyamine was determined in a Beckman LS 5000 scintillation counter. The biological activity of the chromium complexes was also assayed in cultured muscle cells. Culture media and all ingredients were purchased from Sigma Chemical Co., St. Louis, MO. Rat skeletal muscle myoblasts (American Type Culture Collection, Rockville, MD) were initially established for 48 hr in Rulbecco’s modtied Eagle’s medium supplemented with 10% fetal calf serum. Thereafter, the cells were cultured in serum-free treatment medium which was replaced daily. Each of the culture treatment medium contained Dulbecco’s modified Eagle’s medium, 1.0 nM insulin, 50 ng/ml basic fibroblast growth factor, and 50 ng/ml insulin-like growth factor II, and 1.0 PM chromium complex or ,l.O PM chromium chloride. After 72 hr, cells from each treatment medium were washed’and plated at 5.0 x lo4 cells/cm* in 0.5 ml medium that contained additions appropriate for the experiment, To assay the effect of chromium on glucose metabolism, 1.0 nM insulin and 5 nM [6-3H]D-glucose (ICN, Costa Mesa, CA, Sp. Act. = 35 Ci/mmol) was added to Dulbecco’s modified Eagle’s medium. An identical assay was carried out with insulin omitted from the medium. The cells were incubated at 37°C for 1 hr, and removed and rinsed with cold saline over a vacuum filter. The cells were then transferred to vials for radioactivity assay in a Beckman LS 5000 scintillation counter. The biological activity of the chromium complexes was measured in vivo in rats and humans. Weanling, male Long-Evans rats were housed individually in
180
G. W. Evans and D. J. Pouchnik
plastic cages and fed ad libitum a purified diet [16] to which either chromium chloride or one of the chromium complexes was added at a level of 1 pg Cr/g diet. Since we were merely attempting to determine the effect of chromium additions to a basal diet, we did not prepare chromium deficient diets for this group of experiments. The chromium content of the basal diet was 0.8 pg Cr/g diet. After 200 days, 50 PL blood was drawn from the tail of each rat between the hours of 1400 and 1500. Plasma glucose was assayed with the glucose oxidase kit obtained from Sigma Chemical Co. Glycated hemoglobin (HbA,,) was measured with the Glycotest Kit from Pierce Chemical Company (Rockford, IL). For the tests with human volunteers, 12 male and 12 female participants (age 25-36) in a weekly aerobics class were divided into two groups of 6 males and 6 females. Each of the males was given a coded bottle of tablets that contained either 3.3 mg chromium tripicolinate (400 micrograms Cr+3) mixed with 5 mg calcium phosphate or 3.1 mg chromium dinicotinate (400 micrograms Cr”) mixed with 5 mg calcium phosphate. The females were given a coded bottle of tablets that contained either 1.6 mg chromium tripicolinate (200 micrograms Cr+3) mixed with 5 mg calcium phosphate or 1.5 mg chromium dinicotinate (200 micrograms Cr+3) mixed with 5 mg calcium phosphate. The participants were instructed to take one capsule each day at breakfast. Lean body mass was determined prior to beginning the class and exercise regimen and again after 12 weeks. Lean body mass was calculated from total body resistivity measured as described by Segal et al. [171 with a four-terminal portable impedance analyzer (RJL Systems, Detroit, MI). Data are expressed as mean f S.E. Data were analyzed by analysis of variance and Duncan’s multiple range test.
RESULTS Plcollnate Complexes The precipitate formed with a 1:l ratio of chromium and picolinate (pick proved to be homogeneous when assayed with HPLC. A single, symmetrical peak was eluted after 2.92 min with this complex. The composition of this homogeneous complex was deduced from a combination of mass spectral analysis, melting point depression, and ultraviolet spectroscopy. Mass spectral analysis of this precipitate showed a large peak at 52 daltons (chromium ion) and a parent peak at 244 daltons which corresponds with 1 mole chromium coordinated with 1 mole picolinate and four moles’water. Melting point depression (AT = 7.66’0 yielded a molecular mass of 261 f 2 daltons which corresponds with 1 mole chromium, 1 mole picolinate, ,and 5 moles water. When 2.61 mg of this complex was dissolved in 1.0 liter deionized water, peak W absorbance was observed at 264 nm and the absorbance was 0.052 at pH -7.0. Peak absorbance of picolinate is found at 264 nm and the molar absorptivity of picolinate at 264 nm is 4020 L mol-’ cm-’ at pH 7.0 where the pyridine nitrogen would be unprotonated. At pH 3.0, where the pyridine nitrogen would be protonated, the molar absorptivity of picolinate at 264 nm is 7220 L mol-’ cm-‘. Thus, assuming that the molar absorptivity of picolinate coordinated with chromium would fall between 4020 L mol-’ cm-’ and 7220 L mol-’ cm-‘, the molecular formula of the precipitate
CHROMIUM
COORDINATION
COMPLEXES
181
formed with a 1:l ratio of chromium and picolinate was Cr pit (H,O),(OH),. (H,O). The solubility df this precipitated complex in pH 7.0 water was 11.1 PM. The precipitate formed with a 1:2 ratio of chromium and picolinate proved to be homogeneous when assayed with HPLC. A single, symmetrical peak was eluted after 3.59 min with this complex. The composition of this homogeneous complex was deduced from a combination of mass spectral analysis, melting point depression, and ultraviolet spectroscopy. Mass spectral analysis of this precipitate formed with a 1:2 ratio of chromium and showed a large peak at 52 daltons and a parent peak at 331 daltons which corresponds with 1 mole chromium coordinated with 2 moles picolinate and two moles water. Melting point depression (AT = 5.73”C) yielded a molecular mass of 349 f 3 daltons which corresponds with 1 mole chromium, 2 moles picolinate, and 3 moles water. When 3.49 mg of this complex was dissolved in 1.0 liter deionized water, peak UV absorbance was observed at 264 nm and the absorbance was 0.104 at pH 7.0. Thus, the molecular formula of the precipitate formed with a 1:2 ratio of chromium and picolinate was Cr(pic),(H,OXOH)(H,O). The solubility of this precipitated complex in pH 7.0 water was 14.1 PM. Neither the mono- nor the dipicolinate complex was active in the in vitro systems (Tables 1 and 2). The precipitate formed with a 1:3 ratio of chromium and picol,jnate also proved to be homogeneous when assayed with HPLC. A single, @metrical peak was eluted after 4.95 min with this complex. The composition of this homogeneous complex was deduced from a combination of mass spectral analysis, melting point depression, and ultraviolet spectroscopy. Mass spectral analysis of this precipitate showed a large peak at 52 daltons and a parent peak at 418 daltons which corresponds with 1 mole chromium coordinated with 3 moles picolinate. Melting point depression (AT = 4.78”C) yielded a molecular mass of 435 f 2 daltons which corresponds with 1 mole chromium, 3 moles picolinate, and 1 mole water. When 4.35 mg of this complex was dissolved in 1.0 liter deionized water, peak UV absorbance was observed at 264 nm and the absorbance was 0.155 at pH 7.0. Thus, the molecular formula of the precipitate formed with a 1:3 ratio of chromium and picolinate was Cr(pic),-H,O. The solubility of this crystalline complex in pH 7.0 water was 0.6 mM and 2.0 mM in chloroform. Molar absorptivity of this complex at 264 nm was 15546 L mol-’ cm-’ at both pH 3.0 and pH 7.0. As described previously, the molar absorptivity
TABLE 1. The Effect of Chromium Fyridine Carboxylates on Glucose Gxidation in Isolated Fat Cells Addition None Cr mono pit Cr di pit Cr tri pit Cr mono nit Cr di nit Cr chloride
Radioactivity Qm 219 f 34 225 f 45 229*37 231f49 218 f 54 623 f 147’ 243 f 29
Increase 1.02 1.05 1.06 1.0 2.84 1.11
Values are mean f S.E. %lue is significantlydifferent (P < 0.01) from other values in the column.
182 G. W. Evans and D. J. Pouchnik
TABLE 2. Uptake of Glucose Into Skeletal Muscle Cells Cultured With Various Forms of Chromium Addition None Cr mono pit Cr di pit Cr tri pit Cr mono nit Cr di nit Cr chloride PiC
Radioactivity Cpm 369 f 43 3% 408 629 354 367 382 375
f f f f + f f
54 52 6ga 56 45 53 67
Increase 1.07 1.10 1.70 1.0 1.0 1.03 1.01
Values are mean f SE. ‘Value is significantly different (P < 0.01) from other values in the column.
of picolinate at 264 nm is 4020 L mol-’ cm-’ at pH 7.0 but is 7220 L mol-’ cm-’ at pH 3.0. The molar absorptivity of the chromium tripicolinate complex was not altered by change in pH, suggesting that the pyridine nitrogen is involved in the coordination with chromium. The chromium tripicolinate complex was not active in the rat fat pad assay (Table 1). However, chromium tripicolinate added to rat skeletal muscle cultures produced a significant increase in the uptake of glucose (Table 2). Moreover, plasma glucose was significantly less (P < 0.01) in rats fed chromium tripicolinate and this chromium complex also prevented glycation of hemoglobin (Table 3). Chromium tripicolinate supplements resulted in a significant increase (P < 0.01) in lean body mass in both males and females (Table 4) participating in an aerobics class. Nicotinate Complexes The precipitate formed with a 1:l ratio of chromium and nicotinate (nit) proved to be homogeneous when assayed with HPLC. A single, symmetrical peak was eluted after 1.21 min with this complex. The composition of this homogeneous complex was deduced from a combination of mass spectral analysis, melting point depression, and ultraviolet spectroscopy. Mass spectral analysis of this precipitate showed a large peak at 52 daltons and a parent peak at 264 daltons which corresponds with 1 mole chromium coordinated with 1 mole nicotinate
TABLE 3. Plasma Glucose and Gylcated Hemoglobin in Aging Rats Fed Various Chromium Compounds Diet Supplement
Plasma Glucose mM
None
7.9 f 0.11
Cr tri pit Cr di nit Cr chloride
6.6 f 0.12a 7.7 f 0.12 7.8 f 0.11
HfJAI, (o/o) 5.46 3.31 5.19 5.39
f f f f
0.08 OLW 0.09 0.08
Values are mean f S.E. ‘Value is significantiy different (P < 0.01) from other values in the cohmul.
CHROMIUM
COORDINATION
COMPLEXES
183
TABLE 4. Lean Body Mass Development in Volunteers Given Chromium-Pyridine carboxylate supplements Supplement Females Initial resistance (Q) Initial LBM (kg) Final resistance (QI Final LBM (kg) Change in resistance (Q) Change in LBM (kg) Males Initial resistance (Q) Initial LBM (kg) Final resistance (Q) Final LBM (kg) Change in resistance (Q) Change in LBM (kg)
Cr pit
Cr nit
563 f 52 46 f 6 540*45 48 f 4 -24 f 5a + 1.8 f 0.3a
554 f 49 44*5 546 f 47 45 f 5 -7f2 +0.6 f 0.2
455 f 54 64*9 441 f 49 67 f 7 - 15 + 38 +2.1 f 0.3”
461 f 56 62 f 9 454 f 43 63 f 8 -5* 1 0.7 f 0.3
Values are mean f SE. aValue is significantly different (P < 0.01) from other values in the row.
and five moles water. Melting point depression (AT = 7.12”C) yielded a molecular mass of 281 f 2 daltons which corresponds with 1 mole chromium, 1 mole nicotinate, and 6 moles water. When 2.81 mg of this complex was dissolved in 1.0 liter deionized water, peak UV absorbance was observed at 262 nm and the absorbance was 0.031 at pH 7.0. Peak absorbance of nicotinate is found at 262 nm and the molar absorptivity of nicotinate at 262 nm is 3000 L mol-’ cm-’ at pH 6.0 where the pyridine nitrogen would be unprotonated. At pH 3.0, where the pyridine nitrogen would be protonated, the molar absorptivity of nicotinate at 262 nm is 5010 L mol-’ cm-‘. Thus, assuming that the molar absorptivity of nicotinate coordinated with chromium would fall between 3000 L mol-’ cm-i and 5010 L mol-’ cm-i, the molecular formula of the precipitate formed with a 1:l ratio of chromium and nicotinate was Cr nit (H,O),(OH),*(H,O). The solubility of this precipitated complex in pH 7.0 water was 10.9 PM. This complex showed no biological activity (Tables 1 and 2). The precipitates formed with either a 1:2 ratio of chromium and nicotinate or a 1:3 ratio of chromium and nicotinate proved to be homogeneous and identical when assayed with HPLC. A single, symmetrical peak was eluted after 2.32 min when the precipitate from either of these preparations was assayed. The composition of these two homogeneous complexes was deduced from a combination of mass spectral analysis, melting point depression, and ultraviolet spectroscopy. Mass spectral analysis of the precipitates from both of these preparations were identical and showed a large peak at 52 daltons with a parent peak at 369 daltons which corresponds with 1 mole chromium coordinated with 2 moles nicotinate and four moles water. Melting point depression (AT = 5.17”C) yielded a molecular mass of 387 + 2 daltons which corresponds with 1 mole chromium, 2 moles nicotinate, and 5 moles water. When 3.87 mg of this complex was dissolved in 1.0 liter deionized water, peak UV absorbance was observed at 262 nm and the absorbance was 0.060 at pH 6.0. From these observations, the
184
G. W Evans and D. J. Pouchnik
molecular formula of the precipitated complex formed with either a 1:2 or a 1:3 ratio of chromium and nicotinate was C&ric),(H,O),(OH)-H,O. The solubility of the dinicotinate complex in pH 7.0 water was 12.4 PM. The solubility of chromium dinicotinate in chloroform was 52 PM. When the dinicotinate complex was solubilized in 1.0 M HCl and subsequently adjusted to pH 3.0, the molar absorptivity of the complex at 262 nm was 10200 L mol-’ cm-‘. However, as described above, the molar absorptivity of the dinicotinate complex was 6000 L mall’ cm-’ at pH 6.0. These results indicate that the pyridine nitrogen of nicotinate is not involved in coordination of chromium since the molar absorptivity of nicotinate is 5000 L mol-’ cm-’ when the nitrogen is protonated and only 3000 L mol-’ cm-’ when the hydrogen ion dissociates. The chromium dinicotinate precipitate produced a significant increase in glucose oxidation when assayed in the fat pad assay system (Table 1). In contrast, this complex did not enhance glucose uptake into cultured skeletal muscle (Table 2). The precipitated chromium dinicotinate complexes were not effective in either of the in vivo assays (Tables 3 and 4). Chromium picolinate added to rat diets resulted in a decreased concentration of plasma glucose and a concomitant decrease in glycated hemoglobin but chromium nicotinate had no effect on these parameters. In addition, chromium nicotinate supplements given to humans enrolled in an aerobics class failed to produce a significant increase (P > 0.05) in development of lean body mass. DISC1JSSION
The results of our experiments prove that chromium coordination complexes with nicotinate or picolinate differ markedly in both chemical composition and biological action. Our results and those of Cooper et al. [4] provide substantial evidence that chromium is coordinated with nicotinate through the carboxyl oxygen. Green and Tong [18] showed that the molar absorptivity of nicotinate at 262 nm is approximately 5000 L mol-’ cm-’ when the pyridine nitrogen is protonated but decreases to 3000 L mol-’ cm-’ when the proton dissociates. As described above, when the hydrogen ion concentration was decreased in a solution that contained chromium dinicotinate, the molar absorptivity decreased. This observation proves that the pyridine nitrogen in the chromium nicotinate complex is not coordinated. The position of the carboxyl group in the nicotinate molecule obviously precludes formation of a five-membered ring which in turn apparently prevents complexation of chromium with three molecules nicotinate as found in the picolinate complex. Since nicotinate complexes with chromium are monodentate, the remaining coordination sites must be occupied by water or hydroxide unless molecules capable of replacing these are present in the medium. When coordinated with nicotinate and water, the complexes are insoluble at physiological pH. The insolubility at physiological pH arises from the dissociation of the pyridine nitrogens (pK = 4.8) and the conversion of the coordinated water to hydroxide (pK = 41. Once precipitated, chromium nicotinate complexes are extremely difficult to solubilize due to the presence of the coordinated hydroxides which form strong hydrogen bonds in the precipitate. This strong hydrogen bonding may account in part for the ineffectiveness of the chromium nicotinate complexes in the in vivo assays. Although ineffective in vivo, chromium dinicotinate potentiated glucose
CHROMIUM
COORDINATION
COMPLEXES
185
oxidation in isolated fat cells. Similar results have been obtained by other investigators [2-41 but the mechanism whereby chromium dinicotinate enhances glucose oxidation has not been elucidated. Mertz [2] has suggested that a ternary complex is formed with chromium, insulin, and insulin receptors. However, Cooper et al. [4] demonstrated that chromium dinicotinate enhances glucose oxidation in yeast where no insulin is present. Cooper et al. 141 have also presented evidence that the two pyridine nitrogen groups are somehow necessary for activity in the glucose oxidation assay system. The results of our experiments and those of other investigators indicate that picolinate forms bidentate complexes with chromium through the carboxyl oxygen and .the pyridine nitrogen [13, 141. We have recently confirmed this structure by use of x-ray crystallography. The bidentate coordination permits formation of a tripicolinate complex which is uncharged at all hydrogen ion concentrations. Chromium tripicolinate is slightly soluble in water (0.6 mM) but is about three times more soluble in chloroform (2 mM). In a previous series of experiments, we demonstrated that chromium picolinate added to synthetic liposomes increased membrane fluidity. In those experiments, we also discovered that chromium picolinate increased insulin internalization, glucose uptake, and leucine uptake into cultured skeletal muscle. Based on these observations, we have suggested that chromium picolinate potentiates insulin action by increasing membrane fluidity [ 191. In the study described in this paper, chromium picolinate supplements accelerated development of lean body mass in exercising volunteers while chromium nicotinate was ineffective. Several previous studies with both animals and humans prove that chromium picolinate promotes development of muscle with a concomitant loss of body fat. In two separate studies at this university, chromium picolinate supplements resulted in accelerated body fat loss and lean body mass development in males involved in weight-lifting programs [9]. Hasten et al. [lo] detected a significant increase in lean body mass in females and a significant increase in muscle circumference in both males and females given chromium picolinate supplements. Kaats et al. [ll] found significant body fat loss and development of lean body mass in sedentary individuals. During experiments with swine, Page et al. [12] discovered that chromium picolinate supplements in the diet resulted in decreased fat content and increased muscle in rib loins. These investigators also detected significantly lower cholesterol levels in swine supplemented with chromium picolinate [7]. In experiments with human volunteers, Press et al. [6] detected a significant decrease in total cholesterol and LDL-cholesterol during supplementation with chromium picolinate. At the same medical center, investigators detected a significant decrease in serum glucose and glycated hemoglobin in non-insulindependent diabetics [9]. In addition, a marked decrease in blood glucose was produced in native American volunteers taking chromium picolinate supplements 181. As described above, both serum glucose and glycated hemoglobin were significantly decreased in aging rats fed chromium picolinate from the time of weaning which confirms the in vivo effect of chromium picolinate on insulin regulated metabolic functions. Each of the metabolic parameters affected by chromium picolinate supplementation can be linked to the action of insulin. Insulin produces a wide variety of responses that ultimately affect carbohydrate and lipid and protein metabolism
186 G. W Evans and D. J. Pouchnik
[201. Previous studies with muscle cultures suggest that chromium picolinate may
affect the action of insulin through an effect on the rate of insulin internalization, which, by an unidentified mechanism, regulates the synthesis and/or insertion of insulin receptors into the plasma membrane [19]. Because of the lipophilic nature of chromium picolinate, the complex may affect insulin internalization by maintaining a level of membrane fluidity necessary ‘for efficient removal of the insulin-receptor complex from the surface of the plasma membrane. Alternatively, the picolinate complex with chromium may simply provide a delivery system enabling the chromium ion to traverse the,plasma membrane to reach the interior of the cell where it is utilized in some as yet unidentified biochemical pathway. Although the nicotinate and picolinate complexes enhanced insulin action through entirely different mechanisms, in each case chromium was complexed with a pyridine carboxylate. Chromium salts were not effective in either of the in vitro studies and the chromium salt failed to prevent elevated plasma glucose and protein glycation in aging rats. Chromium chloride added to the diets of swine had no effect on body fat or lean body mass [121 and several previous investigations demonstrate that ionic chromium has no effect on insulin action [2,3, 151. Although our results and those of other investigators demonstrate that chromium ion has no effect upon insulin action, positive results have been obtained following oral or intravenous administration of chromium salts [2, 211. Since chromium ion has no apparent function, these observations suggest that chromium can form biologically active complexes within mammalian cells under certain conditions. Several years ago researchers discovered that picolinate carboxylase, the enzyme involved in synthesis of picolinate, is much more active in diabetic animals [22]. This observation suggests that more picolinate is available for combination with chromium in humans and animals with impaired insulin sensitivity. In most, if not all, the successful experiments in which chromium salts were used, the subjects had some form of insulin resistance [2, 211. Chromium coordinates with picolinate readily at 37°C and though not extremely soluble in water is several times more soluble than chromium nicotinate. Chromium picolinate is also soluble in chloroform which suggests that this complex can pass through or function within biological membranes. The formation of chromium nicotinate requires heat, and once formed, these complexes are not extremely soluble in either water or lipid at physiological pH. As Mertz [2] has suggested, in order for a chromium dinicotinate complex to exist and be effective at physiological pH, the complex would have to contain groups that preclude chromium coordination with water. While exogenously synthesized chromium-pyridine carboxylate complexes influence insulin function, whether or not these types of chromium complexes are actually formed in mammalian cells remains to be determined.
REFERENCES 1. K. Schwarz and W. Mertz, Arch. Biochem. BiopIys. 85,292 (1959). 2. W. Mertz, Nutrition Revs. 33, 129 (1975). 3. E. W. Toepfer, W. Mertz, M. M. Polansb, E. E. Roginski, and W. R. Wolf, J. Agric. Food Chem. 25, 162 (1977).
CHROMIUM COORDINATION COMPLEXES
187
4. J. A. Cooper, B. F. Anderson, P. D. Buckley, and L. F. Blackwell, Inorganica Chimicu Actu. 91, 1 (1984). 5. P. Munnis, Bemidji State Uniuersity J Stuknt Res 7, 11 (1987). 6. R. I Press, J. Geller, and G. W. Evans, Western Journul of Medicine 152, 41 (1990). 7. T. G. Page, T. L. Ward, L. L. Southern, and D. L. Thompson, J. Animal Science Suppf. 2 69, 356 (1991). 8. G .W. Evans, Western Journal of Medicine 155,549 (1991). 9. G .W. Evans, Znt. J. Bios. and Med. Res. 11, 163(1989). 10. D. L. Hasten, E. P. Rome, and B. D. Franks, Am. College of Sports Med. Southeast Regional Chupt. Abstmcts 18, 49 (1991). 11. G. R. Kaats, J. A. Fisher, K. Blum, and J. A. Adelman, Am. Aging Assoc. Abstructs 21, 10 (1991). 12. T. G. Page, T. L. Ward, and L. L. Southern, J. Animal Science Suppl. Z 69, 356 (1991). 13. L. Campanella, E. Chiacchierini, G. De Angelis, and V. Petrone, Annali di Chimicu 67, 385 (1977). 14. S. Takata, E. Kyuno, and R. Tsuchiya, Bull Chem. Sot. Japan 41,3416 (1968). 15. R. A. Anderson, J. H. Brantner, and M. M. Polansky, J. A&c. Food Chem. 26,1219 (1978). 16. G. W. Evans and E. C. Johnson, J. Nutr. 111,68 (1981). 17. K. R. Segal, M. Van Loan, P. I. Fitzgerald, J. A. Hodgdon, and T. B. Van' Itallie, Am. J. Clin. Nutr. 47, 7 (1988). 18. R. W. Green and H. K. Tong, J. Am. Chem. Sot. 78,48% (1976). 19. G. W. Evans and T. D. Bowman, J. Inorgan. B&hem., 46, 243 (1992). 20. C. R. Kahn, Ann. Rev. Med. 36,429 (1985). 21. R. A. Anderson, Clin Physiol Biochem 4,31 (1986). 22. A. H. Mehler, E. G. McDaniel, and J. M. Hundley, J. Biol. Chem. 232, 323 (1958). Received March 5, 1992; accepted July 2, 1992