Location of Phosphate Esters in a Wheat Starch Phosphate by 3 " P - Nuclear Magnetic Resonance Spectroscopy '

Cereal Chem. 70(2):145-152 Phosphate esters of D-glucose, methyl c-D-glucopyranoside, and on potato starch were confirmed to be the 6and 3-esters. The wheat maltose, as well as the a,,y-limit phosphodextrins from native potato starch phosphate, prepared by heating starch with sodium tripolystarch (degree of substitution [DS] 0.0033) and phosphorylated amylose phosphate under semidry conditions at an initial pH of 6, contained mainly (DS 0.016) served as model compounds in 3 P-nuclear magnetic resonance 6-monophosphate esters along with lower levels of 3and probably experiments to locate the phosphate groups on a phosphorylated wheat 2-monophosphates. The wheat starch phosphate contained orthophosstarch (DS 0.012). The a,,y-limit phosphodextrins were isolated by ionphate groups at the nonreducing ends of starch molecules, whereas exchange chromatography after exhaustive digestion of amylose or starch endogenous phosphate esters in potato starch occur on inner anhydrophosphates with Bacillus amyloliquifaciens a-amylase followed by glucose units. Aspergillus niger glucoamylase. The endogenous orthophosphate groups The phosphate ester groups on potato starch contribute to its high clarity and viscosity when it is cooked to a paste. Phosphorylation of wheat and corn starches increased paste clarity and consistency, but potato starch remained superior in those properties in spite of the fourfold higher level of phosphate esters on the modified wheat and corn starches and their low level of cross-linking (Lim and Seib, 1993). The location of endogenous phosphate esters on potato starch apparently differs from that on chemically phosphorylated starches. Gramera et al (1966) prepared corn starch phosphate with a degree of substitution (DS) of 0.016 (0.3% phosphorus [P]) by reacting corn starch at 150° C with sodium tripolyphosphate under unspecified conditions. The starch phosphate was subjected to Smith degradation followed by mild acid hydrolysis and anionexchange chromatography to give three main fractions. The components in the main fractions were investigated using paper chromatography, periodate analysis, and optical rotation. The results indicated that 6-, 2-, and 3-phosphates made up 63, 28, and 9% of the fractions, respectively. The location of phosphate esters on native potato starch has been examined extensively over the past 20 years by Hizukuri and his colleagues. Of the total P on potato starch (0.036-0.092%, or 1 atom of P per 209-532 anhydroglucose units [AGUs]), practically all was found on the amylopectin fraction (Abe et al 1982), with two thirds occurring as 6-phosphate, one third as 3-phosphate, and a trace as 2-phosphate (Hizukuri et al 1970, Tabata and Hizukuri 1971). The phosphate esters were not found on the nonreducing terminus of the a-1,4-linked unit chains in potato amylopectin (Tabata et al 1978), and 88% or more were on Bchains (Takeda and Hizukuri 1982). The phosphate esters were located a minimum of 9 AGUs inward from a branch point (Takeda and Hizukuri 1982). Glucoamylase that was free of both a-amylase and phosphatase was unable to bypass phosphorylated AGUs in potato starch (Abe et al 1982). Exhaustive digestion of a potato starch, which contained 1 P per 209 AGU, with glucoamylase gave 17% y-limit dextrin with 1 P per 36 AGU and an average unit-chain length of 14. The 'y-limit dextrin was of high molecular weight, as indicated by its high viscosity in water. Since the P in potato starch was concentrated in the 17% -y-limit dextrin, Abe et al (1982) concluded that either a few amylopectin molecules are highly phosphorylated or that phosphate groups are concentrated locally as a structural feature of the potato starch granule. 'Contribution 92-355-J from the Kansas Agricultural Experiment Station, Manhattan 66502. 2 Research associate, Department of Food Science and Human Nutrition, Iowa State University, Ames 50011. 3 Professor, Department of Grain Science and Industry, Kansas State University, Manhattan 66506. © 1993 American Association of Cereal Chemists, Inc. Recently, Muhrbeck and Tellier (1991) measured the 3 Pnuclear magnetic resonance (NMR) spectra of eight samples of potato starch dissolved in methyl sulfoxide. As the total-P level increased from 0.0153 to 0.0221% of amylopectin, the level of 3-P remained relatively constant (from 0.0044 to 0.0063%), whereas the 6-P level increased (from 0.0 104 to 0.0 165%). Furthermore, the crystallinity and enthalpy of gelatinization decreased with the level of 6-phosphorylation, but not with the level of 3-phosphorylation (Muhrbeck et al 1991). In this work, we used P-NMR spectroscopy on model P compounds to explore the position of phosphorylation on a wheat starch phosphate. MATERIALS AND METHODS Materials All chemicals were reagent grade unless otherwise stated. Potato starch (0.063% P), potato amylopectin, methyl a-D-glucopyranoside (MG), D-glucose, and D-glucose 6-phosphate were purchased from Sigma Chemical Co. (St. Louis, MO). Potato amylose, diphenyl phosphorochloridate, and platinum oxide were from Aldrich Chemical Company, Inc. (Milwaukee, WI). Wheat starch was Midsol 50 provided by Midwest Grain Products Co. (Atchison, KS). Crystalline Bacillus amyloliquifaciens a-amylase (Type IIA), a solution of Aspergillus niger glucoamylase with glucose and preservative, and bovine intestinal alkaline phosphatase (Type I-S) were from Sigma Chemical Co. The enzyme activity was 930 units [U]/mg for a-amylase, 6,100 U/ml for glucoamylase, and 6.8 U/mg for phosphatase. One unit of a-amylase liberated 1 mg of maltose from starch in 3 min (1.0 ,umol/min) at pH 6.9 and 200C; 1 U of glucoamylase released 1 mg of glucose from starch in 3 min (1.9 ,umol/min) at pH 4.5 and 55°C; and 1 U of phosphatase hydrolyzed 1 ,umol/min of p-nitrophenol phosphate at pH 10.4 and 370C. The a-amylase and glucoamylase were found to have no phosphatase activity according to the following methods. D-Glucose 6-phosphate (0.1 g) was mixed with 0.01M acetate buffer (pH 4.5, 6 ml) and glucoamylase (0.1 ml). After 1 hr at 400C, the mixture was adjusted to pH 8, and its P-NMR spectrum was examined. No increase was observed in the intensity of the signal for orthophosphate. It was estimated that 5% hydrolysis of D-glucose 6-phosphate would increase the orthophosphate concentration by 3 mM so that the orthophosphate would be readily detectable. a-Amylase (1 mg) was added to a solution of D-glucose 6-phosphate in water at pH 6-7. After digestion for 1 hr at 700C, the mixture was adjusted to pH 8.0, and its 3 1 P-NMR spectrum was examined. The spectrum showed no increase in the intensity of the orthophosphate signal. The commercial glucoamylase solution was found to contain a trace of inorganic phosphate. Vol. 70, No. 2,1993 145 General Methods Total carbohydrate was measured by the phenol-sulfuric acid method (Dubois et al 1956) and reducing end groups by the modified Park-Johnson method (Hizukuri et al 1981), using D-glucose as the reference standard. Phosphorus in starch and dextrins was measured by the method of Smith and Caruso (1964). NMR spectra were recorded on a Bruker WM400 NMR spectrometer (Bruker Instruments, Inc., Billerica, MA). 3 P-NMR spectra were measured at 162 MHz on aqueous solutions at pH 8.0 ±0.1; the solutions contained 0.02M ethylenediaminetetraacetate to sharpen the signals. Chemical shifts are reported in parts per million from orthophosphate (sodium salt) as internal reference standard. ' C and H-NMR spectra were recorded at 100.6 and 400 MHz, respectively, in CDC13 or D20 solution, and chemical shifts (6) are reported (in ppm) from tetramethylsilane (6 0.0 ppm). Methanol (49.6 ppm) was used as the internal reference for 'C-NMR spectra. Phosphorylation of Amylose and Wheat Starch Wheat starch was phosphorylated at an initial pH of 6 using 5% sodium tripolyphosphate (STPP) in the presence of 5% sodium sulfate as described by Lim and Seib (1993). The P content of the modified wheat starch (0.28%) was corrected for endogenous P in its associated lipids (0.05%). The net 0.23% phosphorus was equivalent to DS 0.012. Potato amylose was phosphorylated as follows. Amylose (5 g) was dissolved in 0.3M sodium hydroxide (200 ml), and STPP (0.5 g) was added, followed by 2M hydrochloric acid to bring the mixture to pH 7. Ethanol (1,200 ml) was added slowly with stirring, and the mixture was kept at 4 C overnight. The precipitate was collected by centrifugation and vacuum-dried at room temperature over calcium sulfate until its moisture level was below 10%. The amylose-STPP mixture was heated at 130'C for 3 hr; then the phosphorylated amylose was separated from contaminating salts by dialysis. The retained material was isolated by precipitation with ethanol and dried in a vacuum desiccator. The yield of amylose phosphate, which contained 0.30% P (DS 0.016), was 3.1 g. D-Glucose 2and 3-Phosphates and Methyl a-D-Glucopyranoside 2and 6-Phosphates D-Glucose 2and 3-phosphates were prepared according to Farrar (1949) and Brown et al (1957), respectively. The esters were converted to their cyclohexylammonium salts using a strongly acidic cation-exchange resin in the cyclohexylammonium form. The cyclohexylammonium salt of a-D-gluCopyranose 3-phosphate (mp 1310C) crystallized from ethanol, whereas the 2-phosphate remained a syrup. Methyl 4,6-O-benzylidene-a-D-glucopyranoside (1 g, mp 168-170'C) (Fletcher 1963) was reacted at 0 C with diphenyl phosphorochloridate (1.2 equivalents) in dry pyridine (2 ml) (Ballou and MacDonald 1963). After 2 hr, water (0.05 ml) was added, and the mixture was evaporated to a syrup. The syrup was dissolved in chloroform (50 ml), and the organic layer was washed repeatedly with water. Vacuum evaporation of the chloroform solution gave syrupy diphenylphosphate ester, from which the last traces of pyridine were removed by high vacuum at 250 C. The phenyl and 4,6-acetal blocking groups were removed by hydrogenation over platinum oxide (0.05 g) at 1 atm in methanolic solution (50 ml). After removal of the catalyst by filtration, the solution was evaporated to a syrup that was dissolved in water (50 ml). The solution was adjusted to pH 8.0 with 0.1M NaOH, and an aliquot (10 ml) of the solution was applied to a column (10 X 200 mm) of strongly basic anion exchange resin (AG1-X8, Bio-Rad Laboratories) in the bicarbonate form. The column was washed with water (500 ml) and developed with 0.4M ammonium bicarbonate. Fractions (5 ml each) were collected and assayed for carbohydrate, and the desired fractions were combined. Ammonium bicarbonate was removed by repeated addition of water and evaporation to a syrup. The residual MG 2-phosphate ammonium salt, which contained a trace of the 3-phosphate ester, was identified by 1 Cand 3 p146 CEREAL CHEMISTRY NMR. MG (1.0 g) was phosphorylated using the same procedure described above. Anion-exchange purification gave an amorphous solid containing predominately MG 6-phosphate mixed with low levels of the 3-phosphate ester and a trace of the 2-phosphate, as shown by 3 Pand ' 3C-NMR. Maltose 6-Phosphate Phenyl 4-0-(4',6'-O-benzylidene-a-D-glUCopyranosyl)-,-Dglucopyranoside (PBG) was synthesized according to the procedure of Takeo et al (1974). PBG was purified on a column of silica gel using chloroform as the developing solvent. The 4',6'acetal was crystallized from acetone-petroleum ether and had a melting point of 139-143C. PBG (0.5 g) was reacted at 25 C with diphenyl phosphorochloridate (1.5 equivalent) in pyridine (5 ml). After reaction for 8 hr, the product was isolated in the usual manner, with care taken to remove all traces of pyridine. The syrupy product was hydrogenated to give a syrup that was purified on a strongly basic anion-exchange column, as described for MG 2-phosphate. Two components were resolved on the column upon eluting with 0.4M ammonium bicarbonate. The first component (30 mg) was maltose 6-phosphate, as determined by 'Cand P-NMR spectroscopy. The second component (32 mg) was cyclohexyl P3-maltopyranoside 6-phosphate. Its 31 P-NMR spectrum showed a singlet at 1.05 ppm, and its 13C-NMR spectrum showed two anomeric signals at 99.2 and 100.6 ppm, two primary carbons at 61.6 and 64.2 ppm, and five cyclohexyl signals at 33.5, 32.0, 25.6, 24.5, and 24.3 ppm. No attempts were made to isolate the ammonium salts of the two derivatives of maltose. Their yields were determined by total carbohydrate assay. a-Limit Dextrins Ten grams of native potato starch or phosphorylated wheat starch (DS 0.012) was dispersed in distilled water (45 ml) containing 2 mM calcium chloride (2.5 ml), and the pH was adjusted to 6.5 with O.1M hydrochloric acid or sodium hydroxide. For phosphorylated amylose (DS 0.016), the polymer (5 g) was dissolved in 0.4M sodium hydroxide (45 ml) containing 2 mM calcium chloride (2.5 ml), and the pH was adjusted to 6.5 with 3M hydrochloric acid. One milliliter of 0.2% aqueous a-amylase (2 mg) was added to the aqueous dispersion of potato starch, wheat starch phosphate, or amylose phosphate. Each mixture was heated with stirring in a boiling-water bath for 10 min and then cooled to 25°C. The pH of a mixture was readjusted to 6.5; an additional 3 ml of a-amylase (3 mg) was added; and the digest was warmed to 700C. The total a-amylase amounted to 465 U/g of starch and 930 U/g of amylose phosphate. After 1 hr at 70°C, the reducing power of the three hydrolysates remained constant, indicating formation of a-limit dextrins. An aliquot (25 ml) was removed from each digest, and a-amylase was inactivated by heating in a boiling-water bath for 15 min. Each mixture was centrifuged at 3,000 X g for 10 min, and any precipitate was removed. The precipitate from the digest of potato starch or wheat starch phosphate contained 3-5% of the carbohydrate and 3-4% of the P present in the a-amylase digests. The precipitate in the digest of amylose phosphate amounted to 24% of original carbohydrate; its P content was not measured. The average degree of polymerization (DP) of the a-limit dextrins in the supernatant solutions was 4.2 for potato starch, 2.0 for phosphorylated amylose, and 4.0 for phosphorylated wheat starch. Each solution was evaporated under reduced pressure to approximately 10 ml, and a portion (5 ml) of the concentrate was mixed with deuterium oxide (2 ml), ethylenediaminetetraacetate dihydrate disodium salt (52 mg), 0.2% sodium azide (70 ,ul), and 1M sodium phosphate buffer (pH 8.0, 50 ,ll). Before measurement of the P-NMR spectra, each mixture was adjusted to pH 8.0 ± 0.1 (measured with a pH meter) by addition of 0.5M sodium hydroxide. a,,y-Limit Phosphodextrin The remaining solution (-25 ml) from an a-limit hydrolysate, which contained approximately 5 g of carbohydrate from potato starch or wheat starch phosphate and 2.5 g from amylose phosphate, was adjusted to pH 4.5 by addition of IM acetic acid. Sodium acetate buffer (0.01 M, pH 4.5, 25 ml) was added, followed by glucoamylase (1 ml, 6,100 U), and digestion was done at 40°C for 1 hr, after which the average DP remained constant at 1.5-2.0. Each hydrolysate was adjusted to pH 7 using 0.5M sodium hydroxide, and the mixture was heated in a boiling-water bath for 15 min. The precipitated solids were removed by centrifugation (3,000 X g for 10 min) and assayed for P. The precipitate from potato starch or wheat starch phosphate contained less than 5% of the initial carbohydrate and P, while that from amylose phosphate amounted to 27% of the starting carbohydrate. P content in the precipitate from amylose phosphate was not measured. After the supernatant was adjusted to pH 7.5 with 0.5M sodium hydroxide, the mixture was added to a column (10 X 200 mm) of strongly basic anion exchange resin (BioRad AG1-X8) in the bicarbonate form. The column was washed with water (500 ml) to remove glucose, and the phosphorylated maltooligosaccharides (a,,y-phosphodextrins) were eluted using 0.4M ammonium bicarbonate. The fractions (5 ml each) testing positive for carbohydrate and P were combined, and ammonium bicarbonate was removed by four successive additions of water followed by evaporation at reduced pressure. The yields of a,-y-phosphodextrins were approximately 50, 210, and 65 mg from 5 g of potato starch, 5 g of phosphorylated wheat starch, and 1.8 g of phosphorylated amylose, respectively. Phosphorus recovered in the a,-y-phosphodextrins was approximately 90% of the initial P in potato starch or wheat starch phosphate. The average DPs of the a,-y-phosphodextrins were 4.4 from native potato starch, 4.5 from wheat starch phosphate, and 4.9 from amylose phosphate. 31 P-NMR measurements were made on the phosphodextrins according to the procedure described for the a-limit dextrins. One half of the solution (-7 ml) that had been used to record the p spectrum of the a,-y-phosphodextrin from potato starch was evaporated under vacuum to a small volume. The concentrate was dissolved in D20 (5 ml), and the solvent exchange was repeated. The residue was dissolved in D2 0 (2 ml), and its 'HNMR spectrum was recorded. The solution was then adjusted to pH 10.4 with IM NaOD, and alkaline phosphatase (30 mg) was added. The digest was held overnight at 250C; its pH was adjusted to 7 with DCl; and the 'H-NMR spectrum was recorded. RESULTS AND DISCUSSION Phosphorylated Model Compounds Phosphate esters of MG served as model compounds for phosphorylated nonreducing ends of a-1,4-linked AGUs in starch, whereas maltose 6-phosphate represented the reducing ends or inner AGUs. Phosphorylated limit dextrins from potato starch and potato amylopectin also served as model substances. The structures of two model substances are shown in Figure 1. Synthesis of Model Compounds Glucose 2and 3-phosphates were prepared by well-established methods, whereas the 2-, 3-, and 6-phosphate esters of MG were prepared here for the first time. Methyl 4,6-0-benzylidene-a-Dglucopyranoside was reacted with diphenyl phosphorochloridate in pyridine, and the intermediate product was catalytically reduced Non-Reducing End Reducing End and Inner AGU's