Article pubs.acs.org/JAFC
Validation and Application of an Improved Method for the Rapid Determination of Proline in Grape Berries Markus Rienth,†,# Charles Romieu,⊥ Rebecca Gregan,§ Caroline Walsh,§ Laurent Torregrosa,# and Mary T. Kelly*,⊗ †
Fondation Jean Poupelain, 30 rue du Gâte-Chien, 16100 F-Javrezac, France Montpellier SupAgro, UMR 1334, Amélioration et Génétique de l’Adaptation des Plantes, 2 place Pierre Viala, F-34060 Montpellier Cedex 02, France ⊥ INRA, UMR 1334, Amélioration et Génétique de l’Adaptation des Plantes, 2 place Pierre Viala, F-34060 Montpellier Cedex 02, France § School of Pharmacy, Trinity College, College Green, Dublin 2, Ireland ⊗ UMR 1083, Sciences pour l’oenologie, Faculté de Pharmacie, Université Montpellier 1, F-34093 Montpellier, France #
S Supporting Information *
ABSTRACT: A rapid and sensitive method is presented for the determination of proline in grape berries. Following acidification with formic acid, proline is derivatized by heating at 100 °C for 15 min with 3% ninhydrin in dimethyl sulfoxide, and the absorbance, which is stable for at least 60 min, is read at 520 nm. The method was statistically validated in the concentration range from 2.5 to 15 mg/L, giving a repeatability and intermediate precision of generally 60 g/L was such that proline concentrations were considerably underestimated in grape juice. The purpose of this study, therefore, was to revisit both methods to devise a simpler spectrophotometric method without the need for evaporation, as in the isatin method cited above, or the extraction step required with traditional ninhydrin methods. The final method developed is based on the ninhydrin reaction in strongly acidic conditions under which other amino acids do not interfere. Furthermore, relative volumes and the solvents were selected such that the need to extract the derived reaction product is obviated. The method was validated statistically, and results for 13 grape juice samples were compared with the HPLC−ninhydrin method for amino acid analysis. It was then applied to the determination of proline accumulation in microvine berries at all stages of development from fruit set to maturity.
■
on a linear regression curve in the concentration range of 2.5−15 mg/ L. Validation. Calibration standards at five concentration points (2.5, 5, 7.5, 10, and 15 mg/L) were prepared in diluted (1 in 5) synthetic grape juice spiked with the analyte mixture. Standard calibration curves were obtained from unweighted least-squares linear regression analysis of the data. The slope and intercept of the calibration graphs were determined through linear regression of the absorbance versus concentration plot. Individual absorbances were then interpolated on the calibration graphs to determine the found (back-calculated) concentrations. The quality of fit was determined using backcalculated-to-nominal concentrations, and the “lack of fit” test was used to confirm the linearity of the method. Within-day and betweenday precision and accuracy of the method were determined by carrying out replicate analyses of the calibration standards. Repeatability was determined by preparing and analyzing each calibration standard four times within a single day (i.e., 20 standards in total) under the same operating conditions. The intermediate precision was determined by carrying out the same operations over 4 days under different operating conditions (different reagents and materials). The precision was given by mean relative standard deviation of the absorbance values, and the accuracy of the method was evaluated as 100 × [mean found concentration/nominal concentration]. The method was crossvalidated by comparing the results obtained by this method with those obtained for the same samples by the classic HPLC post-column derivatization with ninhydrin method. Sugar and Organic Acid Analysis. Hexoses (glucose and fructose) and organic acids (malic, tartaric, citric, and shikimic acid) were determined in a single HPLC injection. The same (5-fold diluted) samples as were used for proline were further diluted 20-fold with 4.375 μM acetate as internal standard. To avoid potassium bitartrate precipitation, 1 mL of diluted sample was mixed with 0.18 g of Sigma Amberlite IR-120 Plus and agitated in a rotary shaker for at least 10 h before centrifugation (13000 rpm for 10 min). The supernatant was transferred into HPLC vials before injection on an Aminex HPX87H column eluted in isocratic conditions (0.05 mL min−1, 60 °C, H2SO4).25 Organic acids were detected at 210 nm with a Waters 2487 dual absorbance detector (Waters Corp., Milford, MA, USA). A refractive index detector Kontron 475 (Kontron Instruments, Rossdorf, Germany) was used to determine fructose and glucose concentrations. Concentrations were calculated according to a previously described method.26
MATERIALS AND METHODS
Reagents and Chemicals. All chemicals were of analytical grade or better, and freshly distilled water stored in a glass container was used for all experiments. Methanol, ethanol, dimethyl sulfoxide, glucose, and acetic, tartaric, malic, and formic acids were obtained from Carlo Erba, Val de Reuil, France. Individual amino acids (Ala, Asn, Arg, Asp, GABA, Glu, Glm, Ile, Leu, Lys, Met, Pro, Ser, Thr, Val) were obtained from Sigma-Aldrich (St Quentin-Fallavier, France). Isatin and ninhydrin were obtained from Fluka, France. Malvidin-3-glucoside was obtained from ExtraSynthèse, France. Solutions. A synthetic grape juice was prepared containing 200 g/ L glucose, 6 g/L malic acid, 8 g/L tartaric acid, and, for studies for the potential interference by anthocyanins, 250 mg/L malvidin 3glucoside. Stock solutions of 1 g/L of each of the amino acids were prepared in 0.1 M HCl and stored at −20 °C when not in use. Standard solutions of proline were prepared by serial dilution of the stock solution. A working mixture was prepared in water or synthetic grape solution containing typical concentrations of the principal grape amino acids: Ala, Asn, Glm, Ser, GABA, 100 mg/L; Arg, 500 mg/L; Asp and Glu, 150 mg/L; and 10 mg/L of the other amino acids. Plant Material. The ML1 microvine that was used in this study shows a dwarf stature and produces inflorescences all along the main axes in the place of tendrils. This phenotype has been proposed as a new model for research in grapevine genetics and ecophysiology.22−24 To study the proline accumulation pattern in berries of these plants, 10 plants were grown under controlled conditions in climate chambers under constant day/night temperatures (30/22 °C) for 3 months. All reproductive organs along the main axes from berry set to maturity were successively sampled within 2 h and immediately frozen in liquid nitrogen. Determination of Proline by the Ninhydrin Method. Ground sample powder was diluted 5-fold with deionized water and frozen at −20 °C until analysis. Reconstituted samples were thawed and centrifuged at 3000g at 4 °C for 10 min. The supernatant was further diluted (5−20 times depending on the stage of development); 750 μL of the diluted sample was added to the same volume of formic acid and vortex-mixed for 2 min. A 750 μL aliquot of ninhydrin 3% in dimethyl sulfoxide (freshly prepared daily) was then added, and following vortex mixing, the reaction mixture was heated at 100 °C in a heating block for 15 min. The absorbance at 520 nm of the resultant salmon pink reaction product was then read on a Biochrom (UK) Libra S12 spectrophotometer. Concentrations were calculated by interpolation
■
RESULTS AND DISCUSSION Method Development. Boctor27 demonstrated that the reaction between isatin and proline is suitable to quantitatively determine proline in a protein hydrolysate and biological fluids. It is a highly specific color reagent for proline, forming a blue derivative, pyrrole blue. The technique adopted in this study was as previously described.21 A 50 mg/L standard (100 μL) was diluted with citrate buffer (0.5 M, pH 4.1, 100 μL). Then 250 μL of a 0.075% (w/v) solution of isatin in acetone (250 μL) and ethanol (500 μL) was added. The tubes were evaporated to dryness in a heating block at 100 °C. The blue proline−isatin residue obtained was dissolved in aqueous 3 mL of acetone/water (2:1, v/v) and the absorbance of the resulting solution measured at 595 nm. This resulted in an absorbance reading of 0.37. Different permutations on the relative volumes and concentration of isatin (increasing to 0.3%) were investigated to obtain an increase in absorbance. However, the results proved inconclusive; it appeared that different absorbance values were obtained depending on the length of time it took the solvents to evaporate in the heating block, and it was difficult to obtain repeatable readings. Furthermore, when applied to real grape juice samples, absorbance reading were 10%) of proline. This observation is borne out in the literature.35 Other adsorbents such as PVPP were found to be less effective in removing the coloring matter or also removed some proline, although to a lesser extent than carbon black. Another approach adopted was to evaluate the extent of interference by anthocyanins in the assay as it had been observed that the pink-purple color of the sample before the addition of the ninhydrin and heating transformed into the characteristic salmon pink of the proline− ninhydrin adduct after heating (Supporting Information Supplemental Figure 1). Therefore, malvidin-3-glucoside was added to blank synthetic grape juice at a concentration of 250 mg/L; following 20-fold dilution and addition of formic acid, one sample was heated with ninhydrin in dimethyl sulfoxide and another with dimethyl sulfoxide only. The latter sample retained its color, whereas the color of the former was transformed to a pale yellow, the same color as is obtained with the blank (Figure 1). It therefore appears that ninhydrin reacts with anthocyanins in such a manner as to render them colorless. This reaction has not previously been reported in the literature and is currently under further investigation. Validation. Calibration standards at five concentration points (2.5, 5, 7.5, 10, and 15 mg/L) were prepared in diluted
Figure 2. Histogram of proline concentrations obtained by using the proposed method and by using amino acid analysis.
dilution of the reaction product, was fully validated and a systematic study of matrix interferences is reported. Interference Study. Different authors have reported various sources of interferences with the proline−ninhydrin assay, including high concentrations of sugars34 and diverse amino acids. Potential Interference from Sugars and Other Grape Juice Components. The sensitivity of the present method is such that the grape juice samples are diluted from 5- to 20-fold, resulting in final sugar concentrations of 10−40 g/L, concentrations that have not been reported to interfere with the assay. However, to verify this, a stock solution of 300 mg/L proline was prepared in synthetic grape juice (composition given under Materials and Methods). This solution was diluted 1 in 20, analyzed according to the developed method, and compared with an aqueous proline solution at 15 mg/L. Measured against blanks of synthetic grape juice or water, respectively, and subjected to the same reaction, the absorbances (n = 5) were not significantly different between the two groups. In all cases the blank (Supporting Information Supplemental Figure 1) is a pale yellow color with an absorbance of approximately 0.03. Potential Interference from Other Amino Acids. Certain basic amino acids, such as ornithine, were reported by Chinard29 to give colored products when heated with ninhydrin at low pH. With the exception of proline, the other commonly naturally occurring amino acids did not react significantly under
Figure 3. Proline and sugar accumulation in microvine berries. Lines are LOESS fits. 3387
dx.doi.org/10.1021/jf404627n | J. Agric. Food Chem. 2014, 62, 3384−3389
Journal of Agricultural and Food Chemistry
■
(1 in 20) synthetic grape juice spiked with the analyte mixture. The repeatability or intraday precision is presented in Table 1. As may be seen, the intraday precision varies from
Validation and Application of an Improved Method for the Rapid Determination of Proline in Grape Berries Markus Rienth,†,# Charles Romieu,⊥ Rebecca Gregan,§ Caroline Walsh,§ Laurent Torregrosa,# and Mary T. Kelly*,⊗ †
Fondation Jean Poupelain, 30 rue du Gâte-Chien, 16100 F-Javrezac, France Montpellier SupAgro, UMR 1334, Amélioration et Génétique de l’Adaptation des Plantes, 2 place Pierre Viala, F-34060 Montpellier Cedex 02, France ⊥ INRA, UMR 1334, Amélioration et Génétique de l’Adaptation des Plantes, 2 place Pierre Viala, F-34060 Montpellier Cedex 02, France § School of Pharmacy, Trinity College, College Green, Dublin 2, Ireland ⊗ UMR 1083, Sciences pour l’oenologie, Faculté de Pharmacie, Université Montpellier 1, F-34093 Montpellier, France #
S Supporting Information *
ABSTRACT: A rapid and sensitive method is presented for the determination of proline in grape berries. Following acidification with formic acid, proline is derivatized by heating at 100 °C for 15 min with 3% ninhydrin in dimethyl sulfoxide, and the absorbance, which is stable for at least 60 min, is read at 520 nm. The method was statistically validated in the concentration range from 2.5 to 15 mg/L, giving a repeatability and intermediate precision of generally 60 g/L was such that proline concentrations were considerably underestimated in grape juice. The purpose of this study, therefore, was to revisit both methods to devise a simpler spectrophotometric method without the need for evaporation, as in the isatin method cited above, or the extraction step required with traditional ninhydrin methods. The final method developed is based on the ninhydrin reaction in strongly acidic conditions under which other amino acids do not interfere. Furthermore, relative volumes and the solvents were selected such that the need to extract the derived reaction product is obviated. The method was validated statistically, and results for 13 grape juice samples were compared with the HPLC−ninhydrin method for amino acid analysis. It was then applied to the determination of proline accumulation in microvine berries at all stages of development from fruit set to maturity.
■
on a linear regression curve in the concentration range of 2.5−15 mg/ L. Validation. Calibration standards at five concentration points (2.5, 5, 7.5, 10, and 15 mg/L) were prepared in diluted (1 in 5) synthetic grape juice spiked with the analyte mixture. Standard calibration curves were obtained from unweighted least-squares linear regression analysis of the data. The slope and intercept of the calibration graphs were determined through linear regression of the absorbance versus concentration plot. Individual absorbances were then interpolated on the calibration graphs to determine the found (back-calculated) concentrations. The quality of fit was determined using backcalculated-to-nominal concentrations, and the “lack of fit” test was used to confirm the linearity of the method. Within-day and betweenday precision and accuracy of the method were determined by carrying out replicate analyses of the calibration standards. Repeatability was determined by preparing and analyzing each calibration standard four times within a single day (i.e., 20 standards in total) under the same operating conditions. The intermediate precision was determined by carrying out the same operations over 4 days under different operating conditions (different reagents and materials). The precision was given by mean relative standard deviation of the absorbance values, and the accuracy of the method was evaluated as 100 × [mean found concentration/nominal concentration]. The method was crossvalidated by comparing the results obtained by this method with those obtained for the same samples by the classic HPLC post-column derivatization with ninhydrin method. Sugar and Organic Acid Analysis. Hexoses (glucose and fructose) and organic acids (malic, tartaric, citric, and shikimic acid) were determined in a single HPLC injection. The same (5-fold diluted) samples as were used for proline were further diluted 20-fold with 4.375 μM acetate as internal standard. To avoid potassium bitartrate precipitation, 1 mL of diluted sample was mixed with 0.18 g of Sigma Amberlite IR-120 Plus and agitated in a rotary shaker for at least 10 h before centrifugation (13000 rpm for 10 min). The supernatant was transferred into HPLC vials before injection on an Aminex HPX87H column eluted in isocratic conditions (0.05 mL min−1, 60 °C, H2SO4).25 Organic acids were detected at 210 nm with a Waters 2487 dual absorbance detector (Waters Corp., Milford, MA, USA). A refractive index detector Kontron 475 (Kontron Instruments, Rossdorf, Germany) was used to determine fructose and glucose concentrations. Concentrations were calculated according to a previously described method.26
MATERIALS AND METHODS
Reagents and Chemicals. All chemicals were of analytical grade or better, and freshly distilled water stored in a glass container was used for all experiments. Methanol, ethanol, dimethyl sulfoxide, glucose, and acetic, tartaric, malic, and formic acids were obtained from Carlo Erba, Val de Reuil, France. Individual amino acids (Ala, Asn, Arg, Asp, GABA, Glu, Glm, Ile, Leu, Lys, Met, Pro, Ser, Thr, Val) were obtained from Sigma-Aldrich (St Quentin-Fallavier, France). Isatin and ninhydrin were obtained from Fluka, France. Malvidin-3-glucoside was obtained from ExtraSynthèse, France. Solutions. A synthetic grape juice was prepared containing 200 g/ L glucose, 6 g/L malic acid, 8 g/L tartaric acid, and, for studies for the potential interference by anthocyanins, 250 mg/L malvidin 3glucoside. Stock solutions of 1 g/L of each of the amino acids were prepared in 0.1 M HCl and stored at −20 °C when not in use. Standard solutions of proline were prepared by serial dilution of the stock solution. A working mixture was prepared in water or synthetic grape solution containing typical concentrations of the principal grape amino acids: Ala, Asn, Glm, Ser, GABA, 100 mg/L; Arg, 500 mg/L; Asp and Glu, 150 mg/L; and 10 mg/L of the other amino acids. Plant Material. The ML1 microvine that was used in this study shows a dwarf stature and produces inflorescences all along the main axes in the place of tendrils. This phenotype has been proposed as a new model for research in grapevine genetics and ecophysiology.22−24 To study the proline accumulation pattern in berries of these plants, 10 plants were grown under controlled conditions in climate chambers under constant day/night temperatures (30/22 °C) for 3 months. All reproductive organs along the main axes from berry set to maturity were successively sampled within 2 h and immediately frozen in liquid nitrogen. Determination of Proline by the Ninhydrin Method. Ground sample powder was diluted 5-fold with deionized water and frozen at −20 °C until analysis. Reconstituted samples were thawed and centrifuged at 3000g at 4 °C for 10 min. The supernatant was further diluted (5−20 times depending on the stage of development); 750 μL of the diluted sample was added to the same volume of formic acid and vortex-mixed for 2 min. A 750 μL aliquot of ninhydrin 3% in dimethyl sulfoxide (freshly prepared daily) was then added, and following vortex mixing, the reaction mixture was heated at 100 °C in a heating block for 15 min. The absorbance at 520 nm of the resultant salmon pink reaction product was then read on a Biochrom (UK) Libra S12 spectrophotometer. Concentrations were calculated by interpolation
■
RESULTS AND DISCUSSION Method Development. Boctor27 demonstrated that the reaction between isatin and proline is suitable to quantitatively determine proline in a protein hydrolysate and biological fluids. It is a highly specific color reagent for proline, forming a blue derivative, pyrrole blue. The technique adopted in this study was as previously described.21 A 50 mg/L standard (100 μL) was diluted with citrate buffer (0.5 M, pH 4.1, 100 μL). Then 250 μL of a 0.075% (w/v) solution of isatin in acetone (250 μL) and ethanol (500 μL) was added. The tubes were evaporated to dryness in a heating block at 100 °C. The blue proline−isatin residue obtained was dissolved in aqueous 3 mL of acetone/water (2:1, v/v) and the absorbance of the resulting solution measured at 595 nm. This resulted in an absorbance reading of 0.37. Different permutations on the relative volumes and concentration of isatin (increasing to 0.3%) were investigated to obtain an increase in absorbance. However, the results proved inconclusive; it appeared that different absorbance values were obtained depending on the length of time it took the solvents to evaporate in the heating block, and it was difficult to obtain repeatable readings. Furthermore, when applied to real grape juice samples, absorbance reading were 10%) of proline. This observation is borne out in the literature.35 Other adsorbents such as PVPP were found to be less effective in removing the coloring matter or also removed some proline, although to a lesser extent than carbon black. Another approach adopted was to evaluate the extent of interference by anthocyanins in the assay as it had been observed that the pink-purple color of the sample before the addition of the ninhydrin and heating transformed into the characteristic salmon pink of the proline− ninhydrin adduct after heating (Supporting Information Supplemental Figure 1). Therefore, malvidin-3-glucoside was added to blank synthetic grape juice at a concentration of 250 mg/L; following 20-fold dilution and addition of formic acid, one sample was heated with ninhydrin in dimethyl sulfoxide and another with dimethyl sulfoxide only. The latter sample retained its color, whereas the color of the former was transformed to a pale yellow, the same color as is obtained with the blank (Figure 1). It therefore appears that ninhydrin reacts with anthocyanins in such a manner as to render them colorless. This reaction has not previously been reported in the literature and is currently under further investigation. Validation. Calibration standards at five concentration points (2.5, 5, 7.5, 10, and 15 mg/L) were prepared in diluted
Figure 2. Histogram of proline concentrations obtained by using the proposed method and by using amino acid analysis.
dilution of the reaction product, was fully validated and a systematic study of matrix interferences is reported. Interference Study. Different authors have reported various sources of interferences with the proline−ninhydrin assay, including high concentrations of sugars34 and diverse amino acids. Potential Interference from Sugars and Other Grape Juice Components. The sensitivity of the present method is such that the grape juice samples are diluted from 5- to 20-fold, resulting in final sugar concentrations of 10−40 g/L, concentrations that have not been reported to interfere with the assay. However, to verify this, a stock solution of 300 mg/L proline was prepared in synthetic grape juice (composition given under Materials and Methods). This solution was diluted 1 in 20, analyzed according to the developed method, and compared with an aqueous proline solution at 15 mg/L. Measured against blanks of synthetic grape juice or water, respectively, and subjected to the same reaction, the absorbances (n = 5) were not significantly different between the two groups. In all cases the blank (Supporting Information Supplemental Figure 1) is a pale yellow color with an absorbance of approximately 0.03. Potential Interference from Other Amino Acids. Certain basic amino acids, such as ornithine, were reported by Chinard29 to give colored products when heated with ninhydrin at low pH. With the exception of proline, the other commonly naturally occurring amino acids did not react significantly under
Figure 3. Proline and sugar accumulation in microvine berries. Lines are LOESS fits. 3387
dx.doi.org/10.1021/jf404627n | J. Agric. Food Chem. 2014, 62, 3384−3389
Journal of Agricultural and Food Chemistry
■
(1 in 20) synthetic grape juice spiked with the analyte mixture. The repeatability or intraday precision is presented in Table 1. As may be seen, the intraday precision varies from
