- Email: [email protected]
Isolation and quantitative determination of some cardioactive glycosides from Digitalis lanata by high-performance liquid chromatography
ANALYTICAL
BIOCHEMISTRY
156.
171-175
(1986)
Isolation and Quantitative Determination of Some Cardioactive Glycosides from Digitalis lanata by High-Performance Liquid Chromatography FERENC OROSZ, MIMI Institute
c?f Enqmology.
Biologicul Budapest
NURIDSANY,
Rexarch Center. t-I-150?. P.O. Bo.r
AND JUDIT OV,&DI Hungarian 7. ffungurj
Acudem~,
($ Sciences,
Received July 23, 1985 A rapid extraction method followed by high-performance liquid chromatographic assay was developed for the quantitative determination of the cardioactive glycosides of Digitalis lanata. The leaf samples were extracted with water or aqueous alcohols. The simple extraction method gives a better yield than the methods described previously. Lanatoside C and its metabolites have been separated on a reversed-phase column with various mixtures of acetonitrile, methanol. and water as mobile phases for isocratic elution. Extraction and quantitative determination of lanatoside C and digoxin from a leaf sample require not more than 30 min. GJ 1986 Academx Press, Inc. KEY WORDS: lanatoside C; digoxin; extraction: quantitative determination; high-performance liquid chromatography.
The digitalis glycosidesare important pharmacologically active substancesfor the treatment of cardiac insufficiency. Since the economic synthesis of these compounds has not been resolved, the separation from leaves and quantitative determination of primary and secondary glycosides possessing different pharmacological effects are needed. Analytical methods in routine use are based on paper, thin-layer, and gas-liquid chromatographic procedures reviewed by Page( 1). However, all of thesemethods are very time consuming and quantitation is difficult. Recently, HPLC procedures have been introduced to separatevarious mixtures of crystallized glycosides (2,3). In fact, the methods are suitable for the separation of only certain groups of cardenolide series;e.g., digitalis glycosideswith high or low polarities cannot be separated in one system (4-6). Although the separation and determination of glycosides from solutions of crystallized glycosides have been resolved relatively well, few attempts have been made at the separation from extracts of Digitalis plants. Wichtl and co-workers (7,8) have carried out detailed chromatographic analysis of leaf ex171
tracts in various solvent systems.This method gives reliable information about the composition of mixtures of cardiac glycosides; however, the pharmaceutically most important glycosides,lanatoside C and digoxin, were not separated. Moreover, the extraction used was cumbersome. Development of a simple extraction method followed by HPLC assaywith isocratic elution suitable for the quantitative determination of the main active substances of Digitalis lanata is the subject of the present paper. MATERIALS
AND
METHODS
The HPLC equipment used in thesestudies was a Liquopump OE 3 12 (Labor MIM Budapest, Hungary). Samples were introduced onto the column through a 20-~1 Rheodyne 7010 injector (Cotati, Calif.). A Labor MIM OE 3 12 ultraviolet spectrophotometer was equipped with a IO-p1 flow cell. Detection wavelength was 230 nm. The 250 X 4.6-mm reversed-phasecolumns (C,,). lo-pm particle size, were obtained from Chromatronix (LiChrosorb, Dimesil). 0003-2697/86
$3.00
Copyright &i 1986 by Academic Press, Inc. All nghfs of reproduction m any form resewed.
172
OROSZ.
NURIDSzkNY.
AND
OV/kDI
The solvents were HPLC-grade Merck products. Cardiac glycosides were purchased from Boehringer (Mannheim, FRG) and Richter G. (Budapest, Hungary). D. lanata leaveswere freshly collected from plants before the flowering stage. RESULTS
AND
DISCUSSION
Separation and quantitation of crystallized lanatoside C and its metabolites. The chromatogram shown in Fig. 1 represents a fast isocratic separation that can be achieved for some active digitalis glycosides on a reversedphase (RP)’ column with acetonitrile:water (2:3). Quantitation evaluation of the peaks is possible since a resolution factor of 2.0 for lanatoside C and digoxin is found. It can be also seenthat the isomers, 01-and P-acetyldigoxin, are separated, although they are very similar in chemical and physical properties. A slight alteration in the composition of the mobile phase results in a dramatic change in the value of the capacity factor, k’. The increase in the polar component of the eluant results in larger k’ values. Therefore, if the separation or isolation of lanatoside C and its metabolites is the aim, it is advantageous to increase the concentration of water in the eluant (Table 1). By decreasingthe polarity of the solvent (i.e., by adding methanol or acetonitrile) the capacity factor of lanatoside A and digitoxin can be significantly decreased.In addition the glycosidesinvestigated could be separated on a RP C-18 column with an isocratic solvent mixture of ethanol:acetonitrile:water (10:45:45) within 8 min (Table 1). It should be noted that the selectivity coefficients (ratio of k’) of the compounds of interest depend on the different packing materials ascalculated from Table 1, lines 5 and 6. Nevertheless, we have found that the differences in the retention times of lanatoside C and digoxin do not change, or only slightly change, with the polarity of the eluants. This may be explained by an increase in polarity due to deacetylation of lanatoside ’ Abbreviation
used: RP, reversed
phase.
: +-15
10 RETEMION
5 TIME (MIN)
FIG. I. Isocratic separation of digitalis glycosides. Sequence of elution: (I) deacetyl lanatoside C: (2) lanatoside C; (3) digoxin; (4) ol-acetyldigoxin; (5) fi-acetyldigoxin: (6) lanatoside A. Column, LiChrosorb RP C- 18: solvent system, water:acetonitrile (3:2): flow rate, 0.86 ml/min.
C, which is proportional to the decreasein polarity due to deglucosylation of lanatoside C. For the separation of these glycosidesit is convenient to selecta system which ensuresrapid separation (cf. Fig 1). In this manner it is possible to achieve rather good resolution. The increasein retention times is parallel with the deterioration of resolution. Various amounts of lanatoside C, digoxin, and their intermediate products were subjected to chromatography using the isocratic system. The calibration curves were drawn according to the peak heights from five chromatograms. The peak height of each glycoside was found to be reproducible with a relative standard error lessthan 1%. Straight lines were obtained by fitting the experimental points in the range
EXTRACTION
mr40000 w-lco*m* rj A ,’ 6 6 V
AND
ANALYSIS
o--m--6 ,’ 6 6
0
OF
DIGITALIS
GLYCOSIDES
173
of 30 to 1000 ng per injection for the compounds investigated. Regression lines were calculated by the method of least squares; the coordinate (0, 0) was not used in the calculation and hence extrapolation of each graph toward the origin may be considered to indicate the overall accuracy of the calibration. Equations for the regression lines are presented in Table 2 in the form Y = mX + c, where XI is nanograms of glycoside and Y is the optical density at 230 nm. Extraction. Glycosides were extracted from D. lanata leaves by maceration in water or in 50% aqueous methanol (or ethanol) for 3 min followed by intensive shaking (Vortex) for 1 min. The extract content was 10% dry wt. In the presence of alcohol the enzymatic conversion of primary into secondary glycosides is prevented. After centrifugation at 60001: the supernatant of the aqueous alcohol extract, being free of proteins, can be directly injected onto the column. Up to approximately 200 injections the parameters of the columns were not changed. If the leaves were pretreated to prevent the enzymatic conversion of primary glycosides (e.g., heat or microwave treatment). extraction could be carried out in water. This was followed by filtration on an Amicon PM 10 filter to remove the compounds with molecular weights higher than 10,000. The reliability and efficiency of extraction procedures will be demonstrated later. The fact that the extracts can be directly applied to HPLC analysis without purification makes the procedure convenient for screening of leaf samples and for simultaneous determination of lanatoside C and digoxin irrespective ofthe total glycoside composition.
Determination of lanatmide C in the e.xtract yf‘Digitalis lanata. To quantify the content of lanatoside C in the extract of D. lanata its complete separation from the other chromophore should be fulfilled. For this purpose two systems, acetonitrile:water (1:2) and methanol: water (3:2), were chosen. As shown in Fig. 2 in and near the exclusion volume of the LiChrosorb column a dominant part of the chromophoric compound is eluted with a
174
OROSZ,
NURIDSANY.
AND
TABLE STATISTICAL
Substances
DATA
OVADI
2
OF CALIBRATION
CURVES
Correlation coefficient
Equation
Deacetyl lanatoside Lanatoside C Digoxin a-Acetyldigoxin
C
Y = 1.1 1 X IO-“ .Y + 4.88 X 10m4 Y = 6.22 x 10-5,Y+ 6.92 x 1om4
0.999
Y = 6.07
X 10m5 A'+
3.28
x 10m4
0.999 0.999
Y = 3.04
X 10m5 A'+
8.91
x
0.999
Note. Chromatographic runs were carried system as the mobile phase.
out on a LiChrosorb
mixture of acetonitrile:water as mobile phase, and lanatoside C can be separated from these polar substances. Moreover, the k’ value for lanatoside C in Table 1 is 1.5 which is below the optimum value of 2 for chromatographic
4015
IO-“
RP C- 18 column
+
+
i
+
-+
10
RETENTION
FIG. 2. Chromatogram
(2: I) solvent
. 10
+
+
.
5
-~+
5
TIME
4.809
using water:acetonitrile
+-~+-“++
15 td~f~t-*--t
7.980 30.506 22.269
methods but within the minimum value of 1 (9). The amount of lanatoside C in the extract. calculated on the basis of the equation in Table 2, was found to be 0.16 -t 0.0 1% of the dry leaf. This value is significantly higher than the value we determined for the extract prepared as described by Potter, 0.11 k 0.0 1% of the dry leaf ( IO).
i ~*
F’ test
(MIN)
of Digitalis lamta extract (extracted with methanol). Peak I. lanatoside C: column, LiChrosorb RP C-l 8: solvent system. water:acetonitrile (2:l): flow rate, 0.86 ml/min.
RETENTION
TIME ,M,N)
FIG. 3. Chromatogram of enzyme-treated Digitalis lanwith 50% aqueous methanol). Peaks identified: ( 1) deacetyl lanatoside C: (2) lanatoside C: (3) digoxin: (4) a-acetyldigoxin: (5) /3-acetyldigoxin. The system was the same as described in Fig. 2. ata (extracted
EXTRACTION
AND
ANALYSIS
Similar analyses have been carried out on Dimesil columns with a mobile phase of methanoI:water (32). Although the capacity factors for Ianatoside C in the two HPLC systems are different (cf. Table l), the same lanatoside C content is found. This indicatesthat the peak is homogeneous. As we already mentioned. simultaneous determination of lanatoside C and digoxin is difficult due to their close retention times. We have overcome this difficulty. As shown in Fig. 3 the extract in which lanatoside C was enzymatically partially converted to its metabelites contains both the primary and secondary glycosides. Their separation by the chromatographic system used was satisfactory. The present method provides a rapid and effective procedure, including extraction. for quantitation of the main glycosides.Ianatoside C and digoxin, of D. lanata.
OF
DlGIl-ALIS
175
GLYCOSIDES
ACKNOWLEDGMENT The columns and the injection valve vided by Chromatronix Inc.. California.
were kindly
pro-
REFERENCES I. 2. 3. 4. 5. 6. 7. 8. 9.
IO.
Page. D. P. (July 1974) FDA p,“J LIlIL.2 ’ in-c a. 6 Castle, M. ( 1. ( 1975) J. c/wow ralojy. 115, 437-446. Evans, F. J. (1974) J. C%romaro.~r. 88, 41 l-412. Lindner. W.. and Frei. R. W. (1976) J Chnvw~ow. 117.81-86. Emi, F.. and Frei. R. W. ( 1977) .J. C’hromurc~,gr. 130, 169-180. Fujii. Y.. Fukuda. H.. Saito. Y.. and Yamaraki. M. ( 1980) .J. C/mvncl~o~r. 202, 139- 143. Wichtl. M.. Mangkudidjojo. M., and Wichtl-Bleyer, W. (1981) J. C‘lrrormrogr. 243, 503-508. Wichtl. M.. Wichtl-Bleyer. W.. and Manghudidjojo. M. ( 1981) J C%rornarcyq 247, 359-365. Johnson. E. L., and Stevenson. R. (1978) Basic Liquid Chromatography. Varian Associates, Palo Alto, Calif. Petter. H. ( 1963) Phurm~:ic~ 18, 554-56 I.
BIOCHEMISTRY
156.
171-175
(1986)
Isolation and Quantitative Determination of Some Cardioactive Glycosides from Digitalis lanata by High-Performance Liquid Chromatography FERENC OROSZ, MIMI Institute
c?f Enqmology.
Biologicul Budapest
NURIDSANY,
Rexarch Center. t-I-150?. P.O. Bo.r
AND JUDIT OV,&DI Hungarian 7. ffungurj
Acudem~,
($ Sciences,
Received July 23, 1985 A rapid extraction method followed by high-performance liquid chromatographic assay was developed for the quantitative determination of the cardioactive glycosides of Digitalis lanata. The leaf samples were extracted with water or aqueous alcohols. The simple extraction method gives a better yield than the methods described previously. Lanatoside C and its metabolites have been separated on a reversed-phase column with various mixtures of acetonitrile, methanol. and water as mobile phases for isocratic elution. Extraction and quantitative determination of lanatoside C and digoxin from a leaf sample require not more than 30 min. GJ 1986 Academx Press, Inc. KEY WORDS: lanatoside C; digoxin; extraction: quantitative determination; high-performance liquid chromatography.
The digitalis glycosidesare important pharmacologically active substancesfor the treatment of cardiac insufficiency. Since the economic synthesis of these compounds has not been resolved, the separation from leaves and quantitative determination of primary and secondary glycosides possessing different pharmacological effects are needed. Analytical methods in routine use are based on paper, thin-layer, and gas-liquid chromatographic procedures reviewed by Page( 1). However, all of thesemethods are very time consuming and quantitation is difficult. Recently, HPLC procedures have been introduced to separatevarious mixtures of crystallized glycosides (2,3). In fact, the methods are suitable for the separation of only certain groups of cardenolide series;e.g., digitalis glycosideswith high or low polarities cannot be separated in one system (4-6). Although the separation and determination of glycosides from solutions of crystallized glycosides have been resolved relatively well, few attempts have been made at the separation from extracts of Digitalis plants. Wichtl and co-workers (7,8) have carried out detailed chromatographic analysis of leaf ex171
tracts in various solvent systems.This method gives reliable information about the composition of mixtures of cardiac glycosides; however, the pharmaceutically most important glycosides,lanatoside C and digoxin, were not separated. Moreover, the extraction used was cumbersome. Development of a simple extraction method followed by HPLC assaywith isocratic elution suitable for the quantitative determination of the main active substances of Digitalis lanata is the subject of the present paper. MATERIALS
AND
METHODS
The HPLC equipment used in thesestudies was a Liquopump OE 3 12 (Labor MIM Budapest, Hungary). Samples were introduced onto the column through a 20-~1 Rheodyne 7010 injector (Cotati, Calif.). A Labor MIM OE 3 12 ultraviolet spectrophotometer was equipped with a IO-p1 flow cell. Detection wavelength was 230 nm. The 250 X 4.6-mm reversed-phasecolumns (C,,). lo-pm particle size, were obtained from Chromatronix (LiChrosorb, Dimesil). 0003-2697/86
$3.00
Copyright &i 1986 by Academic Press, Inc. All nghfs of reproduction m any form resewed.
172
OROSZ.
NURIDSzkNY.
AND
OV/kDI
The solvents were HPLC-grade Merck products. Cardiac glycosides were purchased from Boehringer (Mannheim, FRG) and Richter G. (Budapest, Hungary). D. lanata leaveswere freshly collected from plants before the flowering stage. RESULTS
AND
DISCUSSION
Separation and quantitation of crystallized lanatoside C and its metabolites. The chromatogram shown in Fig. 1 represents a fast isocratic separation that can be achieved for some active digitalis glycosides on a reversedphase (RP)’ column with acetonitrile:water (2:3). Quantitation evaluation of the peaks is possible since a resolution factor of 2.0 for lanatoside C and digoxin is found. It can be also seenthat the isomers, 01-and P-acetyldigoxin, are separated, although they are very similar in chemical and physical properties. A slight alteration in the composition of the mobile phase results in a dramatic change in the value of the capacity factor, k’. The increase in the polar component of the eluant results in larger k’ values. Therefore, if the separation or isolation of lanatoside C and its metabolites is the aim, it is advantageous to increase the concentration of water in the eluant (Table 1). By decreasingthe polarity of the solvent (i.e., by adding methanol or acetonitrile) the capacity factor of lanatoside A and digitoxin can be significantly decreased.In addition the glycosidesinvestigated could be separated on a RP C-18 column with an isocratic solvent mixture of ethanol:acetonitrile:water (10:45:45) within 8 min (Table 1). It should be noted that the selectivity coefficients (ratio of k’) of the compounds of interest depend on the different packing materials ascalculated from Table 1, lines 5 and 6. Nevertheless, we have found that the differences in the retention times of lanatoside C and digoxin do not change, or only slightly change, with the polarity of the eluants. This may be explained by an increase in polarity due to deacetylation of lanatoside ’ Abbreviation
used: RP, reversed
phase.
: +-15
10 RETEMION
5 TIME (MIN)
FIG. I. Isocratic separation of digitalis glycosides. Sequence of elution: (I) deacetyl lanatoside C: (2) lanatoside C; (3) digoxin; (4) ol-acetyldigoxin; (5) fi-acetyldigoxin: (6) lanatoside A. Column, LiChrosorb RP C- 18: solvent system, water:acetonitrile (3:2): flow rate, 0.86 ml/min.
C, which is proportional to the decreasein polarity due to deglucosylation of lanatoside C. For the separation of these glycosidesit is convenient to selecta system which ensuresrapid separation (cf. Fig 1). In this manner it is possible to achieve rather good resolution. The increasein retention times is parallel with the deterioration of resolution. Various amounts of lanatoside C, digoxin, and their intermediate products were subjected to chromatography using the isocratic system. The calibration curves were drawn according to the peak heights from five chromatograms. The peak height of each glycoside was found to be reproducible with a relative standard error lessthan 1%. Straight lines were obtained by fitting the experimental points in the range
EXTRACTION
mr40000 w-lco*m* rj A ,’ 6 6 V
AND
ANALYSIS
o--m--6 ,’ 6 6
0
OF
DIGITALIS
GLYCOSIDES
173
of 30 to 1000 ng per injection for the compounds investigated. Regression lines were calculated by the method of least squares; the coordinate (0, 0) was not used in the calculation and hence extrapolation of each graph toward the origin may be considered to indicate the overall accuracy of the calibration. Equations for the regression lines are presented in Table 2 in the form Y = mX + c, where XI is nanograms of glycoside and Y is the optical density at 230 nm. Extraction. Glycosides were extracted from D. lanata leaves by maceration in water or in 50% aqueous methanol (or ethanol) for 3 min followed by intensive shaking (Vortex) for 1 min. The extract content was 10% dry wt. In the presence of alcohol the enzymatic conversion of primary into secondary glycosides is prevented. After centrifugation at 60001: the supernatant of the aqueous alcohol extract, being free of proteins, can be directly injected onto the column. Up to approximately 200 injections the parameters of the columns were not changed. If the leaves were pretreated to prevent the enzymatic conversion of primary glycosides (e.g., heat or microwave treatment). extraction could be carried out in water. This was followed by filtration on an Amicon PM 10 filter to remove the compounds with molecular weights higher than 10,000. The reliability and efficiency of extraction procedures will be demonstrated later. The fact that the extracts can be directly applied to HPLC analysis without purification makes the procedure convenient for screening of leaf samples and for simultaneous determination of lanatoside C and digoxin irrespective ofthe total glycoside composition.
Determination of lanatmide C in the e.xtract yf‘Digitalis lanata. To quantify the content of lanatoside C in the extract of D. lanata its complete separation from the other chromophore should be fulfilled. For this purpose two systems, acetonitrile:water (1:2) and methanol: water (3:2), were chosen. As shown in Fig. 2 in and near the exclusion volume of the LiChrosorb column a dominant part of the chromophoric compound is eluted with a
174
OROSZ,
NURIDSANY.
AND
TABLE STATISTICAL
Substances
DATA
OVADI
2
OF CALIBRATION
CURVES
Correlation coefficient
Equation
Deacetyl lanatoside Lanatoside C Digoxin a-Acetyldigoxin
C
Y = 1.1 1 X IO-“ .Y + 4.88 X 10m4 Y = 6.22 x 10-5,Y+ 6.92 x 1om4
0.999
Y = 6.07
X 10m5 A'+
3.28
x 10m4
0.999 0.999
Y = 3.04
X 10m5 A'+
8.91
x
0.999
Note. Chromatographic runs were carried system as the mobile phase.
out on a LiChrosorb
mixture of acetonitrile:water as mobile phase, and lanatoside C can be separated from these polar substances. Moreover, the k’ value for lanatoside C in Table 1 is 1.5 which is below the optimum value of 2 for chromatographic
4015
IO-“
RP C- 18 column
+
+
i
+
-+
10
RETENTION
FIG. 2. Chromatogram
(2: I) solvent
. 10
+
+
.
5
-~+
5
TIME
4.809
using water:acetonitrile
+-~+-“++
15 td~f~t-*--t
7.980 30.506 22.269
methods but within the minimum value of 1 (9). The amount of lanatoside C in the extract. calculated on the basis of the equation in Table 2, was found to be 0.16 -t 0.0 1% of the dry leaf. This value is significantly higher than the value we determined for the extract prepared as described by Potter, 0.11 k 0.0 1% of the dry leaf ( IO).
i ~*
F’ test
(MIN)
of Digitalis lamta extract (extracted with methanol). Peak I. lanatoside C: column, LiChrosorb RP C-l 8: solvent system. water:acetonitrile (2:l): flow rate, 0.86 ml/min.
RETENTION
TIME ,M,N)
FIG. 3. Chromatogram of enzyme-treated Digitalis lanwith 50% aqueous methanol). Peaks identified: ( 1) deacetyl lanatoside C: (2) lanatoside C: (3) digoxin: (4) a-acetyldigoxin: (5) /3-acetyldigoxin. The system was the same as described in Fig. 2. ata (extracted
EXTRACTION
AND
ANALYSIS
Similar analyses have been carried out on Dimesil columns with a mobile phase of methanoI:water (32). Although the capacity factors for Ianatoside C in the two HPLC systems are different (cf. Table l), the same lanatoside C content is found. This indicatesthat the peak is homogeneous. As we already mentioned. simultaneous determination of lanatoside C and digoxin is difficult due to their close retention times. We have overcome this difficulty. As shown in Fig. 3 the extract in which lanatoside C was enzymatically partially converted to its metabelites contains both the primary and secondary glycosides. Their separation by the chromatographic system used was satisfactory. The present method provides a rapid and effective procedure, including extraction. for quantitation of the main glycosides.Ianatoside C and digoxin, of D. lanata.
OF
DlGIl-ALIS
175
GLYCOSIDES
ACKNOWLEDGMENT The columns and the injection valve vided by Chromatronix Inc.. California.
were kindly
pro-
REFERENCES I. 2. 3. 4. 5. 6. 7. 8. 9.
IO.
Page. D. P. (July 1974) FDA p,“J LIlIL.2 ’ in-c a. 6 Castle, M. ( 1. ( 1975) J. c/wow ralojy. 115, 437-446. Evans, F. J. (1974) J. C%romaro.~r. 88, 41 l-412. Lindner. W.. and Frei. R. W. (1976) J Chnvw~ow. 117.81-86. Emi, F.. and Frei. R. W. ( 1977) .J. C’hromurc~,gr. 130, 169-180. Fujii. Y.. Fukuda. H.. Saito. Y.. and Yamaraki. M. ( 1980) .J. C/mvncl~o~r. 202, 139- 143. Wichtl. M.. Mangkudidjojo. M., and Wichtl-Bleyer, W. (1981) J. C‘lrrormrogr. 243, 503-508. Wichtl. M.. Wichtl-Bleyer. W.. and Manghudidjojo. M. ( 1981) J C%rornarcyq 247, 359-365. Johnson. E. L., and Stevenson. R. (1978) Basic Liquid Chromatography. Varian Associates, Palo Alto, Calif. Petter. H. ( 1963) Phurm~:ic~ 18, 554-56 I.