- Email: [email protected]
Amino acid analysis of angiotensin I by proton nuclear magnetic resonance spectroscopy
ANALYTICAL
BIOCHEMISTRY
141, 355-360 (1984)
Amino Acid Analysis of Angiotensin I by Proton Nuclear Magnetic Resonance Spectroscopy S.A.MARG~LISAND
B.COXON
Center for Analytical Chemistry, National Measurement Laboratory, National Bureau of Standards, Washington, DC 20234 Received February 13, 1984 The chemical shifts of the isoleucine and histidine protons of angiotensin I were assigned and the chemical shifts of the protons of the other amino acids in this peptide were confirmed at a field strength of 400 MHz. These chemical shift assignments were used to determine the amino acid composition of angiotensin I. These data were then compared to the amino acid composition which was determined by chromatographic analysis of the peptide hydrolysate. The results obtained by the chromatographic method were similar to those obtained by the NMR method. The standard deviations of the results were similar, indicating that these methods are equally precise. The major advantages of the NMR method are that it permits the recovery of the peptide at2er completion of the analysis and improves the quantitation of amino acids which are either partially destroyed by the hydrolysis procedure or require special derivatization methods for detection and quantitation.
Under current analytical protocol at the National Bureau of Standards, the certification of a peptide as a standard reference material (SRM)’ requires that the composition of that peptide be determined by two independent methods. The methods which are currently used for amino acid analysis consist of the hydrolysis of the peptide and subsequent analysis by chromatography of either the amino acids or derivatives thereof. These hydrolytic procedures destroy some of the amino acids and often several procedures must be used to obtain an analysis of a single peptide. The use of proton NMR at 400 MHz provides an alternate nondestructive method for amino acid analysis of peptides. Since the signals in the proton NMR spectrum of angiotensin II (A II) have been assigned to specific protons of the constituent amino acids (l-4), the proton signals of the additional histidine and leucine residues in angiotensin I (A I) can be readily assigned. ’ Abbreviations used SRM, standard reference material; A II, angiotensin II; A I, angiotensin I.
The relative proportion of each amino acid can then be determined by selecting and integrating signals which represent specific protons of a single amino acid and then dividing each integral by the number of protons that it represents. Using this approach, we have determined the amino acid composition of angiotensin I and compared our results with those from an accepted classical chromatographic method of analysis (5). MATERIALS
AND METHODS
Angiotensin I (SRM 998), angiotensin II, and Val 5 angiotensin II (Val 5 A II) were obtained from Beckman Bioproducts Operations, Palo Alto, California 94304. The angiotensins contained significant amounts of acetate. Therefore, before the proton signals 2 Certain commercial equipment, instruments, or materials are identified in this paper to specify adequately the experimental procedure. Such identification does not imply recommendation or endorsement by the National Bureau of Standards, nor does it imply that the materials or equipment identified are necessarily the best available for the purpose. 355
0003-2697/84 $3.00 Copyri@ 0 1984 by Academic Press. Inc. All rights of reproduction in any form reserved.
356
MARGOLIS
of the histidine and the leucine residues of angiotensin I were assigned, the acetate was removed by dissolving the peptide in 0.02 mol/liter hydrochloric acid, transferring the solution to a 5-mm NMR sample tube, and lyophilizing the sample in that tube. The amino acid composition of the SRM was determined on untreated acetate-containing material. The peptide (0.5 mg) was dissolved in 0.5 ml of deuterium oxide (100 at.% D). The spectra were measured with a Bruker Instruments (Biller@ Mass.) Model WM400 spectrometer in the pulse-Fourier transform mode. The instrumental parameters were: size of data set, 16384 points; pulse width, 4 ps (30” flip angle); spectral width, 4 kHz; number of scans, 2000; total relaxation delay, 4.00 s. All chemical shifts are on the delta scale relative to the methyl resonance of sodium 4,4-dimethyl-4-silapentanoate-2,2,3,3-d, which was used as an internal standard. Determination of the amino acid composition by use of the NMR spectra.The proton signals for His H-2 and H-4, the Phe ring H’s, the Tyr m and o H’s, Leu H-2, Val + Ile H-~‘S, Pro H-3 and H-5, Asp H-3, Ile H-4, and Leu H-3 + H-4 + Arg H-4 (Fig. 1) were integrated individually or as groups when necessary. Digital integrals of the proton signals were plotted on chart paper and were then measured manually. The integral for each group of protons listed above was divided by the mean integral for 29 protons which was determined by averaging the integrals of a total of 29 proton signals in the high field region beginning with the signal for Pro H-3b. The resulting value represented the relative proportion of the protons and hence the corresponding amino acid in the angiotensin I. Chemical determination of the amino acid composition.The peptide was hydrolyzed with 6 mol/liter constant boiling HCI in a l-ml borosilicate vial which was flushed with nitrogen and tightly sealed with a Teflon-lined cap. The sample was hydrolyzed for 20 min at 160°C, the solution was cooled, and the
AND COXON
solvent was evaporated under a stream of nitrogen. The sample was then dissolved in 0.5 ml of 0.01 mol/liter HCl, and the solution was analyzed with a Durrum amino acid analyzer (kit) using the Femto Buffer system (Dionex, Sunnyvale, Calif.). RESULTS AND DISCUSSION
Assignment of resonances.The chemical shift for each proton was measured by calculating the median value of the computergenerated shifts for the group of signals which represented each proton. In cases where the signals of several protons overlapped and could not be distinguished from one another we report the range of values for the group of signals. The assignments of the proton resonances for the three angiotensin analogs are summarized in Table 1. In general, they agree with those reported (1) for Asn 1 and Val 5 A II, for both acid and neutral pH. However, the chemical shifts of the protons of Asp are very sensitive to pH in the range of 4-8 (2), and this sensitivity is reflected in the variation of the values reported for these protons. The resonances for several protons or groups of protons could not be resolved. These include the alpha proton of His, located in the region of 4.6-4.8 ppm, the delta protons of Arg, and the beta protons of Phe, His, and Tyr. Our assignments of the resonances of Ile do not agree with those made by Lenkinski et al. (4). Integration of the signals introduced at 1.73, 1.36, and 1.09 ppm by the substitution of Ile for Val at amino acid five in the peptide chain indicates that each signal represents a single proton. The latter two resonances represent the nonequivalent gamma protons of Ile and not the beta and gamma protons as identified by Lenkinski et al. (4). This observation is consistent with the nonequivalence expected for protons attached to a carbon atom which is adjacent to an asymmetric center. Examination of the chemical shifts for the H-2 and H-4 protons of His-6 and the ortho
PROTON
NUCLEAR
MAGNETIC
RESONANCE
2
b
SPECTROSCOPY
OF ANGIOTENSIN
357
1
FIG. 1. Partial proton NMR spectrum at 400 MHz of a solution of angiotensin I (3.6 and acetate (5.22 X 10e4 mol/liter) in 40. The water region is not shown.
and meza protons of Tyr-4 reveals that these are split into two signals under slightly basic conditions. In the case of His, the signals are separated by 20-40 Hz, suggesting that there are two different forms of His. In the case of Tyr, one of the signals is identical to that observed under acid conditions while the other signal is associated with the increase in basic conditions.
I
X
10M4mol/liter)
The assignments of the methyl proton signals of Leu, Ile, and Val were confirmed by homonuclear decoupling of the protons on the beta-carbon atoms of Val and Ile, and the gamma-carbon atoms of Ile and LAX. The assignments of the protons on the alphacarbon atoms of these amino acids were also confirmed by this ==ries of experiments. The assignments fo, rhe protons on the alpha,
358
MARGOLIS
AND COXON
TABLE I CHEMICALSHIFTS (ppm) OFTHEPROTONNMR SIGNAUFORA 11,ValA
II,ANDAI
Peptide pH 2-3 Amino acid proton
pH 7-8
ValSAII
A II
AI
His 6 H-2
8.603
8.584
8.584
His 9 H-2 Phe arom His 6 H-4
7.300 7.295
7.299 7.275
8.561 7.293 7.267
His 9 H-4 Tyr H-m
7.082
7.083
7.248 7.083
Tyr H-o
6.741
6.745
6.745
Tyr H-2 Phe H-2 Pro H-2 Arg H-2 Asp H-2 Leu H-2 Va13 H-2 Val5 H-2 Be H-2 Pro H-5a Pro H-5b Arg H-5 Phe, His, and Tyr H-3 Asp H-3a Asp H-3b Pro H-3a Val H-3 Pro H-4 Pro H-3b Arg H-3a Arg H-3b Be H-3 Leu H-3 & 4 Arg H-4 Ile H-4a Be H-4b Leu Me Leu Me Val3Me Val3Me Val5Me Val5Me Be Me Ile Me
4.591 4.573 4.38 1 4.347 4.313
4.586 4.469 4.379 4.336 4.279
4.082 3.994
4.079
4.588 4.550 4.386 4.333 4.273 4.132 4.082
1.510
1.510 1.356 1.118
0.896 0.855 0.855 0.822
0.901 0.853
4.061 3.800 3.590 Unresolved Unresolved 2.828 2.690 2.248 1.964 0 1.814 1.729 a a 1.598 1.512 1.353 1.111 0.937 0.913 0.905 0.853
0.809 0.790
0.812 0.791
3.763 3.539 2.938 2.814 2.205 1.988 to 1.828 1.721
4.059 3.737 3.556 2.853 2.712 2.212 1.986 to 1.884 1.726 a a
ValSAII
A II
AI
7.83 1 7.726
7.775 7.701
7.294 7.008 6.945
7.305 6.987 6.931
7.038 7.081 6.732 6.780 4.536 4.450 4.369 4.352 3.953
7.037 7.074 6.739 6.780 4.586 4.452 4.383 4.343 3.975
7.700 7.691 7.640 7.280 6.983 6.922 6.894 7.048
4.089 4.004
4.088
4.581 4.562 (I a 3.819 4.160 4.090
4.067 3.785 3.425
4.069 3.753 3.329
2.711 2.561 2.145 I .995 to 1.812 1.736 0 D
3.782 3.438 2.755 2.607 2.158 2.029 (1 1.835
6.761
1.547
1.550 1.340 1.085
0.920 0.842 0.868 0.825
0.919 0.863
2.638 2.487 2.160 1.990 a 1.845 1.738 (1 a 1.532 1.532 1.343 1.086 0.89gb 0.899’ 0.855’ 0.848
0.810 0.786
0.806 0.787
1.748
’ Proton assignment is indistinguishable from that of the previous proton in the list. b These signals cannot be specifically assigned to the methyl groups of the amino acids in each column that this footnote appears.
PROTON
NUCLEAR
MAGNETIC
RESONANCE
beta and delta carbons of proline and on the alpha and gamma carbons of arginine were similarly identified. The formation of the amide bond between Phe and His in angiotensin I results in a downfield shift in the location of the central proton signals of the aromatic complex. At the same time, the sensitivity of the H-2 signal of the Phe to pH is lost as a result of the loss of the free amino group. The additional His and Leu in the angiotensin I molecule also results in a downfield shift of the signals for Pro H-3a, H-Sa, and H-5b. The pH sensitivity of these protons is also different from that of the corresponding protons of angiotensin II. An increase in pH causes a downfield shift for angiotensin II and an upfield shift for angiotensin I in the case of Pro H-3a and H-5a, and an opposite shift for Pro H-5b. Since the H-4 and H-3b signals were not completely resolved from those of the other protons, the effects on these proton signals of lengthening the peptide could not be evaluated. Additional studies are currently in progress to evaluate the effect of the two additional amino acids on the conformation of the angiotensins and the rotamer distribution of the constituent amino acids. The amino acid composition of angiotensin I. The results of the determination of the amino acid composition of angiotensin I are summarized in Table 2. The first column lists the amino acids in angiotensin I and the corresponding proton signals or groups of signals which were integrated are enumerated under Materials and Methods. The second column represents the value determined by dividing the integral for a specific group of protons by the average integral for a single proton and the number of protons in that specific integral. The third column represents the results of the chemical analysis by chromatography of the relative amino acid content normalized to the content of arginine. The proline content could not be determined by this method because it does not detect imino acids when the analysis is performed with the Femto buffers. The histidine content is
SPECTROSCOPY
OF ANGIOTENSIN
I
359
TABLE 2 COMPARISON OF THE AMINO ACID COMPOSITION OF ANGIOTENSIN I AS DETERMINED BY PROTON NMR
AND CHEMICAL
METHODS
Relative content + SD Amino acid His 9 His 6 Phe TY~ LeU PTO
Asp Ile Leu+Arg Ile + Val
NMR method (n = 6) 0.99 + 0.13 0.99 * 0.05 1.00 * 0.03 1.03 * 0.03 0.98 k 0.04 1.00 zk 0.04 1.04 + 0.04 0.98 k 0.04 0.99 Ifr OMb 1.05 k 0.046
Chemical method (n = 10)
2.04 +. 0.10” 1.10 + 0.10 1.02 + 0.02 1.09 + 0.01 N.D. 1.06 + 0.07 1.06 + 0.01 1.00‘ 1.10 + 0.01
“This value represents the sum of the His 6 and 9 contents. b This value represents the relative content of the second amino acid only. ‘The values in this column were normalized to that for arginine.
represented by a single value for both residues and the last two values represent the contents of arginine and valine, respectively. A comparison of the values in columns 2 and 3 indicates that both methods give similar results. The accurate quantitation of the amino acids by the NMR method requires that the spectra be measured under conditions where the total relaxation delay is greater than approximately 5T, ( TI is the spinlattice relaxation time). This value for the relaxation delay insures that the maximum amplitude will be observed for all of the proton signals. In this study, the total relaxation delay, including the acquisition time and the spectrometer relaxation delay, was 4.00 s and the flip angle was 30”. Under these conditions the total effective relaxation delay, taking into consideration the effect of the flip angle, may be expected to exceed the ST, requirement by a comfortable margin. The relaxation times reported for several protons of the angiotensin pentapeptide containing amino acids 4-8 are all less than 1 s (6), relaxation times that are much smaller than the effective relaxation delay for these studies. Furthermore, angiotensin I is a de-
360
MARGOLIS
capeptide and as the length of the peptide increases, so does its correlation time; and, therefore, the relaxation times of the constituent protons may be expected to decrease with increasing chain length. Finally, the methyl protons tend to have the longest relaxation times, but, in this work, only methylene and methine proton signals have been used for the calculations of the relative amino acid composition from the NMR spectra. A comparison of the standard deviations of the measurements obtained by the NMR method and the chemical method indicates that the precision of the two methods is similar. These results support the conclusion that the duration of the relaxation delay was adequate. The objective of this study was to determine the amino acid composition of pure samples of angiotensin. For this reason, we used materials which have been extensively evaluated by HPLC and mass spectrometry as part of the certification of the angiotensin I standard, SRM 998 (available from the National Bureau of Standards). These samples contain no trace of peptide impurities, very low levels of other organic impurities, and less than 2% of D-amino acids. The only significant impurity was the acetate which was associated with the free amino groups. For each mole of angiotensin I, 1.4 mol of acetate was present and the content of this anion was determined by the NMR method described herein. This method would not be suitable for determination of the amino acid compositions of mixtures of closely related peptides. However, it would readily detect the presence of nonamino-acid organic substances which might be present as contaminants. Thus, the proton NMR method rep resents a suitable alternative method for determination of the amino acid composition of purified peptides in the size range of angiotensin I. Although this method requires a relatively large amount of sample (0.5 mg), the sample can be recovered completely and used for other studies. Furthermore, the proton NMR method is a nondestructive method
AND COXON
and can be used to determine the amino acid composition when it is essential to analyze the intact peptide. The amount of peptide used for this study was determined somewhat arbitrarily by the fact that the angiotensin I SRM is packaged as OS-mg samples. The sample size can be reduced by increasing the number of scans used for signal averaging of the NMR data. Under the conditions of our experiments, a three- to fourfold increase in sensitivity could be achieved in this way. The amount of sample used in this type of measurement is also dependent on a number of other variables, including the number of protons per carbon atom, the signal-to-noise ratio of the spectrometer, the level of desired precision of the measurement, and the extent of the spin-spin splitting of the signals used for the area calculations. Although we were able to select successfYly a suitable set of signals for the calculation of the amino acid composition by use of the NMR method, improved resolution of the spectral components would have eliminated the need to calculate the content of Val and Arg by difference. The use of a higher field instrument might have made this possible. ACKNOWLEDGMENTS This research was partially supported by the National Heart, Lung and Blood Institute, National Institutes of Health research agreement No. l-Y-01-HV-70042-00. We thank Dr. P. E. Hare and Mr. R. Weker for their advice and assistance.
REFERENCES 1. Glickson, J. D., Cunningham, W. D., and Marshall, G. R. (1973) Biuchemi~fry 12, 3684-3692. 2. Piriou, F., Lintner, K., Fermandjian, S., Fromageot, P., Khosla, M. C., Smeby, R. R., and Bumpus, F. M. (1980) Proc. Natl. Acad. Sci. USA 77, 8286. 3. Khosla, M. C., Stachowiak, K., Smeby, R. R., Bumpus, F. M., Piriou, F., Lintner, K., and Fermandjian, S. (1981) Proc. Nat/. Acad. Sci. USA 78, 757-760. 4. Lenkinski, R. E., and Stephens, R. L. (1981) J. Inorg. Biochem. 15, 95-111. 5. Benson, J. R., and Hare, P. E. (1975) Proc. NatI. Acad. Sci. USA 72,6 19-622. 6. Niccolai, N., Miles, M. P., and Gibbons, W. A. (1979) Biochem. Biophys. Res. Commun. 91, 157-163.
BIOCHEMISTRY
141, 355-360 (1984)
Amino Acid Analysis of Angiotensin I by Proton Nuclear Magnetic Resonance Spectroscopy S.A.MARG~LISAND
B.COXON
Center for Analytical Chemistry, National Measurement Laboratory, National Bureau of Standards, Washington, DC 20234 Received February 13, 1984 The chemical shifts of the isoleucine and histidine protons of angiotensin I were assigned and the chemical shifts of the protons of the other amino acids in this peptide were confirmed at a field strength of 400 MHz. These chemical shift assignments were used to determine the amino acid composition of angiotensin I. These data were then compared to the amino acid composition which was determined by chromatographic analysis of the peptide hydrolysate. The results obtained by the chromatographic method were similar to those obtained by the NMR method. The standard deviations of the results were similar, indicating that these methods are equally precise. The major advantages of the NMR method are that it permits the recovery of the peptide at2er completion of the analysis and improves the quantitation of amino acids which are either partially destroyed by the hydrolysis procedure or require special derivatization methods for detection and quantitation.
Under current analytical protocol at the National Bureau of Standards, the certification of a peptide as a standard reference material (SRM)’ requires that the composition of that peptide be determined by two independent methods. The methods which are currently used for amino acid analysis consist of the hydrolysis of the peptide and subsequent analysis by chromatography of either the amino acids or derivatives thereof. These hydrolytic procedures destroy some of the amino acids and often several procedures must be used to obtain an analysis of a single peptide. The use of proton NMR at 400 MHz provides an alternate nondestructive method for amino acid analysis of peptides. Since the signals in the proton NMR spectrum of angiotensin II (A II) have been assigned to specific protons of the constituent amino acids (l-4), the proton signals of the additional histidine and leucine residues in angiotensin I (A I) can be readily assigned. ’ Abbreviations used SRM, standard reference material; A II, angiotensin II; A I, angiotensin I.
The relative proportion of each amino acid can then be determined by selecting and integrating signals which represent specific protons of a single amino acid and then dividing each integral by the number of protons that it represents. Using this approach, we have determined the amino acid composition of angiotensin I and compared our results with those from an accepted classical chromatographic method of analysis (5). MATERIALS
AND METHODS
Angiotensin I (SRM 998), angiotensin II, and Val 5 angiotensin II (Val 5 A II) were obtained from Beckman Bioproducts Operations, Palo Alto, California 94304. The angiotensins contained significant amounts of acetate. Therefore, before the proton signals 2 Certain commercial equipment, instruments, or materials are identified in this paper to specify adequately the experimental procedure. Such identification does not imply recommendation or endorsement by the National Bureau of Standards, nor does it imply that the materials or equipment identified are necessarily the best available for the purpose. 355
0003-2697/84 $3.00 Copyri@ 0 1984 by Academic Press. Inc. All rights of reproduction in any form reserved.
356
MARGOLIS
of the histidine and the leucine residues of angiotensin I were assigned, the acetate was removed by dissolving the peptide in 0.02 mol/liter hydrochloric acid, transferring the solution to a 5-mm NMR sample tube, and lyophilizing the sample in that tube. The amino acid composition of the SRM was determined on untreated acetate-containing material. The peptide (0.5 mg) was dissolved in 0.5 ml of deuterium oxide (100 at.% D). The spectra were measured with a Bruker Instruments (Biller@ Mass.) Model WM400 spectrometer in the pulse-Fourier transform mode. The instrumental parameters were: size of data set, 16384 points; pulse width, 4 ps (30” flip angle); spectral width, 4 kHz; number of scans, 2000; total relaxation delay, 4.00 s. All chemical shifts are on the delta scale relative to the methyl resonance of sodium 4,4-dimethyl-4-silapentanoate-2,2,3,3-d, which was used as an internal standard. Determination of the amino acid composition by use of the NMR spectra.The proton signals for His H-2 and H-4, the Phe ring H’s, the Tyr m and o H’s, Leu H-2, Val + Ile H-~‘S, Pro H-3 and H-5, Asp H-3, Ile H-4, and Leu H-3 + H-4 + Arg H-4 (Fig. 1) were integrated individually or as groups when necessary. Digital integrals of the proton signals were plotted on chart paper and were then measured manually. The integral for each group of protons listed above was divided by the mean integral for 29 protons which was determined by averaging the integrals of a total of 29 proton signals in the high field region beginning with the signal for Pro H-3b. The resulting value represented the relative proportion of the protons and hence the corresponding amino acid in the angiotensin I. Chemical determination of the amino acid composition.The peptide was hydrolyzed with 6 mol/liter constant boiling HCI in a l-ml borosilicate vial which was flushed with nitrogen and tightly sealed with a Teflon-lined cap. The sample was hydrolyzed for 20 min at 160°C, the solution was cooled, and the
AND COXON
solvent was evaporated under a stream of nitrogen. The sample was then dissolved in 0.5 ml of 0.01 mol/liter HCl, and the solution was analyzed with a Durrum amino acid analyzer (kit) using the Femto Buffer system (Dionex, Sunnyvale, Calif.). RESULTS AND DISCUSSION
Assignment of resonances.The chemical shift for each proton was measured by calculating the median value of the computergenerated shifts for the group of signals which represented each proton. In cases where the signals of several protons overlapped and could not be distinguished from one another we report the range of values for the group of signals. The assignments of the proton resonances for the three angiotensin analogs are summarized in Table 1. In general, they agree with those reported (1) for Asn 1 and Val 5 A II, for both acid and neutral pH. However, the chemical shifts of the protons of Asp are very sensitive to pH in the range of 4-8 (2), and this sensitivity is reflected in the variation of the values reported for these protons. The resonances for several protons or groups of protons could not be resolved. These include the alpha proton of His, located in the region of 4.6-4.8 ppm, the delta protons of Arg, and the beta protons of Phe, His, and Tyr. Our assignments of the resonances of Ile do not agree with those made by Lenkinski et al. (4). Integration of the signals introduced at 1.73, 1.36, and 1.09 ppm by the substitution of Ile for Val at amino acid five in the peptide chain indicates that each signal represents a single proton. The latter two resonances represent the nonequivalent gamma protons of Ile and not the beta and gamma protons as identified by Lenkinski et al. (4). This observation is consistent with the nonequivalence expected for protons attached to a carbon atom which is adjacent to an asymmetric center. Examination of the chemical shifts for the H-2 and H-4 protons of His-6 and the ortho
PROTON
NUCLEAR
MAGNETIC
RESONANCE
2
b
SPECTROSCOPY
OF ANGIOTENSIN
357
1
FIG. 1. Partial proton NMR spectrum at 400 MHz of a solution of angiotensin I (3.6 and acetate (5.22 X 10e4 mol/liter) in 40. The water region is not shown.
and meza protons of Tyr-4 reveals that these are split into two signals under slightly basic conditions. In the case of His, the signals are separated by 20-40 Hz, suggesting that there are two different forms of His. In the case of Tyr, one of the signals is identical to that observed under acid conditions while the other signal is associated with the increase in basic conditions.
I
X
10M4mol/liter)
The assignments of the methyl proton signals of Leu, Ile, and Val were confirmed by homonuclear decoupling of the protons on the beta-carbon atoms of Val and Ile, and the gamma-carbon atoms of Ile and LAX. The assignments of the protons on the alphacarbon atoms of these amino acids were also confirmed by this ==ries of experiments. The assignments fo, rhe protons on the alpha,
358
MARGOLIS
AND COXON
TABLE I CHEMICALSHIFTS (ppm) OFTHEPROTONNMR SIGNAUFORA 11,ValA
II,ANDAI
Peptide pH 2-3 Amino acid proton
pH 7-8
ValSAII
A II
AI
His 6 H-2
8.603
8.584
8.584
His 9 H-2 Phe arom His 6 H-4
7.300 7.295
7.299 7.275
8.561 7.293 7.267
His 9 H-4 Tyr H-m
7.082
7.083
7.248 7.083
Tyr H-o
6.741
6.745
6.745
Tyr H-2 Phe H-2 Pro H-2 Arg H-2 Asp H-2 Leu H-2 Va13 H-2 Val5 H-2 Be H-2 Pro H-5a Pro H-5b Arg H-5 Phe, His, and Tyr H-3 Asp H-3a Asp H-3b Pro H-3a Val H-3 Pro H-4 Pro H-3b Arg H-3a Arg H-3b Be H-3 Leu H-3 & 4 Arg H-4 Ile H-4a Be H-4b Leu Me Leu Me Val3Me Val3Me Val5Me Val5Me Be Me Ile Me
4.591 4.573 4.38 1 4.347 4.313
4.586 4.469 4.379 4.336 4.279
4.082 3.994
4.079
4.588 4.550 4.386 4.333 4.273 4.132 4.082
1.510
1.510 1.356 1.118
0.896 0.855 0.855 0.822
0.901 0.853
4.061 3.800 3.590 Unresolved Unresolved 2.828 2.690 2.248 1.964 0 1.814 1.729 a a 1.598 1.512 1.353 1.111 0.937 0.913 0.905 0.853
0.809 0.790
0.812 0.791
3.763 3.539 2.938 2.814 2.205 1.988 to 1.828 1.721
4.059 3.737 3.556 2.853 2.712 2.212 1.986 to 1.884 1.726 a a
ValSAII
A II
AI
7.83 1 7.726
7.775 7.701
7.294 7.008 6.945
7.305 6.987 6.931
7.038 7.081 6.732 6.780 4.536 4.450 4.369 4.352 3.953
7.037 7.074 6.739 6.780 4.586 4.452 4.383 4.343 3.975
7.700 7.691 7.640 7.280 6.983 6.922 6.894 7.048
4.089 4.004
4.088
4.581 4.562 (I a 3.819 4.160 4.090
4.067 3.785 3.425
4.069 3.753 3.329
2.711 2.561 2.145 I .995 to 1.812 1.736 0 D
3.782 3.438 2.755 2.607 2.158 2.029 (1 1.835
6.761
1.547
1.550 1.340 1.085
0.920 0.842 0.868 0.825
0.919 0.863
2.638 2.487 2.160 1.990 a 1.845 1.738 (1 a 1.532 1.532 1.343 1.086 0.89gb 0.899’ 0.855’ 0.848
0.810 0.786
0.806 0.787
1.748
’ Proton assignment is indistinguishable from that of the previous proton in the list. b These signals cannot be specifically assigned to the methyl groups of the amino acids in each column that this footnote appears.
PROTON
NUCLEAR
MAGNETIC
RESONANCE
beta and delta carbons of proline and on the alpha and gamma carbons of arginine were similarly identified. The formation of the amide bond between Phe and His in angiotensin I results in a downfield shift in the location of the central proton signals of the aromatic complex. At the same time, the sensitivity of the H-2 signal of the Phe to pH is lost as a result of the loss of the free amino group. The additional His and Leu in the angiotensin I molecule also results in a downfield shift of the signals for Pro H-3a, H-Sa, and H-5b. The pH sensitivity of these protons is also different from that of the corresponding protons of angiotensin II. An increase in pH causes a downfield shift for angiotensin II and an upfield shift for angiotensin I in the case of Pro H-3a and H-5a, and an opposite shift for Pro H-5b. Since the H-4 and H-3b signals were not completely resolved from those of the other protons, the effects on these proton signals of lengthening the peptide could not be evaluated. Additional studies are currently in progress to evaluate the effect of the two additional amino acids on the conformation of the angiotensins and the rotamer distribution of the constituent amino acids. The amino acid composition of angiotensin I. The results of the determination of the amino acid composition of angiotensin I are summarized in Table 2. The first column lists the amino acids in angiotensin I and the corresponding proton signals or groups of signals which were integrated are enumerated under Materials and Methods. The second column represents the value determined by dividing the integral for a specific group of protons by the average integral for a single proton and the number of protons in that specific integral. The third column represents the results of the chemical analysis by chromatography of the relative amino acid content normalized to the content of arginine. The proline content could not be determined by this method because it does not detect imino acids when the analysis is performed with the Femto buffers. The histidine content is
SPECTROSCOPY
OF ANGIOTENSIN
I
359
TABLE 2 COMPARISON OF THE AMINO ACID COMPOSITION OF ANGIOTENSIN I AS DETERMINED BY PROTON NMR
AND CHEMICAL
METHODS
Relative content + SD Amino acid His 9 His 6 Phe TY~ LeU PTO
Asp Ile Leu+Arg Ile + Val
NMR method (n = 6) 0.99 + 0.13 0.99 * 0.05 1.00 * 0.03 1.03 * 0.03 0.98 k 0.04 1.00 zk 0.04 1.04 + 0.04 0.98 k 0.04 0.99 Ifr OMb 1.05 k 0.046
Chemical method (n = 10)
2.04 +. 0.10” 1.10 + 0.10 1.02 + 0.02 1.09 + 0.01 N.D. 1.06 + 0.07 1.06 + 0.01 1.00‘ 1.10 + 0.01
“This value represents the sum of the His 6 and 9 contents. b This value represents the relative content of the second amino acid only. ‘The values in this column were normalized to that for arginine.
represented by a single value for both residues and the last two values represent the contents of arginine and valine, respectively. A comparison of the values in columns 2 and 3 indicates that both methods give similar results. The accurate quantitation of the amino acids by the NMR method requires that the spectra be measured under conditions where the total relaxation delay is greater than approximately 5T, ( TI is the spinlattice relaxation time). This value for the relaxation delay insures that the maximum amplitude will be observed for all of the proton signals. In this study, the total relaxation delay, including the acquisition time and the spectrometer relaxation delay, was 4.00 s and the flip angle was 30”. Under these conditions the total effective relaxation delay, taking into consideration the effect of the flip angle, may be expected to exceed the ST, requirement by a comfortable margin. The relaxation times reported for several protons of the angiotensin pentapeptide containing amino acids 4-8 are all less than 1 s (6), relaxation times that are much smaller than the effective relaxation delay for these studies. Furthermore, angiotensin I is a de-
360
MARGOLIS
capeptide and as the length of the peptide increases, so does its correlation time; and, therefore, the relaxation times of the constituent protons may be expected to decrease with increasing chain length. Finally, the methyl protons tend to have the longest relaxation times, but, in this work, only methylene and methine proton signals have been used for the calculations of the relative amino acid composition from the NMR spectra. A comparison of the standard deviations of the measurements obtained by the NMR method and the chemical method indicates that the precision of the two methods is similar. These results support the conclusion that the duration of the relaxation delay was adequate. The objective of this study was to determine the amino acid composition of pure samples of angiotensin. For this reason, we used materials which have been extensively evaluated by HPLC and mass spectrometry as part of the certification of the angiotensin I standard, SRM 998 (available from the National Bureau of Standards). These samples contain no trace of peptide impurities, very low levels of other organic impurities, and less than 2% of D-amino acids. The only significant impurity was the acetate which was associated with the free amino groups. For each mole of angiotensin I, 1.4 mol of acetate was present and the content of this anion was determined by the NMR method described herein. This method would not be suitable for determination of the amino acid compositions of mixtures of closely related peptides. However, it would readily detect the presence of nonamino-acid organic substances which might be present as contaminants. Thus, the proton NMR method rep resents a suitable alternative method for determination of the amino acid composition of purified peptides in the size range of angiotensin I. Although this method requires a relatively large amount of sample (0.5 mg), the sample can be recovered completely and used for other studies. Furthermore, the proton NMR method is a nondestructive method
AND COXON
and can be used to determine the amino acid composition when it is essential to analyze the intact peptide. The amount of peptide used for this study was determined somewhat arbitrarily by the fact that the angiotensin I SRM is packaged as OS-mg samples. The sample size can be reduced by increasing the number of scans used for signal averaging of the NMR data. Under the conditions of our experiments, a three- to fourfold increase in sensitivity could be achieved in this way. The amount of sample used in this type of measurement is also dependent on a number of other variables, including the number of protons per carbon atom, the signal-to-noise ratio of the spectrometer, the level of desired precision of the measurement, and the extent of the spin-spin splitting of the signals used for the area calculations. Although we were able to select successfYly a suitable set of signals for the calculation of the amino acid composition by use of the NMR method, improved resolution of the spectral components would have eliminated the need to calculate the content of Val and Arg by difference. The use of a higher field instrument might have made this possible. ACKNOWLEDGMENTS This research was partially supported by the National Heart, Lung and Blood Institute, National Institutes of Health research agreement No. l-Y-01-HV-70042-00. We thank Dr. P. E. Hare and Mr. R. Weker for their advice and assistance.
REFERENCES 1. Glickson, J. D., Cunningham, W. D., and Marshall, G. R. (1973) Biuchemi~fry 12, 3684-3692. 2. Piriou, F., Lintner, K., Fermandjian, S., Fromageot, P., Khosla, M. C., Smeby, R. R., and Bumpus, F. M. (1980) Proc. Natl. Acad. Sci. USA 77, 8286. 3. Khosla, M. C., Stachowiak, K., Smeby, R. R., Bumpus, F. M., Piriou, F., Lintner, K., and Fermandjian, S. (1981) Proc. Nat/. Acad. Sci. USA 78, 757-760. 4. Lenkinski, R. E., and Stephens, R. L. (1981) J. Inorg. Biochem. 15, 95-111. 5. Benson, J. R., and Hare, P. E. (1975) Proc. NatI. Acad. Sci. USA 72,6 19-622. 6. Niccolai, N., Miles, M. P., and Gibbons, W. A. (1979) Biochem. Biophys. Res. Commun. 91, 157-163.