Parameters affecting high performance liquid chromatographic

---

420

Anal. Chem. 1980, 52, 420-424

Parameters Affecting High Performance Liquid Chromatographic Separations of Neurohypophyseal Peptide Hormones Dennis D. Blevins, Michael F. Burke, and Victor J. Hruby* Department of Chemistry, University of Arizona, Tucson, Arizona 8572 1

Experimental conditions and parameters involved in high performance liquid chromatographic separations of seven neurohypophyseal hormones on several reverse phase columns were investigated. These peptides Include arginine vasotocin, lysine vasopressin, arginine vasopressin, mesotocin, Isotocin, oxytocin, and glumltocln. The effects of carbon chain length of the reverse phase support and organic solvent were examined. Using the appropriate solvent system and column, all of the peptides were separated from one another. Separation of the peptides was only part of the goal. The beginning of a study to understand the Interaction of the peptides with the stationary phase as a function of structure was also undertaken. This study led to the conclusion that the major parameters which allow effective separation of these structurally and conformationally closely related peptides are both the eluting strength of the mobile phase and the chemical composition of the stationary phase.

T h e separation of moderate size peptides (700-3000 mol wt) of closely similar structure for either analytical or preparative purposes has proved to be a difficult task. A number of classical techniques such as ion-exchange, gel filtration, and partition chromatography, paper and thin-layer chromatography, and countercurrent distribution have been useful in some cases, but continued developments are necessary. Recently several papers have appeared using high performance liquid chromatography (HPLC) for the analytical determination of moderate-sized underivatized and unprotected peptides (1-8) including diastereoisomeric derivatives (S11). However very little has been done to systematically examine the use of HPLC for the separation of closely related moderate sized peptides, the effects of various conditions on these separations, and the physical processes involved in these separations. In this paper we examine the separation of seven of the naturally occurring neurohypophyseal peptides including arginine vasotocin (AVT), glumitocin (GLU), lysine vasopressin (LVP), arginine vasopressin (AVP), mesotocin (MESO), isotocin (ISO), and oxytocin (OXY). The structures of these naturally occurring peptides of which one or more are found in all living animals from teleost and cartilaginous fish t o mammals including man (for a recent review see reference E),are shown in Figure 1. T h e molecular weights of these nonapeptides range from about 950 to 1100. T h e structural variations occur exclusively a t three residues, namely residues 3,4, and 8. Therefore, any separation of these compounds must utilize the differential properties of these molecules originating a t these three positions. We report here the experimental parameters and conditions in reverse phase HPLC which provide an efficient separation of each of these peptides from one another, and which allow the unambiguous analytical identification of these compounds. In addition we have examined the effects of reverse phase 0003-2700/80/0352-0420$01 .OO/O

support and solvent which allows us to obtain insight into the underlying physical-chemical processes which affect the separation of these peptides, and permit us to discuss these processes in terms of the differential properties of the hormones.

EXPERIMENTAL Synthesis a n d Purification of Peptides. All of the neurohypophyseal peptides used in this study were synthesized and purified in our laboratory. Generally the solid phase method of peptide synthesis (13)as used in our laboratory for the synthesis of neurohypophyseal hormones and analogues (14-17) was utilized. An automated instrument for peptide synthesis was used (181, and purification was accomplished by partition chromatography on Sephadex G-25 (19) followed by gel filtration on Sephadex G-25 using 0.2 N aqueous acetic acid. Solvents for chromatography were purified as previously reported (20). The purity of the peptides was established by quantitative amino acid analysis, thin-layer chromatography in at least three solvent systems, optical rotation, elemental analysis, and HPLC under conditions reported here. The individual peptide hormones were prepared as follows: oxytocin as previously reported (15);arginine vasopressin and lysine vasopressin by the method of Yamamoto et al. (17);mesotocin according to the method of Smith (21);glumitocin by the method of Gitu (22); isotocin and arginine vasotocin by modifications of methods previously reported (23). Reagents. The chromatographic mobile phase consists of two solvents: A and B. Solvent A was an aqueous solution of ammonium acetate, 0.05 M, pH 4.0, and was chosen as previously discussed (9, 11; see also 4 and 8). It was made by dissolving 3.85 g of ammonium acetate, reagent grade (Matheson, Coleman and Bell, Norwood, Ohio) in 950 mL of water which was distilled in glass through permanganate. The apparent pH was adjusted to 4.0 by titrating with glacial acetic acid, reagent grade (J.T. Baker Chemical Co., Phillipsburg, N.J.). The final volume was brought to 1 L with the above described water. The resulting aqueous solution was then filtered through a Millipore HAWP 0.45-pm filter (Millipore, Bedford, Mass.) prior to use. Solvent B was the organic portion. Tetrahydrofuran (THF) “distilled-in-glass” (Burdick and Jackson, Muskegon, Mich.) and acetonitrile (CH,CN) reagent grade (Mallinckrodt, St. Louis, Mo.) were used without further purification. Methanol (CH30H) reagent grade (Mallinckrodt, St. Louis, Mo.) was passed through a silica column. Each of the organic solvents was filtered through a Millipore filter, FHLP 0.5 pm, prior to use. All solvents were degassed in vacuo. The samples were prepared by weighing approximately 1 mg of the appropriate peptide and adding one drop of 10% acetic acid in water and then bringing the total volume to about 1mL with water. Sample mixtures were prepared by mixing equal volumes of the appropriate samples together. Apparatus. The HPLC system used has been described previously (24). The three columns used in this study are identified as: C18, C8, and C z The C18column was a FBondapak C18 reverse phase 300 mm long X 3.9 mm i.d. column (Waters Associates, Milford, Mass.); carbon by weight on the stationary phase, lo%, and is endcapped. The mobile phase for the C18 column consisted of the following percentages of solvent B: 14% THF, 22% CH&N, or 36% CH30H. Solvent A constituted the remaining portion. The solvent breakthrough time was 175 s. The chart speed was 1980 American Chemical Society

ANALYTICAL CHEMISTRY, VOL. 52, NO. 3, MARCH 1980

421

7-

64 L'I 1

I '.I

W

2 I-

4

3:

L--

~~

---

-

THF

7

7

-

CH OH 3

CHfN

Figure 2. Plot of relatwe retention ( a )vs. organic modifier in the solvent GLU, systems studied on t h e pBondapak C,, column. (0)AVT. (0) ( A ) LVP, ( ) AVP, ( 0 )MESO, (m) ISO, and (A)OXY

I

I

.

I I - ('>.< - I > 1'- I 1 c - ( , 1 I1 - \..n -I:>..I 7 Y O - I r u-(,1 ? - \ I I I : i 1 ; . (1 !! -

Figure 1.

!?\!

Structures of the naturally occurring neurohypophyseal

/

6 -

peptides arginine vasotocin (AVT), glumitocin (GLU), lysine vasopressin (LVP), arginine vasopressin (AVP), mesotocin (MESO), isotocin (ISO), and oxytocin (OXY). The numbers (1, 2, 3, etc.) correspond to the residue from the N-terminal half-cystine-1 to the Cterminal glycinamide-9

0.33 cm/min. Column efficiencies were calculated in the usual manner (11,25). With 14% THF as the organic modifier portion of the mobile phase, the column efficiency was 990 plates. The CB column was LiChrosorb RP-8,lO pm (EM Laboratories, Darmstadt, Germany) packed in a 250 mm long x 2.1 mm i.d. stainless steel column in our laboratory using a slurry technique; carbon by weight on the stationary phase, 12.2%. The following percentages of solvent B were used on the CBcolumn: 9% THF, 18% CH,CN, or 32% CH30H. Solvent A constituted the remaining portion. The instrumental operating conditons were the same as for the C18column except the chart speed was 0.5 cm/min. The solvent breakthrough time was 46 s. With 9% THF as the organic modifier in the mobile phase, the column efficiency was 690 plates. The C2column was LiChrosorb RP-2,lO pm, (EM Laboratories) 250 mm long X 2.1 mm i.d.; carbon by weight on the stationary phase, 5 % . The following percentages of solvent B were investigated on the C, column: 12% THF, 18% CH,CN, or 32% CH30H. Solvent A constituted the remaining portion. Instrumental operating conditions were the same as for the C8 column. The solvent breakthrough time was 40 s. The data were computed and reported as in our previous work (11). In this study, the relative retention was computed with respect to A V T equal to K l because it eluted first in all solvent systems. Column efficiencies were calculated according to standard methods (26).

RESULTS The elution order of the compounds on all columns (C2,Cg, C18) from least retained to most retained was AVT, GLU, LVP, AVP, MESO, ISO, and OXY in all solvent systems studied. T o allow for comparison from column t o column, the capacity factor ( k ? for oxytocin was held relatively constant by adjusting the percentage of organic solvent in the mobile phase. On each column three different solvent systems were examined. The solvent systems were composed of solvent A, and of solvent B which was either T H F , CH&N, or CH,OH. For purposes of identification, each compound was injected

W 4 1 [L

W

2

2 3-

-a W

. a

i.

--

-

7-

THF

~ 7--C H3C N ~

----CH30H

Figure 3. Plot of relative retention ( a )vs organic modifier in the solvent system studied on t h e LiChrosorb RP-8 column. (0)AVT, (0) GLU, (A) LVP, ( r ) AVP, ( 0 )MESO, (W) ISO,and (A)OXY

both individually and in a mixture. Compounds having similar h'values were combined and injected in order to confirm the order of elution. The relative retentions of these compounds on the CI8 column as a function of solvent are shown in Figure 2. As expected, the peptides with the more lipophilic amino acids were retained longer than those with the hydrophilic amino acids. Thus the following order of elution was seen in all of the solvent systems: AVT > GLU > LVP > AVP > MESO > IS0 > OXY. Although the order of elution remains constant for each solvent, Figure 2 also shows that the relative retention of the components of this mixture is very much a function of the solvent composition. The C18 column was able to provide peak separation under all conditions. It should be pointed out that AVT and GLU as well as IS0 and MESO are resolved only when T H F is used as the organic modifier. As shown in Figure 2, essentially all peaks are separated base line except LVP and AVP which can be identified by their respective capacity factors. The order of elution on the CBcolumn was the same as for the CIS. The relative retentions as a function of solvent are shown in Figure 3. The following order of elution is observed,

422

ANALYTICAL CHEMISTRY, VOL. 52, NO. 3, MARCH 1980

7

A

1

B

7-7-L-

10 TIME

5

-~-r

0

IO

5

0

TIME min

rnin

Figure 4. Chromatogram on LiChrosorb RP-8 of (right to left) (A) AVT, LVP, ISO,and OXY and (B) GLU,AVP, MESO, and OXY with 9% THF as solvent B

AVT = GLU > LVP > AVP > MESO > I S 0 > OXY, except in methanol where LVP and AVP coelute. Typical chromatograms obtained are shown in Figures 4A and 4B. These chromatograms show the separation of mixtures of four native neurohypophyseal peptides with oxytocin being used as a reference in each case. Under conditions used in these experiments, the C8 column was not able to separate all seven peptides. Again, as expected, the C2 column retained the peptides with lipophilic amino acids more than the peptides with hydrophilic amino acids. The relative retentions as a function of solvent are shown in Figure 5. Figure 5 shows that AVT, GLU, LVP, and AVP coelute in methanol and T H F , but are partially separated in acetonitrile. MESO and I S 0 coelute in methanol and acetonitrile, but are separated from each other in T H F except t h a t I S 0 coelutes with OXY.

DISCUSSION Liquid chromatography can be characterized by the mobile phase, the bonded phase, and the stationary phase. The stationary phase consists of solid support with bonded materials plus adsorbed associated mobile phase (27,281. Using this definition of the stationary phase, a change in the composition of the mobile phase produces a change in the modified bonded phase. The chromatographic behavior of the same compound in different solvent systems can be partially explained using this definition (27). The mobile and stationary phases are equally important in HPLC. T h e mobile phase solvents are characterized by their solvent eluting strengths, which are the sum of three types of intermolecular interactions: dispersion, orientation, and hydrogen bonding (25). The role of the stationary phase is not completely understood. The ability of the hydrocarbon stationary phase to interact with the solute has been shown t o cause retention (27). Interactions of the residual silanols with the components of the mobile phase can also affect retention. The conformational changes of the long chain bonded stationary phase with respect to different organic modifiers has also recently been investigated (29,30),and show that the physical structure as well as the effective chemical composition of the stationary phase is a function of the ability of the organic modifier to solvate the bonded hydrocarbon. In addition the strength of the attractive interactions between the solute molecules and the stationary phase is a function of the relative amounts of bonded hydrocarbon vs. free silanol groups

THF

CH3C N

CHSOH

Flgure 5. Plot of relative retention ( a )vs. organic rodifier in the solvent system studied on the LiChrosorb RP-2 column. (0)AVT, (0)GLU, (A)LVP, (0) AVP, ( 0 )MESO, (m) ISO, and (A)OXY

on the surface. Thus the relative retention of the peptides on the bonded columns will be a function of the effective chemical composition of the stationary phase in terms of the available hydrocarbon, the organic modifier, and the water present. I t should be noted, therefore, that an explanation of the chromatographic behavior of reverse phase columns as a function of both mobile phase and stationary phase composition is possible without resorting to a hypothetical “hydrophobic effect” (31). With the appropriate solvent system and column we were able to separate all of the neurohypophyseal hormones used in this study (Figure 1). Separation of the peptides was only part of the goal. The beginning of a study to understand the interactions of the peptides with the stationary phase as a function of structure was undertaken. Separations of structurally similar peptides such as the neurohypophyseal peptides will depend to a significant extent on differences in primary structure. Interaction of the peptides with the stationary phase will involve primarily the side chain moieties of various residues in the peptides and hence any differences in sequence can be important in achieving a separation. In addition, primary structural effects will be modulated by differences in the conformational, dynamic, and topological properties of the peptides. In the latter regard, considerable work has been reported on the aqueous solution properties of these hormones using ‘H nuclear magnetic resonance (NMR) spectroscopy (32-33, 13C NMR spectroscopy (38, 39), circular dichroism (CD) (40-45), and Raman (43-46). For example, the 13C NMR spectra of all seven of the naturally occurring hormones studied here have been examined (38,39). Hruby et al. (39)concluded that since the chemical shifts were virtually the same for the nonvariable residues, all of the hormones have very similar conformations. A recent proton NMR study comparing OXY, LVP, and AVP came t o similar conclusions (37). From the standpoint of dynamic properties, NMR relaxation measurements (33, 47-49) have noted that both OXY and LVP are quite flexible. These studies, as well as the CD and Raman studies indicated that indeed the conformational and dynamic properties are very similar, though oxytocin-like compounds appear to have more compact structures than vasopressins, while the latter compounds have more segmental motion in the C-terminal tripeptide moiety, but a slightly more rigid conformation in the N-terminal region. Since the conformational and dynamic properties of all of these hormones are similar, it is likely these

ANALYTICAL CHEMISTRY, VOL.

Table I. Relative Lipophilicity of Selected Amino Acids" cf(reI)b

phenylalanine leucine is0 leucine arginine

lysine serine glutamine a

2.24 1.99 1.99 ?

0.52 -0.56 -1.09

L/HC L L L H H H H

Reference 50. cf(re1) = relative lipophilicity of the L = lipophilic; H = hydrophilic.

side chain.

properties will have only minor effects on HPLC separations. Instead separations should depend primarily on differences in the properties of side chain groups a t the structurally variable positions of these hormones (Figure 1). As discussed below, t h a t is what we have observed. T h e peptides can be characterized with respect to their lipophilicity. Rekker has suggested (50) t h a t the lipophilic properties of each amino acid are directly related to the relative lipophilicities of the side chain moieties. There are only three positions of change in the amino acid sequence for the peptides studied: positions 3 , 4 , and 8. Table I shows the relative lipophilicities, according to Rekker, for all of the amino acids found in these three positions. Either isoleucine or phenylalanine is found in position 3 of the neurohypophyseal hormones (Figure 1). Position 4 also has two possibilities: glutamine or serine. Position 8 has five possibilities: leucine, isoleucine, lysine, arginine, and glutamine. In discussing the chromatography, the Cz column is used as the base line to show the maximum effect of the mobile phase. T h e limited retention of the solutes and the lack of selectivity illustrate the relative polar nature of the stationary phase. In the T H F and CH30H solvent systems, the C2 column showed only two groups of peptide separation. Group one consisted of AVT, GLU, LVP, and AVP, all of which possess a hydrophilic residue a t a variable position. Group two consisted of MESO, ISO, and OXY, all of which have lipophilic amino acids a t the variable positions. In the CH3CN system, a similar observation is made except that AVP distinctly separates from the former group. In these and subsequent studies, it is unlikely that the differences in the mobile phase selectivity could be due to p H variations since the peptides do not contain aspartic acid, glutamic acid, or a C-terminal carboxylate group. T h e C8 column was also unable to separate all of the peptides. However, this column demonstrated an increase in the attraction of the stationary phase for all the peptides. The greater ability of the stationary phase to interact with the solute is evident in the separation of most of the peptides. Selective modification of the effective stationary phase by a change in the organic portion of the mobile phase is shown by use of THF. Using T H F , MESO and I S 0 were separated with base-line resolution, whereas, in other solvent systems tried, they were not. T h e CI8 column was able to separate all of the peptides. GLU which could not be separated from AVT on the Cz or C8 columns, was separated from AVT with base-line resolution on this column. The use of T H F as the organic component of the mobile phase provided a modified stationary phase with both the greatest retention for all of the peptides as well as the greatest resolution. T h e best explanation is that the chemical composition of the stationary phase which is available t o the peptides, is significantly more hydrocarbon-like with the longer chain bonded phase well solvated by the THF, and thus the additional length of the CI8 bonded group provided the optimum stationary phase.

52, NO. 3, MARCH 1980

423

The structure of the peptides was examined and since the conformational and dynamic properties of these peptides are very similar, sequence changes a t the variable positions were used to explain the separation achieved. The only structural difference between AVT and AVP is in position 3 of the sequence (Figure l ) ,and in this regard it is noted t h a t phenylalanine is both more lipophilic and more bulky than isoleucine. It seems reasonable that the additional interactions of the phenylalanine side chain with the stationary phase are responsible for the longer retention of AVP relative to AVT. I t is doubtful whether position 4 plays an important role in most separations. The only two compounds that have serine in position 4 instead of glutamine are IS0 and GLU (Figure I), but they are eluted a t different ends of the elution order. MESO and IS0 differ only in position 4. The importance of an organic modifier in the mobile phase is emphasized in this case, since only by using T H F as the organic modifier were these two peptides separated with base-line resolution. Position 8 has five possibilities: MESO and IS0 have isoleucine; OXY has leucine; AVP and AVT have arginine; LVP has lysine; and GLU has glutamine. T h e relative lipophilicities of isoleucine and leucine are identical, and it is to be noted that MESO, ISO, and OXY were always in the latter group to elute. The only difference between MESO and OXY is in position 8. The interaction with the stationary phase of the two different amino acids results in their separation, suggesting that the isobutyl group of leucine interacts more strongly with the stationary phase than does the sec-butyl group of isoleucine. LVP and AVP differ only in position 8. At the pH studied, the amine group on the side chain is protonated. The lysine becomes more hydrophilic than the arginine when both are protonated. The arginine, when protonated, can distribute its positive charge through resonance structures to all the nitrogens and effectively increases its lipophilicity. This allows the hydrocarbon portion of the amino acid to interact better with the Stationary phase. This is supported by the observation that in all systems examined, LVP was eluted before AVP when separation was observed. AVT and GLU differ in structure in positions 4 and 8, but are eluted together in most cases. The differences in structure balance in the chromatography. AVT has arginine in position 8 and GLU has glutamine. Glutamine is more hydrophilic than arginine and, therefore, GLU should be eluted first, but is not. In the previous argument it was stated t h a t lysine is more hydrophilic than arginine and glutamine is more hydrophilic than lysine (Table I), therefore glutamine is more hydrophilic than arginine. Position 4 in this case has a predominate influence in the chromatography. The glutamine in AVT is more hydrophilic than the serine in GLU and is responsible for the shorter retention of AVT with respect to GLU. In summary, we have found suitable HPLC procedures for the identification and separation of seven of the naturally occurring neurohypophyseal peptide hormones. Proper choices of reverse phase support and organic modifier are critical t o obtaining effective separations. Finally, we have shown how interactions of the stationary phase and the mobile phase with the variable structural components of the peptide hormones are important in the separations.

LITERATURE CITED Tsuji, K.; Robertson, J. H. J. Chromatogr. 1975, 172. 663-672. Burges, R.; Rivier, J. "Peptides 1976", A. Loffet, Ed.; Editions de 1'Universite Bruxelles: Bruxelles. 1977; pp 85-94. Hancock, W. S.:Bishop, C. A.; Hearn, M. T. W. FEES Lett. 1978, 72, 139- 142. Krummen, K.; Frei, R. W. J . Chromatogr. 1977, 132, 429-436. Hansen, J. J.; Greibrokk, T.; Currie. B. L.; Johansson, K. N.-G.; Folkers, K. J. Chromatogr. 1977, 135, 155-164. Molnar, I.; Horvath, C. J. Chromatogr. 1977, 742, 623-640. Hancock, W. S.;Bishop, C. A.; Prestige, R. L.; Hading, D. R. K.; Hearn. M. T. W. Science 1978, 200, 1168-1 170.

424

Anal. Chem. 1980, 52, 424-427

(8) O'Hare, J. J.; Nice, E. C. J. Chromatogr. 1979, 171,209-226. (9) Larsen, 6.; Viswanatha, V.; Chan, S. Y.; Hruby, V. J. J. Chromatogr. Sci. 1978. 16. 207-210. (IO) Meyers, C: A.; Coy, D. H.; Hwng, W. Y.; Schally, A. V ; Reeding, T. W. Biochemistry 1978, 17,2326-2331. (11) Larsen, 6.; Fox, B. L.; Burke, M. F.; Hruby, V. J. Int. J. Pept. Protein Res. 1979, 13, 12-21. (12) Sawyer, W. H. Fed. Proc. 1977,36, 1842-1848. (13) Merrifield, R. 8. J. Am. Chem. SOC. 1983,85,2149-2154. (14) Hruby, V. J.; Muscio, F.; Groginsky, C. M.; Gitu, P. M.; Saba, D.; Chan, W. Y. J. Med. Chem. 1973, 16,624-629. 115) UDson. D. A.: Hrubv. V. J. J. Ora. Chem. 1976. 41. 1353-1358. H;uby,'V. J.; Upson,*D. A.; A g a r w i , N. S. J. Org.'Chem. 1977,42, 3552-3556. Yamamoto, D. M.; Upson, D. A.; Linn, D. K.; Hruby, V. J. J. Am. Chem. SOC. 1977,99, 1564-1570. Hruby, V. J.; Barstow, L. E.; Linhart, T. Anal. Chem. 1972,44, 343-359. Yamashiro, D. Nature (London) 1984,207, 76-77. Hruby, V. J.; Groginsky, C. M. J. Chromatogr. 1971, 63, 423-428. Smith, C. W., Ph.D. Dissertation, University of Arizona, 1973. Gitu, P. M., Ph.D. Dissertation, University of Arizona, 1974. A. F. . . Hrubv. V. J.: Smith. C. W.: Aaarwal. N. S.: Powers.. S.:. SDatola. . Unpiblished' work, University i f Arizona, 1977. (24) Eckhardt, J. G.; Stetzenbach, K.; Burke, M. F.; Moyers, J. L. J. Chromatogr. Sci. 1978, 16,510-513. (25) Bakalyar, S. R.; McIlwrich, R.; Roggendorf, E. J. Chromatogr. 1977, 142,353-365. (26) Karger, B. L.; Snyder, L. R.; Horvath, C. "An Introduction to Separation Sciences"; John Wiley and Sons: New York. 1973; pp 129-138. (27) Bakalyar, S. R . Am. Lab. 1978, IO, 43-61. (28) Tanaka, N.; Goodell, H.; Karger, B. L. J. Chromatogr. 1978, 158, 233-248. (29) Cline. S.M.; Stetzenbach, K. J.; Burke, M. F. Anal. Len. 1979,in press. (30) Stetzenbach, K. J., Ph.D. Dissertation, University of Arizona, 1980. (31) Hildebrand, J. H. Proc. NaN. Acad. Sci. U . S . A . 1979, 76, 194. (32) Brewster, A. I.R.; Hruby, V. J. R o c . Natl. Acad. Sci. U . S . A . 1973, 70, 3806-3809. (33) Meraldi. J. P.; Hruby. V. J.; Brewster, A. I. R. Proc. Nafi. Acad. Sci. U . S . A . 1977, 74,1373-1377. (34) Boicelli, C . A.; Bradbury, A. F.; Feeney, J. J. Chem. Soc., Perkin Trans. 2 1977,477-482.

(35) Krishna. N. R.; Huang, D. H.; Glickson. J. D.; Rowan, R.; Walter, R. Siophys. J. 1979,26, 345-366, and references therein. (36) Von Dreele, P. H.; Brewster. A. I.; Dadok, J.; Scheraga. H. A,; Bovey, F. A.; Ferger, M. F.; du Vigneaud, V. Proc. Natl. Acad. Sci. U . S . A . 1972,69, 2169-2173. (37) Wyssbroad, H. R.; Fishman, A. J.; Live, D. H.; Hruby, V. J.; Agarwal, N. S.; Upson, D. A. J. Am. Chem. SOC. 1979, 101,4037-4043. (38) Waiter, R.; Prasad, K. U. M.; Deshuriers, R.; Smith, I.C. P. Proc. Nat/. Acad. Sci. U . S . A . 1973, 70,2086-2090. (39) Hruby, V. J.; Deb, K. K.; Spatola, A. F.; Upson, D. A,; Yamamota, D. J . Am. Chem. SOC. 1979, 101,202-212. (40) Beychok, S.; Breslow, E. J. Biol. Chem. lS88, 243, 151-154. (41) Urry. D. W.; Quadrifoglio. F.; Walter, R.; Schwartz, I.L. Proc. Natl. Acad. Sci. U . S . A . 1988, 6 0 , 967-974. (42) Fric, I.; Kodicek, M.; Flegel, M.; Zaoral. M. Eur. J. Biochem. 1975,56, 493-502. (43) Maxfield, F. R.; Scheraga, H. A. Biochemistry 1977, 16,4443-4449. (44) Hruby, V. J.; Deb, K. K.; Fox, J.; Bjarnason, J.; Tu, A. T. J . Bio/. Chem. 1978. 253,6060-6067. (45) Tu, A. T.; Lee, J.; Deb, K. K.; Hruby, V. J. J. B i d . Chem. 1979,254, 3272-3278. (46) Tu, A. T.; Bjarnason, J. 6.; Hruby, V. J. Biochim. Siophys. Acta 1978, 533,530-533. (47) Glasel, J. A.; Hruby, V. J.; McKelvy, J. F.; Spatola, A. F. J. Mol. Biol. 1973, 79,555-575. (48) Deslauriers, R.; Smith, I. C. P.; Walter, R. J. Am. Chem. SOC. 1974, 96,2289-2291. (49) Wyssbrod, H. R.; Balhrdin, A.; Schwartz, I. L.; Walter, R.; Van Binst, 0.; Gibbons, W. A.; Agosta. W. C.; Field, F. H.; Cowburn, D. J. Am. Chem. SOC. 1977,99,5273-5276. (50) Rekker, R. F., "The Hydrophobic Fragmental Constant"; Elsevier: A m sterdam, New York, 1977, p 301.

RECEIVED for review September 28,1979. Accepted December 10,1979. This work supported in part by U.S. Public Health Grant AM-17420 (V.J.H.) and by the National Science Foundation (V.J.H.).

Determination of Fluorescent Compounds by High Performance Liquid Chromatography with Chemiluminescence Detection Shin-ichiro Kobayashi and Kazuhiro Imai * Department of Analytical Chemistry, Faculty of Pharmaceutical Sciences, Univerity of Tokyo, Hongo 7-3- 1, Bunkyo-ku, Tokyo 113, Japan

chemiluminescence (CL) of dansyl amino acids generated with bis( 2,4,6-trichiorophenyI)oxaiate (TCPO) and hydrogen peroxide (H202)was studied for the application to the detection system for high performance liquid chromatography. Not only the concentration of TCPO and H202but also the constkuents of the medium both the intern adle time of CL. The combination of 0.51 mUmln of 5 mM TCPO in ethyl acetate, 1.2 mL/min of 0.5 M H202 in acetone, and 0.18 mUmin of eluent (0.05 M Tris HCI buffer (pH 7.7)-acetonltriie, ""9 'Iv) from a column Of pBondapak '18 was for the fmoi detection Of the fiuorophores. The peak heights of dansyiated alanine, glutamic acid, methionine, and norleucine were proportional to the quantities more than 50 fmoi.

High performance liquid chromatography (HPLC) has been a common tool for the measurement of substances in biological materials. In the determination of minute amounts of substances such as biogenic amines (1-3) and amino acids ( 4 ) in body fluids, fluorescence detection is usually adopted. Generally speaking, the fluorescence detection system for HPLC consists of (a) a light source for the excitation of fluorophores, (b) a flow cell through which the eluate passes and (c) a photomultiplier for the detection of the emitted light from the fluorophores. There are problems to be encountered 0003-2700/80/0352-0424$01 .OO/O

for increasing the sensitivity of the detection system. First, a Part of the Stray radiation comes through the flow cell from the light Source into the PhotomultiPlier to raise the background level. Second, fluctuation of the light source causes variation of the sum of the scattered light and the emitted fluorescence to lower the signal-to-noise ratio. In order to diminish the stray radiation, a droplet of the eluate from a column was directly irradiated by laser beam (5, 6). On the other hand, as a way to excite the fluorophores instead of irradiation, reaction of oxalic esters with hydrogen peroxide (H,O,) has been investigated by several workers (7-9). The proposed mechanism of the chemiluminescence (CL) reaction was as follows: 0

0

1,2-dioxetanedione

0

0

/I IC1 +

r-

it 3-0

fluorophor

-

fluorophor*

fluorophor*

light

(excited state)

+

+

2CO2

fluorophor

I t has been used mainly for the determination of H,Oz in solution (10-12), and only two papers have appeared on its 0 1980 American Chemical Society