Determination of TetrahydroisoquinolineAlkaloids in Biological Materials with High Performance Liquid Chromatography Ralph M. Riggin’ and Peter 1.Kissinger* Department of Chemistry, Purdue University, West Lafayette, Ind. 47907
A new approach is described for the selective determination of tetrahydroisoquinolines(TIQ’s) formed by condensation of
catecholamines with aldehydes In vivo. Hlgh performance cation exchange chromatography In comblnatlon with amperometric detection permits simultaneous quantltation of tetrahydropapaverollne (THP) and salsollnol (SAL) at 2 ng/g in rat brain and 2 ng/mL In urine. Examples of application of the new method are given for studies pertainingto the possible Involvement of TlQ’s in alcohol abuse.
The primary metabolic pathways of the catecholamines have been extensively studied and their involvement in neurotransmission and other processes in the peripheral and central nervous systems has long been recognized. There exists, however, another possible metabolic route for these compounds which has only recently received attention. This pathway involves the nonenzymatic condensation of catecholamines with aldehydes to form tetrahydroisoquinoline alkaloids (TIQ’s) which can act as “false” neurotransmitters and result in catecholamine depletion. Such a reaction scheme is shown below. R
Although it is conceivable that this reaction could occur normally by utilizing endogenous aldehydes (e.g., glyceraldehyde or glyoxylate), these compounds are generally present a t low concentrations within the mitochondria or other organelles of the cell and thus never encounter the catecholamines. TIQ formation is believed to have its most probable significance under abnormal circumstances in which the body is presented with an external stimulus which results in significant levels of circulating aldehydes. The most obvious example of such a stimulus is alcohol consumption. During alcohol intoxication, significant levels of circulating acetaldehyde are present and it has been proposed by various workers that the condensation of acetaldehyde with catecholamines to form TIQ’s may account in some way for the addicting properties of alcohol. The literature adequately describes the rationale for this hypothesis (1-4). I t appears as though no definite conclusions have been reached as to the credibility of this idea; however, the balance of evidence reported to date is negative. A recent encouraging report by Myers and Melchior indicates that drinking behavior in rats can be influenced by ventricular perfusion of TIQ’s in brain (5).There are also reports that TIQ‘s form during various drug therapies (e.g., E-Dopa treatment of Parkinson’s disease) and may play some role in the pharmacological effects of these drugs (6). Present address Battelle Columbus Laboratories, 505 King Avenue, Columbus, Ohio 43201. 530
ANALYTICAL CHEMISTRY, VOL. 49, NO. 4, APRIL 1977
It has long been recognized that available analytical methods for catecholamines are quite poor, especially when dealing with endogenous levels in biological samples. Since TIQ levels will always be significantly less than their catecholamine precursors, it is obvious that their assay will be even more difficult. This is indeed the case and has hampered workers in this area to a large extent. Recently we have described a simple and sensitive assay procedure for urinary and tissue catecholamines (7,8).This method utilizes high performance liquid chromatography (HPLC) coupled to a thin-layer electrochemical detector (LCEC). The advantages of using liquid chromatography as opposed to gas-liquid and other forms of chromatography or fluorometry to assay catecholamines are obvious and have been adequately described by us and others (7,9).The same advantages apply to the present TIQ assay. The fundamental problem with HPLC analysis of TIQ’s is the relatively poor sensitivity of the commonly used detectors (e.g., UV-VIS absorption). The use of thin-layer electrochemistry as a detection system has numerous advantages in terms of sensitivity, selectivity, and cost. The sensitivity using electrochemical detection can sometimes be two to three orders of magnitude better than for UV and its inherent selectivity makes it a much more feasible approach for assays of many biological materials. A fundamental advantage (in many cases a limitation) of LCEC is the fact that only electroactive compounds can be detected. For assay of TIQ’s, this turns out to be a great advantage since all the compounds of interest contain at least one phenolic substituent, thus making them electrooxidizable. Figure 1illustrates the electrochemical behavior of two typical TIQ’s. By selecting a detection potential in the range of +0.7 to +0.8 V, one can oxidize the TIQ’s (and catecholamines) while not detecting compounds with much larger oxidation potentials. Using the LCEC technique, we have detected the presence of TIQ’s in plant matter and in the urine of individuals following consumption of bananas or cocoa-based products (10, 11).This report describes analytical procedures for ppb levels of TIQ’s in body fluids and tissues. Several preliminary studies on TIQ formation in vivo are also discussed. EXPERIMENTAL The present assay for urinary TIQ’s is a modification of our earlier procedure for Catecholamines ( 7 ) .Urine was collected over acid (to pH 2) using 6 M HCl and stored at -35 “C prior to analysis. Four milliliters of acid-hydrolyzed (7) urine were placed in polyethylene centrifuge tubes with 1.5 g of (NHJzS04, capped and shaken, and centrifuged to remove solids and precipitated proteins. Acid hydrolysis is necessary to release the alkaloids from sulfate and glucoronide conjugation. The supernate was transferred to 12-mL glass centrifuge tubes and extracted twice with ethyl acetate and once with hexane (5-mL aliquots). The nonaqueous layers were discarded. The aqueous layer was transferred to a 10-mLbeaker with 100 gL each of 5% sodium metabisulfite and 10%EDTA. The stirred sample was then adjusted to pH 8.5 and placed in a 5-mL conical vial containing 80 mg of alumina (12) and shaken for 12 min on a reciprocating shaker. The urine was aspirated off and the alumina was washed three times with distilled water and dried in vacuo for 3 min at 30 “C. The compounds were eluted from the alumina with 400 KLof 1 M
DA
/7
0.4
0.2
1nA
1
THP
A
I
-
20
-
40
-
60
B
I
I
I
20
10
0
MINUTES CIA
Flgure 1. Cyclic voltammetry of
tetrahydroisoquinolines
LCEC chromatogram of an extract from acid-hydrolyzed human urine (spiked)
Flgure 2.
= 200 mV/s: 1 mM in 0.1 M HCI04; 3-mm diameter carbon paste electrode. (A) Salsoline (l-methyi-6-methoxy-7-hydroxy-TlQ). (6) Salsolinol (l-methyl6,7-dihydroxy-TIO)
NE, 305 ng/mL; EPI, 180 ng/mL; DA, 720 ng/mL; SAL, 200 ng/mL; and THP, 200 ng/mL. Detector potential: 0.72 V vs. Ag/AgCI. Column: Vydac SCX, 50 cm. Mobile phase: Mcllvaine buffer, pH 3.8
acetic acid. After shaking for 10 min, the eluate was assayed using a liquid chromatographequipped with an electrochemicaldetector.The system was modified from that previously described by replacing the syringe injection port with an external-loop nonmetallic valve. This modification was necessary to eliminate an artifactual peak arising from the syringe needle. A pellicular Vydac SCX stationary phase was dry packed in a 50 cm X 2 mm i.d. glass column. The mobile phase consisted of a mixture of 480 mL of 0.1 M citric acid, 320 mL of 0.2 M NaZHP04, and 200 mL of distilled deionized water. A drop of toluene per liter was included to prevent microbial growth. The detector potential was maintained at +0.7 V vs. Ag/AgCl and the flow rate was 0.4 mL/min. Tissue TIQ’s. Brain TIQ’s and catecholamines were assayed as follows. Male Wistar rats, 250-350 g, were anesthetizedwith ether and killed by neck fracture. The whole brain was rapidly removed and frozen on dry ice. The brain was weighed and then homogenized in 6 mL of a solution 0.5 M in HCI, 0.1 M in HClOd, and 1%in sodium metabisulfite. The homogenate was centrifuged at 2000 X g for 10 min at 0 “C and the supernate was stored at -35 “ C prior to assay. The tissue supernate was assayed for TIQ’s and catecholamines in the same manner as described above for urine. Animals. Male Wistar rats, 250-350 g, were housed in metabolic cages and given a totally liquid diet containing 0.25 g Sustagen (Mead Johnson and Co.) per mL. Rats were given the diet as either a 10% ethanol solution or as a pure water solution containing an isocaloric amount of sucrose. The use of a totally liquid diet maximizes the quantity of ethanol which rodents consume (13). Urine specimens (24 h) were collected into a glass vial containing 0.5 mL 6 M HC1 and frozen at -35 OC prior to analysis. L-Dopa was administered orally as a suspension in soy oil.
completely resolved from each other as well as from their catecholamine precursors. T h e detection limit for THP and SAL is approximately 2 ng/mL in urine using this procedure and n o interfering components are found in human or r a t urine samples. Rats given t h e 10% alcohol liquid diet for 25 days excreted n o measurable amount of free or conjugated THP or SAL even though they consumed 8-13 g/kg of ethanol per d a y during that period. These results d o not, however, disprove the hypothesis that alkaloids are formed during alcohol intoxication for several reasons. If formation of the TIQ’s takes place, the compounds would undoubtedly be present in or near t h e nerve endings of t h e peripheral or central nervous system where catecholamines are in highest concentration. TIQ’s and catecholamines compete for a very efficient uptake mechanism in the nervous tissue ( 4 ) and t h u s t h e TIQ’s t h a t form will be protected from metabolism after being taken u p by t h e nerve endings. Upon stimulation of the nerve, the TIQ’s and catecholamines will be released and will either be re-uptaken or metabolized (primarily by catechol-0-methyl transferase) with very small quantities of the free TIQ’sbeing released into circulation and eventually excreted. For this reason our inability to detect unmetabolized TIQ’s does not rule out their formation in vivo. A more satisfactory means of studying T I Q formation is by use of excised tissue samples, especially brain, where quantities of t h e TIQ’s could be stored. A gas chromatographic procedure developed by Collins and Bigdeli (14) is very sensitive toward salsolinol, but much less sensitive toward THP. Their procedure is useful for monitoring salsolinol in whole brain or brain parts. However, t h e need for extensive sample preparation (lyophilization, derivatization, etc.) makes t h e procedure time consuming a n d
Y
RESULTS AND DISCUSSION Figure 2 illustrates a chromatographic separation of several TIQ’s and catecholamines which were extracted from human urine, Tetrahydropapaveroline ( T H P ) and salsolinol (SAL), two TIQ’s hypothesized to form during alcohol intake are
ANALYTICAL CHEMISTRY, VOL. 49, NO. 4, APRIL 1977
531
NL DA
SAL
20
IO
0
MINUTES
Flgure 4, Chromatogram of an add-hydrolyzed urlne extract from a rat given 100 mg/kg L-Dopa. Condltlona same as Flgure 3 32
24
16
8
0
MlNUllS
Flgure 3.
LCEC chromatogram of an extract from rat braln containlng
35 ng/g aalsollnol Detector potentlal: 0.72 V vs. Ag/AgCI. Column: Vydac SCX, 50 cm. Moblle phase: Mcllvaine buffer, pH 3.6
difficult to perform quantitatively, The present technique eliminates many of the sample preparation steps and is remarkably sensitive toward both TIQ’s and catecholamines. The LCEC technique can permit quantitation of concentrations of alkaloids (e.g., T H P and salsolinol) at a level of 10 pmlg in whole rat brain. Figure 3 illustrates the separation of salsolinol at a level of 36, ng/g in whole rat brain. Using this procedure, dopamine, norepinephrine, THP, and salsolinol can be determined simultaneously and quite rapidly (20-30 samples/day). Eight rats (four in each of two separate trials) were given a 10%alcohol liquid diet for 25 days during which time they consumed 8-13 g/kg/day of ethanol. The brains of these animals were assayed for TIQ’s and catecholamines by the previously described procedure. Catecholamine levels were not significantly different from nonalcoholic controls and neither salsolinol nor THP was detectable (Le., less than 2 ng/g brain tissue). Previous work has shown salsolinol to be formed in small quantities in rat brain when large acute doses of ethanol are administered (only when acetaldehyde levels are elevated by administration of pyrogallol) but chronic alcohol studies have always given negative results (15).Such negative findings are not conclusive proof that alkaloid formation is not an important process since it is conceivable that small amounts do form in one discrete brain region giving rise to a pharmacological effect but not contributing a substantial quantity of alkaloid to whole brain studies. The excellent detection limit of the LCEC technique combined with the relatively convenient sample workup procedure make it ideally suited to studies of alkaloid formation in small tissue samples. There have been reports that certain drugs may react in vivo to form TIQ’s. Urinary excretion of salsolinol and THP during L-Dopa therapy has been reported in humans (6)although the quantitative aspects of these experiments have been brought into doubt (16).Rats were administered a 100 mg/kg dose of L-Dopa and 24-h urine samples were collected. Figure 4 shows a typical chromatogram of a urine extract from animals given L-Dopa. Both THP and salsolinol were present as was an additional component having a retention time identical to 6,7-dihydroxytetrahydroisoquinoline(DHTIQ). DHTIQ is 532
ANALYTICAL CHEMISTRY, VOL. 49, NO. 4, APRIL 1977
rapidly synthesized from the reaction of formaldehyde with dopamine and its formation in vivo is quite reasonable, since endogenous sources of methanol (a precursor of formaldehyde) have been reported (171, The formation of DHTIQ in vitro has been recently reported (18) but its existence in vivo has not been previously demonstrated. Amounts of THQ, salsolinol, and DHTIQ excreted were quite variable but were on the order of 2 to 5 pglday for a 100 mglkg dose of L-Dopa. The significance of these amounts of TIQ’s in L-Dopa therapy remains to be determined.
CONCLUSION A simple liquid chromatographic technique provides a convenient, sensitive means for monitoring tetrahydroisoquinoline alkaloid formation in vivo. This technique provides workers in the field with a means to study the role of these compounds in alcoholism and drug therapy in a much more effective manner than was heretofore possible. Recent work with microparticle reverse-phase CIS packing materials indicates that improvements in chromatographic efficiency will permit significantly better sensitivity and resolution. The reverse-phase approach is ideally suited to amperometric detection as has been demonstrated for the metanephrines (19) and the catecholamines (20). ACKNOWLEDGMENT The authors thank Michael Collins, Gerald Cohen, and Henry Weiner for helpful comments and encouragement. LITERATURE CITED (1) V. G. Davis and M. J. Walsh, Science, 167, 1005 (1970). (2) 0.Cohen and M. Collins, Science, 167, 1749 (1970), (3) C. A. Collins, J. L, Cashaw, and V. E. Davis, Biochem. Pharmcob, 22, 2337 (1973). (4) 0. Cohen, In “Alcohol Intoxication and Withdrawal: Experimental Studies I”, M. M. Gross, Ed., Plenum Press, New York, N.Y., 1973, p 33, (5) R. D. Myers and C. L. Meichior, Sclence, In press. (6)M. Sandler, S. 8. Carter, K. R. Hunter, and G. M. Stern, Nature (London), 241, 439 (1973). (7) P. T. Kisslnger, R. M. Rlggin, R. L. Alcorn, and L. D. Rau, Blochem. Med., 13. 299 11975). (8) C. Refshauge,P. T. Kisslnger, R. Drelllng, L. Blank, R. Freeman,and R . N. Adams, Llfe Scl., 14, 311 (1974). (9) I. Moinar and C. Horvath, Clln. Chem. (Wlnston-Sa/em, N.C.), 22, 1497 (1976). (10) R . M. Rlggln, M. J. McCarthy, and P. T. Klssinger, J. Agrlc. Food Chem., 24, 189 (1976). (11) R . M. Riggln and P. T. Kissinger, J. Agric. Food Chem., 24, 900 (1976). (12) A. H. Anton and D. F. Sayre, J. Phafmacol. Exp. Ther., 136,360 (1962). (13) 0. Freund, Arch. Neurol. (Chicago), 21, 315 (1969). (14) M. A. Collins and M. G. Bigdeli, Llfe Scl., 16, 585 (1975).
(15)M. A . Collins and M. G. Bigdeli, in “Alcohol Intoxication and Withdrawai:
Experimental Studies Ii”, M. M. Gross,Ed., Plenum Press, New York, N.Y.,
1975.D 79. (16)G. Cotien, Biochem. pharmacol., 25, 1123 (1976). (17)J. Axelrod and J. Daly, Science, 150,892 (1965). (18) W. J. A. Vanderheuvel, V. F. Gruber, L. R. Mardei, and R. W. Walker, J. Chromatogr., 114, 476 (1975). (19)R. E. Shoup and P. T. Kissinger, Clin. Chem. ( Winston-Salem, N.C.),sub-
mitted.
(20) R. M. Riggin and P. T. Kissinger, Clin. Chem.( Winston-Salem,N.C.),sub-
mitted.
RECEIVEDfor review November 11,1976. Accepted January 11, 1977, This work was supported by the National Institute Of Sciences, the Science Foundam tion, and the Showalter Trust Fund.
Preparative Gel Permeation Chromatographic Separation of Solvent Refined Coal W. M. Coleman, D. L. Wooton, H. C. Darn,* and L. T. Taylor* Department of Chemistry, Virginia Polytechnic institute and State Universlty, Blacksburg, Va. 2406 1
A preparative quantltatlve Separation of THF- and CHCIasoluble solvent reflned coal (Pittsburgh No. 8) has been demonstrated employlng gel permeatlon chromatographic techniques utlllzlng three column packlngs: a styrene-dlvlnyl benzene packlng (Blo-Beads S-X4), a cross-linked poly(Nacryloylmorphollne) polymer (Enzacryl Gel K-l), and a modlfled alkylated dextran (Sephadex LH-20). Each packlng material Is evaluated based on the extent and time of separatlon and the cost of materlals. The percent recovery In the Blo-Bead case Is greater than 95%. Average molecular weights vla vapor phase osmometry on the four slzed fractions of THF-soluble SRC from Blo=Beadssuggests that a separatlon accordlng to molecular welght has been achieved. The description of a short tapered glass column to economlcally carry out these preparatlve separatlons Is descrlbed.
The search for materials having pore sizes that would permit separation of high molecular weight substances by effective molecular size has been extensive and has given rise to the widely used technique of gel permeation chromatography (GPC) (1).A variety of materials from numerous manufacturers have been developed to effect separations via GPC (2). In many instances rather complex mixtures have been efficiently size-separated on an analytical scale (3).These separations are not without some problems, not the least of which is one related to cost ( 4 ) . One, therefore, is prohibited from even considering preparative-size separations in many situations. In this light we wish to report the results of an evaluation of three GPC packing materials for the separation of coal liquids. A preparative-size separation of a solvent refined coal ( 5 ) (SRC) is also described employing one of these packing materials at a cost of $&lO/column including column and packing material. The SRC process is one of several research pilot plant operations under way for converting coal to a low sulfur, high Btu content fuel (6). Samples from the pilot plant include solid SRC product, process solvent, and light organic liquid product. The solid product is the material under examination in this report.
EXPERIMENTAL The SRC solid product on which our separations were accomplished was obtained from an Electric Power Research Institute funded Southern Services Inc. pilot plant operated by Catalytic Inc. a t Wilsonville, Ala. A bituminous coal, Pittsburgh No. 8, served as the SRC source. The liquid chromatograph used was a Spectra-Physics Model 3500 B equipped with a thermostated refractive index detector. Bio-Beads
S-X4 (200-400 mesh) and S-X4-400 (25 rm) were obtained from Bio-Rad Laboratories, Rockville Centre, N.Y. Enzacryl Gel K-1, medium, was obtained from Aldrich Chemical Co., Milwaukee, Wis. Sephadex LH-20 was obtained from Sigma Chemical Co., St. Louis, Mo. Each column was packed with either a T H F or CHCls slurry of the material (Figure 1) using an occasional 20 psi Nz to facilitate packing. The glass column, Figure 2, was prepared by the glass fabrication facility of this Department. Tetrahydrofuran and chloroform were obtained from Burdick & Jackson Laboratories, Muskegon, Mich., and were used as received. Molecular weight determinations via vapor phase osmometry were performed by Galbraith Microanalytical Laboratories, Knoxville, Tenn.
RESULTS AND DISCUSSION Large scale pilot plants for the liquefaction of coal are currently in operation. Since coal liquids are expected to become commercially available during the next decade, it is important to establish their suitability as sources of petrochemicals and power generation fuels. A thorough characterization of such materials by a variety of physicochemical probes (Le., I3C NMR, lH NMR C-H-N-S and trace metal analysis, molecular weight analysis, mass spectrometric analysis, etc.) normally requires gram quantities of material. Since the number of compounds in a coal liquid sample is staggering, a preliminary separation of the material into numerous fractions based on functionality, polarity, or molecular size is highly desirable prior to any type of characterization. The scale for GPC separations varies widely. For analytical determinations, sample sizes are small (1-2 mg); whereas, for preparative-size separations, larger quantities may be employed (0.5-1.0 g). The application of GPC to polymer chemistry is rather routine (7);however, its use in coal research has been mostly limited to analytical separations (8). For example, a supercritical-gas (toluene) extract, comprising 17%of a low-rank coal, has been separated (9) by a combination of solvent fractionation and both silica-gel and gel permeation (Sephadex-LH20) chromatography. The n-hexane solubles extracted by Soxhlet extractor of a Yahari coal subject to mild hydrogenation conditions are reported to be separable into 15 fractions via GPC (IO);however, the elution did not occur in the order of high to low molecular weight. The THF-soluble product after two alkylation steps with an Illinois coal is reported to be separable into 40 fractions via GPC with quantitative recovery (11);however, these workers do not reveal the conditions under which the separation was achieved. It should be pointed out that the employment of GPC to separate petroleum-type products has been more extensive (12) both on an analytical and a preparative scale. ANALYTICAL CHEMISTRY, VOL. 49, NO. 4, APRIL 1977
533
