Charles N. FalanyS, Mitchell D. Green, and Thomas R. Tephly. From the ... (1) and 1-naphthol (2) as aglycone acceptors in rat hepatic or intestinal ... Enzymatic Mechanism. A of Glucuronidation. B o 041. 0.03i. O"1. 7. 1219. /. 0 o,w, v-. W)". '1.
THEJOURNAL OF BIOLOGICAL CHEMISTRY 0 1987 by The American Society of Biological Chemists, Inc.
Vol. 262, No. 3, Issue of January 25, pp. 1218-1222,1987 Printed in V.S.A.
The Enzymatic Mechanism of Glucuronidation Catalyzed by Two Purified Rat Liver Steroid UDP-Glucuronosyltransferases* (Received for publication, August 7, 1986)
Charles N. FalanyS, Mitchell D. Green, and Thomas R. Tephly From the Department of Pharmacology, University of Iowa, Iowa City, Iowa 52242
A kinetic analysisof two homogeneous rat liver ste- such as p-nitrophenol and 1-naphthol are concerned (5). In roid (3a-hydroxysteroid and 178-hydroxysteroid) addition, more than one UDP-glucuronosyltransferase may UDP-glucuronosyltransferases was conducted using catalyze the glucuronidation of certain endogenous combisubstrate kinetic analysis,product inhibition studies, pounds, as has been suggested for androsterone (6). In many and dead-end competitive inhibition studies. Double cases where more than one enzyme catalyzes the glucuronide reciprocal plots of initial velocityuersus substrate con- conjugation of a given substrate, the apparent kinetics for centration, using bisubstrate kinetic analysis, gave a that substrate are different with the separate enzymes (5). sequential mechanism. Product inhibition studies were compatible with either a rapid equilibrium, random- This could be expected to complicate the determination of order kinetic mechanism or an ordered Theorell- the mechanism of glucuronidation in microsomal preparaChance mechanism. Results of dead-end competitive tions. The purpose of this study was to determine the enzymatic inhibition studies excluded an ordered TheorellChance mechanism. The cumulative results are con- mechanism of glucuronidation catalyzed by two homogeneous sistent witha rapid equilibrium random-order sequen- rat liver UDP-glucuronosyltransferases. 3a-Hydroxysteroid tial kinetic mechanism for the glucuronidation of tes- UDP-glucuronosyltransferase catalyzes the glucuronidation tosterone by purified 178-hydroxysteroid UDP-glucu- of the 3a-hydroxyl position of steroids and bile acids (7), as ronosyltransferase and of androsterone by purified 3a- well as the 3-hydroxy position of short chain bile acids (8, 9) hydroxysteroid UDP-glucuronosyltransferase. and certain arylamines (10). 17P-Hydroxysteroid UDP-glucuronosyltransferase catalyzes the conjugation of 17P-hydroxysteroids, p-nitrophenol,and1-naphthol (5). The reaction mechanisms of these two steroid UDP-glucuronosyltransferThe conjugation of endogenous and exogenous compounds ases were determined using androsterone and testosteroneas with glucuronic acid is catalyzed by UDP-glucuronosyltrans- the aglycone substrates, since these compounds represent ferases (UDP glucuronate P-D-glucuronosyltransferase (ac- important physiologic substrates for these enzymes. ceptor-unspecific), EC 2.4.1.17). These enzymes catalyze the EXPERIMENTALPROCEDURES transfer of glucuronic acid from uridine 5’-diphospho-a-~Materials-[4-14C]Testosterone (50 mCi/mmol) and [1,2-3H]anglucuronic acid to a nucleophilic site on a suitable aglycone acceptor molecule. Conjugation of the aglycone usually dimin- drosterone (40-60 Ci/mmol) were purchased from New England Nuclear. UDP-glucuronic acid (ammonium salt), testosterone, anishes its biological activity and enhances its elimination from drosterone, uridine 5’-disphosphate, uridine 5’-monophosphate, testhe body. tosterone-8-D-glucuronide,androsterone-@-glucuronide, epiandrosSeveral laboratories have investigated the enzymatic mech- terone (5n-androstan-3B-ol-17-one), epitestosterone (4-androstenanism of glucuronidation using various substrates in micro- 17a-ol-3-one), and L-a-phosphatidylcholine (eggyolk, type XI-E) somal preparations from different species. Using morphine were obtained from Sigma. All other chemicals wereof analytical (1)and 1-naphthol (2) as aglycone acceptors in rathepatic or grade. Enzyme Preparations and Assays-3n-Hydroxysteroid and 178intestinal microsomal preparations, respectively, the conjuhydroxysteroid UDP-glucuronosyltransferaseswere purified to hogation reaction was reported to have an ordered sequential mogeneity from hepatic microsomal preparations of female Spraguemechanism. In guinea pig hepatic microsomes, using p-nitro- Dawley rats using chromatofocusing and affinity chromatography (5). phenol as substrate, arapid equilibrium random-order mech- Protein determinationswere performed using the method of Bradford anism was obtained (3). In contrast,bilirubin glucuronidation (11)with a Bio-Rad protein assay kit. Androsterone and testosterone glucuronidation assays were conin guinea pig liver microsomal preparations was determined to exhibit an ordered sequential mechanism (4).It is now ducted as previously described (5, 12). Variable concentrations of appreciated, at least in rat liver, that some UDP-glucurono- UDP-glucuronic acid and steroid substrates were added as described under “Results.” When variable concentrations of radioactive steroid syltransferases exhibit overlapping substrate specificities, es- were used, unlabeled steroid was added to the reaction mixture after pecially where the conjugation of small phenolic substrates the assay had been terminated, but prior to the extraction step, to
* This work was supported by National Institutes of Health Grant GM26221. Part of this work was published in abstract form (Green, M. D., Falany, C. N., and Tephly, T.R. (1986) Fed. Proc. 45, 934) and presented in St. Louis, MO at the annualmeeting of the Federation of American Societies for Experimental Biology in April, 1986. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. $Present address: Dept. of Pharmacology and Cancer Center, University of Rochester, Rochester, NY 14627.
bring the concentration of the steroid in the mixture to 50 pM for testosterone and 56.25 p~ for androsterone. This was done to prevent differential extraction efficiencies during the extraction step caused by different steroidconcentrationsin the reaction mixtures. All reactions were performed a t 37 “C in the presence of 100 pgof phosphatidylcholine. This amount of phospholipid provides for maximal enzymatic rates of glucuronidation of the substrates used in this study. For product inhibition studies and dead-end inhibition studies, the inhibitors were added to the reaction mixtures as aqueous solutions (UDP, UMP, and steroid glucuronides) or indimethyl sulfoxide (epi-steroids). When the epimeric steroids were used, the concentration of dimethyl sulfoxide in the reaction mixture was 1 2 % and was
1218
1219
Enzymatic Mechanism of Glucuronidation
B
A
o
041
0.03i
O"1
7
0
o,w,
v-
/
'1
W)"
FIG. 1. Bisubstrate kinetic analysis of purified 3a-hydroxysteroid UDP-glucuronosyltransferase. Androsterone (ANDRO) glucuronidation activities were determined using variable concentrations of the steroid (3.75-15.0 p ~ and ) different concentrations of UDP-glucuronic acid (UDPGA) (0.1-3.0 mM) using purified 3ahydroxysteroid UDP-glucuronosyltransferase.Each reaction mixture (1.0 ml) contained 1.2 pg of purified enzyme (specific activity: 560 nmol/min/mg), and thereactions were incubated a t 37 "Cfor 10 min in the presence of 100 pg/ml phosphatidylcholine, as described under "Experimental Procedures." Double reciprocal plots of initial velocity versus substrate concentration are presented in panels A and B. Each point represents the mean value obtained from three reaction mixtures. Below each panel are graphs which represent the appropriate secondary plots of intercepts and slopes for the datashown above. Velocity is expressed as nmol of glucuronide formed/min/ mg of protein. also included in the control reactions. Kinetic Analysis--Reciprocal initial velocities were plotted graphically against reciprocals of the substrate concentrations. In allcases, a linear relationship was obtained. The methods of Cleland (13,14) were used for the determination of Michaelis constants andinhibition constants.
the reaction mechanism of Sa-hydroxysteroid and l7p-hydroxysteroid UDP-glucuronosyltransferases. Nonsaturating substrate concentrations were used because of the low solubility of the steroids in thereaction mixtures. The appropriate steroid glucuronide and UDP were used as theproduct inhibitors. RESULTS The results of product inhibition studies for 3a-hydroxysteBisubstrate Kinetics-Bisubstrate kinetic analysis of puri- roid UDP-glucuronosyltransferase are presented in Fig. 3. At fied 3a-hydroxysteroid UDP-glucuronosyltransferase was a constant concentration of UDP-glucuronic acid, UDP was performed by varying the androsterone concentration in the a noncompetitive inhibitor with respect to androsterone. presence of several concentrations of UDP-glucuronic acid. When UDP-glucuronic acid was varied at a fixed concentraResults from initial velocity studies are shown in Fig. 1. Fig. tion of androsterone, the inhibition by UDP was competitive. lA is a Lineweaver-Burk plot for androsterone glucuronida- Androsterone glucuronide inhibition was competitive with tion at different UDP-glucuronic acid concentrations. In Fig. respect to androsterone and noncompetitive with respect to lB, thesame data areplotted as function a of UDP-glucuronic UDP-glucuronic acid. These patterns of product inhibition acid at different concentrations of androsterone. The lines are suggestive of a Theorell-Chance bireactant mechanism or converge to the left of the vertical axis. This pattern is of a rapid equilibrium random-order mechanism with abortive characteristic of a sequential mechanism in which both reac- complex formation (16). Productinhibitorstudies using testosterone glucuronide tants must add to the enzyme before any products can be released (15).Similar data were obtained from the bisubstrate and UDP were also conducted using preparations of homokinetic analysis of 17&hydroxysteroid UDP-glucuronosyl- geneous 17/3-hydroxysteroid UDP-glucuronosyltransferase. transferase when testosterone was used as the aglycone (Fig. The results of these experiments are summarized in Table 11. 2). Secondary plots of intercepts and slopes were used to Testosterone glucuronide was determined to be a competitive determine the kinetic constants: KUDp.glucumnic acid, Ksteroid, inhibitor with respect to both testosterone and UDP-glucuronic acid. UDP was a competitive inhibitor with respect to Ki stemid and Ki ~ ~ p .acid.~ These l ~ results ~ ~ are ~ summarized ~ i ~ in Table I. The K u D P - acid ~ Iand ~ ~Kstemid ~ ~ ~ obtained ~ in this UDP-glucuronic acid, but was a noncompetitive inhibitor study using bisubstrate kinetic analysis are very similar to toward testosterone. These patternsof product inhibition are consistent with a Theorell-Chance reaction mechanism, a the apparentK,,, values reported previously (5). Product Inhibition Studies-Studies were conducted at rapid equilibrium ordered mechanism with abortive complex nonsaturating substrate concentrations to further elucidate formation, or a rapid equilibrium random-order reaction
Enzymatic Mechanism of Glucuronidation
1220
S O YY UDCOA 200 W UDCOA $00 YY UDCOA ZOO0 W UDCGA
aQUYAW4 7
-0.I
‘1
“‘“1
m0”
!A=0”
FIG.2. Bisubstrate kinetic analysis of purified 17g-hydroxysteroidUDP-glucuronosyltransferase.. Testosterone (TEST)glucuronidation activities were determined using variable concentrations of the steroid (5.050.0 PM) and different concentrations of UDP-glucuronic acid (UDPGA) (0.1-5.0mM) using purified 178hydroxysteroid UDP-glucuronosyltransferase.Each reaction mixture (1.0 ml) contained 1.0 pg of purified enzyme (specific activity: 320 nmol/min/mg), and the reactions were incubated at 37 “Cfor 10 min in the presence of 100 pg/ml phosphatidylcholine, as described under “Experimental Procedures.” Double reciprocal plots of initial velocity versus substrate concentration are presented in panels A and B . Each point represents the average value obtained from two reaction mixtures. Below each panel are graphswhich represent the appropriate secondary plots of intercepts and slopes for the datashown above. Velocityis expressed as nmol of glucuronide formed/min/mg of protein. TABLEI Kinetie constants of 30-hydroxysteroid and 17~-hydroxysteroid U D P - ~ l u c u r o n o s y l t r a n r ~(UDPGT) es Constant
Description
Kinetic constants 3a-Hydroxyste- 178-Hydroxysteroid UDPGT roid UDPGT PM
K, Kb
K,
340 K- for UDPGA” K,,, for steroid
Dissociation220 constant of E . UDPGA complex Kib Dissociation constant of E . steroid complex Kip Dissociation constant of E . steroid glucuronide complex Kk Dissociation constant of E . UDP comdex a UDP-glucuronic acid.
530 13 210
38
5
25
80
167
1
7
mechanism with abortive complex formation (16). Kinetic constants obtained from the product inhibition studies are given in Table I. Dead-end Inhibition Studies-Dead-end competitive inhibition studies were performed in order to distinguish between a Theorell-Chance, a rapid equilibrium ordered, or a rapid equilibrium random-order sequential mechanism. If the reaction mechanism were ordered or if a Theorell-Chance mechanism pertained, then, one uncompetitive inhibition pattern should be observed in the presence of a dead-end inhibitor (17). Competitive dead-end inhibitors, when bound to the
active site of the enzyme, do not allow product formation to occur (17). UMP does not participate in the glucuronidation reaction, but has been shown to be a competitive inhibitor of UDP-glucuronic acid in guinea pig hepatic microsomal preparations (4).We have previously reported that the epimeric steroids 3p-hydroxyandrosterone and 17a-hydroxytestosterone inhibit steroid conjugation catalyzed by 3a-hydroxysteroid and 17P-hydroxysteroid UDP-glucuronosyltransferases, but are not substrates for these enzymes (18).UMP and the epimeric steroids were employed and found to be competitive dead-end inhibitors toward either UDP-glucuronic acid or the steroid substrates (Fig. 4 and Table 11). UMP was found to be a . noncompetitive inhibitor with respect to androsterone for glucuronidation catalyzed by purified 3a-hydroxysteroid UDP-glucuronosyltransferase (Fig. 4B).Androsterone glucuronidation was also noncompetitively inhibited by epiandrosterone when UDP-glucuronic acid was the varied substrate. Epitestosterone was found to be a noncompetitive inhibitor with respect to UDP-glucuronic acid for the 17p-hydroxysteroid UDP-glucuronosyltransferase (Table 11). When testosterone was the varied substrate, noncompetitive inhibition was observed in the presence of UMP. The results of these dead-end competition inhibition studies are consistent with a random sequential reaction mechanism, but not with an ordered sequential or aTheorell-Chance reaction mechanism. DISCUSSION
The reaction mechanism for both 3a-hydroxysteroid and 178-hydroxysteroid UDP-glucuronosyltransferases, purified
1221
Enzymatic Mechanismof Glucuronidation A
-0.1
0
0.0
0:1
0.2
VANDRO @A)-l
0.02,
/’
FIG.3. Produet inhibition studies. Product inhibition studies using homogene& 3ru-hydroxysteroid UDPglucuronosyltransferase were conducted under nonsaturating concentrationsof the nonvaried substrate. Reaction mixtures (1.0 ml) contained 1.0-1.2 pg of purified enzyme (specific activity: 450-600 nmol/min/mg), and the reactions were incubatedat 37 ‘C for 10 min in the presenceof 100 pg/ml phosphatidylcholine,as described under “Experimental Procedures.” In panels A and C, the concentration of UDP-glucuronic acid was held constant (5 mM), andthe concentration of androsterone was varied. Inpanels B and D, the concentration of androsterone was constant (56.3 p ~ ) and , the concentration of UDP-glucuronic acid was varied. Each point represents the mean value obtained from three reaction mixtures and are presented as double reciprocal plots of initial velocity versus substrate concentration.Velocity is expressed as nmol of glucuronide produced/min/mg of protein. TABLEI1
of UDP-glucuronosyltransferase in guinea pig liver is not known. In rats, however, phenolphthalein and bilirubin glucuronidation have different developmental patterns (19, ZO), suggesting that these substrates may be conjugated by differSubstrate ent enzymes. Investigations of the glucuronidation reaction Testosterone UDP-glucuronic Inhibitor mechanism in rat liver microsomal preparations, using moracid phine as the aglycone (l),and in rat intestinal microsomal Testosterone glucuronide Competitive Competitive preparations (2), using 1-naphthol as the aglycone, have reoncompetitive ve UDP ported ordered sequential reaction mechanisms. In both reEpitestosterone Noncompetitive Competitive ports, however, complete product inhibition studies were not UMP Noncompetitive Competitive performed, nor were the results of complete studies using dead-end competition inhibitors or isotope exchange studies from rat liver microsomes, was determined to be a sequential reported. substrate binding mechanism on the basis of bisubstrate P both steroid UDPOur results show that the K ~ D for kinetic analysis when androsterone or testosterone, respecglucuronosyltransferases is very low. These data would sugtively, was used as theaglycone acceptor molecule. Thus, both gest that these enzymes are very sensitive to theintracellular UDP-glucuronic acid and thesteroid must bind to theenzyme concentration of UDP and the energy charge of the hepatobefore any products are released. Product inhibition studies cyte. It has recently been shown that compounds which are alone were inconclusive for determining whether the substrates bound in an ordered or random manner. However, the extensively glucuronidated (such as valproic acid, acetuse of epimeric steroids and UMP, as dead-end competitive aminophen, and salicylamide) can cause a depletion of hepatic inhibitors of the enzymatic activity, gave inhibition patterns UDP-glucuronic acid levels in vivo (21, 22). This may result which support a random binding order of the substrates. The in a decrease in in vivo glucuronidation due to an increase in data, therefore are consistent with a random sequential rapid intracellular UDP levels. Similarly, agents such as ethanol equilibrium reaction mechanism for the glucuronidation of (23, 24) andanesthetics (25, 26), which indirectly deplete steroids by these enzymes. Vessey and Zakim (3) have also hepatic UDP-glucuronic acid levelscould, by altering the reported that theglucuronidation of p-nitrophenol in hepatic energy charge of the cell, cause increases in the intracellular microsomal preparations from guinea pig and cows exhibited concentration of UDP and inhibit glucuronidation. It is una rapid equilibrium random sequential reaction mechanism. clear whether a decrease in UDP-glucuronic acid or a possible is the most These conclusions were based on product inhibition studies increase in intracellularUDPconcentrations to decreases in glucuronide using UDP and p-nitrophenol glucuronide, as well as isotope importantfactorcontributing exchange studies to rule out aTheorell-Chance reaction mech- formation because UDP levels have not been determined in UDP-glucuronic acid-depleted livers. anism. This is the firststudy to investigate the reaction mechanism Potrepkaand Spratt (4) reported an ordered sequential mechanism for bilirubin glucuronidation in guinea pig hepatic of glucuronidation using homogeneous, purified UDP-glucumicrosomal preparations. This conclusion was based on bi- ronosyltransferases. It is not clear, however, whether this substrate and product inhibition studies. In the product in- reaction mechanism can be extrapolated to all forms of UDPhibition studies, however, phenolphthalein glucuronide was glucuronosyltransferase, or whether the same reaction mechused instead of bilirubin glucuronide. Whether or not phen- anism would hold for all substrates for the same enzyme, olphthalein and bilirubin are conjugated by the same isoform especially if these substrates were to be nonphysiologic comProduct and dead-end inhibition studies on 170-hydroxysteroid UDP-glucuronosyltransferase
1222
Enzymatic Mechanism of Glucuronidation
A
B
O.OM
0.001 0 COWlROL
a
#, -0.m
O.tmY E ~ A M O I O mY W l A I D I O
-0.06
0.06 0.00
0.10
0.S
0.20
-0:m
-0:zo
t/M
C
0
0.010,
0.020
//
0.015
o
mW UMP
0 1
mW UUP
0
-
a 2 nu -d
0
s
10
I
5
0
10
WDPGA (mM) FIG.4. Dead-end competitive inhibition studies using purified Sa-hydroxysteroid UDP-glucuronosyltransferase. Androsterone glucuronidation activities were determined in the presence or absence of UMP or epiandrosterone undernonsaturatingconcentrations of the nonvaried substrate. Reaction mixtures (1.0 ml) contained 1.0-1.2 pg of purified enzyme (specific activity: 450-600 nmol/min/mg), and thereactions were incubated at 37 "C for 10 min in the presence of 100 pg/ml phosphatidylcholine, as described under "Experimental Procedures." In panels A and B , the concentration of androsterone was varied (5.6-56.3 p ~ at) a constant concentration (5 mM) of UDP-glucuronic acid. In p a d s C and D, the concentration of androsterone was constant (56.3 p ~ ) and , the concentrations of UDP-glucuronic acid were varied (0.1-5.0 mM). Each point represents the mean value of three reaction mixtures and are presented as double reciprocal plots of initial velocity uersus substrate concentration. Velocity is expressed as nmol of glucuronide formed/min/mg of protein.
pounds. However, it can be concluded from our study and others (1-4)that thesubstrates for the glucuronidation reaction bind the enzyme in a sequential manner as opposed to a ping-pong mechanism. Acknowledgments-Wewould like to acknowledge the excellent technical assistance of Elizabeth Swain and thehelpful advice of Dr. Yacoub Irshaid. We also wish to acknowledge the help of Arlene Grace in the preparation of the manuscript. REFERENCES
1. Sanchez, E., and Tephly, T.R. (1975)Mol. Pharmacal. 11,613-620 2. Koster, A. S., and Noordhoek, J . (1983)Biochirn. Biophys. Acta 761, 7685
,
UWP
3. Vessey, D. A., and Zakim, D. (1972) J. Biol. Chem. 247,3023-3028 4. Potrepka, R.F., and Spratt,J. L. (1972)Eur. J. Biochem. 29,433439 5. Falany, C. N., and Tephly, T. R. (1983)Arch. Biochern. Biophys. 227,248258 6. Matsui, M., and Nagai, F. (1985)J. Pharmacobio-Dyn. 8,679-686 7. Kirkpatrick, R. B., Falany, C. N., and Tephly, T. R. (1984)J. Biol. Chem. 259,6176-6180 8. Kirkpatrick, R. B., Green, M. D.,,Hagey,L. R., Hofmann, A. F., and Tephly, T. R. (1985)Hepatology (Baltrmore) 5,1009 9. Radominska-Pyrek, A,, Green, M., Lester, R., and Tephly, T. R. (1986) Fed. Proc. 45,933 10. Green, M.D., Irshaid, Y., and Tephly, T. R. (1986)Biological Reaetiue
Intermediates 111 (Kocsis J. J., Jollow, D. J., Witmer, C. M., Nelson, J. O., and Synder, R., eds)'pp. 399-405, Plenum Publishing Corp., New
Ynrk 11. Bradford, M. M. (1976)A d . Biochem. 72,248-254
12. Tukey, R. H.,Billings, R. E., and Tephly, T. R. (1978)Biochem. J. 171,
-"-
RKQ-RR'l """
13. Cleland, W. W. (1967)Adu. Enzymol. 29,l-32 14. Spector, T., and Cleland, W. W. (1981)Biochem. Phurmacol. 30,l-7 15. Rudolph, F. B., and Fromm, H.J. (1983)Contemporary Enzyme Kinetics and Mechanism (Purich. D. L..ed) DD. 53-73. Academic Press. New York 16. Rudolph, F.B. (1983) Contemporaj; Enzyme Kinetics and Mechanism (Purich, D. L., ed) p. 207-232,Academic Press, New York 17. Fromm, H. J. (1983)8ontemporary Enzyme Kznetics andh4echanism (PurAcademic Press, New York ich, D. L., ed) p. 233~251, 18. Falany, C. N., 8rkpatnck. R. B., and Tephly, T. R. (1985)Aduances in Glucuronide Conjugation (Matern. S., Bock K W , and Gerok, W., e&) 41-49,MTP Press Limited, Lancaster &reatBritain 19. W%art, G. J., Mossman, S., Donald, A., and Dutton, G. J. (1977)Biochem Soc. Trans. 6,721-723 20. Fu_chs, M., Rao, M. L., and Breuer, H. (1977)J. Steroid Biochem. 8, 235841
21. Hjel!e, J. J., Hazelton, G . A., and Klaassen, C.D. (1985)Drug Metab. DU s. 13,3541 22. HoweY1,O S. R., Hazelton, G. A., and Klaassen, C. D. (1986)J. Phurmncol. Ex Ther 236,610-614 2:3. Mol&us, P.; Andersson, B., and Norling, A. (1978)Biachem. Phurmncol. 27,2583-2588 2.4. Minnigh, M. B., and Zemaitis, M.A. (1982)Drug Metab. Disps. 10, 183188
25. Eriksson G and Strkth, D. (1981)FEES Lett. 124,39-42 26. Christen&og, P. I., and Eriksson, G. (1985)Acta AnoesthesioL Scand. 29, 629-631
