Synthesis and Benzodiazepine Receptor Affinities of Rigid Analogues of 3-Carboxy-/?-carbolines: Demonstration That the Benzodiazepine Receptor Recognizes Preferentially the s-Cis Conformation

8/12/2019 Synthesis and Benzodiazepine Receptor Affinities of Rigid Analogues of 3-Carboxy-/?-carbolines: Demonstration Th

1/6

J. Med. Chem. 1989, 32, 1799-18 04 1799

and assayed for adenylate cyclase to determine if basal activitywas elevated.

Effects of Com poun ds on Light Em ission. The procedureswere as previously described,8 except that drugs were dissolvedin insect saline containing no calcium and 20 mM manganesechloride. These salt conditions increased the specificity of theassay by blocking any potential effects of drugs on endogenousrelease of octopamine from nerve terminals.

D e t e r m i n a t i o n o f I r r e v e r s i b l e B i n d i n g o f 3H-NC-5Z.Washed light-organ membranes (0.04 mL of 100 mg wet wt/mL)were added to 0.16 mL of a mixture containing (final concentration): 80 mM Tris maleate, pH 7.4; 10 mM theophylline; 8mM MgCl2; 0.5 mM ethylene glycol bis(0-aminoethyl ether)-N,iV,iV',iV'-tetraacetic acid; 2 mM ATP; and 0.1 mM GTP.Displacing agents were added and the mixture was incubated at4 C for 15 min, after which 3H-NC-5Z (17 Ci/mmol) was addedto a final concentration of 1 ix M and the incub ation co ntinuedfor an additional 60 min. The mixture was then transferred toquartz m inicuvettes (1-mm pa th length, total volume 0.25 mL)and photolyzed for 15 min at a distance of 2 cm from a bank ofthree 15-W Sylvania GT E germicidal lamps (G15T8). Cuv etteswere turned over every 2 min during t he expo sure and, followingphotolysis, the co ntents of the cuv ettes were transferred to tubes

0-Carbol ines possess ing a ca rboxyl g roup a t the 3 -pos-ition [for example, ethyl /3-carboline-3-carboxylate 03-CCE,1 ; see Ch ar t I ) ] , a re know n to b in d wi th h ig h a ff in ity tot h e c e n t r a l b e n z o d i a z e p i n e r e c e p t o r s ( B Z R )1 ,2 an d have

been pro posed as endogenous l igands of these recep tors .1,3

M o r e o v e r, s t r u c t u r e - a c t i v i t y s t u d i e s i n t h i s s e r i e s h a v eshown that a carboxyl group implicated in an ester l inkage[e.g., 1, IC5Q (co nc en trat i on of l igand inh ibi t i ng 5 0% oft r i t i a ted benz odiazep ine b ind ing) = 7 nM ]1 d e m o n s t r a t e sa h igher a ff in i ty fo r the BZR t ha n one th a t i s pa r t o f anamide l inkage [e .g. , FG 7142 (2); IC5 0 = 657 n M ] .4 ' 6 O nthe bas i s o f the X-ray c rys ta l s t ruc tures o f a homologousester methyl /3-carboline-3-carboxylate ( /3-CCM, 3)6, 7 a n dam ide [N -e thy l ' /S -carbo l ine-S-carboxamide ( /3 -CEA, 4) ] ,8

f Institut de Chimie des Substances Naturelles.' Laboratoire de Physiologie Nerveuse.5 Laboratoire Gen etique, Neurogen etique et Com portemen t.

containing 10 mL of 6 mM T ris maleate, pH 7.4. The tu bes werecentrifuged at 120000g for 30 min, the pellet was resuspendedin 10 mL of buffer an d washed twice more. Th e final pellet wassolubilized in 1 mL of Protosol (New En gland Nuclea r) for 15h at 30 C, transferred to a scintillation vial, and neutralized with1 mL of IN HC1. Eighteen milliliters of Aquasol-2 (New EnglandNuclear) was then added and the radioactivity was measured byscintillation counting in a Serle Delta 300. In some exp eriments,membranes (along with 0.2 mL of 0.63% bovine serum albuminwas carrier protein) were washed, instead, by five cycles of precipitation with 7.5% cold trichloroacetic acid and solubilizationof the centrifuged pellet with 0.2 mL of 1 M NaOH.

Acknowledgmen t . We wish to thank Edw ard J.Hunnicutt and Dr. Kenneth Kirshenbaum for technicalassistance. This work was suppo rted by USDA 8600090and 88-37153-3617 and by grants from the Upjohn Company and the Katharine Daniels Research Fund.

Re gi str y No . 1, 77472-97-0; 2, 5391-39-9; 3, 579-66-8; 4,86861-31-6; 5, 4751-48-8; 6, 86861-32-7; 7, 86861-20-3; 8,121158-53-0; 9, 121158-54-1; 10, 121158-55-2; 11, 121158-56-3;adenylate cyclase, 9012-42-4.

i t has been proposed by Codding tha t one poss ib le reasonfor this obser ved difference in bin din g aff ini t ies of th e

(1) Braestru p, C ; Nielsen, M.; Olsen, C. E. Proc. Natl. Acad Sci.U.S.A. 1980, 7 7, 2288.

(2) Guzman, F.; Cain, M.; Larscheid, P.; Hagen, T.; Cook, J. M .;Schweri, M.; Skolnick, P.; Paul, S. M. J. Med. Chem. 1984, 27,564.

(3) Pena, C; Medina, J. H.; Novas, M. L.; Paladini, A. C; DeRobertis, E. Proc. Natl. Acad Sci. U.S.A. 1986, 83 , 4952.

(4) Locock, R. A.; Baker, G. B.; Micetich, R. G.; Coutts, R. T.;Benderly, A. Prog Neuro-Psychopharmacol. Biol. Psychiatry1982, 6 , 407.

(5) The IC5Q value of FG7142 (compound 2) which we report inTable II (197 nM) differs somewhat from the litera ture value4657 nM). This may be due to minor differences in the in vitro

procedures used . However, all the values of Table II weredetermined by ourselves under exactly the same experimentalconditions and are consistently reproducible.

(6) Muir, A. K. S.; Codding, P. W. Can J. Chem. 1985, 63, 2752.

Synthesis and Benzodiazepine Receptor Aff ini t ies of Rigid Analogues of3-Carboxy-/?-carbol ines: D em ons trat ion Th at the Be nzod iazepin e Re cepto rRecognizes Pre fe ren t ia l ly the s-Cis Con forma tion of the 3-Carboxy Gr oup

G i l b e r t D o r e y,f G u i l l a u m e P o i s s o n n e t ^ M a r i e - C l a u d e P o t i e r,1 L i a P r a d o d e C a r v a l h o , ' P a t r i c e Ve n a u l t ,8

G e o rg e s C h a p o u t h i e r, Jean Ross ie r, ' P ie r re Po t ie r,* and Rober t H. Dodd*^

Institut de Chimie des Substances Naturelles and Laboratoire de Physiologie Nerveus e of the Centre National de la RechercheScientifique, F 91198 Gif-sur-Yvette, Cedex, France, and Laboratoire Genitique, Neurogene tique et Comp ortement, URA 216,CNR S, Universite Paris V, 45, rue des Saints-Peres, 75006 Paris, France. Received A pril 19, 1988

LfMn dolo[3',2':4,5]pyrido[3,2-6]-2-penten-5-olide (6) and ltf,5ff-indolo[3 ',2'-c]-6,7-dihydro-2-p yridone (7), rigidanalogues of methyl 4-ethyl-0-carboline-3-carboxylate (8) and N-methyl-4-ethyl-/3-carboIine-3-carboxamide (9),respectively, were synthesized and their in vitro binding affinities to the central typ e benzodiazepine rec eptors werecom pared . Th e IC50 values of 6 and 8 were approxim ately eq uivalen t (42 and 27 nM , respectively). Th e am idederivative 9, for which theoretical energy calculations indicate th at the s-trans carbonyl conformation is the preferredone, displayed very low affinity (IC50 > 104 nM). However, when the carbonyl group of 9 was forced to ado pt th es-cis conformation as in lactam 7, binding to the benzodiazepine receptor was largely restored (IC50 = 150 nM), indicatingthat the s-cis carboxy conform ation at C-3 of (3-carbolines is preferentially recognized by this rec eptor. In vivo,compo und 6 showed neither convulsant, proconvulsant, nor anticonvulsant activity in mice. Moreover, 6 did notantagon ize meth yl /3-carboline-3-carboxylate induc ed conv ulsions in mice. This lack of activity of 6 was attrib utedto its inability to cross the blood-brain barrier since no significant displacement of [ 3 H]Ro 15-1788 from mousebrain benzodiazepine receptors by 6 could be observed in vivo.

0022-2 623/89/18 32-1799 $01.50/0 1989 Am erican Chem ical Society


2/6

1800 Journal of Medicinal Chem istry, 1989, Vol. 32, No.

Scheme I

C H J C O J C H J C H J C H O + 0 , N C H:C O , E INaOAc

EtOH / H: 0

HO

Dorey et al

*. .CO: E t

N0 2

[ H ,, Raney Nickel J IN H , ^ ^ ^ N

Figure 1. Confo rmations of the C-3 side chains of various 3-carboxy-/3-carbolines as determin ed by X-ra y crystallography: (A)/3-CCM (3) (refs 5 and 6); (B) /3-CEA (4) (ref 8); (C) DMCM (5)(ref 9).

Chart I

1 R = OCH2CHj

2 R NHCHj

3 R = OCH3

4 R = NHCH2CH3

( P-CCE )

( F G 7 1 4 2 )

( P - C C M )

((3-CEA)

6 X = 0

7 X = NH

8 R = OCHj

9 R = NHCHj

esters and am ides resides in the different conformationsof the respective carbonyl groups apparently favored byeach type of molecule. Th e carbonyl group of 0-CCM (3)in the solid state adop ts a conformation in which its oxygenatom is cis to the pyridine nitrogen (Figure 1A) while thecarbonyl group of /3-CEA (4) favors the opposite, transconformation (Figure IB). Preferential recognition of thes-cis conformation by the BZR would then explain thehigher binding affinity of /3-CCM compared to /3-CEA.

This argument is however tempered by the fact thatsome high-affinity esters [e.g., methyl 6,7-dimethoxy-4-ethyl-/3-carboline-3-carboxylate (DMCM, 5), Figure 1C]possess an s-trans conformation in the crystalline sta te.9-11

(7) Bertolasi, V.; Ferretti, V.; Gilli, G.; Borea, P. A. Acta Crys-tallogr.. Sect. C: Cryst. Struct. C ommun. 1984, 40 , 1981.

(8) Muir, A. K. S.; Codding, P. W. Can. J. Chem. 1984, 62, 1803.(9) Gilli, G.; Borea, P. A. Personal comm unication.

(10) Borea, P. A.; Gilli, G.; Bertolasi, V.;macol. 1987, 31 , 334.

Ferretti, V. Mol. Phar-

Assuming that only one conformation of the carbonylgroup is recognized by the BZR, identical for both /3-CCMand DMCM, then it is obvious that the X-ray crystalstructure of one of these molecules does not reflect the realnatur e of the ligand-recep tor interaction. Extension ofconformational preferences of compounds in the solid state

to those in solution may not always be justified; a normallystable conform ation of a moiety ma y be forced into a lessstable one by reason of the energy imp arted by its interaction with a receptor.

In an effort to resolve this question as to which conformation of the carbonyl group of 3-carboxy-/3-carbolinesis recognized by th e BZR, we have synthesized /3-carbolinesin which this critical functionality is forced to a dop t a rigidfixed s-cis conformation as part of either a cyclic ester(lactone 6) or cyclic amide (lactam 7) linkage.

The affinities of 6 and 7 for the BZR were determinedand compared to the affinities of the structurally relatedbut conformationally free 4-ethyl ester and amide ana-

(11) A referee has pointed ou t that the high affinity of DMCM forthe benzodiazepine receptor as determined in vitro with rathippocampal membranes (IC50 = 1.1 nM using [3H]DMCM asradioligand;26 IC50 = 6.6 nM using ^Hjflunitrazepam27) wouldtend to indicate that the trans conformation of the carbonylgroup, which this compound adopts in the crystalline state(Figure 1), is actually that recognized preferentially by thereceptor. However, our results with compound 8 show thatthis is unlikely. This compound, which represen ts DMCMwithout its dimethoxy groups, demonstrates a reduced affinityfor the benzodiazepine receptor (ICso = 27 nM using [3H]-flunitrazepam as radioligand, Table II) compared to bothDMCM and /3-CCM, precisely as a re sult of this com pound'spreference for the s-trans conformation, as our calculationsshow. The high affinity of DMCM is thus more likely due tothe presence of the dimethoxy groups than to its favored s-trans carbonyl conformation.


3/6

BZR Affinities of Analogues of 3-Car boxy-@ -carbolines

Scheme II

Journal of Medicinal Chem istry, 1989, Vol. 32, No. 8 1801

Table I. Calculated0 Energies and Torsion Angles of the Various3-Carboxy-/3-carbolines

C O j E t

logues, 8 and 9, respectively. Th e most stable conformersof these 4-substituted and 4-unsubstituted /3-carbolineswere also determined by theoretical calculations using anenergy minim ization program. Th e results strongly suggestthat the s-cis conformation of the carbonyl group (Figure1A) or one approaching this position is indeed preferentially recognized by the BZR.

Chemis t ry. The lactone 6 has reportedly been syn

thesized in low yield via microbial hydroxylation of 4-ethyl-0-CCE (8, R = OEt)12 while 5-membered lactonesand lactams analogous to 6 and 7 have been produced asside products in the reactions of 4-(bromomethyl)-/3-CCE(16)13 with potassium hydroxide and prim ary am ines, respectively. No binding data for these compounds weregiven.

We chose to synthesize the six-membered lactone 6 sincethis rigid structure disposes the carbonyl group and theadjoining oxygen atom in a geometry most closely resembling that observed in the crystal structure of /3-CCM (3)(see below). The methodology of Neef and co-workers13developed for the synthesis of 4-substituted /3-carboline-3-carboxylic esters was utilized to produce the precursor15 (Scheme I). Thu s, condensation of ethyl nitroacetate

with 2-formylethyl acetate (10)14

gave the unsta ble /3-hy-droxy ester ad duct 11, which was reacted d irectly withindole in toluene-ace tic acid, yielding the addition product12. Red uction of the nitro group of 12 with Rane y nickelin ethanol followed by P ictet-Sp engler -type cyclization ofthe resulting tryptop han derivative 13 with paraformaldehyde (Sandrin procedure)15 led to the 1 2 3 4-tetra-hydro-|8-carboline 14. This compound was subsequentlydehydrogenated to the fully aromatic 0-carboline 15 byheating under reflux in xylene in the presence of palladiumon charcoa l. W hen a solution of /3-carboline 15 in etha nolwas treated w ith a catalytic quan tity of sodium, the desiredlactone 6 precipitated almost quantitatively.

The 5-lactam 7 was prepared according to Scheme II.The protected 4-(bromomethyl) derivative of /3-CCE (16)13

reacted cleanly with potassium cyanide in dimethyl sulfoxide to give the 4-(cyanomethyl)-/3-carboline 17. Hy-drogenation of the cyano group using rhodium on aluminaas catalyst led, in one step, to the lactam 7.

The N-methylamide 9 was prepared by treatment of4-ethyl-/3-CCE (8, R = OE t)13 with methylamine in a sealed

(12) Neef G.; Eder, U.; Petzoldt, K.; Seeger, A.; Wieglepp, H. J.Chem. Soc, Chem. Commun. 1982, 366.

(13) Neef G.; Eder, U.; Huth , A.; Rah tz, D.; Schmiechen, R.; Sei-delmann, D. H eterocycles 1983, 20 , 1295.

(14) Eyjolfsson, R. Acta Chem. Scand 1970, 24 , 3075.(15) Sand rin, J.; Soerens, D.; Hutc hins, L.; Richfield, E.; Ungemach,

F.; Cook, J. M. Heterocycles 1976, 4 , 1101.

no.

3322889

967

C = 0conformer

s-transs-ciss-transs-ciss-transs-ciss-transs-ciss-ciss-cis

E, k c a l / m o lb

(0.1)d

37.441.641.651.041.747.946.9

59.443.851.7

a, deg*( i l 0 ) *

0000

16135

172

3216 11

By using the program MACROMODEL.16 b Lowest energy afterminimization. c Torsion angle formed by N2 C3C=0.d Estimated stand ard error.

tube at room temperature for 7 days.Resul ts and Discuss ion

Energy Calcula t ions . The MACROMODEL moleculargraphics program was used to calculate the energies E)and the torsion angles a, formed by N2 C3C=0) ofthe most stable conformations of the C-3 ester and amideside chains of both the unsu bstituted (compounds 3 and2, respectively) and 4-ethyl-substituted /?-carbolines(compounds 8 and 9, respectively) (Table I).

16 Th e resultsshow that, for the ester derivatives 3 and 8, the s-trans

conformers are always the most stable (i.e., lowest energy).Moreover, whereas for the nonsubstituted ester 3 the energy difference between the s-trans and s-cis conformersis relatively small (A = 4.2 kcal), that for the 4-substituted ester 8 is somewhat higher (AE = 6.3 kcal), and thisregardless of the conformation adopted by the 4-ethylgroup (not shown).11 Moreover, the carbonyl group of 8is forced out of the plane of the heterocycle a = 161),no doubt d ue to the proxim ity of the sterically bulky ethylmoiety.

In the case of the am ides 2 and 9, the s-trans conformersare again the most stable. In contras t to the esters, however, a much highe r energy difference (10-12 kcal) is seenbetween the s-trans amide and its s-cis conformer. As withthe e sters, the 4-ethyl substitue nt h as, as a principle effect,an increase in the energy difference between the s-cis ands-trans conformers and a forcing of the carbonyl groupoutside the plane of the heterocycle.

It is interesting to note that this theoretical approachto molecular geometry reproduces, in the case of the 4-unsubstituted amide derivative 2, almost exactly thecrystal structure of the iV-ethyl analogue of 2 (compound4) as determined by X-ray diffraction.8 On the other han d,the calculated conformational preference of the carbonylgroup of /3-CCM (3) (s-trans) is the opposite of that determined by X-ray crystallography (s-cis).6'7 However, asothers have inferred,10,17 our theoretical calculations showthat the energy differences between the cis and trans estersare sufficiently small that both forms can be expected toexist in solution in approximately equal amounts.

Energy minimizations of the rigid structures 6 and 7show a maximum deflection of the carbonyl group of, respectively, 16 and 10.8 from the plane of the carbolineheterocycle, which is more tha n th at of the unsubs tituteds-cis carbolines 3 an d 2 a = 0) but considerably less than

(16) Still, W. C; Richards, H. G. J.; Guida, W. C; Lipton, M.;Liskamp, R.; Chang, G.; Hendrickson, T. MACROMODEL V1.5,Department of Chemistry, Columbia University, New York,NY.

(17) Loew, G. H.; Nienow, J. R.; Paulson, M . Mol Pharmacol 1984,26 , 19.


4/6


5/6

BZR Affinities of Analogues of 3-Carbox y-B-carbo lines Journal of Medicinal Chem istry, 1989, Vol. 32, No. 8 1803

good in vi tro aff ini ty for the benzodiazepine receptor, weinves t iga te d th e ab i l i ty of com poun ds 6 and 7 to in h ib i tthe b ind ing of [ 3 H]Ro 15-1788 in vivo in mice . T hu s,accord ing to a procedure tha t we have prev ious ly descr ibed ,22 the lac tone 6 or the lac tam 7 was adm inis te re d(20 m g/ kg in Twee n 80, ip) in mice followed, 30 min later,by [ 3 H]Ro 15-1788. After 3 mi n, the mice were s acrificed,and a f te r d i ssec t ion , the rad ioac t iv i ty re ta ined in thecortex, cerebellum, and hippocam pus was determ ined. Th einhib i t ion of [ 3 H]Ro 15-1788 by compounds 6 and 7 wasfound to be pract ical ly insignif icant (between 5 and 10% )in the three brain regions studied . In contras t , /3-CCM (3)displaces 50% of specif ical ly bound [ 3 H]Ro 15-1788 at asl i t t l e as 0 .72 mg/kg in mouse cerebe l lum under the samecondi t ions .2 2 b Thus, the lack of in vivo act ivi ty displayedby 6 may be expla ined by i t s s low or incomple te penet ra t io n in to the bra in , the resu l t of unfavorab le pha rmacokine t ic or pha rma cod yna mic fac tors .

In conc lus ion , the presen t s tudy demons t ra tes tha t theBZR recognizes , p re fe ren t ia l ly, the s-cis conformat ion of3-carboxy-/3-carbolines or one approaching this geometry.This resu l t thus cor robora tes severa l recent ly proposedbenzodiazepine receptor l igand b inding models1 0 , 1 9 , 2 3 inwhich th i s par t icu la r conformat ion was on ly tac i t ly assum ed. T he effect of a fixed carbo nyl conform ation on thein vivo act ivi ty of /3-carbolines is currently being invest igated via the synthesis and p harmacological test ing of mo rel ipophi l ic ana logues of 6 a n d 7.

E x p e r i m e n t a l S e c t i o nChemistry. Melting points were determined on a Biichi ap

para tus and are uncorrected. IR spectra were taken in Nujol orneat with a Perkin-E lmer 297 instrum ent. Proton NM R spectrawere determined on Varian T-60, Bruker WP 80, or Bruker WP200-MHz instrum ents. Chemical shifts are given as 8 values withreference to Me4Si as internal standard. Thin-layer chromatography was performed on Merck silica gel 60 plates withfluorescent indicator. Th e plates were visualized with UV light(254 and 366 nm ). Merck silica gel (230-400 mesh) w as used forall column chromatograp hy. Electron impact (EI) mass spectrawere done on an AEI MS-50 spectrom eter. Chemical ionization

(CI) and fast atomic bombardment (FAB) mass spectra wereobtained respectively on a modified24 AEI MS-9 and on a KratosMS-80 spectrometer. High-performance liquid chromatography(HPLC) was performed on a Waters 6000 A instrument equippedwith a 10 X 250 mm column of Spherisorb ODS2 (5-jim) and aUV440 detector. Elem ental analyses were performed at the ICSN ,CNRS, Gif-sur-Yvette, France.

(2iS,3#S)-Ethyl5-Acetoxy-3-hydroxy-2-nitropentanoate(11). To a solution of 2-formylethyl acetate (10)14 (17.2 mmol)in ethanol (15 mL) containing sodium acetate (0.2 mmol) in water(1.5 mL) at 0 C was added dropwise ethyl nitroacetate (17.2mmol) under a nitrogen atmosphere. The reaction mixture wasstirred a t 10 C for 1 h and allowed to stand at room tem pera tureovernight. The ethan ol was then removed in vacuo, and theresidue was extracted with ether (40 mL ). Th e extract was washedwith saturated aqueous sodium chloride, dried over anhydrous

sodium sulfate, and concentrated to leave 3.5 g of an oily residue(11), which was used without further purification in the followingsteps: IR (neat) 1560 (N02), 1740 (OC=0), 3500 cm1 (CH-OH); EIMS, m /z 232 (MH+ - H20), 207 (MH+ - CH3CO); NMR(CDC13) 8 1.28 (3 H, t, C tf3CH20), 1.28 (2 H, d oft, Ctf2CH20),2.02 (3 H, s, CH3CO), 3.62 (1 H, bs, OH), 4.12 (5 H, m, CH3Ctf20,C # O H C H2, Cff2OAc), 5.33 (1 H, m, N02CtfC02Et).

(22) (a) Potier, M .-C; Prad o de Carvalho, L.; Dodd, R. H.; Brown,C. L.; Rossier, J. Life Sci. 1988, 43, 1287. (b) Potier, M .-C;Prado de Carvalho, L.; Dodd, R. H.; Besselievre, R.; Rossier,J. M ol. Pharmacol. 1988, 34 , 124.

(23) Tebib, S.; Bourguignon, J.-J.; Wermuth, C.-G. J. Comput.-Aided Mol. Des. 1987, 1 , 153.

(24) Varen ne, P.; Bardey, B.; Longevialle, P.; Das, B. C. Bull. Soc.Chim. Fr. 1977, 886.

2RS,ZRS) -Ethyl 5 -Acetoxy-3- ( indol -3-y l ) -2-n i t ro-pentanoate (12). A solution of indole (50 mg; 0.43 mmol), hy-droxynitroester 11 (314 mg) (containing 1/3 of unreactive ethylnitroacetate by NMR), toluene (5 mL), and glacial acetic acid (0.5mL) was heated under reflux at 110 C under a nitrogen atmosphere for 2 h. After this period ano ther 314 mg of hydroxy-nitroeste r 11 was added , and the reaction mix ture was heatedunder reflux for a further 2 h. Evaporation of the solvents in vacuogave a brown syrup. Traces of acetic acid were removed byrepeate d codistillation with toluene in vacuo. The residue waschromatographed on silica gel by using 1:1 toluene-ethanol asdeveloper to yield 58 mg ( 39% , based on indole) of 12 as a yellowoil: IR (CHC13) 1560 cm'1 (N02) ; EIMS, m /z 348 (M+), 301 (M+- HN02); NMR (CDC13) 0.90 [1.8 H, t, J = 7 Hz, R (or S),C tf3C H20 ], 1.27 [1.2 H, t, J = 7 Hz, S (or R) , C tf3C H20], 2.00(3 H, s, Ctf3CO), 2.20 (2 H, m, H-4), 3.77-4.53 (5 H, m, H-3, H-5,OCtf2CH3), 5.57 [1 H, t, J = 10 Hz R an d S) , H-2], 6.87 (3 H,m, Ar), 7.40 (1 H, d, J = 8 Hz, Ar), 7.67 (1 H, d, J = 7 Hz, Ar),8.33 (1 H, bs, NH indole, D20 exchangeable). Anal. (C17H2oN206)C, H, N.

2RS ,3RS ) - E t h y l 5 -Ace toxy-3- ( indo l -3 -y l ) -2 -amino-pe nt an oa te (13). A vigorously stirred mixture of indole 12 (240mg; 0.69 mmol), ethanol (35 mL), and Raney nickel (1 g) washydrog enated at atmo spheric pressure for 4 h. Th e nickel wasthen removed by filtration over Celite and washed with hot ethanol(250 mL). Evaporation of the filtrate in vacuo afforded a yellowsyrup which was chromatographed on a silica gel column usingdichlorometha ne-ethanol (7:0.5) as developer to give 107 mg (48%)of the desired amine 13: IR (CHC13) 3400 cm 1 (N H2); EIMS,m/z 318 (M+), 245 (M+ - C 02Et, M+ - CH2OAc), 216 (M+ -N H2C H C 02Et); NMR (CDC13) 5 1.13 [1.8 H, t, J = 7 Hz, R (orS) , C tf3CH20], 1.33 [1.2 H, t, J = 7 Hz, S (or R) , C tf3CH20], 2.13(2 H, s, NH2, D20 exchangeable), 2.15 (3 H, s, Ctf3CO), 2.15 (2H, m, H-4), 3.53 (1 H, m, H-3), 3.80 (1 H, d, H-2), 4.00 (4 H, m,H-5, O C#2CH3), 8.20-8.60 (5 H, m, Ar), 9.33 (1 H, bs, NH indole,D20 exchangeable); HRMS, m /z calcd for C17H22N204 318.1558,found 318.1569.

Ethyl 4-(Acetoxyethyl)-|3-carboline-3-carboxylate (15). Amixture of amine 13 (570 mg; 1.79 mmol) and paraformaldehyde(65 mg; 0.72 mmol) in toluene (75 mL) was heated unde r refluxunder a nitrogen atmosphe re for 3 h. Th e water formed wasremoved by means of a Dean -Stark apparatu s. The reactionmixture was then filtered to remove excess solid paraformaldehyde, and the toluene was evaporated in vacuo. The resultingcrude ethyl 4-(acetoxyethyl)-l,2,3,4-tetrahydro-|3-carboline-3-carboxylate (14) (HRMS, m/z calcd for C18H22N204 330.1593,found 330.1586) was dissolved in anhydro us xylene (150 mL ), and10% palladium on charcoal catalyst (300 mg) was added . Th emixture was heated under reflux w ith vigorous stirring for 2.5 huntil no tetrahydro-/3-carboline 14 remained , as indicated by T LCusing tolue ne-eth ano l (9:2) as developer. The cataly st was removed by filtration over Celite and washed w ith warm xylene (500mL). Eva pora tion of the filtrate gave an oily residu e which waspurified by chromatography on silica gel with chloroform -ethylacetate (1:1) as developer, yielding 230 mg (39% in two steps)of crystalline /3-carboline 15: mp 190-191 C; EIMS, m/z 326(M+); NMR (CDC13) 5 1.28 (3 H, t, J = 7.5 Hz, OCH2CH3), 1.73(3 H, s, Ctf3CO), 3.75 (2 H, t, J = 7 Hz, Ctf2CH2OAc), 4.40 (2H, q, J = 7.5 Hz, OCtf2CH3), 4.54 (2 H, t, J = 7 Hz, CH2CH2OAc),7.33 (1 H, t, J = 7 Hz, Ar), 7.57 (2 H, m, Ar), 8.38 (1 H, d, J =

8 Hz, Ar), 8.90 (1 H, s, Ar), 10.28 (1 H, s, NH indole, D20 exchangeable). Anal. (C18H18N204) C, H, N.lIMndolo[3',2':4,5]pyrido[3,2-ft]-2-penten-5-olide (6). To

a stirring solution of sodium (2.5 mg; 0.1 mmol) in ethanol (80mL) was adde d th e /3-carboline 15 (320 mg; 0.98 mm ol). After23 h at room temperature, the white solid that precipitated wascollected by filtration and recrystallized from dichloromethane-ethanol to give the desired lactone 6 (200 mg; 86%) as a whitepowder: mp 315-318 C dec; EIM S, m/z 238 (M+); NMR(M e2SO-de) S 3.82 (2 H, t, J = 6 Hz, Ctf2CH20), 4.73 (2 H, t, J= 6 Hz, CH2Ctf20), 7.43 (1 H, t, J = 8 Hz, Ar), 7.70 (1 H, t, J= 8 Hz, Ar), 7.78 (1 H, d, J = 8 Hz, Ar), 8.33 (1 H, d, J = 8 Hz,Ar), 9.03 (1 H, s, Ar), 12.28 (1 H, bs, NH indole, D20 exchangeable). Anal. (C14H10N2O2) C, H, N.

Eth yl 4-(Cyanomethyl)-/9-carboline-3-carboxylate (17). T oa suspension of potassium cyanide (5 mg; 0.077 mmol) in an-


6/6

1804 Journal of Medicina l Chem istry, 1989, Vol. 32, No. 8 Dorey et al.

hydrous dimethyl sulfoxide (2 mL) was added dropwise ethyl9-acetyl-4-(bromomethyl)-0-carboline-3-carboxylate (16) (preparedas described in ref 13) (20 mg; 0.053 mmol) dissolved in 1 mL ofDMSO. The stirring mixture was heated at 70 C for 40 min andthen allowed to stand at room temperature for 1.5 h until nostarting material remained, as indicated by TLC. After evaporation of the solvent in vacuo, the residue was extracted withdichloromethane (5 mL), and the organic solution was washedwith saturated aqueous sodium chloride, dried over sodium sulfate,and concentrated to leave a yellow solid (13 mg) which was re-crystallized from dic hloro meth ane-h exane to yield 17 (11 mg;74%) as pale yellow needles: mp 183-187 C dec; CIMS (C4H10),m/e 280 (M + 1, 100); NM R (CDC13) 8 1.47 (3 H, t, J = 7 Hz,OCH2Ci f3), 4.60 (2 H, q, J = 7 Hz, OCtf2CH3), 4.88 (2 H, s,C/f2CN), 7.47 (1 H, m, Ar), 7.65 (2 H, d, J = 5 Hz, Ar), 8.35 (1H, d, J = 8 Hz, Ar), 9.03 (1 H, s, Ar), 9.88 (1 H, bs, NH indole);H R M S , m/z calcd for C16 H13 N302 279.0950, found 279.0979.

l ff-Indolo[3'':4,5]pyrido[2,3-c]-6,7-dihydro-2-pyridone(7). A mixtu re of 4-(eyanom ethyl)-/3-carboline 17 (20 mg; 0.072mmol), 10% ethanolic ammonia (10 mL), and 5% rhodium onalumina (10 mg) was hydrogenated on a Parr a pparatu s at 45 psifor 15 h. The catalyst was removed by filtration over a glass filterand washed with warm ethanol (100 mL), and the filtrate wasevaporated to give 14 mg of crude product. This m aterial waspurified by high-performance liquid chromatography on a re-versed-phase silica gel column using acetonitrile-water-tri-ethylamine (28:72:0.1, flow rate = 3 mL/min) as developer, affording, with a retention tim e of 12 min, the desired lactam 7 (9mg; 53%) as a white powder: mp >300 C; FABM S (Gly, DM F+),238 (M + 1); NMR (Me2SO-d6) 8 3.67 (4 H, s, CH 2CH 2), 7.39 (1H, t, J = 8 Hz, Ar), 7.66 (1 H, t, J = 8 Hz, Ar), 7.76 (1 H, t, J= 8 Hz, Ar), 7.99 (1 H, s, NH amide, D20 exchangeable), 8.31 (1H, d, J = 8 Hz, Ar), 8.94 (1 H, s, Ar), 12.07 (1 H, bs, NH indole,D20 exchangeable); HRM S, m/z calcd for C1 4HnN sO 237.0902,found 237.0906.

iV-Methyl-4-ethyl-/J-carboline-3-carboxamide (9). Into atube cooled to -78 C was introduced ethyl 4-ethyl-/3-carboline-3-carboxylate (51 mg; 0.19 mmol) (prepared as describe in ref 13)and liquid methylam ine (3 mL ). The amp ule was sealed andallowed to stand at room tempe rature for 1 week. The tu be wasthen opened, and the mixture was poured into cold ethanol (10mL). Evap oration of the solven ts gave a white solid which wasrecrystallized from ethanol-hexane to afford the desired /3-car-bolinecarboxam ide 9 (20 mg; 42%) as colorless needles: mp216-217 C; EIMS, m/z 253 (M+); NMR (Me2SO-d6) 8 1.52 (3H, t, J = 7 Hz, Ctf3CH2), 2.97 (3 H, d, J = 4 Hz, NHCff3), 3.80(2 H, q, J = 7 Hz, CH3Ctf2), 7.48 (1 H, t, J = 8 Hz, Ar), 7.80 (2H, m, Ar), 8.42 (1 H, d, J = 8 Hz, Ar), 8.72 (1 H, d, J = 4 Hz,NH amide) , 8.92 (1 H, s, Ar), 11.80 (1 H, bs, NH indole). Anal.(C15H16N3O.V4H20) C, H, N .

Biological Methods. In Vitro Benzo diazepine R eceptorBinding Assays. The techn ique described by Rehavi et al.18 wasused. Briefly, male Sprague-Dawley rats were decap itated, th ebrains were excised, and the cortex was dissected. Each cortexwas homogenized in 5 mL of ice-cold Tris-HCl (50 mM, pH 7.4)(corresponding to 6 m L/ g of tissue) with a Polytron . The hom -ogenate was centrifuged a first time at 460g for 3 min, and thesup erna tant was recentrifuged at 224O0g for 20 min. The resultingsupernatant was discarded, and each pellet was resuspended in3 mL of buffer an d centrifuged again at 22400g for 20 min. Th e

resulting pellet was again suspended in 3 mL of buffer, reho-mogenized, divided into 1-mL fractions, and stored at -20 C forat least 24 h before use. For the inhibition stu dies, the thawedmembrane preparations were diluted with 20 volumes of ice-coldbuffer, and 900-ML aliquots containing approximately 60 f ig ofprotein, as determined by the Lowry method,25 were incubated

(25) Lowry, O. H.; Rosebrou gh, N. J.; Farr , A. L.; Rand all, R. J. J.Biol. Chem. 1951, 193, 265.

(26) Braes trup, C; Nielsen, M.; Honore, T. J. Neurochem. 1983, 41,454.

at 0 C for 60 min with 50 nL of [3H]flunitrazepam (65 Ci/mm ol,CEA, final concentration 1 nM) an d 50 nL of varying concentrations of the test compound ranging from 105 to 5 X 10~10 M(final concentrations for a total volume of 1 mL) . Nonspecificbinding was measured in the presence of 1 nM nonradioactiveflunitrazepam and represented 10-15% of the total binding.Incubations were terminated by adding 3 mL of cold buffer toeach incubation tube, filtering immediately through WhatmanGF/B glass fiber filters, and washing each filter four times with3 mL of ice-cold buffer. These op erations were performed in lessthan 15 s per sample. Radioactivity retain ed on the filters wascounted in 10 mL of Aquasol (NEN) scintillation solution withan LKB W allac 1215 Rackbeta 2 counter. Each value was determin ed in triplicate . IC50 values (the concentration of ligandinhibiting 50% of flunitrazepam binding) were determined byHofstee analysis. Results are given in Table II.

In Vivo Studies . Drug s. Pentylenetetrazole (PTZ) (M.Richard Laboratories, Sauzet, France) was dissolved in saline.Methyl /3-carboline-3-carboxylate (/3-CCM), synthesized by oneof us (R.H.D.), was dissolved in 0.1 N HC1 (1 mg in 100 ML) anddiluted to volume with saline. [ 3 H]Ro 15-1788 (87 Ci/mm ol) wasfrom Du Pont NEN (Boston, MA). The test drugs were used asa suspension in Tween 80 and saline. Unless otherwise stated ,all drugs were administered subcutaneously (sc) at a dose volumeof 0.05 mL/10 g of body weight.

Convulsion Studies with Compound 6. The convulsantactivity of compound 6 was studied by treating m ale Swiss mice(25 g) with 5 and 20 mg/kg, ip, of 6 (lots of five mice). Th e micewere observed for 30 min for convulsions, indications of myoclonicactivity, tremo rs, or other behavioral anomalies. To test foranticonvulsant activity, mice (10 per experiment) were treatedwith compound 6 (5 and 10 mg /kg, ip) before ad ministr ation ofPTZ (70 mg/kg) or /3-CCM (10 mg/kg), 10 and 5 min later,respectively. The same protocol using subconvulsive doses of PTZ(30 mg/kg ) or (3-CCM 5 mg/kg) was used to test for proconvulsantproperties of 6 (10 mg/kg). For each series of experiments, parallelcontrols were run with animals injected with saline and Tweenonly.

In Vivo Inhibition of [ H]Ro 15-1788 Binding. Inhibitionof the in vivo binding of [ 3 H]Ro 15-1788 to mouse brain benzodiazepine receptors was performed as previously described.22Briefly, mice were treated with the test compound (6 or 7, 20mg/kg, sc) 30 min before the iv administration of [ 3 H]Ro 15-1788(50 fiCi/kg). After 3 min, the mice were decap itated, and theirbrains were rapidly excised and dissected on ice. Cortex fromone hemisphere, cerebellum, and hippocampus were homogenizedin approximately 30 volumes (3 mL for cerebellum and cortexand 1.5 mL for hippocampus) of ice-cold Tris-HCl (50 mM, pH7.4) using a Kinematica Polytron (type PT 10/35) for 5 s at halfspeed. Aliquots of the homogenates (600 ML) were immediatelyfiltered through W hatman glass fiber filters (GF/B ), and the filterswere rinsed with two 5-mL portions of ice-cold Tris-HCl buffer.Membrane-bound radioactivity was retained on the filters andwas counted by liquid scintillation in an LKB Rackbeta 2 counterafter addition of 10 mL of aqueous counting scintillant (Am-ersham, Les Ulis, France). Total binding values were obtainedfrom mice treated only with [ 3 H]Ro 15-1788. Nonspecific bindingvalues were determined from mice injected with diazepam (10mg/kg) 60 min before decapitation.

A c k n o w l e d g m e n t . This work was suppor ted by Gr ant87094 from D R ET . We also tha nk DR E T for fel lowships(M.-C.P., G.D.), Dr. C. Ouannes for some preliminary work,and Dr. R. Breckenridge for cr i t ical appraisal of themanu script . Th e valuable assistance of Mrs. M.-T. Adeline(HP LC ) and Mr. J . Cam ara (com puter g raphics ) isgra te fu l ly acknowledged .

(27) Braestru p, C; Schmiechen, R.; Neef G.; Nielsen, M.; Peter sen,E. N. Science 1982, 216, 1241.

Documents

Synthesis and Benzodiazepine Receptor Affinities of Rigid Analogues of 3-Carboxy-/?-carbolines: Demonstration That the Benzodiazepine Receptor Recognizes Preferentially the s-Cis Conformation