Synthesis of anthracyclinones. Part 1. Regioselective alkylation of 5

0 downloads 0 Views 334KB Size Report
acetylation followed by hypochlorite oxidation and ester- ification gave ... alkylation methods. ... aldol condensation we attempted to prepare an enamine from.
Synthesis of anthracyclinones. Part 1. Regioselective alkylation of 5-hydroxyquinizarin LAURENCE M. HARWOOD, LESLIEC. HODGKINSON, JAMESK. SUTHERLAND,' AND PATRICK TOWERS Chemistry Department, The Victoria University of Manchester, Manchester, England MI3 9PL Received November 9, 1983 This paper is dedicated to Professor Paul de Mayo on the occasion of his 60th birthday

Can. J. Chem. Downloaded from www.nrcresearchpress.com by 80.188.127.134 on 05/28/13 For personal use only.

K. SUTHERLAND, and PATRICK TOWERS. Can. J. Chem. 62, 1922 LAURENCE M. HARWOOD, LESLIE C. HODGKINSON. JAMES (1984). Condensation of leuco-5-hydroxyquinizarin with aldehydes under Marschalk conditions (sodium hydroxide - water) gives 2-alkyl-5-hydroxyquinizarins with high regioselectivity, while the 3-alkyl compounds are obtained under Lewis conditions (piperidinium acetate - isopropanol). Reaction of the quinone with nitronate anions in methanol yields 2-alkyl derivatives exclusively.

LAURENCE M. HARWOOD, LESLIE C. HODGKINSON, JAMES K. SUTHERLAND et PATRICK TOWERS. Can. J. Chem. 62, 1922 ( 1984).

La condensation de la leuco hydroxy-5 quinizarine avec les aldthydes, en utilisant les conditions de Marschalk (hydroxyde de sodium et eau), conduit aux alkyl-2 hydroxy-5 quinizarines avec une grande rCgiosClectivitC; toutefois, si on utilise les conditions de Lewis (acCtate de pipiridinium et isopropanol), on obtient les composCs alkylCs en position 3. La rCaction de la quinone avec les anions nitronates dans le mCthanol conduit exclusivement aux dCrivCs alkylCs en position 2. [Traduit par le journal] Our choice of the readily available 5-hydroxyquinizarin ( I ) as a starting point for the synthesis of the aglycones of the anthracycline antibiotics daunomycin and carminomycin (1) necessitated the development of methods for the regioselective substitution of 1 at C-2 and (or) C-3. Prior work by Marschalk et al. (2) had shown that aldol condensation between leucoquinizarin (12) and aldehydes was an effective method for the preparation of 2-alkylquinizarins. Brockrnann and Miiller (3) carried out Marschalk condensations with leuco-5-hydroxyquinizarin (13) and presented evidence for some regioselectivity in favour of C-2. We first examined the Marschalk condensation of propionaldehyde with the leuco-compound 1 3 and obtained the 2-propyl derivative (2). The structure of 2 was established by conversion to the tri-0-methyl ether which was cleaved with KOBu'-H20 using Hodge's method (4) to give, after methylation of the acidic products with CHIN2, ~n-methoxybenzoic acid methyl ester (22%) and the ester 14 (28%). An authentic sample of the ester 14 was synthesized starting with propionylation of 1,4-dimethoxybenzene followed by Clemmensen reduction to 2-n-propy I- l,4-dimethoxybenzene; Friedel-Crafts acetylation followed by hypochlorite oxidation and esterification gave material identical with the degradation product. Reaction of the leuco-compound 1 3 with propionaldehyde in isopropanol containing piperidinium acetate (conditions described by Lewis (5)) gave predominantly the 3-propylquinone (3). Since we have not found any chromatographic system which will separate the 2- and 3-propyl isomers it is necessary to convert the product to the tri-0-methyl ethers to determine isomeric purity; these isomers can be separated by tlc. Degradation of the tri-0-methyl ether of 3 by the Hodge procedure was much more complex than in the previous case but did yield m-methoxybenzoic acid (9%) and the ester 15 (6%). The tri-0methyl ethers of 2 and 3 show minor differences in their 'H nmr spectra which we have found to be present in other isomeric pairs. We believe these differences are reliable enough to allow assignment of structures. In the 2-alkyl series the proton resonates at 8,, 7.00 with the methoxyl singlets at 3.94, 3.91, and 3.83 while in the 3-alkyl series the 2-proton absorbs at 6.82 and 'Author to whom correspondence may be addressed.

9 R2 = (CH2)2C(OCH2)2Me, R' = H 10 R2 = (CH2)2COMe,R3 = H 11 R~ = CH(OH)Et, R3 = H

OMe 14 R1 = Pr", R2 = H

IS R'

=

H, R' = Pr"

the methoxyls at 3.94, 3.90, and 3.88. Table I summarizes the isomer ratios obtained with a variety of aldehydes using the two alkylation methods. We have no definitive experiments pertaining to the origin of the regioselectivity in these reactions; however, titration of leuco-5-hydroxyquinizarin indicates that in dilute alkaline

HARWOOD ET AL.

TABLE1. Isomer ratios in Marschalk and Lewis condensations Aldehyde EtCHO EtCHO HO[CH2],CHO HO[CH2],CHO MeCH(OH)[CH2I2CHO HOCH2CH0

Conditions"

A B A A B A

Total yield (%)'

Ratio 2- :3alkylation"

50 46 44 44 49 50

I

1923

12+$.l=;9 NR2

0

NR~

HOS cNR2

H

0

1 :9

7: 1 9: 1 I :7

L e v . I-Hydroxyanthraquinone

-

Can. J. Chem. Downloaded from www.nrcresearchpress.com by 80.188.127.134 on 05/28/13 For personal use only.

"A. Marschalk conditions; B, isopropyl alcohol - piperidinium acetate.

'Pure recrystallized material. "Ratios determined after conversion to methylethers. "No 3-alkylation.

solution the bulk species present is a monoanion. This is the likely species present under the conditions of the Marschalk condensation. If the predominant anion present is 16' then the kinetically active species must arise by proton abstraction from position 2 which will generate an enolate stabilized by hydrogen bonding; no such stabilization is available to the enolate arising from proton abstraction from position 3, since the hydrogen bond is stabilizing the oxyanion. Under the Lewis conditions the kinetically reactive species may well be 17' derived from the neutral leuco-compound to give, by proton abstraction from position 3, a more stable enolate than that derived from position 2 due to more extensive hydrogen bonding. With the regioselectivity of the reactions established we turned to the problem of introducing side-chains containing functional groups. Table 1 shows that a-,y-, and S-hydroxyaldehydes can be used successfully. Krohn and Radeloff (6) have defined conditions where a wide variety of aldehydes have been successfully condensed with leuco-quinizarin and the reaction intercepted at the aldol stage to give 2-(1-hydroxyalky1)quinizarins after aerial oxidation; we have found this method to be an excellent one using leuco-quinizarin but not generally applicable to the 5-hydroxy derivative. However, the protocol described by Shaw and co-workers (7) is an effective method using leuco-5-hydroxyquinizarin (13). Reaction of 2,3-epoxybutanal with 13 under controlled conditions gave a compound formulated as 18 (v,,,: 1665, 1625 cm-'; SH: 12.32

rolidine in refluxing toluene gave 1-hydroxyanthraquinone in excellent yield. Application of these conditions to leuco-5hydroxyquinizarin yielded 1,5-dihydroxyanthraquinone. We formulate this reaction as shown in Scheme 1. At lower temperatures amines did not react with leuco-quinizarin, but quinizarin (produced by adventitious oxidation) gave amino derivatives. This was confirmed by reaction of quinizarin with pyrrolidine, piperidine, and benzylamine to give the amines of the type 19. Since a likely mechanism of formation of these compounds is nucleophilic addition followed by aerial oxidation, we investigated the reactions of the quinizarins with carbanions.

21 (double bond) 22 (without =)

(lH,s),11.93(lH,s),7.56(lH,dd,J=8.0and1.3Hz),7.22 Reaction of 5-hydroxyquinizarin with an excess of l-nitro(IH, t , J = 8.0Hz),7.00(1H,dd, J = 8.0and 1.3Hz),6.98 ( l H , s), 4.93 (IH, td, J = 8.0 and 3.0 Hz), 4.40 (IH, qd, J = 8.0 and 3.0 Hz), 3.72 (IH, bs), 3.17 (2H, m), 1.22 (3H, d, J = 8.0 Hz)). On oxidation with pyridinium chlorochromate a methyl ketone was formed (v,,,: 1730, 1660, 1620 cm-'; SH: 5.26(1H,t, J = 7Hz),3.45(2H,d, J = 7Hz), 2.38(3H,s)). The conditions of the Marschalk condensation are not compatible with the presence of certain functionalities in the aldehyde; in particular, leaving groups p to the aldehyde function do not survive. In an effort to define milder conditions for the aldol condensation we attempted to prepare an enamine from leuco-quinizarin (12); to our surprise, reaction of 12 with pyr-

'The huge difference in acidity between 2-acyl-I-naphthols and 2-acyl-l,8-naphthalenediolswas established during the structure determination of terrarnycin (8). 3Selective formation of the 3,4-enarnine as intermediate instead of 17 was considered but rejected in the expectation that the adduct would undergo arornatization eithcr to the 3-hydroxyalkyl-1.5-dihydroxyanthraquinone (see later) or to the 3-alkyl-1,s-dihydroxy-4pyrrolidinylanthraquinone. However, the electrophilc in the reaction might well be the irnrnoniurn salt of the aldehyde.

propane and half that concentration of sodium methoxide in methanol gave the 2-propyl compound 2 in excellent yield. By conversion to the trimethyl ether it was shown, within our detection limits, to be uncontaminated with any of the 3-isomer. 2-Nitropropane and nitromethane gave similar results; in the latter case prolonged reaction gave 2,3-dimethyl5-hydroxyquinizarin. Reaction with 1,4-dinitrobutane gave the tetracycle 20, and the ethylene acetal of 4-nit.robutan-2-one produced 9. We postulate that the regioselectivity is determined in the addition of the nitronate anion to the unionized 5-hydroxyquinizarin to give the more stable 3-enolate. In support of this suggestion, no reaction occurs when enough base is present to convert both quinizarin and nitroalkane to their respective anions. Similar arguments can be used to account for regioselectivity in an electron transfer mechanism. When the reaction with 1-nitropropane is carried out in 95% ethanol the likely intermediate 23 can be trapped by water to give, after aerial oxidation, the hydroxypropyl compound 11.4 4This reaction was discovered by Mr. A. Kurnar in these laboratories.

CAN. J . CHEM. VOL. 62. 1984

In order to test some of the later stages of our proposed synthesis we carried out some model experiments. The alcohol 5 could be oxidized to the aldehyde with pyridinium chlorochromate without oxidation of the quinizarin ring. Dithionite reduction of the aldehyde under Marschalk conditions gave the tetracycle 20. Hydrolysis of the acetal 9 generated the ketone which, under Marschalk conditions, gave a mixture of 21 and 22; dithionite reduction of 21 gave 22. Experimental

Can. J. Chem. Downloaded from www.nrcresearchpress.com by 80.188.127.134 on 05/28/13 For personal use only.

All sarnples on which accurate mass measurements were determined show a single spot on tlc. 2-n-Prop~~l-5-l1yc/ro,~yqr~i11izcrri11 (2) 5-Hydroxyquinizarin (512 mg) was dissolved in 5% NaOH-H20 and the bluc solution adjusted to pH 10 with 10% HCI-H,O. After warming to 40°C under a N, atmosphere a solution of Na,CO, (2 g) and Na2S,0, ( I g) in H,O (20 mL) was added until the orange leucoanion was formed. Propionaldehyde (348 mg) in H,O (10 mL) was added over 30 min and the mixture boiled with reflux for 2 h. Heating then ceased and air was passcd through thc solution for 15 min. After acidification with I0 M HCI the precipitatc was filtcred off, washed with H 2 0 , dried, and extractcd with boiling CHCI, (4 X 20 mL). Thc CHC13 extract was dried and evaporated to give a red solid which on recrystallization from CCI, gave the propyl der-ivtrtive (2) (300 mg), mp 187-190°C; A,,,;,, (EtOH): 233 (40 OOO), 253 (24 500), 292 (7 800), 466 (9 700). 48 1 ( I l 400). 493 ( 13 OOO), 5 14 (9 5001, and 527 nm (9 000); 6 , : 1.03 (3H, t, J = 7.5 Hz), 1.73 (2H. m), 2.72 (2H, t , J = 7 . 5 H z ) , 7 . 0 9 ( l H , s ) . 7 . 2 4 ( l H . d d , J = 8and 1.3 Hz),7.63 ( I H , t , J = 8 H z ) , 7 , 8 2 ( l H , d d , J = 8and 1.3Hz), 12.18(1H.s), 12.25 ( I H , s). 13.46 (1H. s). AII~II. calcd. for C17HI,05: C 68.5, H 4.7%: found: C 68.3, H 4.7%. Thc following 2-substituted 5-hydroxyquinizarins wcre prepared from the aldehydes indicatcd. in a similar way. 2-(5'-Hydroxypentyl) (4) from 5-hydroxypentanal (43%). mp 187- 191°C (CCI,). Ancrl. calcd. for Cl,HIKO(,:C 66.7, H 5.3%; found: C 66.4. H 5.4%. 2-(4'-Hydroxybutyl) (5) from 4-hydroxybutanal (44%) (CCI,), mp 195- 199°C (CCI,). An(11. calcd. for ClxHl(,Or,: C 65.9, H 4.9%; found: C 66.0, H 4.9%. 2-(2'-Hydroxyethyl) (7), from hydroxyacctaldchyde (50%) mp >300°C. ClhHI20(, requires M': 300.0634; found: 300.0629. 2-(2'-Methyl-3'-furyl)methyl (8) from 2-mcthylfuran-3-carboxaldehyde (72%). mp 2 15-218°C (CCI,). A~rcrl.calcd. for CZ0HI,Oh:C 68.6. H 4.0%; found: C 68.6. H 4.0%. 2-Ethyl- from acetaldehyde (67%), mp 125- 128°C (Me,CO). Ancrl. calcd. for CI(,HII05:C 67.6, H 4.3%: found: C 67.9, H 4.4%. 2-(2'-Mcthylpropyl) from 2-methylpropanal (63%), mp 196199°C (Me2CO).A~~crl. calcd. for C I X H l h 0C 5 :69.2, H 5.2%; found: C 69.6. H 5.6%. The ether (18) from 2.3-epoxybutanal (80%). mp >205"C (dec.) (CCI,). ClxH,,Ohrequires M': 326.0790; found: 326.0785. Oxidation of 18 with pyridinium chlorochromatc gave a ketone, mp > 180°C (dec.). v ,,,,,,: 1720 c m - I ; 2.40 (3H, s). 3.45 (2H, d, J = 7 Hz). 5.28 (IH, t, J = 7 Hz), 12.1 l (IH. s). 12.54 (IH, s). 2-(Cyclopcnt-3'-eny1)mcthylfrom cyclopcnt-3-enc carboxaldchyde (42%). mp 199-20 1 "C (sublimation). Atlerl. calcd. for CzoHl,O,: C 7 1.4, H 4.8%; M': 336.0998; found: C 71.3, H 4.8%; M': 336.099 I. 2-n-Propyl-l,4,5-tri1r1etl~o.q\~n1rtl1r~~ce11e-9.I0-cliut1e 2-11-Propyl-5-hydroxyquinizarin (2) (122 mg) was boiled with rcflux in dimethoxycthanc (15 mL) containing NaH (400 mg of 50% dispcrsion in oil) and Mcl ( I g) for 12 h. Addition of Et,O, filtration. and evaporation of solvcnt gave a brown pastc which was washcd with pentane, 2 M HCI. and extracted with Et?O. Rcmoval of solvcnt gavc a gum which was purificd by plc on silica gcl GF254 eluting with EtOAc-Et20 ( 5 : 3) to give thc ether (58 mg). mp 49-52°C; 6,,: 1.00 ( 3 H . t . J = 7.5Hz). 1 . 6 6 ( 2 H . n 1 ) . 2 . 6 6 ( 2 H . t . J = 7.5Hz).3.83

S a n d 2 H z ) , 7 . 4 6 ( 1 H , t , J = 8Hz),7.61 (1H.dd.J CZl1H2~,OS requires M': 340.1310; found: 340.1304.

=

8and2Hz).

Cleavcrge of 2-n-propyl-l ,4,5-tritnetI1o,~~~n1?1I~1'nce1re-9, 10-(Iiotle The ether (56 mg) in din~ethoxyethane( I 5 mL) containing KOBu' (250 mg) and H,O (12 mg) was boiled with rcflux under a Nz atmosphere for 4 h. Brine (20 mL) was added and the solution extracted with Et,O (2 X 30 mL). Acidification of the aqueous layer, extraction with EtOAc. and conccntration of the dried extract gavc a gum which was treated with CHIN,-Et20. Removal of solvent gave a gum which was resolved by plc on silica gel GF254 by eluting with pcntane-Et,O (4:3) into mcthyl 3-n~ethoxybenzoate(6 mg) (identical to an authentic ( I l mg); 6": specimen) and tnerhyl2,5-clirnetho,~y-4-n-prop!'lbet1zo~e 0.96 (3H, t. J = 7.5 Hz), 1.62 (2H, m), 2.57 (2H, t , J = 7.5 Hz),, 3.69 (3H, s), 3.71 (6H, s), 6.65 (IH, bs), 7.27 (IH, s). CI,H,KO, requires M': 238.1205; found: 238.1205. Synthesis of methyl 2,5-dit~1e~hoxy-4-n-propylbe11zonte I ,4-Dimethoxybenzene (2.76 g) and propionic acid (2.22 g) were dissolved in polyphosphoric acid (120 mL) and the solution stirred at 70°C for 2 h. H,O (100 n1L) was cautiously added and the mixture extracted with EtzO (2 x 100 mL). The ethereal layer was washed with saturated NaHC0,-H20, H20. dried. and concentrated to give 2-propionyl-l,4-dirnetlzos~~herrzer~e (3.69 g) as a low melting (ca. 8°C) solid. Anal. calcd. for C , ,HI,O,: C 68.0. H 7.3%; found: C 68.2, H 7.3%. An amalgam was prcpared from HgC12 (150 mg), Zn dust (2 g) in 10 M HCI (5 mL) was added to the ketone (970 mg) suspended in H,0 (20 mL), and the mixture boiled with reflux for 5 h. During this time 10 M HCI was added in portions (5 X I mL). Aftcr 18 h at ambicnt temperature the mixture was extractcd with EtzO (3 X 15 mL), the extract washed with saturated NaHC0,-H,0, HzO, dried, and concentrated to give a brown oil which on bulb-to-bulb distillation (430 mg). A~rnl.calcd. for gave 2,5-cli1nethox~~-/-n-pr-opylhenzerle C I I H l d O ZC: 73.3, H 9.0%; found: C 73.7, H 9.0%. The ether (400 lng) in CS, (10 mL) was slowly added to powdered AlCl, (355 mg) in CSI (20 n1L) at 0°C. After 17 h the mixture was pourcd on to ice and 10 M HCI ( I mL) added. Extraction with EtzO (3 X 15 n1L) gave, after washing with H20, drying, and concentration, an oil which was chromatographed on silica gel; elution with pentane-Et,0 (3: I ) gave starting ether (149 mg) and 2,5-dir11etho,~~~-4-n-p1'opyIc1~'etopI1e11o mp 38-39°C (pentanc). A I I ~ calcd. ~. for C13HI,03:C 70.3, H 8.1%; found: C 70.7, H 8.1%. The ketone (30 mg) in dioxan (0.3 n1L) was added to 5% NaOCI-H20 (15 mL) and 10% NaOH-H20 ( 1 mL). After heating for 2 h on a steam bath. thc cooled solution was cxtracted with EtzO and the aqueous layer acidified and extracted with EtOAc (3 x 10 mL). Drying and concentration of the EtOAc extract gave a gum which was treated with CHzN,-Et30. Preparative layer chromatography of the product on silica GF254 and elution with pentane-Et10 (4: 3) gave methyl 2,5-dimethoxy-4-11-propylbcnzoate identical with the material obtained from degradation of the ether by nnir, ir, and mass spectroscopy and tlc.

3-n-Propyi-l,4.5-trihycfru,~a11thrn(1~1i1~011e le~cco-5-Hydroxyquinizarin(13) (1.087 g) was suspended in EtOH (70 mL) and the mixture warmed to 60°C under a NZ atmosphere. Propanal (3.6 g) was addcd, followed by piperidinium acetate (665 mg) in EtOH (5 mL). After boiling with rcflux for 3 h, the cooled solution was carefully acidified with 5 M HCI. The red precipitate was filtered off. dried (828 mg), and recrystallized from CH2CI2to give the 3-n-propyly~linorle (3) (744 mg), mp 193°C. AIICII. calcd. for CI7H1,O5:C 68.5. H 4.7%; found: C 67.9. H 4.4%.

I-Hydrosynt~threrq~tit~or~e 1e~re.o-Quinizarin(12) (242 lng) in PhMe (150 mL) was refluxed undcr a N, atmosphere in a Dean-Stark apparatus. Pyrrolidine (7 1 mg) in PhMc (20 mL) was added and the solution boiled with reflux until no more H 2 0separated (ca. 4 h). The solid obtained aftcr evaporation of solvcnt was extractcd with hot Mc,CO (3 x 50 mL). Evaporation of thc Mc,CO cxtracts gavc a dark ycllow solid which was cxtractcd with hot CCI, ( 3 X 50 mL). Evaporation of this cxtract and rccrystallization of thc residue from Mc2C0 gave thc quinone ( 1 86 mg). (3H.s).3.91(3H.s),3.94(3H.s),6.98(lH.s).7.00(lH.dd.J=

HARWOOD ET AL.

was acidified with 10% HCI-H20 and extracted with CHCI, (3 x 25 niL). Concentration of the dried solution gave the c~lcohol11 (320 mg), mp216°C(CHC13);61.1:14.03 ( I H , s). 13.46(lH, s), 13.29 (IH, s), 7.83 ( I H , dd.J = 8and 1.8 Hz), 7.65 ( I H , t,J = 8 Hz). 7.41 ( I H , t,J=0.6Hz),7.25(1H,dd,J=8and1.8Hz),4.91(lH,bd,J= 5.6 Hz), 1.78 (2H. m), 0.94 (3H. t , J = 8 Hz). CI7HI4Oh requires M': (IH.s).14.28(1H,s),8.2(2H.m),7.4(2H,m),5.93(1H,s),3.6 314.0790;found:314.0785. (4H, m), 1 95 (4H, m). Cl,HliNO, requires M': 309.1001; found: Hydro1y.si.s oJ' oc.etl11(9) 309.1000. This product was shown to arisc from quinizarin produced The acetal 9 (370 mg) was dissolved in hot MezCO (150 mL) and by oxidation of the leuco-compound. 5% HCI-H,O (70 mL) added. After 10 h at ambient temperature the Me2C0 was removed it, vacno and the solid recrystailized from Rec7ctiorl.s oJ'5-hj~elroxyqiri11izc1ri11 with nitrorlote otliotzs CHC13-pentane to give the ketot~e10 (287 mg), mp 189- 190°C; 5-Hydroxyquinizarin (256 mg) was suspended in EtOH (100 mL) v,,,.,: 1720 cm-I; 661: 2.99 (2H, t , J = 8.2 Hz), 2.84 (2H, t , J = 8.2 and a NaOEt solution (ex. 112 mg Na and 20 niL EtOH) added, forming a blue solution. I-Nitropropane (890 mg) was added and the Hz), 2.18 (3H, s). C I ~ H I , O ,requires M': 326.0790; found: 326.078 1 . solution boiled with reflux until no starting quinone remained (ca. 24 h). The cooled solution was acidified