2210
COMMUNICATIONS TO
THE
EDITOR
Vol. 83
The generality of this method for preparing other P-arabinosides for biological evaluation is currently under investigation. 12
tionally recrystallized three times from acetonitrile to give 3 g. of material, the analytical sample, m.p. 161-163' (Fisher-Johns block, uncorrected). (12) The synthesis of l-~-D-arabinofuranosy1-5-fluorouracil (Vc) A n d 6 Calcd. for CsHjCIPN: P, 19.66; N , 8.90: by another route w a s announced by J. J. Fox, N. Yung, I. Wempen, C, 45.74; H, 3.20; C1, 22.51; mol. wt. calcd. for R. Duschinsky and L. Kaplan, Abstr. Intl. Union Pure and Applied ( C C H ~ ( C I ) P N473. ) ~ , Found: P, 19.63: N, 9.01: Chemistry (Symposium on Natural Products), Australia, August C, 45.67: H, 3.44; C1, 22.38; mol. wt.,6445, 448. 1960, P. 66. Their synthesis begins with 5-fluorouridine (prepared The principal P-N ring infrared absorptions are a t from 5-fluorouracil) which is converted I J ~ Q a 2,2'-anhydronucleoside intermediate to Vc in 26% yield based upon 5-fluorouracil (personal 1180 cm.-l (s) and 1210 cm.-1 (s), with no discommunication from Fox and Yung). cernible absorptions a t the reported values2 for the LIFE SCIENCES DIVISION ELMERJ. REIST tetramer. Infrared analysis indicated t h a t the reH. OSIECKI mainder of the reaction product contained addiJEANNE STANFORD RESEARCH INSTITUTE LEONGOODMANtional trimer and also tetramer. Although recovery MENLOPARK,CALIFORNIA B. R. BAKER of trimer from this residue is difficult, procedures RECEIVED FEBRUARY 9, 1961 for its accomplishment are being investigated. Using a different procedure, the tetramer reSYNTHESIS OF cently has been prepared in this laboratory in PHENYLCHLOROTRISPHOSPHONITRILE better than 60% yield and this work will be reSir: ported shortly. Both trimer and tetramer now We wish to report the direct preparation and are being studied with respect to alkylation, arylapositive identification of phenylchlorotrisphospho- tion, and other substitution reactions. We wish to thank the Armstrong Cork Comnitrile, rPh(C1)PN 13. Although previous workers have studied the synthesis of the [Ph(CI)PNl, pany, Lancaster, Penna., for generous support of system, none has reported isolation of the trimer. this work. Thus Bode and Bach' treated PhPC14 with ammo(5) Schwartzkopf Microchemical Laboratories. nium chloride and could isolate only a partially (6) Ebulliometric measurement in benzene. We are indebted to hydrolyzed derivative for which analysis indi- Dr. Ralph G r i 5 t h , Sinclair Research Laboratories, for this measurecated the formula N3P3Ph3C1(OH)2. Shaw and ment. Stratton2 repeated the reaction and isolated MELLON INSTITUTE F. S. HUMIEC I. I. BEZMAX 13, PA. [Ph(CI)PN14 in two isomeric forms. Herring,3 PITTSBURGH RECEIVED FEBRUARY 27, 1961 using the novel procedure of treating NaN3 with PhPC12, isolated a mixture of phenylchlorophosphonitriles with an average molecular weight of STEREOCHEMISTRY OF SUBSTITUTION TO 5000. Recently, Tesi4 reported synthesis of ASYMMETRIC SILICON [Me(CI)PN13 by treatment of N3P3C13(NMe2)3 Sir : with MeMgBr to give N3P3Me3(NMe2)3which was Since the removal of the barrier to synthesis of converted to [Me(CI)PNI3by treatment with HCI. optically active organosilicon compounds having Although this procedure gives an [R(Cl)PW] reactive groups bonded to asymmetric silicon, and compound, i t is not a direct synthesis. I n our preparation PhPC14 was made by chlori- the discovery of many stereospecific reactions a t , ~ ~of~ ~the ~ major remaining tasks has nation of PhPCl2 (Victor Chemical Works) in ~ i l i c o n one carbon tetrachloride then recrystallization from been stereochemical correlation of configuration the same solvent under dry nitrogen. A solution for a few key compounds containing the cr-naphof 121 g. (0.484 mole) of PhPC14 in 250 ml. of dried thylphenylmethylsily1 group (a-XpPhMeSi-, deand redistilled s-tetrachloroethane was added over signated RaSi*-below), in order t h a t the stereoa period of 28 hr. to a slurry of 197 g. (3.71 moles) chemistry of many reactions of these compounds of NHrCl in 50 ml. of dry xylene. Reflux was might become known. One of the most widely used methods for correlatthen maintained for an additional 24 hours: unreacted ammonium chloride was filtered off, washed ing configurations of optically active compounds with dry benzene, and the washes were combined having similar structures is the Fredga method with the filtrate. Concentration of the solution based on diferences in phase b e h a ~ i o r . ~This gave a gummy solid (I) and solution (11) which method as applied by K. hlislow has provided could not be separated effectively by filtration. many fruitful results in recent years, and the pertiHowever, addition of petroleum ether converted nent case observed in the present work is his (I) to a solid which was filtered off and recrystal- "case 2"j in which pure optical isomers of two diflized from acetonitrile to give 8 g. of crude trimer, (1) L. H. Sommer and C. L. Frye, J . A m . ChPm. SOL.,81, 1013 m.p. 135-150° (A). Solution (11) was distilled to ( 19.59). (2) L. H. Sommer and C. L. Frye, ibid., 82, 3796 ( l Q f , O ) . dryness to give a hard gum which was crystallized (3) L.H. Sommer and C. L, Frye, ibid., 82, 4118 (1960). fractionally from acetonitrile to give an additional (4) See A. Fredga in "The Svedberg," Almqviut and Wikesells, 12 g. of crude trimer, m.p. 130-150' (B). The Uppsala, 1944, p. 261, and J. Timmermans, J . rhlin. p h y s . , 4 9 , 162 infrared curves of A and B were similar, showing (1952). Conclusions drawn on the basis of a diffeyence in phase bestrong absorptions in the 1200 cm.-' region, typical havior without exception have proved accurate. K. Mislow and M. Hefler, J . A m . Ckem. SOC.,74, 3668 (1952). of the trimeric phosphonitrile ring. A was frac- For( 5a) recent application of "case 2" for determination of the coufigura(1) H. Bode and H. Bach, B e y . , 7 6 , 215 (1942). (2) R . A. Shaw and C. Stratton, Chem. & I n d . , 52 (1959). (3) D. L. Herring, Chem. &Znd., 717 (1960). (4) G. Tesi, Proc. Chcm. Soc., 404 (1960).
tional relationships between the pure enantiomers of 3-thioloctanedioic acid and 3-methyloctanedioic acid see K. Mislow and W. C. Meluch, ibid.. 78, 5920 (1956). For other examples see J. Timmermanr,
ref. 4.
COMMUNICATIONS TO THE EDITOR
May 5, 19G1
ferent substances t h a t are isomorphous give solid solutions when they are of the same configuration and a eutectic when they are of opposite configuration. I n the present work X-ray diffraction provided clear and consistent answers for R3Si*F, RaSi*H and R3Si*CI: (1) Solid solutions are formed by (-)R3Si*F and ( -)R3Si*H; by (-)R3Si*F and (+)R3Si*C1; and by (-)R3Si*H and (+)R&3i*Cl. ( 2 ) Eutectic mixtures are formed by (+)R3Si*F and (-)RsSi*H; by (+)R3Si*F and (+)R3Si*C1; and by (+)R3Si*H and (+)RaSi*CI. Optically active R3Si*H, R3Si*F and R3Si*Cl crystallize individually in the orthorhombic system with space group P 2 1 2 ~ 2 ~an ; example of perfect isomorphism is observed. The three mixtures (1) also crystallize in the orthorhomic system and are isomorphous with the pure component compounds ; no doubling of the unit cell dimensions or change in the symmetry is observed. Extension of the above stereochemical correlations to include R3Si*OH and some of its derivatives was achieved through use of optical rotatory dispersion curves obtained through the kindness of Professor Stanley Kirschner, who made his apparatus available to us for use a t Wayne State University.6 Very similar curves showing negative Cotton effects' were obtained for (-)RsSi*F, (-)RaSi*H, (+)R3Si*C1, (-)R3Si*OCH3, (-)R3Si*OCOCH3, and (-)R3SiOH. Since the first three compounds have the same configuration from X-ray application of the Fredga method, and the last three are rigorously known to have the same configuration from chemical correlations of configuration,2 it is concluded t h a t all six compounds have the same configuration. Rotatory dispersion data for four of the compounds in cyclohexane solvent a t 24' and c. 0.2 are given: (- )RaSi*F;
-
- 33',
[alt~o 108', [ a j 3 6 0 - 164O, [cy1310 -22O0, [ c ~ I 3 3 3 - 2 4 8 ' , [~~]sze-160', [~~]31s-60' (-)RsSi*H; [01]689-30°1 [0(1400-100~, [rU1ss0-140°, [aluo - 160°, [ -230°, [a]816 54' (-)R3Si*OH; f01]589-21', [altoo-131', [a13so-218°, [alssr [a1589
-
HpO in EtzO
(1) ( +)RaSi*C1 ____) ( +)R3Si*OH LiAIH, (2) (+)R3Si*Cl -+ ether
(+)R3Si*OCH3
pentane; amine
LiAIHl (4) (+)R3Si*F 4(-)RsSi*H ether
KOH(s) Hz0 (5) (+)R3Si*Cl ___) --+(+)R3Si*OH xylene
NaB( OCH3),
(6) ( +)R3Si*Cl
-
ether
( - )R3Si*C1
KOH(s) xylene
xylene (meso) RaSi*OSi*R3
LiAlH,
-
(9) (+)RaSi*-O-COCeH5
___f
ether
c 1 2 _3
cc14
(12) (+)R3Si*OCI&
pentane; amine (-)R,Si*OCHa
(+)R3Si*C1
LiAlH,
-+ ether
(13) (+)R3Si*OSi*R3
( -)R3Si*H
CHsOH
(10) (4- )R3Si*-O-COCH2C1
(11) (-)R3Si*H
( +)R3Si*OCHs
____f
(7) (+)R,Si*OH-
(+)R3Si*H
LiAIHl
___f
ether
(+)R3Si*H
KOH(s) HzO (14) (+)RsSi*H
xylene
___f
+(+)R3Si*OH
+ t-C,HgMgCl -+ (+)R3Si*H + 24 hr.
(15) Grignard reduction: (+)R3Si*OCH3 95O
-
(6) S. Kirschner, A. J. Sonnessa, D. C. Bhatnagar and D. Moy, Abstracts of the 138th Meeting of the American Chemical Society, September, 1960, p. 14-N. (7) As has been pointed out by C. Djerassi, "Optical Rotatory Dispersion," McGraw-Hill Book Co., New York, N. Y.,1960, pages 102-108, correlation between rotatory dispersion curves and canfiguration for acyclic compounds must take into account the possibility of free-rotationnl "conformational" isomerization. However, this kind of uncertainty of interpretation certainly should not obtain for five of the compounds studied on the basis of examination of accurately scaled molecular models and, in fact, does not obtain even in the case of R&i*C1.
-
(+)R.%Si*H
MeOH
(3) (+)R,Si*Cl
-314'9 [Or]S26-44' (-)RaSi*OCHs; [a]~g-15', [a]4oo-6O0, [a]~30-102', f a1340 - 133', [ a l s a z 157", [a]sm-63', [a]soa-4O0
All six compounds studied by rotatory dispersion have the first maximum in the ultraviolet a t 317 mp (E = 310). The stereochemical correlations of configuration resulting from the above experimental data permit the definitive formulation of the stereochemistry of the reactions of R3Si*X already published5'53 plus some additional reactions not yet published. Reactions 1-10 proceed with inversion, and reactions 11-15 with retention of configuration.
2211
CHz=C( CHa)2
(16) (
- )RaSi*OCHs
CHaONa(10-'M) _____f
CHsOH racemization; too fast to measure
For reactions with essentially nucleophilic reagents, the above stereochemical data may be summarized in terms of leaving groups : (1) For good leaving groups such a s -Cl or -OCOR, whose conjugate acids have p K a smaller than ca. 5, the favored stereochemistry i s inversion of configuration. (2) For poor leaving groups such a s -H, -0CH3, or -OH, whose conjugate acids have pK, larger than ca. 10, the predominant stereorhemistry depends upon the relative importance of several factors and m a y be either retention or inversion of configuration in individual cases. The interesting complexities engendered in organosilicon stereochemistry by use of hybrid orbitals containing a 3d component, as revealed by
.>'I 12 --
COMI.\IGIUIC.ITIOXS TO THE EDITOR
1-01. s3
the present work :ind previous studies of bridge- C I I E M O T H E R A P E U T I C EVALUATIOX O F STEREOISOMERS O F head silicon, will be dealt with at length in a later " 2 ' - (ETHPLEXEDIIMIXO)-DI-~-BUTANOL.~HC~ -Oral t r e a t m e n t a p Subcutaneous treatmenta full article. EffiEffib5-ethank Dow Corning Corporation for conMan. cacy Max. Sacy tol. intol. Intinued generous support. EDsob dosee dexd EDaob dosec dexd L. 11. SUMMER(-)-Isomer Inactive' 6400 Inactivee 800
C. I,. FRCE meso Form 500 6400 500 800 ILi C. MVSOLF 90 6400 90 800 COLLEGEOF CHEMISTRY A X D PHESICS C; A PARKER (&)-Form THF,PEYSSYLVASIA STATEUNIVERSITY P G RODEWALII(+)-Isomer -15 6100 120 45 SO0 18 I I I 5 E R S I T C P.4RK, P A K TV ~ I I C H A F IStreptomycin , 80 400 5 lr. OKAC.4 Isoniazid 1.2 100 so R.PEPIVSKV a ildiriinistration by daily single dosage. Estimated RECEIVED MARCH 30, 1961 median effective dose (in mg./kg./day administered for 14 days from day of infection) required for 60 day survival STEREOSPECIFICITY IN A NEW TYPE OF SYNTHETIC where all infected untreated control mice died in an average t h e of 17 days. The infecting organism was the human ANTITUBERCULOUS AGENT',2 The strain of Mycobacterium tuberculosis, H37R,. Sir: maximum tolcrated dose is the amount in mg./kg./day adWe wish t o report the synthesis of a new highly- ministered for 14 days which gave about 10% weight loss Ratio of maximum tolerated dose t o after O n e week. active antituberculous compound, the dextroro- median effective dose. e Inactive a t the maximum tolerated tatory form of 2,"- (ethylenediimino) -di-1-butanol dose.
[ethambutol),:' which is four times as active as streptomycin against an established infection with the human strain of Xycobactcrium tuberculosis in mice. It is also fully active against isoniazidresistant infections as well as against streptomycinresistant infections in mice. The sharp stereospecificity of the activity of this synthetic chemotherapeutic agent (dextro = 12 x rneso = >200 X lero) contrasts with the equality in acute toxicity of the stereoisomers in mice. The high efficacy index (ratio of tolerance to potency) of this cornpound has warranted its clinical trial, the results of which will be reported elsewhere. The experimental chemotherapeutic study of the stereoisomers of " 2 ' - (ethylenediimino) -di-1butanol in comparison with streptomycin and isoniazid is summarized in the table. The antituberculous activity of these isomers varied considerably with the configuration a t the asymmetric carbons, the leilo isomer being inactive a t the maximum tolerated dose (at least a 200-fold difference in :wtivity'). The one center with the dextrorotatory configuration in the ineso isomer appears to be the source of its activity even though it is considerably l r s s than half as active as the dextro form, both in iiiizlo and iiz ~lliiro. I n contrast, the acute lethal toxicity in mice of the three isomers and the racemate is the same within experimental error. The reason for the remarkable stereospecificity in this synthetic antibacterial agent is under study. (+)L',?'-(Ethy1enediimino)-di-1 -butanol showed no activity against lethal infections with Gram-negative and Gram-positive organisms in mice. The high antibacterial specificity also was confirmed by the lack of activity in vitro against these organisms or against various fungi. (+) -2,2'-( Ethylenediimino)-di-1-butanol also was highly effective using delayed treatment of an established mycobacterial infection of mice. I n addition, the activity against infections with inycobacteria resistant to current drugs has been demon(1) Arlclitional data will be published by Thomas, Baughn, XVilkinson and Shepherd, A m . Rea. Res?. D i s . , 1961. (2) \Ve wish t o acknowledge the valuable assistance of Drs. 5. Kushner and H. J. White of these Laboratories. (3) Ethambutol is the generic name reserved for this isomer by the I.rderlr Lahoratories Division of the American Cyanamid Company.
strated clearly. Against a lethal infection in mice with a human strain resistant to the maximum tolerated dose of streptomycin, this new agent was fully as active parenterally as against the strains sensitive to this drug. T h e same was true of oral treatment in a lethal infection with a bovine strain resistant to the maximum tolerated dose of isoniazid. Equally important is the fact that repeated growth in zlitro in its presence has failed so far to show the development of resistance to this substance by the human strain of Mycobacterium tuberculosis, H37Rv. The highly active (+)-isomer was prepared b y brief heating of ethylene dichloride with excess A -4270yield of purified (+)-2-aniino-l-b~tanol.~ (+) -2,2'-(ethylenediimino) -di-1-butanol (m.p. base5 87.5-88.8', m.p. dihydrochlorideG 198.5200.3') was obtained after removal of the less soluble rneso isomer (m.p. base 135.8-136.5'; m.p. dihydrochloride 203.5-204.6 ') corresponding to approximately the amount of lcz~oimpurity in the (+)-2-amino-l-butanol. The Zevo diamine (m.p. base7 S9-90'; m p . dihydrochlorideY200.qj201.5") was prepared in the same way from ( - ) 2-amino-1-butanol4 in 52% yield. When ( + ) 2-aminobutanol reacted with either ethylene chloride, ethylene bromide or ditosyl glycol, the condensation occurred most rapidly a t the temperature obtained without solvent, although a longer time a t lower temperature (alcohol) has given comparable results. The main products, each isolated in about 40%, yield, are the racemic base (m.p. 7,5-7Go : b.p. 160- 170' (0.3 mm.) : dihydrochloride m.p. 179-ISO') and the meso diamine. These were separated readily as a result of the low, solubility of the latter in a number of solvents. The racemic and meso isomers of ,3,p'-diethyl-l,4-piperazine( 4 ) F r o m commercial (+)-7-aminobutanol by the tartrate resolution procedure of F. H. Radke, R. B. Fearing and S. W. Fox, J . .4m.C h m .
,58,8>%C , 11.8% H, 13.7'; S . Found: 58.8% 755 K. ( 0 ) a%5 +5.5' =t0.4 (HzO). A i d . Calcd.: 43.3% C, 9.5% H, 10.1% 3 , 25.09'a C1. Found: 43.5% C,9.7% H,10.4% N , 25.6% C1. (7) A n a l . Found: 58.5% C, 12.0% H,13.8% N. (8) as; -4.7 i 0.4" (HzO). Anal. Found: C, 43.5; H, 9.5;
C , 1 2 1% H, 13
s.i n 2 7 : ; a.2 5 . 4 ~ " .
