J. Am. Chem. Soc.

1985, 107, 3891-3898 3891

A Stereospecific Route to Aziridinomitosanes: The Synthesis

of Novel Mitomycin Congeners

Samuel Danishefsky,* Ellen M. Berman, Marco Ciufolini, Sarah Jane Etheredge, and
Brigitte E. Segmuller
Contribution from the Department of Chemistry, Yale University,
New Haven, Connecticut 06511. Received July 23, 1984

Abstract: A stereospecific route to the aziridinomitosanes is described. The key phases of the synthesis are the following:
(i) the coupling of a phenol with a functionalized allylic alcohol by an adaptation of the Mitsunobu technology (see 6 + 7
-*·
8); (ii) a Claisen rearrangement to establish the full side chain required for the synthesis (see 8 —* 9); (iii) a stereospecific
seleno-amination-alkylation-oxidation sequence (see 15b -* 20, 21 and 22); and (iv) a stereospecific aziridination sequence
which begins with a concave face hydroxylation (see 19 —24 -*· 31). The preparations of novel mitosene fV-oxide derivatives
(see compounds 34 and 35) as well as quinone ketal related to the 9a-deoxymitomycins (see compound 37) are also described.
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
Downloaded via NATL LBRY OF SERBIA on September 24, 2024 at 13:32:47 (UTC).

The structural novelty of the mitomycins and their chemical

lability constitute a formidable challenge to those who would
undertake their synthesis.1,2 Interest at the purely chemical level
was augmented by early demonstrations, in various experimental Me/" >V/ "-N·"
models, of antibiotic and antitumor potential of some of these S
i
compounds.3 The emergence of mitomycin C as a clinically useful '-H
anticancer chemotherapeutic resource43,13,5 promotes continuing mitomycin A X = 0Me; R = H
research in this field.6,7 However, in spite of the enormous amount mitomycin C® X=NH2; R =H
of synthetic activity which the mitomycins have inspired, the only porf ¡romycin X =
NH2; R=Me
successful total syntheses of naturally occurring mitomycins were
those described by Kishi8,9 and collaborators.
Our efforts have been directed toward several goals. Of course,
it was hoped to achieve a total synthesis of one or more of the
naturally occurring mitomycin congeners (see structures below).
8
Also, de novo synthesis could lead to new kinds of mitomycin
10

analogues which would not be readily obtained by modification

of the natural products themselves.
Compounds of the type c were selected as intermediary targets.

IH-pyrrole ¡1,2-0 indolej

(1) Kametani, T.; Takahashi, K. Heterocycles 1978, 9, 293.

(2) Franck, R. W. Fortschs. Chem. Org. Naturst. 1979, 38, 1.
(3) Remers, W. A. “The Chemistry of Antitumor Antibiotics”; Wiley-In-
terscience: New York, 1979; Yol. I, p 221 ff.
(4) Remers, W. A. In “Anticancer Agents Based on Natural Product
Models”; Cassady, J. M., Douros, J. D., Eds.; Academic Press; New York,
1980; p 131 ff.
(5) Lown, J. W. In “Molecular Aspects of Anticancer Drug Action”;
Neidle, S., Waring, M. J., Eds.; Verlag Chemie: Weinheim, 1983; p 283 ff. 5,
(6) For some notable contributions to this field since the review articles,
see: (a) Rebek, J.; Shaber, S. H. Heterocycles 1981,16, 1173. (b) Nanita, Such systems might lend themselves to introduction of the aziridine
Y.; Nagai, N.; Maruyama, K. Chem. Lett. 1983,1383. (c) Naruta, Y.; Arita, linkage in a stereospecific fashion. A riskier venture would involve
Y.; Nagai, N.; Uno, H.; Maruyama, K. Ibid. 1982, 1859. (d) Luly, J. R.; an attempt to introduce the required 9a helero functionality
Rapoport, H. J. Am. Chem. Soc. 1983, ¡05, 2859. (e) Luly, J. R.; Rapoport,
. Org. Chem. 1984, 49, 1671. through a position-specific oxidation of these or related compounds.
(7) (a) Danishefsky, S.; Regan, J.; Doehner, R. J. Org. Chem. 1981, 46,
The chemical issues to be explored were: (i) the feasibility of the
5255. (b) Danishefsky, S.; Regan, J. Tetrahedron Lett. 1981, 22, 3919. Claisen rearrangement14 with such an extensively functionalized
(8) (a) Nakatsubo, F.; Cocuzza, A. J.; Keeley, D. E.; Kishi, Y. J. Am. phenyl allyl ether; (ii) the feasibility of, and stereoselectivity of,
Chem. Soc. 1977, 99, 4835. (b) Nakatsubo, F.; Fukuyama, T.; Cocuzza, A. pyrroline ring construction via the format suggested in the
J.; Kishi, Y. Ibid. 1977, 99, 8115. (c) Fukuyama, T.; Nakatsubo, F.; Cocuzza, transformation of b — c; and (iii) the feasibility and stereo-
A. J.; Kishi, Y. Tetrahedron Lett. 1977, 18, 4295.
specificity of introduction of the aziridine ring. An account of
(9) Kishi, Y. J. Nat. Prod. 1979, 42, 549. our findings is provided herein.
(10) For the recently re-formulated absolute configuration of mitomycin
C, see: Shirahata, K.; Hirayama, N. J. Am. Chem. Soc. 1983, 105, 7199. Discussion of Results
(11) (a) Yahashi, R.; Matsubara, I. J. Antibiot. 1976, 29, 104. (b) Ya- Partial reduction of the benzylidene acetal l15 provided the
hashi, R.; Matsubara, I. Ibid. 1978, 31, 6. The absolute configuration of
mitomycin B presented in this paper has recently been questioned by Shir- monobenzyl ether 2. This compound has also served as a valuable
ahata10 and disproven by Hornemann and Heins (Hornemann, U.; Heins, M.
J. J. Org. Chem. 1985, 50, 1301).
(14) An account of earlier work in this area is provided in: Regan, J.
(12) Morton, G. O.; van Lear, G. E.; Fulmor, W. J. Am. Chem. Soc. 1970, Dissertation, University of Pittsburgh, 1980.
92, 2588.
(15) Eliel, E. L.; Badding, V. G.; Rerick, . N. J. Am. Chem. Soc. 1962,
(13) Urakawa, C.; Tsuchiya, H.; Nakano, K.-I. J. Antibiot. 1981, 34, 243. 84, 237.

0002-7863/85/1507-3891 $01.50/0 © 1985 American Chemical Society

3892 J. Am. Chem. Soc., Vol. 107, No. 13, 1985 Danishefsky et al.
Scheme I

8,—[R =(CH2)2OSIMe2,Bu;
R'
=CH20Bn]
—) 10 12a R=Bn; R'=H; X=OSi'BuMe,
9 b R=Ac-, R
—) =H;X=OSiTBuMe2
—

R1 =
[R=CH20Bn; (CH2)20Si Me2 BiQ
R
Bn;R'=N02;X=0Si,BuMe2 and OH
=
I3g
Jj, R=Ac; R'=N02;X=0SitBuMe2 and OH

14a, R°Bn; R'=NH2; X = OH

¿R=Ac; R'=NH2; X = OH
R=Bn; R'=NH2; X = Br
R'=NH2; X Br
=
b, R=Ac;

differentiated functionalized intermediate for other activities.16 to be correct. Compound 8 was converted to its Claisen rear-
Oxidation of 2 with pyridinium chlorochromate17 afforded the rangement product, presumed to be 10, and thence, via its benzyl
useful enal 3. In the course of the oxidation, the Z double bond ether, 12a, through a series of steps which served to corroborate
had undergone isomerization. By a simple Reformatsky-like the structural assignments (vide infra). The minor product 9 also
transformation,18 compound 4 is available from 3 in quantitative suffers Claisen rearrangement. However, when the resultant
(crude) yield. Compound 4 is reduced with lithium aluminum compound 11 is carried through the same steps, it produces
hydride. The diol 5 is thus available in 80% yield from 3. Reaction products which are no longer related to those obtained starting
of 5 with fert-butyldimethylchlorosilane19 in the presence of di- with 10. Accordingly, in practice it was not necessary to effect
methylaminopyridine provided the monosilyl ether 6 (80%). the separation between 8 and 9.
Little need be said about the synthesis of the required phenol Compound 10 was converted to the bisbenzyl ether 12a, and
7 8,9,20,21 initia] attempts to attach the phenol to the allylic carbinyl to the ether acetate 12b. The bisbenzyl ether 12a was converted
center by a derived allylic mesylate or bromide met with failure. by nitration (90% fuming nitric acid-acetic anhydride-mercuric
Some years ago, Manhas22 et al. described the application of the acetate)25 to 13a, which was accompanied by some of its free
Mitsunobu condensation23 reaction to the synthesis of cholesteryl alcohol arising from cleavage of the silyl group. Reaction of this
phenyl ether. Several other instances24 of the use of the azodi- material with zinc dust and aqueous HC1 afforded compound 14a.
carboxylate-phosphine mediated “dehydrations” in the synthesis The primary hydroxyl function underwent smooth displacement
of phenolic ethers had also appeared during the course of our upon reaction with carbon tetrabromide-triphenylphosphine.26
investigations. The amino group did not seem to complicate the transformation.
The extendability of this reaction to the allylic alcohol system In this sequence, compounds 12a, 14a, and 15a were fully
present in 6 was investigated. In the event, coupling was achieved characterized.
in ca. 80% yield through the use of tri-n-butylphosphine and ethyl The corresponding transformations were conducted on large
azodicarboxylate in ether. Examination of the crude coupling scale in the acetate series (12b-15b) without chromatography and
product indicated a ca. 10:1 ratio of two closely related compounds. without full characterization of intermediates between the coupling
These products were separated by preparative scale HPLC, though product mixture (10 + 11) to the tricyclic olefin 22. As in the
not without considerable loss of tempo. The NMR spectra of benzyl ether series, nitration was carried out with nitric acid and
compounds 8 and 9 are quite similar and do not rigorously reveal mercuric acetate in acetic anhydride, and reduction was conducted
the structures of the allylic isomers. with zinc dust in aqueous HC1. Once again, the primary hydroxyl
It was presumed that the major isomer is 8 while the minor group thus generated (see structure 14b) underwent clean con-
product is the SN2'-derived system 9. This surmise was shown version to the anilino bromide 15b.26
The stage was now set for the critical biscyclization of the
properly functionalized, properly positioned hexenylaniline system.
(16) Ko, S. Y.; Lee, A. W. M.; Masamune, S.; Reed, L. A., Ill; Sharpless, As expected, the trans double bond in 15a and 15b mitigated
K. B.; Walker, F. J. Science 1983, 220, 949.
against intramolecular alkylation of the amino group by the
(17) Corey, E. J.; Suggs, J. W. Tetrahedron Lett. 1975, 16, 2647.
(18) (a) Rathke, M. W. Org. React. 1975, 22, 423. (b) Rathke, M. W.; primary carbon bearing the bromide. It was hoped that attack
Lindert, A. J. Org. Chem. 1970, 35, 3966. of a suitable electrophile E+, upon the double bond, leading to
(19) Corey, E. J.; Venkateswarlu, A. J. Am. Chem. Soc. 1972, 94, 6190. establishment of the indoline ring, would also set the stage for
(20) Roger, R.; Demerseman, P.; Laval-Heantet, A.-M.; Rossignol, J.-F.; a second alkylation to generate the complete pyrroloindoline
Cheutin, A. Bull. Soc. Chim. Fr. 1968, 1026. See also ref 8 and 9.
(21) Our somewhat modified route to phenol 7 is provided in the supple- system required. In fact, we have reduced this concept to practice
mentary material. with several electrophiles.27 However, here we describe the
(22) Manhas, M. S.; Hoffmann, W. H.; Lai, B.; Bose, A. K. J. Chem. Soc., outcome when the electrophile is the very interesting Nicolaou
Perkin Trans. 1 1975, 461.
(23) Mitsunobu, O. Synthesis 1981, 1.
(24) (a) Cooper, R. D. G.; Jose, F.; McShane, L.; Koppel, G. A. Tetra- (25) Stock, L. M.; Wright, T. L. J. Org. Chem. 1977, 42, 2875.
hedron Lett. 1978,19, 2243. (b) Bittner, S. B.; Assaf, Y. Chem. Ind. (London) (26) Hooz, J.; Gilani, S. S. H. Can. J. Chem. 1968, 46, 86.
1975, 6, 281. (c) Nakano, J.; Mimura, M.; Hayashida, M.; Kimura, K.; (27) For an account of these efforts, see: Berman, E. M. Dissertation, Yale
Nakanishi, T. Heterocycles 1983, 20, 1975. University, 1983.
Stereospecific Route to Aziridinomitosanes J. Am. Chem. Soc., Vol. 107, No. 13, 1985 3893

reagent, fV-phenylselenophthalimide (N-PSP).28

Contrary to earlier precedents,29 reaction of 15a with N-PSP
in methylene chloride followed by basic extraction afforded a high
yield of a diastereomeric mixture of selenides formulated as 16
and 17. At this stage, the stereochemistry of these products was

that the formation of a cyclic selenonium ion through the action

of N-PSP on the double bond is highly reversible and that the
process critically depends on anchimeric displacement by the
nitrogen. It can further be postulated that such displacement is
more ready in the diastereomer leading to 20 since the two large

not known. The relationship of the phenylseleno function of the groups emerge trans relative to the indoline ring. The corre-
sponding pyrroline ring which would lead to 18 had its large groups
junction hydrogen was assumed to be trans by the mechanism of
the process (i.e., trans addition of “amino” and “phenylseleno” disposed in a syn fashion. The reasons for the much greater
to the E double bond). It was therefore assumed that the two specificty in the case of the acetate 15b relative to the benzyl ether
15a must lie in subtle conformational differences which are not
substances were “facial isomers” differing in the configurations
of the chiral centers at C[ and C9a, relative to that at C9. Indeed, readily understood in detail. In any case, a stereospecific and
oxidative deselenylation produced a 1:4 mixture of olefins 18-19. highly processable route to compound 19 was now available.
With a practical route to 19 and 22 having been achieved, the
Again, at this stage, the stereochemistry of these compounds could
not be assigned with certainty.30 However, subsequent work with remaining issues in unspecified order involve (i) quinone formation,
compound 19 also produced via a stereospecific route (vide infra),
(ii) installation of the aziridine, and (iii) introduction of C9a
established it to be the compound which bears the exo-oriented heterofunctionality. The optimal stage, a priori, for attempting
the most difficult of these challenges, i.e., oxidation of the 9a
benzyloxymethyl group, as shown.
The practicalities of synthetic processing would be well served position, can be debated. We chose to demonstrate that the
if the cyclization were more specific. When the N-PSP reaction aziridine can, in fact, be fashioned from the double bond of
was carried out with the acetate 15b, a single tricyclic product compound 19 and to learn the stereochemical course of such a
was obtained. When this product, formulated as 21, was subjected process. Of the various early attempts to functionalize the double
to the action of m-chloroperoxybenzoic acid, a single tricyclic bond of 19 (without effecting aromatization),32 the reaction with
osmium tetroxide33 seemed to be the most promising. At this stage,
olefin, 22,31 was produced. The overall yield of compound 22 from
the 10:1 mixture of coupled isomers (8-9), without chromatog- it was not possible to deduce the stereochemistry of the diol.
Indeed, even the configurational relationship between C10 and C9a
raphy of any intermediates, was 31%. The Experimental Section
was not known with certainty although it was assumed30 to be
provides the details of this “flask-to-flask” conversion.
as shown. The stereochemistry of the eventual aziridine which
Since a benzyl ether at the phenolic center would provide more
was obtained, 32 (vide infra), shows the diol to be structure 24,
suitable protection in the installation of the aziridine, the acetate
was cleaved with potassium carbonate-methanol. The phenol 23,
wherein hydroxylation has actually occurred from the concave
face anti to the exo-disposed benzyloxymethyl group. Mono-
upon benzylation with benzyl bromide, afforded compound 19.
The relative configurations at carbons 9a and 10 in these com- mesylation34 was achieved through reaction of 24 with mesyl
chloride and triethylamine in methylene chloride. The mono-
pounds was not known at this stage,30 but became rigorously
defined at the stage of the aziridine derivative, 32. mesylate 25 was not well characterized. Rather, in crude form,
The reasons for the stereospecificity in the case of the acetate it was subjected to the action of tetra-A'-butylammonium azide35
15b and, indeed, the reasons for the direction of stereoselectivity in benzene under reflux. Examination of the infrared and mass
in the case of the benzyl ether 15a are not clear. It can be argued spectra of the product indicated that an azidohydrin, presumably
compound 26, had been produced. This product, upon treatment
with excess methanesulfonyl chloride and triethylamine in
(28) (a) Nicolaou, K. C.; Clarmon, D. A.; Narnette, W. E.; Sietz, S. P. methylene chloride gave rise to a crude product whose infrared
J. Am. Chem. Soc. 1979, 101, 3704. (b) Nicolaou, K. C. Tetrahedron 1981, and NMR spectra are consistent with its being an azidomesylate.
37, 4097. Reaction of this product, formulated as 27, with trimethyl
(29) (a) Clive, D. L. J.; Wong, C. K.; Kiel, W. A.; Menchen, S. M. J.
Chem. Soc., Chem. Commun. 1978, 379. (b) Clive, D. L. J.; Farina, V.; phosphite in tetrahydrofuran,8·9 led to a crude product, presumed
Singh, A.; Wong, C. K.; Kiel, W. A.; Menchen, S. M. J. Org. Chem. 1980, to be 28, which was treated with sodium hydride in tetrahydro-
45, 2120.
(30) The NMR spectrum of the major olefin isomer is very similar to that
of an analogous compound previously7·14 obtained by Regan in a very lengthy (32) An extensive discussion of these attempts is provided in: Berman, E.
route using activated cyclopropane chemistry. In that work, the stereochem- M. Disseration, Yale University, 1983, and unpublished results.
istry at C9 was formulated as exo on mechanistic considerations. Therefore, (33) Van Rheenen, V.; Kelly, R. C.; Cha, D. Y. Tetrahedron Lett. 1976,
the stereochemistry of the major olefin was formulated as shown above. That 1973.
this assignment is rigorously proven via compound 32 also supports the pre- (34) Considerable difficulty was experienced in various attempts to produce
vious proposal. the bis mesylate of diol 24, reflecting the unreactive nature of the C-l hy-
(31) A portion of crude acetoxy olefin 22 was hydrolyzed and O-benzylated droxyl, and allowing a ready and clean entry into the monoactivated species.
as described in the Experimental Section. HPLC analysis of the crude di- (35) Brandstrom, A.; Lamm, B.; Palmertz, I. Acta Chem. Scand., Ser. B
benzyl olefin thus obtained revealed no contamination by the isomer 18. 1974, 28, 699.
3894 J. Am. Chem. Soc., Vol. 107, No. 13, 1985 Danishefsky et al.

Preliminary forays toward this most challenging goal have only

resently been initiated. Some possible future directions which these
pursuits might follow a foreshadowed by the following three
experiments. It was found that oxidation of the now readily
available compound 19 with m-chloroperoxybenzoic acid affords,
very cleanly, the A-oxide 34.39·40
Also of interest with the finding that the A-methylaziridine 31
reacts with the same reagent to afford the anilino A-oxide 35,
without any apparent interference from the tertiary aziridino
28 X =0S07Me;
nitrogen.
’
r**' c.

Y =
N-P(OMe)2

furan. The presumed resultant, A-phosphorylaziridine, was

Finally, we note that partial catalytic debenzylation of com-
subjected to reduction with lithium aluminum hydride. There
pound 30 occurs specifically, such as to liberate the phenolic ether.
was thus obtained, after silica gel chromatography, the homo-
Reaction of compound 36 with DDQ in methanol affords the
geneous secondary aziridine, now known to be 30. Unfortunately,
the yields of pure 30 from diol 24 were in the range of 30-40%.
quinone ketal 37 in high yield. It is possible that quinone ketals
We sought to relate our synthetic intermediates to the 9a-
might prove to be more stable than quiñones themselves toward
some of the operations which might be useful for introduction of
deoxymitomycin series36 by methylation of the aziridino “N-H” the C9a hetero function.
group. Methylation was accomplished albeit only in 60-65% yield The results of ongoing investigations which seek to develop these
through the agency of methyllithium and methyl iodide in ether.
findings more fully will be disclosed is due course.
The resultant compound 31 underwent debenzylation via reaction
with sodium and liquid ammonia-THF. A highly crystalline Experimental Section41
intermediate, mp 155.5-156.5 °C, was produced in essentially (Z)-2-Butene- 1,4-diol Benzylidene Acetal (1). A solution of benz-
quantitative yield. There need be no uncertainty about either the aldehyde (106 g, 102 mL, 1 mol), benzene (500 mL), (Z)-2-butene-
gross structure of stereochemistry of this aziridino diol since an 1,4-diol (97 g of 91% pure cis compound, equivalent to 88 g of cis diol,
1 mol), and camphorsulfonic acid (l g) was heated under reflux for 18
X-ray crystallographic determination revealed its structure to be
that shown in 32.37 Barring some most improbable happenings h with continuous azeotropic removal of water. The dark reaction mix-
ture was cooled and washed with water (3 X 150 mL) and with brine (1
in going from the diol to the aziridine, the stereochemical as-
X 200 mL). Removal of the solvent under reduced pressure left a residue
signments of the various intermediates must be those shown in which was vacuum-distilled to give 139 g of pure acetal (79%), bp 99-100
formulas 24-31. Thus, the relationship C¡-C2 imino linkage °C (0.25 mmHg). A forerun was also obtained (18 g) as a 1:1 mixture
relative to the C9 hydroxymethyl group of this stereospecifically of benzaldehyde and derived acetal. NMR (CDC13, 90 MHz)
generated aziridine is one of the mitomycin A,C-porfiromycin 7.55-7.20 (m, 5 ), 5.81 (s, 1 H), 5.72 (m, 2 ), 4.30 (m, 4 H). ,3C
family (vide supra). NMR (CDClj, 22.5 MHz) 138.7, 129.7, 128.1, 127.9, 126.2, 101.9, 64.2.
The oxidation of p-methoxyphenol derivatives with 2,3-di- IR (CDClj, Cm"1) 3000, 2940, 2850. m/e 176 (P+). Caled for
chloro-5,6-dicyano-1,4-benzoquinone in methanol to afford p- CnHl204: m/e 176.0837. Found: m/e 176.0820.
quinone monoketals is well known.38 Indeed, this reaction was (Z)-4-Benzyloxy-2-buten-l-ol (2). A solution of LiAlH4 (8.5 g, 0.22
used in our work to prepare quinoketal 37 (vide infra). For the mol) in ether (200 mL) was carefully added (cannula) to a solution of
A1C13 (90 g, 0.67 mol) in ether (500 mL). The temperature of the
purpose at hand, it was of interest to develop a direct oxidation resultant mixture was maintained at 0 °C by cooling with an ice bath.
of compound 32 to its corresponding quinone. It was reasoned A solution of the acetal 1 (60 g, 0.34 mol) in ether (400 mL) was added
that this might be accomplished if aqueous rather than methanolic dropwise to the mixture of LiAlH4/AlCl3, with good stirring and cooling.
solvent were used. In the event, reaction of diol 32 with DDQ The addition required 30 min, after which time the reaction was stirred
in aqueous THF furnished a 61% yield of a purple powder, mp (0° to room temperature) for an additional 2 h. The reaction was
146-152 °C. The spectral properties of this compound in con- quenched by addition of 500 mL of 10% aqueous H2S04 (0 °C). The
junction with the rigorously defined structure of its precursor, 32, mixture was transferred to a separatory funnel and diluted with more
define it to be the hydroxymethylquinone 33, related in stereo- water to dissolve the aluminum salts. The organic phase was removed,
and the aqueous solution was extracted with more ether. The combined
chemistry to the major classes of mitomycins. ether extracts were dried (MgS04) and concentrated to give 58 g of
Future synthetic work in this area will require a more concise
product (96%). The compound was used without further purification.
and higher yielding route for installation of the aziridine. Fur- ‘H NMR (CDCI3, 90 MHz) 7.33 (s, 5 ), 5.80 (m, 2 H), 4.43 (s, 2 H),
thermore, to reach the actual mitomycins, a workable solution 4.05 (m, 4 ), 2.70 (br s, 1 H). ,3C NMR (CDC13, 22.5 MHz) 137.6,
to the C9a functionalization problem, at an appropriate stage of 132.2, 218.1, 127.5 (two overlapping resonances), 127.3, 71.9, 65.3, 57.8).
the synthesis, must be devised.

(39) The olefinic linkage of 7V-oxide 34 proved to be entirely resistant to

(36) Kinoshita, S.; Uzu, K.; Nakano, M.; Shimizu, M.; Takahasi, T.; peracid epoxidation even under forcing conditions.
Matsui, M. J. Med. Chem. 1971, 14, 103. (40) jV-Oxide 34 was produced by Dr. William H. Pearson of our labo-
(37) Crystallographic data for this compound are provided as supplemen- ratories. The same experimental procedure may be applied to the synthesis
tary material. of the N-oxide corresponding to compound 19 (95% yield after chromatog-
(38) Buchi, G.; Chu, P.-S.; Hoppmann, A.; Mak, C.-P.; Pearce, A. J. Org. raphy).
Chem. 1978, 43, 3983. (41) Protocols are provided as supplementary material.
Stereospecific Route to Aziridinomitosanes J. Am. Chem. Soc., Vol. ¡07, No. 13, 1985 3895

IR (CHClj, cm'1) 3400, 3000, 2350. m/e 160 (P+ 18). Caled for
-

been used and all of the phenol solution had been added. The reaction
CnH120: m/e 160.0888. Found: 160.0898 m/e. mixture was stirred for 2 h following the last addition of phenol solution.
(£ )-4-Benzyloxy-2-butenal (3). Alcohol 2 (150 g, 0.84 mol) was The reaction mixture was warmed to room-temperature and the white
added in one portion to a cool, rapidly stirring suspension of pyridinium precipitate of diethyl hydrazodicarboxylate was removed by suction fil-
chloroohromate17 (279 g, 1.29 mol) and Celite (300 g) in dichloro- tration through a pad of Celite supported by a bed of Merck 70-230
methane (2.5 L). After stirring for 3 h, the resulting dark mixture was mesh silica gel. The solid was washed well with ether and the filtrate
diluted with 1.5 L of ether. Filtration through a pad of Florisil (suction) was concentrated to a syrupy residue. Scrutiny of the crude reaction
left a dark solid residue which was washed with more ether. The com- mixture by HPLC (µ-Porasil, 5% EtOAc/hexane) established the pres-
bined filtrates were concentrated to give crude aldehyde 3 (95 g, 64%) ence of compound 8 contaminated by 10-11% of compound 9. Chro-
as an oil, which was used without further purification. NMR, matography of the crude aryl ether (150 g of Merck 230-400 mesh silica
(CDC13, 90 MHz) 9.59 (d, H, J = 8 Hz), 7.33 (s, 5 H), 6.81 (dt, H,
1 1 gel, 2-5% EtOAc/Hex) provided 11.8 g of purified product (80%), still
7, =
4, 72 =
15 Hz), 6.42 (dd, 1 H, 7, = 8, 72 = 15 Hz), 4.57 (s, 2 H), contaminated with 10-11% of the SN'-derived ether. Separation of the
4.23 (m, 2 H). ,3C NMR (CDC13, 22.5 MHz) 192.7, 152.8, 137.2, two isomers, while readily achieved by HPLC, proved impractical for
131.2, 128.1, 127.5, 127.2, 72.5, 68.2. IR (CHC13, cm'1) 1684. m/e 176 preparative purposes. Therefore, the mixture of regioisomeric ethers 8
(P+). and 9, as prepared above, was used for further reaction.
(£)-6-Benzyloxy-3-hydroxy-4-hexenoate (4). Methyl bromoacetate Compound 8. NMR (CDC13, 500 MHz) 7.31-7.37 (m, 5 H), 6.75
(103.6 g, 0.68 mol) was added in a single portion to a mixture of aldehyde (d, 1 H, 7 = 8 Hz), 6.49 (d, 1 , 7BA = 8 Hz), 5.80 (m, 2 ), 4.80 (m,
3 (92 g, 0.52 mol), trimethyl borate (160 mL), and zinc dust (53 g, 0.81 1 H), 4.42 (s, 2 ), 4.04 (dd, 1 H, 7, = 5, 72 = 13 Hz), 3.99 (dd, H, 1

mol) in THF (160 mL). A rapid exothermic reaction was observed. J¡ = 5, 72 = 13 Hz), 3.75-3.88 (a broad 2 H multiple! containing two
After the mixture was stirred for 6 h, the reaction was quenched by the 3 H singlets at 3.82 and 3.77, 8 ), 2.15 (s, 3 ), 2.10 (m, 1 ), 1.90

sequential addition of glycerol (200 mL), saturated aqueous NH4C1 (200 (m, 1 ), 0.90 (s, 9 ), 0.06 (s, 3 ), 0.04 (s, 3 H). 13C NMR (CDC13,
mL), ether (500 mL), and water (300 mL), with continuous stirring. 22.5 MHz) 152.6, 149.3, 145.1, 138.1, 133.0, 128.8, 129.1, 127.5, 127.3,
Separation of the organic layer and extraction of the aqueous phase with 120.7, 113.9, 105.0, 76.4, 71.5, 69.7, 60.1, 59.2, 55.5, 38.8, 25.8, 18.1,
ether (2 X 300 mL), followed by concentration of the combined ether 8.9, 5.5. IR (CHC13, cm'1) 3000, 2950, 1590, 1480, 1250. m/e 486 (P+),
extracts, afforded a residue of crude /3-hydroxy ester. Dilution with 300 471, 378, 321, 291, 207, 168. Anal. Caled for C28H42Si05: C, 69.16;
mL of CH2C12 and filtration through a pad of anhydrous Na2S04 gave , 8.70. Found: C, 68.90; , 8.51.
a solution of /3-hydroxy ester which was concentrated in vacuo. The Compound 9. NMR (CDCl, 500 MHz) 7.33-7.36 (m, 5 H), 6.75
product thus obtained (131 g, 100%) was used without further purifica- (d, 1 , 7ab = 9 Hz), 6.49 (d, 1 , 7BA = 9 Hz), 5.75 (dt, 1 H, 7 = 6,
tion. NMR (CDC13, 90 MHz) 7.32 (s, 5 ), 5.75-5.90 (m, 2 H), 7=14 Hz), 5.61 (br dd, 1 H, 7, = 14 72 = 7 Hz), 4.73 (m, 1 H), 4.63
4.45-4.70 (broad 1 H multiple! containing a 2 H singlet at 4.50, 3 H), (s, 2 H), 3.88 (s, 3 H), 3.79 (s, 3 H), 3.75 (m, 1 H), 3.65 (dd, 1 H, 7,
4.00 (m, 2 H), 3.67 (s, 3 H), 3.07 (br s, 1 H), 2.53 (d, 2 H, 7 = 6 Hz). =
6, 72 = 10 Hz), 3.60 (t, 2 H, 7 = 6 Hz), 2.22-2.29 (m, 2 ), 2.15 (s,
IR (CHC13, cm"1) 3480, 1720. m/e 250 (P+). 3 ), 1.56 (s, 9 ), 0.06 (s, 6 H). IR (CHC13, cm"1) 3000, 2950, 1590,

(£)-l-Benzyloxy-2-hexene-4,6-diol (5). A solution of /3-hydroxy ester 1480, 1250. m/e 486 (P+) 471, 378, 321, 291,276, 261, 233, 207, 168.
4 (66.5 g, 0.27 mol) in ether (500 mL) was added dropwise to a solution (±)-(9/?*,9aS*)-8-Acetoxy-9-[(benzyloxy)methyl]-9,9-dihydro-5,7-
of lithium aluminum hydride (15.2 g, 0.40 mol) in ether (500 mL), with dimethoxy-6-methyl-3//-pyrrolo[l,2-a]indole (22) without Purification
vigorous mechanical stirring. The temperature of the reaction was kept of Intermediates.42 A mixture of isomeric aryl ethers 8 and 9 (34.9 g,
below 10 °C. Addition of the ester required 1.5 h, after which time 72 mmol) was dissolved in 105 mL of AyV-dimethylaniline and the
stirring was continued for another 15 min. The reaction was quenched resulting solution was thoroughly purged with N2. The mixture was
by the sequential addition of water (15 mL), 20% aqueous NaOH (15 heated under reflux for 2 h. After cooling, the solution was poured into
mL), and water (45 mL). The granular precipitate of aluminum salts a separatory funnel containing ice (600 g) and concentrated HC1 (116
was separated by suction-filtration and concentration of the filtrate gave mL). Extraction with ether produced an organic phase which was
the crude diol (52 g, 88%). The crude compound was suction-filtered washed with H20 (1 X 300 mL) and brine (1 X 300 mL). The ether
through a pad of Merck 230-400 mesh silica gel, which was further phase was further dried (MgS04) and concentrated to afford crude
washed with ethyl acetate. Concentration gave 47 g of clear, colorless phenol 10 in quantitative yield. Dissolution of phenol 10 in pyridine (100
diol (80%) which was submitted directly to silylation. NMR (CDC13, mL) and treatment with neat Ac20 (23 g, 225 mmol) at 25 °C for 8 h
90 MHz) 7.34 (s, 5 H), 5.72-5.87 (m, 2 H), 4.52 (s, 2 H), 4.40 (m, 1 produced acetate 12b (36.1 g, 95%) after an extractive workup (2.4 N
), 4.30 (m, 2 ), 3.80 (br t, 2 H, 7 = 6 Hz), 2.50-2.88 (br, 2 H), HC1, ether, Na2S04) followed by filtration of the combined extracts
1.60-1.83 (m, 2 H). IR (CHC13, cm"1) 3380. m/e 131 (P+ C6H5CH2).
-

through a pad of Merck 230-400 mesh silica gel (suction). Nitration was
(£)-l-Benzyloxy-6-ferf-butyldimethylsilyloxy-2-hexen-4-ol (6). To effected by dissolving a portion of acetate 12b (24.5 g, 46.4 mmol) in
a solution of crude diol 5 (20.0 g, 90.1 mol) in 590 mL of dry methylene Ac20 (250 mL). To that mixture column added a solution of Hg(OAc)2
chloride was added 4-dimethylaminopyridine (440 mg, 3.6 mmol) and (1.42 g, 4.4 mmol) in glacial AcOH (125 mL). The mixture was stirred
at 0 °C during addition of fuming 90% HN03 (12.4 mL, 16.5 g of
triethylamine (13.8 mL, 99.2 mmol). To this system was added, dropwise
over 1.5 h, a solution of Zerr-butyldimethylsilyl chloride (11.56 g, 76.6 HN03) over 5 min. Stirring at 0 °C was continued for another 10 min.
mmol) in 140 mL of methylene chloride. The reaction was stirred at The mixture was poured over 1.5 kg of ice and extracted with ether. The
room temperature under nitrogen for 16 h. The solution was washed with ether extracts were thoroughly washed with aq 5% K2C03 (removal of
3% aqueous sulfuric acid (1 X 300 mL), saturated sodium bicarbonate AcOH and Ac20), then with brine. Drying over Na2S04, filtration, and
(1 X 300 mL), and then brine. The organic layer was dried over sodium concentration furnished crude nitro compound 13b (25 g) which was
sulfate, filtered, and concentrated in vacuo to afford a yellow oil. directly submitted to reduction. The compound was thus diluted with
116 mL of methanol and 116 mL of concentrated HC1. Stirring at 25
Chromatography on silica gel with 15% ethyl acetate in hexanes afforded
24.39 g (80%) of allylic alcohol. NMR (CDC13, 90 MHz) 7.33 (s, °C was continued for 15 min. Zinc dust (60 g, 0.91 mol) was added in
5 H), 5.84 (m, 2 ), 4.50 (br s, 2 H), 4.45 (br m, 1 ), 4.05 (d, 2 H, portions, and with good stirring (overhead), in such a way as to maintain
7=4 Hz), 3.82 (dt, 2 H, 7, = 3, J2 = 6 Hz), 3.27 (d, 1 H, 7 = 3 Hz), a vigorous, but not violent, reaction. The mixture was allowed to reflux
1.74 (m, 2 ), 0.89 (s, 9 ), 0.10 (s, 6 H). 13C NMR (CDC13, 22.5 (without external application of heat) whereby the orange-red color of
MHz) 138.5, 135.9, 128.5, 127.8, 127.7, 126.5, 72.2, 70.9, 70.3, 61.5, the nitro compound slowly disappeared. After 20 min, following addition
39.1, 26.1, 18.3, 5.3. IR (CHC13, cm"1) 3470, 3000, 2940, 1470, 1360, of the last portion of Zn, the solution was filtered through a plug of glass
1260. m/e 279 (p+ Z-Bu). Anal. Caled for C]9H32Si03: C, 67.81; H,
- wool. The zinc was washed well with methanol, then with CH2C12 to
9.58. Found: C, 67.59; H, 9.45. ensure complete removal of adsorbed product. The clear, nearly colorless
filtrates were diluted with 2 L of EtOAc. The mixture was washed well
(£)-l-Benzyloxy-6-ferf-butyldimethylsilyloxy-4-[(2,4-dimethoxy-3-
methylphenyl)oxy]-2-hexene (8) Plus Allylic Isomer 9. A 500-mL, with saturated aqueous NaHC03, whereby a pink color developed. The
three-neck flask fitted with overhead stirrer was charged with alcohol 6 organic layer was also washed with brine and dried over Na2SG4. Fil-
(12.3 g, 36.6 mmol) and ether (70 mL). Freshly distilled tri-n-butyl- tration and concentration afforded crude amino alcohol 14b (20 g) which
was used as such for the following bromination reaction. The compound
phosphine was added (12.2 g, 60 mmol) and the mixture was cooled in
a dry ice/CCl4 bath (-25 °C). Neat diethyl azodicarboxylate (DEAD) was dissolved in 460 mL of CH2C12 and diluted with 920 mL of ether.

was injected (2.1 g, 12.0 mmol), 1.9 mL), whereupon the orange color Solid CBr4 was added (30.4 g, 91.8 mmol) followed by triphenyl-
of the diester faded rapidly. Stirring at -25 °C was continued for 10 min, phosphine (14.4 g, 55 mmol). Stirring at 25 °C was continued for 30
after which time exactly one-fifth of a solution of phenol 7 (5.1 g, 30.5 min, during which time a precipitate of triphenylphosphine oxide formed.
mmol) in 30 mL of ether was added dropwise. The resulting mixture was The solution was concentrated to one-third of the original volume.
stirred for 1 h at -25 °C. The same sequential additions of DEAD [1.9
mL)/10 min at -25°/one-fifth of phenol solution/1 h at -25°] was (42) The preparation of bromide 15a via the fully characterized interme-
repeated four more times until a total of 10.5 g (60 mmol) of DEAD had diates 12a, 13a, and 14a is provided as supplementary material.
3896 J. Am. Chem. Soc., Vol. 107, No. 13, 1985 Danishefsky et al.

Filtration through a column of 500 g of Florisil (gradient hexane -» 15%, i H, JA,W = 12.5 Hz), 4.45 (d, JWA, = 12.5 Hz, 1 H), 3.78 (s, 3 H),
EtOAc/hexane) removed phosphine-derived materials and afforded 16 3.77-3.75 (m, 2 ), 3.70 (s, 3 ), 3.30-3.50 (m, 4 H), 2.83 (m, 1 H),
g of bromide 15b as a golden oil. 2.17 (s, 3 ), 2.17 (multiple! under methyl, 1 ), 1.90 (m, 1 h). IR
This unstable compound was immediately cyclized using the following (CHC13, cm-1) 2960, 1587, 1470, 1450, 1410, 1047. m/e 615 (P+).
procedure. To a solution of 16 g of bromide 15b in 2 L of CH2C12 was Caled for C35H37NO480Se: m/e 615.1888. Found: m/e 615.1887.
added in one portion at room temperature and under N2 12 g of NPSP.28 Minor Selenide 17. NMR (CDC13, 500 MHz) 7.60-7.20 (m, 15
After 40 min the mixture was concentrated and taken up with ether. ), 5.06 (d, 1 H, 7AB = 10 Hz), 4.97 (d, 1 , 7AB = 10 Hz), 4.39 (br
Treatment with 0.5 N aqueous NaOH (3 X 200 mL) was necessary to s, 2 ), 4.19 (dd, 1 H, 7, = 8, J2 = 10 Hz), 4.55 (m, 2 H, J¡ 2, J2 =
=

remove phthalimide from the desired material. The organic layer was 10 Hz), 3.77 (s, 3 H), 3.75 (m, 1 H, 7, = 8 Hz), 3.71 (s, 1 H), 3.55 (br
further washed with distilled water (2 X 300 mL) and brine (1 X 300 t, 1 H, 7 = 8 Hz), 3.49 (m, 1 H), 3.23 (m, 1 ), 2.18 (s, 3 H, overlapping
mL), then dried over Na2S04. Concentration furnished 18.2 g of crude with a m, 1 ), 2.04 (m, 1 H). m/e 615 (P+), 337, 261, 246, 218.
tricyclic selenide 21. The compound was immediately submitted to (±)-(9R*,9aS*)-3-Benzyloxy-9-[(benzyloxy)methyl]-9,9-dihydro-5,7-
oxidative deselenation. To this end, the material was dissolved in CH2C12 dimethoxy-6-methyl-3i/-pyrrolo[l,2-a]indole (19) Plus Stereoisomer 18.
(500 mL) and cooled to -25 °C (dry ice/CCl4 bath). Portions of a The mixture of isomeric selenides (1.56 g, 2.54 mmol) was dissolved in
solution of MCPBA (commercial 85% pure reagent, 6.6 g) in CH2C12 32 mL of dry methylene chloride and cooled to -23 °C (C02/CC14). A
(60 mL) were added until TLC analysis of the reaction mixture failed solution of m-chloroperbenzoic acid (85%, 0.570 g, 2.80 mmol) in 32 mL
to indicate the presence of selenide 21. Diispropylamine was added (10 of dry methylene chloride was added via a dropping funnel over about
mL) and the cold solution of selenoxide was rapidly poured into 1 L of 10 min. The reaction mixture stirred at -23 °C for 30 min before 540
refluxing CC14. Heating under reflux was continued for 10 min, after µ . of diisopropylamine was added. The faintly yellow solution was
which time the solution was cooled and washed well with 1 N NaOH. transferred to boiling carbon tetrachloride (ca. 100 mL) and the solution
The solution was also washed with brine and dried over Na2S04. Fil- refluxed for 5 min. The reaction was cooled to room temperature with
tration and concentration afforded a dark residue which was chromato- the aid of a cool water bath. After further diluting with 300 mL of
graphed over 700 g of Merck 230-400 mesh silica (gradient hexane - methylene chloride, the solution was washed with 1.0 N sodium hy-
15% EtOAc/hexane). The pure acetoxy olefin 22 thus obtained droxide (1 X 100 mL) and then with brine (1 X 100 mL). The organic
amounted to 6.4 g (31% overall yield from the Mitsunobu product). portion was dried over sodium sulfate and concentrated in vacuo at room
NMR (CDClj, 270 MHz) 7.40-7.32 (m, 5), 5.80 (br s, 2 H), 4.77 (br temperature; the residue (1.65 g) was chromatographed on silica gel
d, l H, J = 4.6 Hz), 4.56 (br d, 2 H), 4.19 (dd, 1 H, f = 4.3, J2 = 15.6 (10% ethyl acetate in hexanes). Scrutiny of the olefin so obtained (0.938
Hz), 4.04 (br dd, 1 H, 7, = 4.7, J2 = 15.6 Hz), 3.82 (s, 3 H), 3.74 (dd, g, 82%) using analytical HPLC (Waters µ-Porasil, 8% ethyl acetate in
1 H, 7 = 4.6, J2 = 8.9 Hz), 3.66 (s, 3 ), 3.61 (ddd, 1 , 7[ = 72 = 4.6, hexanes) revealed a mixture of isomeric olefins in a ratio of 3.8 to 1.
73
= 8.9 Hz), 3.46 (dd, 7, = 72 = 8.9 Hz), 2.18 (s, 3 ), 2.17 (s, 3 H). Analytical samples at both olefins were obtained by HPLC separation
13C NMR (CDC13, 22.5 MHz) 168.3, 144.1, 143.4, 143.2, 138.8, 136.8, of the mixture. However, the minor isomeric olefin was most easily
130.6, 128.1, 127.4, 125.0, 121.8, 76.6, 72.8, 72.0, 60.3, 60.0, 58.9, 57.5, removed after the mixture had been subjected to the action of Os04.
46.9, 20.1, 14.0, 9.3. IR (neat, cm"1) 2925, 2350, 1760, 1600, 1450, Apparently, osmylation of the minor olefin is less facile.
1400, 1195, 725, 680. m/e 409 (P+), 366, 352, 300, 286, 258, 244, 214, Major Olefin 19. NMR (CDC13, 500 MHz) 7.28-7.40 (m, 10 H),
91. 5.78 (m, 1 H), 5.75 (m, 1 ), 5.04 (d, 1 , 7AB = 11 Hz), 4.96 (d, 1 H,
(±)-(9R*,9aS*)-8-Benzyloxy-9-[(benzyloxy)methyl]-9,9-dihydro-5,7- 7ba = 11 Hz), 4.78 (br m, 1 H), 4.54 (d, 1 , 7A,B, = 12 Hz), 4.48 (d,
dimethoxy-6-methyl-3//-pyrrolo[l,2-a]indole (19). Acetate 22 (1.04 g, 1 H, JWA, = 12 Hz), 4.18 (dddd, 1 , 72 = 16, J2 = 4.5, J3 = 2 Hz), 4.04
2.54 mmol) was dissolved in 25 mL of dry methanol and treated with (br d, 1 H, 7 = 16 Hz), 3.98 (dd, 1 H, 7 = 9, 7 = 4 Hz), 3.81 (s, 3 H),
solid K2C03 (439 mg, 3.18 mmol). The suspension was stirred at room 3.79 (s, 3 H), 3.55 (ddd, 1 , 72 = 10, 72 = 4, 73 = 4 Hz), 3.45 (dd, 1
temperature for 2 h, after which time the reaction was quenched by H, 7, = 10, J2 = 8.6 Hz), 2.17 (s, 3 H). 13C NMR (CDC13, 22.5 MHz)
addition of 5 mL of water which had been saturated with lH (dry ice). 145.1, 144.6, 143.0, 141.4, 138.2, 137.6, 130.9, 128.0, 127.6, 127.4, 127.1,
The solvent was removed under vacuum and the residue was passed 126.7, 124.8, 121.4, 76.2, 74.2, 72.6, 72.3, 67.4, 60.0, 58.8, 57.6, 46.8,
through a short plug of silica gel (EtOAc). Concentration of the eluate 25.2, 9.0. IR (CHC13, cm"1) 2985, 2857, 1597, 1460, 1450, 1408, 1052.
left 896 mg of product (96%). This material was dissolved in 7 mL of m/e 459 (P+ + 2), 458 (P+ + 1), 457 (P+), 442, 336, 348, 258. Anal.
DME and 7 mL of DMF. Solid KH (300 mg; thoroughly washed with Caled for C29H3lN04: C, 76.12; H, 6.85; N, 3.06. Found: C, 76.02;
pentane) was added to the solution and the resulting suspension was H, 6.85; N, 2.95.
allowed to stir at 25 °C for 10 min. Neat benzyl bromide was added (500 Minor Olefin 18. NMR (CDC13, 500 MHz) 7.30-7.45 (m, 10 H),
mg, 2.9 mmol) and the resulting mixture was stirred at 25 °C for 3 h. 6.03 (ddd, 1 H, 7, = 5.5 , 72 = 4 , 73 = 2 Hz), 5.85 (ddd, 1 H, 7, = 5.5,
The reaction was quenched by addition of 5 mL of water which had been j2 = 4, 73 = 2 Hz), 5.06 (d, 1 , 7AB = 11 Hz), 5.00 (d, 1 , 7BA = 11
saturated with C02 (dry ice). Dilution with 20 mL of water and 20 mL Hz), 5.03 (m, 1 H), 4.47 (d, 1 , 7A,B, = 12 Hz), 4.41 (d, 1 , 7B-A- =
of saturated aqueous NaCl and extractive workup (ether) gave a crude 12 Hz), 4.14 (br dddd, 1 H, 7! = 16, J2 = 4, 73 = 2, 74 = 1 Hz), 4.07
product which was chromatographed (30 g of Merck 70-230 mesh silica (dddd, 1 H, 7, = 16, J2 = 5.5 , 73 = 2 , 74 = 2 Hz), 4.05 (dd, 7, = 10,
gel, 50% EtOAc/hexane) to afford 1.06 g of dibenzyl olefin (91%). 72
= 4 Hz), 3.80 (s, 3 H), 3.77 (s, 3 H), 3.66 (dddd, 1 H, 7! = 13, J2
NMR (CDC13, 500 MHz) 7.28-7.40 (m, 10 H), 5.78 (m, 1 H), 5.75 (m, =
n.4, 73 = 4,7„ = 1 Hz), 3.34 (dd, 1 H, 7, = 11.4, 72 = 10 Hz), 2.19
1 ), 5.04 (d, 1 , 7AB= 11 Hz), 4.96 (d, 1 , 7BA = 11 Hz), 4.78 (br (s, 3 H). m/e 459 (P+ + 2), 458 (P+ + 1), 457 (P+).
m, 1 H), 4.54 (d, 1 H, JA,W = 12 Hz), 4.48 (d, 1 , 7B,A, = 12 Hz), 4.18 (±)-(lS*,2R *,9R *,9aR*)-8-Benzyloxy-9-[(benzyloxy)methyl]-
(dddd, 1 H, 7, = 16, J2 = 4.5, 73 = 2 Hz), 4.04 (br d, 1 H, 7 = 16 Hz), 2,3,9,9a-tetrahydro-5,7-dimethoxy-6-methyl-17/-pyrrolo[l,2-a]indole-
3.98 (dd, 1 H, 7, = 9, 72 = 4 Hz), 3.81 (s, 3 H), 3.79 (s, 3 H), 3.55 (ddd, I, 2-diol (24). To a solution of iV-methylmorpholine iV-oxide (238 mg,
1 , 7[ = 10, 72 = 4, 73 = 4 Hz), 3.45 (dd, 1 H, 7, = 10, J2 = 8.6 Hz), 1.56 mmol) in 2.7 mL of ferf-butyl alcohol, 950 µ . of tetrahydrofuran,
2.17 (s, 3 H). 13C NMR (CDC13, 22.5 MHz) 145.0, 144.6, 143.0, 141.4, and 300 µ * of water was added 85 µ ^ of a solution of Os04 in tetra-
138.2, 137.6, 130.9, 128.0, 127.6, 127.4, 127.1, 126.7, 124.8, 121.3, 76.2, hydrofuran (1 g in 10 mL). The olefinic substrate (630 mg, 1.38 mmol),
74.2, 72.6, 72.2, 67.4, 60.0, 58.8, 57.6, 46.8, 25.2, 9.0. IR (CHC13, cm"1) contaminated by ca. 15% of the stereoisomeric olefin, was added as a
2985, 2900, 1597, 1460, 1450, 1408, 1050. m/e 459 (P + 2), 458 (P + soltuion in 3.4 mL of tetrahydrofuran. The reaction mixture was stirred
1),457 (P), 442,336, 348,258. Anal. Caled for C29H31N04: C, 76.12; at room temperature for 8 h. The solution was diluted with ethyl acetate
H, 6.85; N, 3.06. Found: C, 76.02; H, 6.85; N, 2.95. and washed with water and with brine. The organic portion was filtered
(±)-(lS*,9R*,9aS*)-l-Phenylseleno-8-benzyloxy-9-[(benzyloxy)- through a short column of Florisil and the filtrate was concentrated to
methyl]-2,3,9,9a-tetrahydro-5,7-dimethoxy-6-methyl-l//-pyrrolo[l,2-a]- 593 mg of a white, foamy solid. Chromatography on 15 g of silica gel
indole (16) Plus Stereoisomer 17. The bromide 15a42 (3.67 g, 4.74 mmol) with 50% ethyl acetate in hexanes gave 413 mg (61%) of diol 24 as a
was dissolved in 56 mL of dry methylene chloride and a solution of single isomer. !H NMR (CDC13, 500 MHz) 7.30-7.40 (m, 10 ), 5.04
TV-phenylselenophthalimide (N-PSP) (1.43 g, 4.74 mmol) in 45 mL of (d, 1 H, 7ab = 12 Hz), 4.96 (d, 1 , 7BA = 12 Hz), 4.53 (d, 1 H, 7A,B,
methylene chloride was added dropwise over 1 h. After stirring for 30 =
11 Hz), 4.51 (d, 1 H, JWA, = 11 Hz), 4.35 (m, 1 ), 4.12-4.60 (m,
min, the reaction was diluted with 350 mL of methylene chloride and 2 H), 4.0 (m, 1 H), 3.90 (m, 1 H), 3.78 (s, 3 H), 3.73 (s, 3 H), 3.63 (m,
washed with 0.5 M sodium hydroxide (1 X 200 mL) and then brine (1 1 H), 3.39 (dd, 1 H, 7, = 11, 72 = 8 Hz), 3.17 (dd, 1 H, 7, = 11, 72 =
X 200 mL). The organic phase was dried over sodium sulfate, filtered, 6 Hz), 2.16 (s, 3 H). 13C NMR (CDC13, 22.5 MHz) 145.3, 144.6, 143.2,
and concentrated in vacuo to 3.90 g of a reddish oil. Rapid chroma- 140.9, 138.1, 137.9, 128.4 (three overlapping resonances), 128.2, 127.9,
tography on 47 g of silica gel first with 5%, then 15% ethyl acetate in 127.7, 125.3, 122.3, 74.7, 73.7, 73.1, 72.7, 72.3, 60.5, 58.6, 55.8, 42.5,
hexanes yielded 1.56 g (57%) of a mixture of stereoisomeric selenides 9.4. IR (CHC13, cm'1) 3380, 1574, 1430. m/e 492 (P+ + 1), 491 (P+),
(estimated HPLC ratio ca. 4:1). 400, 372, 309.
Major Selenide 16. NMR (CDC13, 500 MHz) 7.20-7.60 (m, 15 (±)-(laR *,8R *,8aR *,8bS*)-7-Benzyloxy-8-[(benzyloxy)methyl]-
), 5.03 (d, 1 H, 7ab = 10 Hz), 4.92 (d, , 7BA = 10 Hz), 4.62 (d,
1 1,1 a,2,8,8a,8b-hexahydro-4,6-dimethoxy-5-methylazirino[2',3':3,4]-
Stereospecific Route to Aziridinomitosanes J. Am. Chem. Soc., Vol. 107, No. 13, 1985 3897

pyrrolo[l,2-a Jindole (30). To a solution of the diol 24 (390 mg, 0.79 overlapping resonances), 127.9, 127.4, 125.1, 123.4, 74.7, 73.6, 73.1, 72.3,
mmol) in 4.4 mL of dry dichloromethane at 0 °C under argon was added 60.5, 58.1, 55.2, 50.1, 48.4, 46.9, 45.4, 9.3. IR (CHC13, cm'1) 2906,
methanesulfonyl chloride (68 µ , 0.88 mmol, freshly distilled from P205) 1590, 1440, 1398. m/e 486 (P+). Caled for C30H34N2O4 m/e 486.2519
followed by triethylamine (170 µ , 1.22 mmol). After stirring at 0 °C (P+). Found: m/e 486.2509.
for 1 h 30 min, the reaction mixture was diluted with ether and washed (±)-(laí? *,8R *,8aR *,8bS *)-7-Hydroxy-8-hydroxymethyl-
with cold saturated sodium bicarbonate (2X10 mL) and then with brine l,la,2,8,8a,8b-hexahydro-4,6-dimethoxy-l(5-dimethylazirino[2,,3':3,4]-
(1 X 20 mL). The ethereal portion was dried over sodium sulfate, pyrrolo[l,2-a Jindole (32). To a solution of the dibenzyloxy aziridine 31
filtered, and concentrated to 471 mg (100%) of a faintly tan foamy solid: (35 mg, 0.072 mmol) in 250 µ of dry tetrahydrofuran was added liquid
NMR (CDClj, 90 MHz) 2.95 (s, 3 H), m/e 569 (P+). ammonia (ca. 1 mL). Sodium metal (ca. 60 mg) was added to the cold
A solution of tetrabutylammonium azide (4.047 g, 14 mmol, previ- solution. The blue color was permitted to persist for about 10 min, before
ously dried by azeotropic distillation of a benzene solution in vacuo) in the reaction was quenched with the addition of solid ammonium chloride.
22 mL of dry benzene was added to the crude monomesylate. The The mixture was evaporated at room temperature under a flow of ni-
solution was gently refluxed for 2.5 h under an argon atmosphere. After trogen. The residue was layered with ethyl acetate and water, then
cooling to room temperature, the reaction was diluted with ether and transferred to a separatory funnel. The mixture was diluted with ethyl
washed thoroughly with water (6 X 50 mL) and then brine (1 X 50 mL). acetate and the organic layer was separated. The aqueous portion was
The organic layer was dried over sodium sulfate and concentrated to 412 extracted twice with ethyl acetate. The combined extracts were dried
mg of crude azido alcohol 26 as a tan foamy solid: IR (CHC13, cm™1) over sodium sulfate and then concentrated in vacuo. Chromatography
2095; m/e 516 (P+), 488 (P+ N2), 474 (P+ N3).
- -
of the residue on silica gel (8% methanol/ethyl acetate) afforded white
To a solution of the crude azido alcohol in 4.3 mL of dry methylene crystalline diol 32 (22 mg, 100%) mp 155.5-156.6 °C. NMR
chloride at 0 °C under argon were added freshly distilled methanesulfonyl (CDC13, 500 MHz) 4.02 (dd, 1 H, 7, = 10, 72 = 4.2 Hz), 3.88 (dd, 1
chloride (87 uL, 1.1 mmol) and then triethylamine (180 µ , 1.3 mmol). H, J\ = 10, J2 = 10 Hz), 3.75 (s, 3 H), 3.75 (m, 1 H, under methoxy),
After stirring for 2 h at 0 °C, the mixture was diluted with ether, washed 3.69 (s, 3 H), 3.65 (m, 1 H, under H3), 3.64 (d, H, 7= 11.2 Hz), 3.04
1

with saturated sodium bicarbonate and then brine, and dried over sodium (dd, 1 H, 7, = 11.2, J2 = 3.5 Hz), 2.41 (dd, 1 H, 7, = 5.2, J2 = 3.5 Hz),
sulfate; the solvent was removed in vacuo to leave 447 mg of crude azido 2.35 (s, 3 ), 2.19 (d, 1 H, 7 = 5.2 Hz), 2.16 (s, 3 H). IR (CHC13, cm™1)
mesylate: >H NMR (CDC13, 90 MHz) 2.75 (s, 3 H), IR (CHC13, cm™1) 3174, 2898, 1250. m/e 306 (P+). Anal. Caled for C16H22N204: C,
2095. The crude azide was dissolved in 130 mL of tetrahydrofuran and 62.73; H, 7.24; N, 9.14. Found: C, 62.56; H, 7.33; N, 8.91.
heated at reflux with trimethyl phosphite (340 µ , 2.9 mmol) for 4 h. (±)-(lafi*,8R*,8aJ?*,8bS*)-l,la,2,8,8a,8b-Hexahydro-8-(hydroxy-
After the solution had cooled to room temperature, the solvent was methyl)-6-methoxy-l,5-dimethylazirino[2',3':3,4]pyrrolo[l,2-a]indole-
removed in vacuo and the residue was dissolved in 5 mL of anhydrous 4,7-dione (33). Phenol 32 (9 mg, 0.029 mmol) was dissolved in 1 mL
tetrahydrofuran. The solution was then transferred (via syringe) to a of 10% aqueous tetrahydrofuran and ca. one crystal of p-nitrophenol was
suspension of sodium hydride (250 mg, 60% dispersion in oil, previously added. To this colorless solution was added 2,3-dichloro-5,6-dicyano-
washed with dry pentane) in 10 mL of tetrahydrofuran. After stirring I, 4-benzoquinone (7.3 mg, 0.032 mmol, added as a solution in 225 µ
at room temperature under argon for 4 h, the reaction was quenched by of THF). Immediately on addition, the solution became deep purple.
the addition of saturated sodium sulfate solution. The mixture was After the mixture was stirred at room temperature for 40 min, the vol-
layered with methylene chloride and washed with saturated sodium bi- atiles were removed in vacuo to leave a black, purple powder. The
carbonate. The organic portion was dried over sodium sulfate, filtered, powder was dissolved in chloroform and washed thoroughly with cold
and concentrated to 575 mg of crude phosphoryl aziridine 29. The 10% aqueous potassium carbonate. The organic portion was washed once
aziridine was dissolved in 17 mL of anhydrous ether and subjected to the with brine and dried over sodium sulfate. Concentration in vacuo af-
action of lithium aluminum hydride (0.90 mL, 1 M solution in ether, forded 5.2 mg (61%) of quinone as purple needles, mp 146-152 °C dec.
Aldrich) at 0 °C under an argon atmosphere. After 1 h, an additional ‘H NMR (CDC13, 500 MHz) 5.15 (br, s, 1 H), 3.95 (dd, 1 H, 7, = 13,
5 Hz), 3.89 (br m, 1 H), 3.86 (s, 3 ), 3.81 (d, 1 H, 7 = 10 Hz),
portion of LiAlH4 solution (0.90 mL) was added and the reaction con-
=
72
tinued at 0 °C for 45 min. The cooling bath was removed and the 3.75 (dd, 7 = 11, 72 = 9 Hz), 3.53 (m, 2 H, containing a d, 1 H, 7 =
mixture warmed to room temperature over 15 min. The reaction was 13 Hz), 2.62 (t, 1 H, 7 = 5 Hz), 2.41 (s, 3 ), 2.30 (d, 1 H, 7 = 5 Hz),
finally quenched at 0 °C with 3.6 mL of 20% aqueous sodium hydroxide 1.99 (s, 3 H). IR (CHC13, cm™1) 3290, 2960, 1660, 1590, 1503, 1457,
solution. After stirring at room temperature for 30 min, the mixture was 1440, 1260. m/e 292 (P+ + 2), 290 (P+), 262, 261, 232, 231, 190, 100.
diluted with 0.1 N sodium hydroxide and extracted into methylene N-Oxides 34 and 35. A solution of olefin 22 (165 mg, 0.403 mmol)
chloride. The combined extracts were washed once with brine and then in CH2C12 (4 mL) was cooled to 0 °C. Solid MCPBA was added in
dried over anhydrous potassium carbonate. Concentration in vacuo af- small portions (82 mg total, 0.403 mmol assuming 85% pure reagent).
forded 347 mg of crude aziridine 30. Chromatography of the residue on The reaction mixture was warmed to room temperature and stirred for
silica gel (10 g) with ethyl acetate-hexane-triethylamine 8:1:0.5) af- 30 min. The solution was diluted with more CH2C12 and washed with
forded 145 mg of pure aziridine (39% overall from diol 24). NMR saturated aqueous NaHC03, dried (Na2S04), and concentrated in vacuo
(CDC13, 500 MHz) 7.31-7.26 (m, 10 H) 4.93 (d, 1 , 7AB = 11 Hz), to afford 171 mg of product (99.7%). NMR (CDC13, 250 MHz)
4.85 (d, 1 , 7AB = 11 Hz), 4.48 (d, 1 , 7AB = 12 Hz), 4.43 (d, 1 H, 7.47-7.24 (m, 5 H), 5.79 (br m, 2 H), 5.64 (br s, 1 H), 5.38 (br d, 1 H,
7AB
= 12 Hz), 4.18 (m, 1 H, J¡ = 1 Hz), 3.95 (d, 1 H, J = 7 Hz), 3.71 7ab = 16 Hz), 5.12 (br d, 1 , 7AB = 16 Hz), 4.64 (d, 1 , 7AB = 12
(s, 3 ), 3.70 (s, 3 H), 3.57 (d, 1 H, J = 11 Hz), 3.44-3.41 (m, 2 H), Hz), 4.50 (d, 1 H, 7ab = 12 Hz), 4.10 (s, 3 ), 3.90-3.76 (m, 2 H), 3.74
2.97 (dd, 1 H, J¡ = 3, J2 = 11 Hz), 2.82 (br s, 1 H), 2.72 (br d, 1 H, (s, 3 H), 3.43 (br t, 7, = 7, 72 = 2 Hz), 2.24 (s, 3 ), 2.17 (s, 3 H). IR
J = 3 Hz), 2.10 (s, 3 H). IR (CHC13, cm™1) 2975, 1600, 1460, 1438, (film, cm™1) 2900, 1760, 1660, 1465, 1185. m/e 425 (P+), 409, 407, 300,
1398, 1350, 1250. m/e 473 (P+ + 1), 472 (P+), 458, 457, 381. 244.
(±)-(la/?*,8/?*,8a/Z*,8bS*)-7-(Benzyloxy)-8-[(benzyloxy)methyl]- In a similar manner, 9 mg of TV-oxide 35 was obtained from 9 mg of
l,la,2,8,8a,8b-hexahydro-4,6-dimethoxy-l,5-dimethylazirino[2',3,:3,4]- aziridine 31. NMR (CDC13, 270 MHz) 7.43-7.26 (m, 10 H), 4.99
pyrrolo[l,2-8 jindole (31). To a solution of aziridine 30 (30 mg, 0.064 (d, 1 H, 7ab = 11 Hz), 4.91 (d, 1 , 7AB = 11 Hz), 4.55 (d, 1 H, JAB
mmol) in 380 µ of anhydrous ether at room temperature under argon = 12 Hz), 4.53 (d, 1 , 7AB = 12 Hz), 4.21 (d, 1 H, 7 = 5 Hz), 4.05
was added MeLi (74 µ , 0.115 mmol, 1.55 M solution in ether; Aldrich) (d, 1 H, 7 7 Hz), 3.77-3.75 (m, H), 3.76 (s, 3 H), 3.73 (s, 3 H), 3.63
=
1

via syringe in a single portion. After the reaction had stirred for 20 min (d, 1 H, 7= 11 Hz), 3.46 (m, H, 7, = 6 Hz), 3.06 (dd, 1 H, 7, = 11,
1

at room temperature (turning slightly cloudy), methyl iodide (7.1 µ , 72

= 4 Hz), 2.41 (dd, 1 H, 7! = 5, 72 = 4 Hz), 2.39 (s, 3 H), 2.29 (d,
0.114 mmol) was added. After 60 min, the mixture was diluted with 1 H, 7 = 5 Hz), 2.15 (s, 3 H). m/e 486 (P+ -

16), 395, 367, 274, 259.

methylene chloride and layered with a small amount of water. The Quinone Ketal 37. A solution of aziridine 30 (25 mg, 0.053 mmol)
mixture was washed once with 0.1 N aqueous sodium hydroxide and then in dry MeOH (1 mL) was added to a suspension of 10% Pd/C (20 mg)
with brine. The organic portion was then dried over anhydrous potassium in 1 mL of MeOH. The mixture was saturated with H2 (bubbling), and
carbonate. Evaporation of the volatiles left 28.1 mg of crude N- after 20 min at 25 °C the solution was filtered through Celite. Con-
methylaziridine. Chromatography on silica gel (ethyl acetate-hexane- centration of the filtrate left 20 mg (99%) of phenol 36. The compound
triethylamine 8:1:0.5) afforded 20 mg (64%) of pure aziridine 31 and 4 was dissolved in 0.5 mL of MeOH and treated with 0.6 mL of a stock
mg of recovered starting material. NMR (CDC13, 500 MHz) solution of DDQ (340 mg) and p-nitrophenol (13 mg) in 15 mL of
7.25-7.40 (m, 10 ), 5.01 (d, H, JAB =11,5 Hz), 4.93 (d, 1 H, JBA
1 MeOH. Stirring at 25 °C was continued for 45 min, then the solvent
=
11.5 Hz), 4.56 (d, 1 H, 7A,B, = 11.7 Hz), 4.54 (d, 1 H, JB.A. = 11.7 was removed in vacuo. The solid residue was taken up with CH2C12 and
Hz), 4.12 (brd, 1 H, 7 = 4.8 Hz), 3.91 (d, 1 H, J = 6.9 Hz), 3.77 (s, washed with 1 N NaOH and brine. The organic phase was dried over
3 H), 3.75 (s, 3 ), 3.50 (d, 1 H, J= 11.1 Hz), 3.37 (m, 2 H), 2.94 (dd, K2C03 and evaporated, leaving 22 mg of crude product. Chromatog-
1
H.7, = 11.1, 72 = 3.6 Hz), 2.35 (s, 3 H), 2.30 (dd, H, 7, = 3.6, J2
1
raphy over 0.5 g of Merck 230-400 mesh silica gel (95:10:5 EtOAc/
=
5.1 Hz), 2.16 (s, 3 ), 2.12 (d, H, J = 5.1 Hz). 13C NMR (CDC13,
1
hexane/Et3N) afforded 9 mg of ketal 37 as a yellow oil (40%). NMR
22.5 MHz) 146.0, 144.4, 143.9, 143.3, 142.0, 138.6, 137.7, 128.3 (three (CDC13, 500 MHz) 7.37-7.34 (m, 5 H), 4.63 (d, 1 , 7AB = 12 Hz), 4.3
3898 J. Am. Chem. Soc. 1985, 107, 3898-3902

(d, 1 ,
/AB = 12 Hz), 4.21 (dd, 1 H, /, = 9, J2 = 4 Hz), 4.03 (d, H, 1
compound 32 as well as the procedures related to the solution of
J =
Hz), 3.87 (dd, 1 H, Jx = 12, J2 = 4 Hz), 3.65 (s, 3 H), 3.59 (m,
8 the crystal structure, thermal parameters (Table I), bond distances
1 H), 3.49 (m, 1 H), 3.42 (m, 1 H), 3.23 (s, 3 H), 3.22 (s, 3 ), 3.08 (Table II), and bond angles (Table III). Experimental protocols
(br s, 1 H), 2.87 (br s, 1 ), 1.87 (s, 3 H). for the preparative work, as well as modifications in the synthesis
Acknowledgment. This work was supported by PHS Grant CA
of phenol 7, are also provided. In addition, experimental proce-
28824. NMR spectra were obtained through the auspices of the dures for the conversion of purified ether 8 to 15a through fully
Northeast Regional NSF/NMR Facility at Yale University, characterized intermediates 12a, 13a, and 14a are included, as
which was supported by NSF Chemistry Division Grant CHE well as the spectral properties of the crude intermediates 12b, 13b,
7916210. 14b, 15b, and selenitic 21. Also included are the preliminary results
of transformations of tricyclic amine oxides (17 pages). Ordering
Supplementary Material Available: An ortep drawing of information given on any current masthead page.

Comparative Tests of Theoretical Procedures for Studying

Chemical Reactions1
Michael J. S. Dewar* and Donn M. Storch2
Contribution from the Department of Chemistry, The University of Texas at Austin,
Austin, Texas 78712. Received October 29, 1984

Abstract: A simple procedure is described for estimating the effective errors in molecular energies calculated by ab initio
methods with respect to use of the latter in studies of chemical reactions. The procedure is illustrated by application to the
STO-3G, 3-21G, and 6-31G* models. Parallel results from semiempirical models (MINDO/3, MNDO, AMI) are included
for comparison.

Introduction Table I. Errors” in Total Energies (£) and Heats of Atomization

The most basic problem in chemistry is to find out how chemical (HA), Calculated by the RH Method Using the 6-31G* Basis Set"
reactions take place. It is also, unfortunately, a problem that molecule error in £ error in HA HA (obsd)
cannot be solved by experiment, because the time a chemical
acetylene 324 106 392
reaction takes is too short (<10~13 s) for its course to be
~
578 189 955
propane
observed.3 Current “experimental” approaches rely on theory 494 176 814
cyclopropane
to delineate possible mechanisms for a reaction. Experiments are cyclopropene 554 166 656
then devised to distinguish between the various possibilities. Such diacetylene 738 190 676
an approach is naturally limited by the efficacy of the theory on 1,3-butadiene 682 221 972
which it is based, and until recently only qualitative theories were hydrazine 444 162 412
used in this connection. Studies of reaction mechanisms would hydrazoic acid 699 205 320
acetonitrile 523 51 478
clearly be much more effective if they were based on a theoretical 799 207 478
cyanogen
approach able to reproduce the properties of molecules, in par- methanol 515 130 487
ticular their energies, quantitatively. Reaction mechanisms could
dimethyl ether 704 94 671
indeed be predicted unambiguously if the corresponding potential ozone 631 68 145
energy surfaces could be calculated accurately. acetone 896 225 938
To be chemically useful, such a calculation must be carried out dimethylamine 554 203 825
properly, i.e., with full geometry optimization, etc., and without benzene 1301 649 1320
making any assumptions, and the method used must also be “All values in kcal/mol. The ab initio energies were taken from ref
sufficiently accurate. Unfortunately, no current ab initio procedure 14. These refer to calculations carried out with full geometry optimi-
comes anywhere near to achieving the needed accuracy, in an a zation, using the 6-31G* basis set. Experimental total energies of at-
priori sense. The errors in molecular energies, calculated by oms, relative to nuclei and electrons, were estimated from ionization
standard Hartree-Fock-type ab initio models, are indeed com- energies (Weast, R. C. “CRC Handbook of Chemistry and Physics”,
65th Ed.; CRC Press: Boca Raton, FL, 1984-5; pp E-63,64. Energies
parable with the corresponding heats of atomization.4 Since this
(eV): H, 13.598; C, 1030.080; N, 1486.029; O, 2043.794. For con-
point is still not generally appreciated, some additional examples version factors, see Table II (footnote). Experimental heats of atomi-
are shown in Table I. Indeed, since the calculated values refer
zation (standard state, gas phase, 25 °C) were estimated from ther-
to molecules at equilibrium geometries, without zero point or mochemical data listed by Cox and Pilcher. [Cox, J. D.; Pilcher, G.
thermal energy, the real errors are greater than indicated, by ca. “Thermochemistry of Organic and Organometallic Compounds”; Aca-
5%. Some improvement is possible in “beyond HF (Hartree- demic Press: New York, 1970]. The experimental total energy of a
Fock)” methods, but these are confined to small molecules and molecule refers to the sum of its heat of formation and the total ener-
the residual errors are still enormous, by chemical standards. gies of the component atoms. It therefore includes kinetic energy
terms. Since the calculated values for the molecular energies refer to
equilibrium geometries, without corrections for zero point or thermal
(1) Part 75 of a series of papers reporting the development and use of energy, the errors listed above are correspondingly too small; see text.
quantum molecular models. For part 74, see; Dewar, M. J. S.; Grady, G. L.;
Merz, K. M., Jr.; Stewart, J. J. P. J. Am. Chem. Soc., in press. The errors due primarily to neglect of electron correlation.
are
(2) Present address: USAFA/DFC, Colorado Springs, Colo. 80840. If the correlation energy of a set of atoms did not change sig-
(3) Our inability to observe reactions is due not merely to lack of tech-
niques but to the limits set by the Uncertainty Principle. nificantly when they combine to form a molecule, the errors might
(4) Dewar, M. J. S.; Ford, G. P. J. Am. Chem. Soc. 1979, 101, 5558. then cancel in calculating the differences in energy (heats of

0002-7863/85/1507-3898S01.50/0 &copy; 1985 American Chemical Society