вход по аккаунту


N-Halocarbamate Salts Lead to More Efficient Catalytic Asymmetric Aminohydroxylation.

код для вставкиСкачать
[5] a ) Chioramme-M is readily prepared from methanesulfonamide in water by
addition of the stoichiometric amount of sodium hydroxide and rerr-butylhypochlorite [Sb]. This method is adapted from a procedure developed by
Campbell et 211. for the preparation of sodium N-chloro-N-carbdmates [5c-e]
and proved general for synthesis of both sodium N-chloro-A-aryl- and -alkylsulfonamides Substitution of the recommended [5c-el solvent methanol by
water proved to be highly advantageous. b) rert-Butylhypochlorite is commercially available: for a simple preparative procedure see: M. J. Mintz, C.
Walling, Org. . S j ~ i i r h .Call V d V. 1983, 183-187: c) M. M. Campbell, G
Johnson. C h i i Rciz. 1978, 78.65-79; d ) E. Herranz, S A. Biller. K. B. Sharpless, J Ain. Chwn. Sor. 1978, 100. 3596-3598. e) E. Herranz, K. B. Sharpless,
J Org. Chrm. 1978. 43, 2544-2548
[6] Alkyl sulfonamide derived chloramine salts appear to have been little studied
In fact. we found only one earlier report on this class of compounds: F. E.
Hardy. J C%eni.Soc 1970, 2087-2089.
(71 Only wrr-butyl rriin.7-crotonate and isopropyl !runs-cinnamate did not appear
in the firs1 papei-. The isopropyl {runs-cinnamateis included here in addition to
the inethql iruii.\-cinnamate because it gave a considerably worse result when
Chlorainine-T in CH,CN/H,O was employed: regioselectivity: 76.24. ee.
[8] The n-propanol water solvent system was also shown to give better selectivities
with C'hloramine-T. although i n some cases the reaction became loo slow to be
useful A Houi-I. K. B. Sharpless, unpublished results.
[9] Temperatures lower than 10 'C led to longer reaction times and problems with
ligand solubilit).
1101 D. J. Berrisford. C. Bolm. K . B. Sharpless. A I ~ ~ Y H
1995. 107. 11591171. Anycii. ( . / i w . Inr W . Engl 1995, 34, 1059-1070
[ I I] Considrriible evidence for the existence of a second cycle has been recognized
and examined i n the closely related catalytic asymmetric dihydroxylation process. For detailed discussion of problems in the second cycle in the A D process
see: a ) J. S. M. Wai, I . Marko, J. S . Svendsen. M G . Finn, E. N. Jacobsen.
K . B. Sharpleas. J. Am. C/iem.Sac. 1989. t / / ,1123-1125; b) H. L. Kwong,C.
Sorato. Y. Ogino. H. Chen, K. B. Sharpless. EJtruliedron Lert. 1990,31, 2999 3002.
[I21 H.-T Chang, K. B. Sharpless, unpublished results.
113) P. C. Sennhenn. .I Rudolph, C. P Vlaar, M. Bruncko, S . Immel, H. C Kolb,
K . B. Sharpless. unpublished results.
[14] We helievc thal there are also "indirect" effects since the presence of the ligand
dramatically shifts the IV - I1 equilibrium (step b2) in favor of 11; H:T. Chang.
K. B. Sharpless. unpublished results.
1151 H. C. Kolb. M S. Van Nieuwenhze. K . B. Sharpless. Cliem. Re&,.1994, 94.
2483 2547
1161 a ) E. H. Gold. E. Bahdd, J. Org. Cliem. 1972,37, 2208-2210 This method has
recently heen applied succeasfully by Seebach et al. for the deprotection of
derived from &-amino alcohols: b) D. J. Ramon. G .
Guilleiia. D Scebach, Heli.. Chini. A r m 1996, 79, 875-894.
[I71 K . B. Sharpless. A . O . Chong, K. Oshima, J Org Clieni. 1976. 41. 177-179;
see also [5d. e]
[I81 G. Li. H Angert. K . B . Sharpless. Angew. Cliem. 1996, 108, 2995-2999:
Angcii. Client ln! Ed Engl 1996. 35, 2813-2817.
N-Halocarbarnate Salts Lead to More Efficient
Catalytic Asymmetric Aminohydroxylation""
G u i g e n Li, Hubert
H. Angert, and K. Barry Sharpless*
Dedicated to Professor Koji Nakanishi
The recently discovered catalytic asymmetric aminohydroxylation (AA) process"] has been easier to improve than anticipated. In the preceding paper, we described how substantial
Prof. K . 8. Sharpless, Dr. G Li, Dr. H. H. Angert
Department of Chemistry. The Scripps Research Institute
10550 N Torrry Pines Road. La Jolla, CA 92037 (USA)
Fax: Int code +(619)784-7562
e-mail: sharplewr
[**I This research was supported by the National Institutes of Health (GM 28384).
the National Science Foundation (CHE 9531152), and the W. M. Keck Foundatlon. We thank Dr. Pui Tong Ho for many helpful discussions. H. H A. is
grdleful to the Deutsche Forschungsgemeinschaft (DFG) for providing a postdoctoral fellowship.
AnPew. (%em. Inl. Ed. E n d . 1996. 35, N o . 23,24
increases in enantio- and regioselectivity were achieved by replacing Chloramine-T (TsNClNa) with analogs bearing substituents smaller than thep-tolyl group (e.g. CH,SO,NClNa) .[']
We report here that the process becomes even more effective
when the chloramine salts of sulfonamides are replaced by those
of alkyl carbamates. Application of the new procedure to 2vinylnaphthalene (1) with BnOCONClNa as the oxidant and
nitrogen source (Scheme 1) proceeds with exceptional enantioselectivity and regioselectivity (> 10: 1). In addition, in virtually all the entries in Tables 1 and 2 better selectivities were
achieved than those realized in the sulfonamide-based procedures for the AA.['-31 Also noteworthy is the increased scope,
hinted at here by the very good results obtained with methyl
acrylate and vinylnaphthalene 1 (Table 1 , entries 4 and 7),
which will be further documented
Styrene-type olefins such as 1 are conspicuously absent from
the first sulfonamide-based AA reports".'] for the simple reason that they were poor substrates. The reaction of styrene itself
produced an approximately 2: 1 mixture of regioisomers, 5070 % ee each, in low yield.['] This was quite unexpected, since in
terms of both turnover and enantioselectivity styrenes are by far
the best substrates in the catalytic asymmetric dihydroxylation
(AD) process.[61 However, even more surprising than the low
selectivities were the abysmal rates. Styrenes, the fastest reacting
A D substrates, were among the slowest reacting substrates in
the sulfonamide-based AA reactions,". 'I and since the reactions
were even slower when the ligand was added than when it was
omitted, this is an example of ligand-decelerated catalysis.[61By
contrast, in this new carbamate-based process the reactions of
all olefins studied to date are ligand-accelerated, as has been the
universal experience in the A D process.[']
In comparing Tables 1 and 2, it becomes clear that the ethyl
carbamate series is superior to the benzyl carbamate series in
terms of rate, enantioselectivity, regioselectivity. and yield. The
effect of the smaller ethyl substituent may be analogous to that
assumed on replacement of a methyl substituent for the p-tolyl
substituent on the sulfonamide-suppression
of the second
cycle and better fit in the ligand binding pocket of the catalyst.L81
Despite the greater effectiveness of the ethyl reagent, the benzyl
reagent will often be preferred because it can be cleaved easily by
Since the other results in the tables need no comment, we now
consider what general insight this new system offers for catalysis
research. Such lessons are often more valuable than any particular catalytic process can ever be.
The original report on carbamate-based, osmium-catalyzed
aminohydroxylation of olefins appeared in 1978,['' two years
after publication of the analogous sulfonamide-based process.l'ol The latter process['*] was used many times over the past
twenty years, especially in syntheses of amino sugars. The old
carbamate version[93''I was used much less frequently, but it
has the distinction of being a key step in the first synthesis of the
tert-butoxycarbonyl (BOC) protected side chain analog of tax01, now known as Taxotere.["]
Those familiar with the original, nonasymmetric, carbamatebased procedures will realize that the sodium N-chlorocarbamate salts were never used directly in the process, either by us or
by others."31 They were always first converted in situ to the
corresponding silver or mercury salts. So why did we fail to
discover earlier that the original sodium salts are superior to
their silver and mercury
The answer is that far too
little water was present, so we observed no turnover in the first
few attempts to extend the effective process based on TsNClNa
to one based on ROCONClNa. Scott Biller (then an undergraduate at MIT) came up with the idea of using the silver salt, which
VCH Verlugsgesellschaft mhH. 0-6945t Weinheim.19Y6
057O-(i833,96;3523-28/3 $ 15.UU+ 3 d J
4 Yo K ~ O S O ~ ( O H ) ~
n-PrOH/H20, 0 "C
Scheme 1 AA reactions of 2-vinylnaphthalene (1) with henzyl carbamate. DHQD-H
dihydroquinine, PHAL = 1,3-phthalazined1yl,2 = benzyloxycarhonyl
proved highly successful even under our water-starved conditions. No one ever tried the sodium salt again. Even if it had
been tried, "no reaction" would have been the verdict unless a
lot more water were present than described in any of the published procedures.
Following the encouraging results of the catalytic AA reaction with sulfonamides,['. we reexamined the carbamatebased catalytic aminohydroxylation process, primarily because
carbamate protecting groups are easier to remove from nitrogen. Optimization of the AA procedure using ArS0,NClNa
salts had led to the choice of cosolvent systems containing 50%
water. The high water content is thought to speed the slow
hydrolytic steps and thereby suppress involvement of the rateand selectivity-damaging second cycle.r81At this point, good
fortune struck.
One of us (G. L.)-unaware that the standard carbamate process called for acetonitrile solvent with only a few equivalents of
water and for the silver, not the sodium salt of the chloramine-ran the first experiment using 50 % aqueous n-propyl
alcohol as solvent, EtOCONClNa as the chloramine salt,
(DHQ),PHAL as the chiral ligand, and methyl cinnamate as the
substrate. The result was spectacular (Table 2, entry 1). For
even though the system was homogeneous, the reaction was
almost free of the deleterious effects on selectivity that arise
whenever second-cycle catalysis participates.['] Recall that in all
homogeneous versions of the AD, the second cycle always intrudes to some extent and can only be completely suppressed by
using inconvenient two-phase systems.16]
Further controls revealed which variables are crucial for
maintaining the exceptional catalyst activity and selectivity of
this new AA system.[161The two most important factors are
water concentration and the nature of the metal salt of the
N-chlorocarbamate, and the two factors are linked. The interdependence is dramatic for the alkali metal salts of N-chlorocarbamates: from essentially no turnover with a few equivalents of
water, to better and better reactivity as the percentage of solvent
water increases. The effect of water concentration on reactivity
in the silver N-chlorocarbamate system is, as expected, less dramatic. In any case, whatever the water concentration, in the
system examined (stilbene, sodium or silver benzyl N-chlorocarbamate, (DHQ),PHAL, K,OsO,(OH),, CH,CN, and H,0),['61
reactions with the silver carbamate always gave lower yields,
lower enantioselectivities, and more by-products than reactions
with the sodium carbamate.
0 VCH Verlugsgesellschaft mbH, 0-69.151 Weinhelm, 1996
With this better catalytic asymmetric aminohydroxylation process now in hand, we and others are
busy trying to map out its scope.[41
There may also be a message here
of broader significance for catalyst(q-2
99 Yoee
discovery research. As noted in the
64 O/o
brief historical account, this trivially different, but successful, system
was passed by long ago simply beHYZ
cause we failed to examine the effect of water concentration with every change of chloramine salt. This
joins the list of many important lessons in catalysis we have learned
99 Yoee
the hard way over the last 20 years.
70 Yo
The common theme of these lessons is that catalytic cycles are sen= dihydroquinidine, DHQ-H =
sitive to far too many reaction variables for even the best catalyst
discovery endeavors to have a high
success rate. Whether one is just trying to improve existing, but
feeble catalysis, or prospecting for the first hint of a new catalytic transformation, unknown obstructions, usually interconnected in sinister ways, lie everywhere. Fast (automated) screening
techniques, along with the chance discoveries they engender,
would seem to offer the best chance of getting through such
minefields quicker and more often. At present, it could be argued that attempts to discover new catalysts fail, because of the
limited endurance of human researchers.
Finally, the significance of entry 7 in Table 1 will not be lost
on chemists interested in the synthesis of unusual amino acids.
This result opens a simple two-step route from vinyl arenes to
enantiomerically pure aryl glycines of either R or S configuration." '1
Experimental Procedure
General Procedure- Unless otherwise noted, all the carhamate AA reactions shown
in Tables 1 and 2 were carried out on a 1 mmol scale as described here for methyl
trans-cinnamate and the benzyl carhamate derived chloramine salt (Table 1, entry 1). Ethyl carbamate or tert-hutyl carbamate can also he used in place of benzyl
Benzyl carhamate (0.469 g, 3.10 mmol) was dissolved in 4 mL of n-propyl alcohol
in a 20-mL scintillation vial equipped for magnetic stirring. To this stirred solution
was added a freshly prepared solution of NaOH (NaOH (0.122 g, 3.05 mmol) in
7.5 mL of water). followed by a freshly prepared solution of tert-hutyl hypochlorite
(0.331 g, 3.05 mmol, ca. 0.35 mL) [lS]. Then a solution of the ligand (DHQ),PHAL
(40 mg, 0.05 mmol, 5 mol%) in 3.5 mL of n-propyl alcohol was added. (It is important that the ligaud be completely dissolved in the n-propyl alcohol before it is
added; the DHQ analog is more readily soluble than the DHQD analog.) The
reaction mixture should be homogeneous at this point. The vial was then immersed
in a room-temperature water bath, and stirred for a few minutes. Then the olefin
(trans-methyl cinnamate, 0.162 g, 1 mmol, 0.067 M) was added, followed by the
osmium catalyst (K,OsO,(OH),, 14.7 mg, 0.04 mmol, 4 mol%). The reaction mixture was stirred for 40 min (the approximate reaction time is given for each entry in
the tables), and the light green color of the solution gave way to light yellow at the
end. After TLC analysis confirmed the absence of starting material, 7 mL of ethyl
acetate was added and the phases were separated. The lower, aqueous phase was
extracted with ethyl acetate (3 x 5 mL). Thecombined organic extracts were washed
with water and brine, dried over MgSO,, and concentrated to dryness to afford the
crude product contaminated by some excess henzyl carhamate. Purification by flash
chromatography (hexane/chloroform/methanol6/4/1, v/v/v) provided (233,35)-3
(0.204 g. 65% yield, 94%ee ) a s a colorless soiid; m.p. 101-103 "C; [z];" = 4.4
(c = 0.32 in 95% EtOH).
Simplified procedures for cases in which the product crystallizes directly from the
reaction mixture:
1) (S)-2(Table 1. entry 7): This is one of only three cases in which the reaction was
conducted at O T , but the procedure is otherwise identical (up to the generation of
the chloramine salt) to the one used for methyl cinnamate. It then diverges as
follows: The vial was immersed in a 0 ° C icebath and 2-vinylnaphthalene (1)
(0.154 g. 1 mmol) was added, followed by K,OsO,(OH), (14.7 mg, 0.04 mmol,
Angew. Chem. Int. Ed. Engl. 19%. 35. No. 23124
Table 1. Results of AA reactions with BnOCONClNa
Product [a]
Config. [a]
ee [%I
[ " / . I[bl
t [hl
3 [c]
7 [dl
8 [dl
Id] Major product from the reaction using (DHQ),PHAL. [b] Yields of the reactions using (DHQ),PHAL. [c] CH,CN/H,O was used as the solvent. [d] Reactions were
performed at 0 C
4 mol%). The green homogeneous mixture was stirred at O'C for 3 h by which time
it had transformed into an almost colorless slurry. A crystalline precipitate, which
was nearly pure by HPLC and NMR, was removed by filtration. One wash with
cold (ca. 5 C) nPrOH/H,O ( l / l . 3 mL) yielded pure (S)-2 (0.460 g, 70% yield,
>99% ee). A small sample of this solid was dissolved in ethyl acetate and filtered
through a small plug of silica gel; m.p. 133-134°C; [a]iO = 50.9 (c = 0.32 in
95% EtOH). (A practical note: Very similar results were realized in an otherwise
identical experiment in which the solvent volume and the amounts of ligand and
osmium were the same. but 2 mmol of olefin and 4 mmol of chloramine were used.
One-half of the total amount of olefin was added at the beginning of the reaction.
After that portion was consumed, the remainder was added and the reaction completed as usual.)
2) (lS,2S)-8 (Table I , entry 6): This procedure for stilbene is reminiscent ofearlier
AD [19] and AA [ I ] processes, which take advantage of the fact that both the
stilbene and its oxidation product are insoluble in the medium. As in the preparation
of 2. workup consists of merely filtering the crude reaction mixture. The following
procedure differs from that for 2 in that it i s a straight sixfold scale-up and requires
a 120-mL boftle. (Anothersmall differenceis that inaddition tofinishingasaslurry,
the reaction mixture starts as a slurry.) However, since the reaction sequence is
otherwise identical to that just described for 2 (even down to the reaction time of 3 h
at O'C), the only new details describe the treatment of the product after its isolation
from the reaction mixture. Filtration of the cold slurry provided the crude solld
A n p n Cliem. Ini. Ed. Engl. 1996, 35, No. 23/24
product which was washed once with cold (ca. 5"C) nPrOH/H,O (111, 12 mL) to
give (1S,2S)-8 (1.85 g. 92% yield, 91 YOe e , nearly pure by NMR and HPLC); m.p.
135-137°C; [2]i0 = -7.8(c = 0.70in95% EtOH).Asinglerecrystallizationfrom
methanol raised the ee to >99%; m.p. 149-151 "C; [4i0= - 10.0 (c = 0.57 in
95%' EtOH).
Determinution ofenantiomeric excesses and optical rotations: Rotations were determined for the products in Tables 1 and 2 at the stated enantiomeric excesses obtained with the (DHQ),PHAL ligand unless otherwise indicated, and were measured in 95% EtOH.
3: ChiralcelOD-H, iPrOH/hexane (15/SS), 0.60 mLmin-'; 16.9 min (2S,3R),
19.7 rnin (2R,3S);
4.4 (c = 0.32).
4: Chiralcel AD, iPrOH/hexane (317). 0.60 mLmin-'. 10.4 min (2S,3R), 8.6 min
(2R.39; [ l p O = + 8.75 (c = 0.13).
5 : Chiralcel AD. iPrOH/hexane (119). 0.80 mLmin-', 40.7 rnin (ZS.3S). 27.0 min
(2R,3R); [2]bo = - 5.4 (c = 1.2, mother liquor after two recrystallizations from
6: ChiralcelOB,iPrOH/hexane(l/9).0.70mLmin-', 54.7 min(2H),68.9 min(2S):
[z]io = -17.6 (c = 0.80).
7: ChiralcelAD. iPrOH/hexane (5/95),0.60 mLmin-', 44.2 rnin (lS.2R). 25.6 rnin
= +19.8 (c =1.5).
8: ChiralcelOD-H. iPrOH/hexane (3/7), 0.70 mLmin-'. 10 7 min (lS.ZS),
12.6min (1R.2R); [2]i0 = -7.8 (c = 0.57).
8 VCH Verlagsgesellschaft mbH, 0-69451
Weinheim, 1996
0570-0833/96/3523-281SS 15.00+ .2S!O
Table 2. Results of AA reactions with EtOCONClNa
Product [a]
Config [a]
t [h]
[a] Major product from the reaction using (DHQ),PHAL [b] Yields of the reactions using (DHQ),PHAL. [c]CH,CN/H,O was used as the solvent.
2: ChiralcelAD, iPrOH/hexane (3/7), 0.70 mLmin-', 9.7 min (2s).13.8 min (2R):
[XI;' = + 50.9 (c = 0.32).
9. ChiralcelAD, iPrOH/hexane (119). 0.70 mLmin-'. 18.8 rnin (2s). 27.0 min
(2R); [a]:' = 30.0 (c = 0 55).
10: ChiralcelOD-H. lPrOH/hexane (15/85), 0.60 mLmin-'; 11.6 min (2S,3R),
12.6 min (2R,3S); [a];' = 2.78 (c = 0.90).
11: ChiralcelAD, rPrOH/hexane (1/9), 0.70 mLmin-', 30.3 min (2S,3R), 19.5 rnin
(2R.3S); [a];' = 17.9 (c = 0.43).
12: ChiralcelAD, iPrOH/hexane (1/9), I.OmLmin-'. 40.1 min (2S.3R). 23.9 min
16.5 ( c = 0.67).
13. Chiralcel AD, iPrOH/hexane (119). 1.0 mLmin-'. 26.2 min (2R.3R). 18.0 min
(2S,3S); [a];' = - 38.2 (c = 0.10).
Assi~nmentof the ahsolute configurations: The methyl cinnamate derivative 3 was
converted into an authentic sample of the methyl ester of the taxol side chain by
hydrogenolysis (Pd/C) and subsequent Schotten-Baumann acylation [3] [HPLC:
ChiralcelAD, 10% iPrOH/hexane (1/9), 1 mLmin-', 19.1 mm (2R.3S). 22.1 min
(2S.3R)I [I]. The methyl 2,6'-dimethylcinnamate derivative 4 was assumed to have
the same configuration as the methyl cinnamate derivative 3 The signs of the
rotations for the Z-protected derivatives of the dimethyl fumarate derivative 5 and
the methyl acrylate derivative 6 were compared to the literature values; (2R,3R)-5:
[al,y = - 22.6 (c = 2.0 in CHCI,); (2S.3.S)-5: ':IX[
24.2 (c = 1.0 in
CHCI,) [21]; (2R)-6: [a];' = -17.7 (c =1.3 in CH,OH); (2S)-6: [a];;" = 18.8 (c
= 1.42 in CH,OH) [22]. Cyclohexene derivative 7 and trans-stilbene derivative 8
were correlated with authentic samples of their p-toluenesnlfonamide analogs by
hydrogenolysis (Pd/C), followed by N-tosylsulfonylation [HPLC for sulfonamide
from 7: Chiralcel OG, 1 5 % iPrOHihexane, 1 mLmin-'. 14.3 rnin (1S,2R),
17.6 rnin (1R,2S). HPLC for sulfonamide from 8. Chiralcel OD-H, 1 5 % iPrOH/
hexane, 1 mLmin-', 14.3 rnin (l.S,2R), 17.6 rnin (1R,2S)]. Styrenederivative9 was
converted into (S)-(+)-2-phenylglycinol, [a];' = 29 8 (c = 0.5 in 1 N HCI), [a]:5
33 (c = 0.75 in 1 N HCI) (Aldrich). The RuCIJHJO, oxidation of 9 also
provided the known N"-(benzyloxycarbony1)-L-phenylglycine.Vinylnaphthalene
derivative 2 was assumed to have the same configuration as styrene derivative 9.
Methyl cinnamate derivative 10 was compared with a sample that was prepared
from the known toluenesulfonamide analog 11.31. Compounds 11- 1 3were assumed
to have the same configuration as 10.
Received: July 1, 1996 [Z92751E]
German version: Angrw. Chem. 1996. 108. 2995-2999
Keywords: amino alcohols * asymmetric aminohydroxylation
carbamates catalysis
[l] G . Li, H.-T. Chang, K. B. Sharpless, Angeu. Chem. 1996, 108, 449-452;
Angen,. Chem Int. Ed. Engl. 1996, 35. 451-454.
[2] K. B. Sharpless, J. Rudolph, P. C. Sennhenn, C . P. Vlaar. Angew. Chem. 1996,
108. 2991-2995; Angew Chem Int. Ed. Engl. 1996.35.2810-2813.
131 G. Li, K. B. Sharpless, Acra Chem. Scand. 1996, 50, 649-651
[4] K. B. Sharpless et al., unpublished results.
VCH Verlagsgesellschajr mhH. 0-49451 Weinheim, 1996
[5] H.-T. Chang, K. B. Sharpless, unpublished results.
[6] H. C. Kolb, M. S. VanNieuwenhze. K. B. Sharpless, Chem. Rev. 1994. 94,
2483 -2547.
[7] D. J. Berrisford, C. Bolm, K. B. Sharpless, Angeu. Chem. 1995, 107, 11591171; Angeu Chrm In!. Ed. Engl. 1995.34.1059-1070. With more hindered
olefins in the original NMO-AD procedure and also in the two-phase ferricyanide procedure, the ligand may not noticeably increase the turnover rate.
but to the best of our knowledge it does not slow it down either.
[S] See Scheme 2 in ref. [2]. In the present case X = COOR. This assumption is
supported by our finding (not in tables) that methyl trans-cinnamate gives the
BOC-protected product in only 78 % er when terl-BuOCONCINa is used as the
oxidant (cf. 99Yoee and 94%ee in Tables 1 and 2 for the reaction of the same
substrate with the corresponding benzyl and ethyl carbamates (R = Bn and
Et), respectively)
[9] E. Herranz. S . A. Biller, K. B. Sharpless. J: Am. Chem. Soc. 1978, 100. 35963598.
[lo] K. B. Sharpless. A. 0. Chong, K Oshima. J Org. Chem. 1976, 41, 177-179.
[ I l l a) E. Herranz. K. B. Sharpless, J Org. Chem. 1980. 45, 2257-2259, b) Org.
Synth. 1983. 41. 85-93; c) ihrd. 1983. 61, 93-97.
[12] a) D. Guenard. F. Gueritte-Voegelein, P. Potier. Ace. Chem. Res. 1993. 26,
160-167; b) F. Gueritte-Voegelein. D. Guenard. F. Lavelle, M.-T. Le Goff. L.
Mangatal, P. Potier, J: Med. Chem. 1991, 34, 992-998.
[13] The literature appears to be in error: M. M. Campbell, G. Johnson, Chern. Rev.
1978. 78.65-79
1141 In reexamining the results of the old experiments 19,111 one sees that only trace
amounts of water were used (2-4.5 equiv based on olefin (1 % of total solvent)). probably due to concern that the O,Os=NCOOR species might hydrolyze to OsO, faster than i t reacted with the olefin. in any case. it is frustrating to realize how very close, and on how many different occasions, one can
come to the "right" conditions over years of experimenting, and still miss the
target zone or zones within which the reaction conditions support useful, or at
least detectable, catalytic turnover. - Metal-catalyzed redox processes are inherently multistep reactions, and the principal challenge in developing them is
to find reaction conditions under which each of the steps can function at an
acceptable rate. The search is guided above all by the knowledge that compromises are inevitable. A given set of reaction conditions can never be optimal for
all the component reactions of the cycle. The slowest step constitutes the
turnover-limiting step and defines the throughput (turnover rate) of the entire
system [15]. In terms of the utility of a catalyst, the turnover rate (activity) is
"not everything. it is the only thing" (paraphrased from a quote about the
importance of winning. attributed to Vince Lombardi). - T h e complexity and
interdependence of catalytic cycles means that "apparently" minor variations
in the reaction conditions sometimes have major effects on the catalytic activity. This has certainly been the case for osmium-catalyzed oxidations of olefins,
for which the maximum turnover rates (umJ are virtually always identical with
the rates for hydrolytic release of the chelated diol or amino alcohol product
from the osmium coordination sphere. This is probably the least "interesting"
0570-0833jU413523-2814$15.00+ ,2510
Angeu. Chem. In!. Ed Engl. 1996.35. No.23124
step in the entire sequence, but in catalysis every step in the cycle is by definition
of equal importance.
B. Bosnich, personal communication.
G . Schlingloff, H. H. Angert, G . Li, K. B. Sharpless. unpublished results.
The oxidation of the Z-protected amino alcohol from styrene (see Experimental Procedure) gave the Z-protected aryl glycine in roughly 80% yield and with
no loss of enantiomeric purity. The scope of this two-step amino acid synthesis
will be reported soon: K. L. Reddy, G. Li, K. B. Sharpless. unpublished results.
We strongly recommend using tert-butyl hypochlorite that has been freshly
prepared following the procedure described in M. J. Mints, C. Wdlhng, Org.
S p h . Colt. Vul. V 1983, 183-187. This material has proveii reliable for up to
two months if it is kept and dispensed under cold (at roughly 4 "C) and dry
conditions. Some commercially obtained samples of tert-butyl hypochlorite
have given substandard enantioselectivities and yields, probably because they
were stored for longer periods a t or above room temperature and/or not under
dry conditions.
Z.-M. Wang, K. B. Sharpless, J Org. Chem. 1994,59. 8302-8303.
a) M. T. Nuiiez, V. S. Martin, J. Org. Chem. 1990, 55, 1928-1932; b) P. H.
Carlsen. T. Katsuki, V. S. Martin, K. B. Sharpless, ihid. 1981. 46, 3936-3938.
S. Hanessian, B. Vanasse, Can. .
Chem. 1993, 7i,1401-1406.
U. Schmidt, R. Meyer, V. Leitenberger, F. Stibler, A. Lieberknecht, Synthesis
1991, 409- 413.
Scheme 1. Reaction course for the ci5-dihydroxylation of olefins with OsO, as
suggested by Sharpless.
Theory Rules Out a [2+ 21 Addition of Osmium
Tetroxide to Olefins as Initial Step of the
Dihydroxylation Reaction**
Ulrich Pidun, Christian Boehme, and
Gernot Frenking"
The addition of OsO, to olefins yielding cis-diols has become
an intensively investigated reaction in stereoselective synthesis
since Sharpless developed the catalytic version of the reaction
into a powerful method for highly enantioselective functionalization of olefins using cinchona alkaloids as ligands."] The
transfer of the stereochemical information of the chiral ligand to
the substrate was explained by Sharpless with a two-step mechanism for the addition reaction (Scheme I), which should occur
rather than a concerted [3 + 21 addition as originally proposed
by Criegee."] It was suggested that the initial step of the reaction
course is a [2 + 21 addition yielding the metallaoxetane 1, which
then rearranges to the cyclic ester 2.[31The barrier for the bond
migration 1 + 2 must be very low, because intermediates of
structure 1 have never been observed, whereas base adducts of
2 are well known.[41
Strong support for a two-step mechanism of the osmylation
reaction came from kinetic studies, which revealed a nonlinear
correlation between the temperature and the enantioselectivity
of the reaction.15a1More recently kinetic studies of the addition
of OsO, to various olefins both in the presence and the absence
of a base (pyridine) have been carried out.[sb1In all cases the
Eyring plots showed tw'o linear regions. It follows that even the
base-free osmylation appears to proceed by a two-step mechanism.
In a previous theoretical study at the QCISD(T)//HF level of
theory we could show that the postulated intermediate 1 and the
elusive base adduct 1 'NH, are indeed minima on the potential
[*] Prof. Dr. G. Frenking, Dipl. Chem. U. Pidun, Dipl. Chem. C. Boehme
Fachbereich Chemie der Universitit
Hans-Meerwein-Strasse, D-35032 Marburg (Germany)
Fax: lnt. code +(6421)28-2189
Theoretical Studies of Organometallic Compounds, Part 24. 'This work has
been supported by the Deutsche Forschungsgemeinschaft (SFB 260.
Graduiertenkolleg Metallorganische Chemie) and the Fonds der Chemischen
Industrie. We thank Dr. P.-0. Norrby for helpful comments. Part 23: S. Dapprich, G . Frenking, OrgannmetaNics 1996. 15, 4547.
Chem. Int. Ed. End. 1996, 35. N o . 23/24
energy surfaces.[6] The metallaoxetane 1 was calculated to be
30.5 kcalmol-1 higher in energy than the osmate ester 2. The
reaction of ethylene with osmium tetroxide yielding 1 was predicted to be endothermic by + 18.3 kcalmol-l, while the [3+2]
addition yielding 2 was calculated to be exothermic by
- 12.2 kcalmol-' .r61 The activation barriers for the two reactions and for the rearrangement 1 + 2 were not given, however.
In this paper we present the calculated transition states for the
[2+2] and [3+2] additions of OsO, to C,H, and for the rearrangement 1 + 2 using density functional theory (DFT)[71in
conjunction with a relativistic effective core potential (ECP) for
osmium.[*]We also calculated the transition states for the basecatalyzed reaction using OsO,(NH,) as model compound. The
reaction energies and activation barriers for the base-free reaction are predicted at the CCSD(T) level of theory['] with DFT
optimized geometries. The nature of the transition states was
analyzed by IRC (intrinsic reaction coordinate) calculations.
The NH, complexed structures were studied only at the DFT
level.[' 1'
Figure 1 shows the calculated geometries of the molecules and
the transition states, optimized at B3LYP with basis set II.[lol
The theoretically predicted 0 s - 0 distance for OsO, (1.716 A) is
in excellent agreement with the experimental value (1.71 3 A)
The calculated geometry of the complex 2.NH3is also in accord
with the X-ray structure analyses of related cinchona alkaloid
complexes of OS"'.['~~The observed 0 s - N distances are
2.243(5) and 2.27(2) A; the calculated value is 2.264 A.
Figure 2 shows the calculated reaction profile for the basefree reaction. The theoretical reaction energy at CCSD(T)/II//
B3LYP/II for the reaction C,H, +OsO, + I is + 11.1
kcalmol-', while the formation of 2 is exothermic
(- 21.2 kcal mol- '). The [2 + 21 and [3 + 21 reactions are calculated to be thermodynamically slightly more favorable than determined in our previous work,@]which is due to the different
methods used in the two studies. The energy difference between
1 and 2 calculated here (32.3 kcalmol-') is very similar to our
previous result (30.5 kcalmol- ') .161
The most important results of the present study are the activation barriers for the reactions. Figure 2 shows that the activation barrier at CCSD(T)/IIr'O1 for the [2 + 21 addition yielding
1 is very high (44.7 kcalmol- '). In contrast, the activation barrier for the [3 21 addition is much lower (9.6 kcalmol-'). The
;Ci VCH Verlag.sge.reI1.schnft mhH, D-69451 Weinheim, 1996
0570-0833j9613523-2817 i5.00+ .25,'0
Без категории
Размер файла
604 Кб
salt, aminohydroxylation, efficiency, asymmetric, halocarbamate, catalytic, morel, leads
Пожаловаться на содержимое документа