Journal of the American Chemical Society
Page 6 of 10
1
and are a testament to the privileged nature of the pad-
observed by H NMR analyses of the crude and purified
1
2
3
4
5
6
7
8
dlewheel structure of dirhodium catalysts (entries 1, 2,
and 4). This phenomenon should find widespread appli-
cations to late stage pyrrolidine synthesis.
materials. We believe that the failure of this particular
substrate could be due to an unfavorable conformational
effect or an unidentified attribute of this insertion. None-
theless, our data suggest a need for the rational design of
catalysts able to favor HAT over HDT to further improve
reaction scope and efficiency. In light of the high stereo-
specificity observed with 1c, 1d, and 3j, as well as the high
diastereoselectivity of 3a-e, it appears that a stepwise
HAT with fast radical rebound leads to productive δ-
insertion while HDT results in side-product formation.
With the limited amount of data, a change to a more con-
certed mechanism for less activated sp3 centers cannot be
completely ruled out.
As shown in Scheme 3 (2f) and Table 2 (4d), the kinetic
preference of the nitrenoid provides excellent regioselec-
tivity for unactivated over activated sp3 centers. To fur-
ther explore the site-selectivity of this methodology, in-
ternal competition experiments were conducted (Scheme
4, Panel a). The availability of highly diastereoselective
9
1
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
catalysts was beneficial in simplifying the H NMR anal-
yses of these experiments. Using the cis-selective catalyst
Rh2(S-tertPTTL)4, 3g yielded a 1.0:1.2 mixture of 4g and 4g'
with a slight preference for the 3° versus benzylic 2° cen-
ter. Notably, 4g was isolated as a single diastereomer.33
On the other hand, a more significant preference for 3°
(4.2:1.0) was observed when internal competition was
against an unactivated 2° center (3h). These results sug-
gest a selectivity profile of 3°≥benzylic>2°>>1° and corre-
sponds well with the selectivity observed with Rh2(OAc)4
catalyzed nitrene insertions of sulfamate esters.34
As a working hypothesis, we propose the stereo- and
regio-chemical outcomes of the preceding nitrenoid in-
sertions arise from non-bonding interactions within the
substrate and between the substrate and ligands of the
catalyst, the combination of which determines the pre-
ferred geometry at the δ-carbon (Figure 2). Intermediate
A (Panel a) depicts the acyclic Rh2-nitrenoid complex
with the substrate in a staggered, anti-conformation (an
energy minimum). When the catalyst ligand R3 is bulky
(upper pathway) and in the chiral-crown conformation
with all quadrants sterically hindered, steric repulsion
between R3 and R1 favors rotamer B where R1 is positioned
away from the catalyst's ligands and Hb is suitably posi-
tioned for nitrene insertion leading to the cis-
diastereomer by way of transition state C. As discussed
above, the extent of the steric effect is dependent on the
cavity size of the chiral crown catalyst; this relationship is
critical for the design of highly cis-selective catalysts
when R1 is smaller than phenyl. For non-chiral crown cat-
alysts, R1 approaches the least hindered quadrant of the
catalyst resulting in low or no cis-selectivity due to lim-
ited interaction with R3. When R3 is small (lower path-
way) and/or a non-chiral crown catalyst, e.g, R3 = CH3- as
in Rh2(OAc)4, is utilized, nitrene insertion into Ha via
transition state D is favored resulting in the trans-
diastereomer. In conformer A (Figure 2, Panel b), R1 is
positioned farthest from the nitrenoid and assumes an
anti-orientation with respect to the C(3)-C(4) bond. This
favors formation of five-membered rings. For larger ring
sizes (>5) to be formed, R1 would need to be positioned
closest to the nitrenoid, i.e., E. This stereochemical model
is similar to that proposed by Doyle and co-workers42 for
diastereoselective carbene C-H insertions.
The observed site-selectivities suggest a build-up of
positive charge or radical character at the reaction site
since 3° and benzylic centers provide better stabilization
for both radicals and carbocations. To gain further in-
sights on this, the kinetic isotope effect (KIE) for 3i was
1
measured (Scheme 4, Panel b). H NMR analysis of the
crude reaction mixture and isolated 4i/4i' gave a KIE of
5.3, which suggests significant C-H bond breakage in the
transition state. This value is similar to that observed with
Fe-dipyrrinato (KIE 5.3),17 [Ru2(hp)4Cl] (KIE 4.9),35 and
Cu-diketiminato (KIE 5.3-6.6)36 which are believed to oc-
cur through a stepwise mechanism, but higher than val-
ues observed with those believed to occur through a con-
certed mechanism (KIE 1-3).25,37,38 Further, our observed
KIE highlights a potential difference in the reactivity of
unstabilized and stabilized nitrenes since nitrene transfer
with sulfamate esters using Rh2(esp)2 gives a KIE of
2.9.37,39. In agreement with the stereospecific transfor-
mations of 1c and 1d, insertion into the 3° benzylic center
in 3j (Scheme 4, Panel c) occurred with no detectable loss
in enantiomeric excess. However, the formation of signifi-
cant quantities of olefin 4j' suggests a competitive hydride
transfer mechanism is operative. Indeed, the formation of
TFE adducts (observed as side products during insertions
into benzylic centers)40 and α-insertion products to form
aldehydes or ketones may be explained by a hydride
transfer (HDT) mechanism. For the latter case, it was of
interest41 to determine the influence of an α-deuterium in
selecting between hydrogen or hydride transfer mecha-
nisms. Due to the significant KIE observed, it is reasona-
ble to expect a slower α-deuterium atom transfer (DAT)
or deuteride transfer (DDT) coupled with a corresponding
faster δ insertion when all competing processes occur. To
this end, the mono-and di-deuterated forms of 3k were
evaluated under the insertion conditions (Scheme 4, Pan-
el d). Although a significant difference in reaction rates
was observed, all three substrates gave the corresponding
aldehyde as the major product, and no pyrrolidine was
CONCLUSION
The rational design of regio- and stereoselective cata-
lysts for direct sp3 C-H functionalization remains chal-
lenging.43 The results presented here along with other
recent developments should provide alternatives to the
use of auxiliaries and directing groups and help accom-
plish these goals. We anticipate these results will expedite
the development of predictive tools for the design of se-
lective catalysts for other catalyst-controlled transfor-
mations involving unprotected nitrenoids (or electro-
philic
aminating
agents).
ACS Paragon Plus Environment