Origins of regioselectivity in the reactions of α-lactams with nucleophiles. Two distinct acid-catalyzed pathways involving O- and N-protonation

Article abstract of DOI:10.1021/jo020246h

Sterically stabilized α-lactams react by two distinct acid-catalyzed pathways. Protonation on oxygen results in nucleophilic attack at the acyl carbon and gives C-2 products. Protonation on nitrogen leads to nucleophilic attack at the C-3 carbon and yield

Full text of DOI:10.1021/jo020246h

Or igin s of Regioselectivity in th e Rea ction s of r-La cta m s w ith
Nu cleop h iles. Tw o Distin ct Acid -Ca ta lyzed P a th w a ys In volvin g O-
a n d N-P r oton a tion
Robert V. Hoffman,* Zhiqiang Zhao, Abran Costales, and David Clarke
Department of Chemistry and Biochemistry, North Horseshoe Drive, New Mexico State University,
Las Cruces, New Mexico 88003-8001
rhoffman@nmsu.edu
Received April 8, 2002
Sterically stabilized R-lactams react by two distinct acid-catalyzed pathways. Protonation on oxygen
results in nucleophilic attack at the acyl carbon and gives C-2 products. Protonation on nitrogen
leads to nucleophilic attack at the C-3 carbon and yields C-3 products. The mechanism thus
developed is very useful for understanding the changes in rates and product distributions in the
reactions of sterically stabilized R-lactams with nucleophiles. It can also be extrapolated to other
R-lactams so that a more coherent picture of R-lactam reactivity can be developed.
SCHEME 1
In tr od u ction
R-Lactams (aziridinones) 1 received a significant amount
of attention in the 1960s and early 1970s because of their
interesting and potentially useful reactivity. Their chem-
istry has been summarized in a review by Sheehan and
Lengyel1 in 1968 and in a shorter summary by L’abbe´2
in 1980. A review of the reactions of R-lactams with
nucleophiles has appeared recently as well.3 R-Lactams
are reactive compounds that react with nucleophiles at
room temperature to give two regioisomers. Either the
nucleophile becomes attached to the acyl carbon (C-2 of
the R-lactam) and a 2-amino acid derivative is formed
(C-2 product) or the nucleophile becomes attached to the
saturated ring carbon (C-3 of the R-lactam) and a
2-substituted secondary amide results (C-3 product)
(Scheme 1).
Two generalizations, which were first suggested by
Sheehan1 and reiterated by L’abbe´,2 have generally
served as the starting point for most mechanistic discus-
sions of R-lactam chemistry. Efforts to understand the
factors that influence the regioselectivity of the reaction
are also based on these generalizations. The first is that
the nucleophile attacks the C-2 and C-3 carbons com-
petitively.4 For example, the groups of Quast,5 Maran,6
and D’Angeli7 have used the stereospecific incorporation
of nucleophiles at C-3 as evidence that simple nucleo-
philic attack at C-3 competes with nucleophilic addition
to C-2.
The second is that good, anionic, nonprotic nucleophiles
tend to attack the C-2 carbon while weaker, protic
nucleophiles tend to attack the C-3 carbon. It was
reported that N-tert-butyl-3,3-dimethylaziridinone reacts
with water, methanol, tert-butyl alcohol, benzylamine,
2-toluenethiol, and hydrobromic acid, among others, to
give C-3 products.8 Conversely, sodium methoxide and
potassium tert-butoxide react at the acyl carbon to give
C-2 products. The same general trend was found for
several other R-lactams.9 Recently, Talaty surveyed the
reactions of a large number of proton-bearing and non-
proton-bearing nucleophiles with 1,3-di-tert-butylaziri-
dinone and found a significant number of exceptions. In
particular, amines were found give both C-2 and C-3
products depending on the nucleophilicity and the steric
bulk.10
(1) Sheehan, J . C.; Lengyel, I. Angew. Chem., Int. Ed. Engl. 1968,
7, 25.
(2) L’abbe, G. Angew. Chem., Int. Ed. Engl. 1980, 19, 276.
(3) Hoffman, R. V. In The Amide Linkage: Structural Significance
in Chemistry, Biochemistry, and Materials Science; Greenberg, A.,
Breneman, C. M., Liebman, J . F., Eds.; Wiley-Interscience: New York,
2000; pp 137-155.
(5) (a) Quast, H.; Leybach, H. Chem. Ber. 1991, 124, 2105. (b) Quast,
H.; Leybach, H.; Wu¨rthwein, E.-U. Chem. Ber. 1992, 125, 1249.
(6) Maran, F. J . Am. Chem. Soc. 1993, 115, 6557.
(4) In the 1968 review,1 Sheehan was careful to include the pos-
sibility that dipolar ion i was the source of C-3 products. Stereochemical
studies have largely discounted this possibility.
(7) (a) D’Angeli, F.; Marchetti, P.; Cavicchioni, G.; Maran, F.;
Bertolasi, V. Tetrahedron: Asymmetry 1991, 2, 1111. (b) D’Angeli, F.;
Marchetti, P.; Rondanin, R.; Bertolasi, V. J . Org. Chem. 1996, 61, 1252.
(8) Sheehan, J . C.; Lengyel, I. J . Am. Chem. Soc. 1964, 86, 1356.
(9) Shimazu, M.; Endo, Y.; Shudo, K. Heterocycles 1997, 45, 735.
(10) (a) Talaty, E. R.; Yusoff, M. M. J . Chem. Soc., Chem. Commun.
1998, 985. See also: (b) Yusoff, M. M.; Talaty, E. R. Tetrahedron Lett.
1996, 37, 8695. (c) Talaty, E. R.; Dupuy, A. E., J r.; Utermoehlen, J .
Chem. Soc., Chem. Commun. 1971, 16.
10.1021/jo020246h CCC: $22.00 © 2002 American Chemical Society
Published on Web 06/22/2002
5284
J . Org. Chem. 2002, 67, 5284-5294

Reactions of R-Lactams with Nucleophiles
SCHEME 2
SCHEME 3
Our interest in R-lactams arose from a study of the
chemistry of N-sulfonyloxy amides 2. In the presence of
amine bases and a variety of added nucleophiles, these
compounds were found to give 2-substituted amides.11
The first step in the process, as indicated by kinetic
deuterium isotope effects and leaving group effects,12 is
a concerted 1,3-elimination to give an R-lactam 3. The
R-lactam reacts with nucleophiles to give products. Good
nucleophiles attack the acyl carbon, giving C-2 products,
while poor nucleophiles attack the C-3 carbon and give
C-3 products. Competitive attack by the amine at either
the C-2 or C-3 position was ruled out by experiments
using tert-butylamine as the nucleophile. A mixture of
products from both C-2 (4) and C-3 attack (5) was
obtained (Scheme 2).13
Moreover, it was found that the ratio of C-2 to C-3
product is dependent on the concentration of tert-buty-
lamine. Thus, there are different rate laws for the
formation of the C-2 and C-3 products and the C-2
product increases in a first-order fashion with respect to
tert-butylamine but the C-3 product does not. This finding
effectively rules out the traditional mechanism invoking
direct C-2 vs C-3 attack by the nucleophile (which would
differ by the rate constant but not the order of the
reaction). Furthermore, it suggests that there are com-
peting pathways for the two products and that there is a
second species besides 3 present in solution that leads
to the C-3 product. This species was thought to be
N-protonated R-lactam 6, which is capable of reacting
with even weak nucleophiles at C-3.13 The ammonium
salt produced in the formation of 3 from 2 is likely to be
the proton source since it is the only known acid present
in the reaction mixture. Theoretical studies indicate that
N-protonation is favored slightly over O-protonation.14
The mechanism consistent with these results is shown
in Scheme 3.
insight into the process is desirable. In the first place, 3
is generated in situ from N-mesyloxyamide 2. Despite the
data that show that 2 undergoes a concerted 1,3-elimina-
tion to (presumably) an R-lactam, the fact that 3 cannot
be isolated from reactions of 2 has led to the suggestion
that some other intermediate is produced by elimination
from 2.10a,15 This is unlikely in our opinion, but it can be
addressed experimentally.
Second, the regioselectivity of nucleophilic addition to
R-lactams has previously focused on the nature of the
nucleophile, and little attention has been paid to the
reaction environment. If protonation of the R-lactam
results in C-3 attack, then the presence proton donors
in the reaction mixture could play a significant role in
the regioselectivity.
This study was undertaken with two goals in mind.
We wished to confirm that the 1,3-elimination reaction
of N-sulfonyloxyamides with amine bases does indeed
produce R-lactams in situ, and we wished to determine
if the acidity and/or protic nature of the reaction mixture
influences the regioselection as suggested in Scheme 3.
Resu lts
If the reaction of an N-sulfonyloxyamide with base
produces an R-lactam in situ, then the chemistry of that
R-lactam ought to be identical to the chemistry of the
same R-lactam prepared and reacted as a pure compound
(assuming that the reaction conditions are the same).
Thus, N-sulfonyloxyamide 6 was chosen as the precursor
for R-lactam generation in situ since it would produce
R-lactam 7 upon reaction with base. R-Lactam 7 can also
be isolated and reacted as a pure compound for compari-
son purposes. The regioselectivity of the reaction of pure
7 with amine nucleophiles compared with the product
distribution obtained when 6 is reacted with the same
amine nucleophiles could be used to confirm the identity
of the intermediate as an R-lactam (Scheme 4). Moreover,
the response of the regioselectivity to changes in reaction
conditions would also lend mechanistic insight into the
pathways leading to the C-2 and C-3 products.
Although this mechanism appears to account for the
reaction of 3 with nucleophiles satisfactorily, further
(11) (a) Hoffman, R. V.; Nayyar, N. K. J . Org. Chem. 1995, 60, 7043.
(b) Hoffman, R. V.; Nayyar, N. K. J . Org. Chem. 1995, 60, 5992. (c)
Hoffman, R. V.; Nayyar, N. K.; Chen, W. J . Org. Chem. 1993, 58, 2355.
(d) Hoffman, R. V.; Nayyar, N. K.; Klinekole, B. W. J . Am. Chem. Soc.
1992, 114, 6262. (e) Hoffman, R. V.; Nayyar, N. K.; Chen, W. J . Org.
Chem. 1992, 57, 5700.
(12) Hoffman, R. V.; Nayyar, N. K.; Chen, W. J . Am. Chem. Soc.
1993, 115, 5031.
(13) Hoffman, R. V.; Nayyar, N. K.; Chen, W. J . Org. Chem. 1995,
60, 4121.
(14) Tantillo, D. J .; Houk, K. N.; Hoffman, R. V.; Tao, J . J . Org.
Chem. 1999, 64, 3830. The calculations were carried out on 1,3,3-
trimethylazridinone, which does not have a conjugating phenyl sub-
stituent at C-3. Simple amides significantly favor O-protonation by
8-11 kcal/mol. For an excellent discussion of amide basicity, see:
Greenberg, A. In The Amide Linkage: Structural Significance in
Chemistry, Biochemistry, and Materials Science; Greenberg, A., Bren-
eman, C. M., Liebman, J . F., Eds.; Wiley-Interscience: New York, 2000;
pp 47-83.
N-Mesyloxyamide 6 was prepared by the literature
procedure.11e The R-lactam 7 was prepared from N-tert-
butyl-2-bromo-2-phenylacetamide 8 and potassium tert-
butoxide in ether.16 It was found that rigorous exclusion
of moisture was necessary for success; thus, freshly dried
(15) No alternatives were suggested in ref 9a, but one possibility
would be dipolar ion i, which was first suggested by Sheehan.1,4
(16) Baumgarten, H. E.; Chaing, N.-C. R.; Elia, V. J .; Beum, P. V.
J . Org. Chem. 1985, 50, 5507.
J . Org. Chem, Vol. 67, No. 15, 2002 5285

Hoffman et al.
SCHEME 4
TABLE 1. Regioisom er s fr om th e Rea ction of Am in es
w ith Mesyloxya m id e 6
entry
amine
product ratio C-2/C-3
1
2
3
9a , isopropylamine
9b, pyrrolidine
9c, piperidine
27.6
21.1
7.0
4
5
6
9d , morpholine
6.7
5.5
0.2
9e, 1-adamantylmethylaminea
9f, diethylamine
F IGURE 1. Regioselectivity in the reactions of 6 and 7 with
diethylamine as a function of temperature. Open squares )
7; filled squares ) 6.
7
8
9
10
11
9g, tert-amylamine
9h , diisopropylamine
9i, dicyclohexylamine
9j, 1-adamantylamine
9k , N-methyl-tert-butylamine
0.18
0.16
0.1
produce the reactive intermediate, and the second reacts
as a nucleophile toward the intermediate. In addition, 1
equiv of the corresponding ammonium mesylate salt is
produced during the conversion of 6 to the intermediate.
Thus, for reactions of 6, 2 equiv of the amine was used.
For reactions of R-lactam 7, 1 equiv of the amine was
used and 1 equiv of the ammonium mesylate salt was
added slowly to the reaction mixture to simulate the
formation of the ammonium salt over the course of the
reaction of 6 with the amine. The two reaction mixtures
are not exactly the same since the amine concentration
during the reaction of 6 is higher than for 7 (they become
equal only at the end of the reaction). Moreover, the
addition of the ammonium salt to 7 is somewhat faster
than the rate of its formation from 6; thus, its concentra-
tion is higher during the reactions of 7 (they become
equal only at the end of the reaction). However, the
reaction conditions, if not identical, are certainly com-
parable.
Reactions of N-mesyloxyamide 6 and R-lactam with
diethylamine 9f were compared as a function of temper-
ature, amine concentration, and ammonium salt concen-
tration. The results are shown graphically in Figures
1-3. Reactions of 6 and 7 exhibit almost identical
changes in regioselectivity as a function of temperature
(Figure 1). In addition, it is seen that lower temperatures
favor formation of the C-2 product while higher temper-
atures favor the formation of the C-3 product.
0.08
<0.01b
9e ) 1-C6H10CH2NH2. b Only 11k was detected in the product.
a
ether and freshly sublimed potassium tert-butoxide were
required for good results. Furthermore, the potassium
tert-butoxide was added slowly (as an ether slurry) to the
reaction mixture, the reaction was filtered and evapo-
rated rapidly, and dry pentane was used to recrystallize
the product. Given these precautions, reasonable yields
of 7 can be obtained reliably.
Eleven amines 9a -k were reacted with N-mesyloxy-
amide 6 to determine which (if any) gave mixtures of C-2
and C-3 products (Table 1). Generally high yields (80-
90%) were obtained for regioisomeric mixtures. The
individual regioisomers were separated and character-
ized. The C-2 products 10a -k all had sharp singlets for
the N-tert-butylamine group at about δ 1.1, while the C-3
products 11a -k all had sharp singlets for the N-tert-
butylamide group at about δ 1.3. N-tert-Butylphenylac-
etamide also gives an N-tert-butyl amide singlet at δ 1.3,
which corroborates the assignment. These characteristic
tert-butyl singlets were used to quantitate the C-2/C-3
ratio in the crude products. Primary and cyclic secondary
amines gave mostly C-2 product (Table 1, entries 1-5),
while bulkier amines gave mostly C-3 product (Table 1,
entries 6-11). Diethylamine 9f was chosen for further
study since it gives C-2/C-3 mixtures, reacts at reasonable
rates, and has a relatively simple NMR spectra.
Reactions of 6 and 7 were next compared as a function
of the number of diethylamine equivalents at 0 °C. To
maintain similar reaction conditions, reactions of 6
contained one extra amine equivalent, and reactions of
7 had 1 equiv of diethylammonium mesylate added slowly
to the reaction mixture. The results are shown in Figure
Valid comparison of the reactions of N-mesyloxyamide
6 and R-lactam 7 with 9f requires that the reaction
conditions be the same. The reaction of 6 with an amine
requires 2 equiv of the amine. One is used as a base to
5286 J . Org. Chem., Vol. 67, No. 15, 2002

Reactions of R-Lactams with Nucleophiles
TABLE 2. Effect of Am m on iu m Mesyla te Sa lt on
Rea ction Tim e of 7 a n d Am in esa
amine
reactant ratiob
reaction time
9f
9f
9g
9g
1:1.25:0
1:1.25:1
1:1.25:0
1:1.25:1
90 min
30 min
6 h
2.5 h
a
Reactions were carried out in CD2Cl2 at room temperature.
Reactant ratio is 7/amine/ammonium salt.
b
6 since 1 equiv of the salt was produced over the course
of the reaction. The plot for 6 in Figure 3 has been
corrected for this difference by shifting the salt equiva-
lents by one. As before, 7 gave slightly higher amounts
of the C-2 product due to the higher initial salt concen-
tration. The responses of the regiochemistry to changes
in salt concentration are very similar for 6 and 7. Figure
3 also shows that increasing the amount of ammonium
salt favors the C-2 product by a significant amount.
The data in Figures 1-3 show that the intermediate
produced from 6 and R-lactam 7 show nearly identical
chemical responses to changes in temperature, amine
concentration, and ammonium salt concentration. Al-
though not reported here, similar responses are observed
when tert-amylamine is the amine reactant. Thus, it is
clear that the reaction of 6 with amines does indeed
produce R-lactam 7 in situ by a concerted 1,3 elimina-
tion.12 The reason that R-lactam 7 cannot be isolated is
because its rate of reaction with nucleophiles in solution
is faster than its rate of formation from 6; thus, its
concentration remains low. Moreover it is reasonable to
conclude that the reaction of other N-mesyloxyamides
with bases is a general method for the generation of
R-lactams in situ.
F IGURE 2. Regioselectivity in the reactions of 6 and 7 with
diethylamine at 0 °C as a function of amine eqivalents. Open
squares ) 7; filled squares ) 6.
A most interesting result of these studies was the large
change in regioselectivity wrought by the ammonium salt
(Figure 3). This suggests that some type of acid-base
interaction between the R-lactam and the ammonium ion
plays an important role in the formation of at least one
of the regioisomers. It was thus of interest to further
characterize the effects of acids on the reactions of
R-lactam 7 with nucleophiles.
F IGURE 3. Regioselectivity in the reactions of 6 and 7 with
diethylamine at 0 °C as a function of diethylammonium
mesylate equivalents. Open squares ) 7; filled squares ) 6.
2. The plots in Figure 2 are offset by one amine equiva-
lent since reactions of 6 contained one more amine
equivalent at the beginning of the reaction. Moreover, 7
gave systematically higher amounts of the C-2 product,
probably because the rate of salt addition to 7 (1.5 h) was
faster than the rate of salt generation from 6 (3 h); thus,
the salt concentration was higher in 7. Nevertheless, the
change in regiochemistry as a function of amine concen-
tration, given by the slope of each plot, is very similar
for 6 (slope ) 0.037) and 7 (slope ) 0.031). The data also
show that increasing amine concentrations favor the C-2
product but the increase is much less than first order.
Finally, reactions of 6 and 7 were compared as a
function of the ammonium mesylate salt concentration
at 0 °C. It had been proposed that protonation of an
R-lactam was the first step in the production of the C-3
product (Scheme 3).13 In the experiments that led to the
proposed mechanism, the only proton source in solution
was the ammonium mesylate salt produced during the
formation of the R-lactam from an N-mesyloxyamide.
Thus, it was of interest to determine if an ammonium
salt could actually have a significant effect on the
regioselectivity. To achieve similar reaction conditions,
2.25 equiv of diethylamine was used with 6, whereas 1.25
equiv was used with 7. One equivalent less of diethyl-
ammonium mesylate was added initially to reactions of
The effect of the ammonium salt on the rate of reaction
was studied by reacting 7 with diethylamine and tert-
amylamine. For each amine, two reactions were runs
one with 1 equiv of the corresponding ammonium mesy-
late salt present and one without. The reaction time was
taken as the time required for the complete disappear-
1
ance of the starting material by H NMR. The results
presented in Table 2 show that a single equivalent of the
ammonium salt more than doubles the rate of reaction.
Presumably, the ammonium ion functions as a general
acid toward the R-lactam.
The effect of a specific acid on the rate of reaction was
examined next using a similar approach. Two reactions
were carried out using a solution of 7 in dichloromethane-
d2 containing 1.25 equiv of methanol as the nucleophile.
To one reaction was added about 10 µL of methane-
sulfonic acid. The reaction without the added acid
required 24 h for the complete disappearance of 7,
whereas the reaction to which methanesulfonic acid was
added was complete in less than 15 min. Similar results
were obtained using 2-propanol-d8 as both the solvent
and nucleophile. The reaction without added acid re-
J . Org. Chem, Vol. 67, No. 15, 2002 5287

Hoffman et al.
TABLE 3. Effect of Sp ecific Acid Ca ta lysis on th e
Rea ction Tim e of 7 w ith Wea k Nu cleop h ilesa
TABLE 5. Rea ction of 7 w ith Dieth yla m in e in Alcoh ol
Solven ts a t Room Tem p er a tu r ea
methanesulfonic
solvent
product Ratio C-2/C-3
nucleophile
acid (µL)
solvent
reaction time
t-BuOH
i-PrOH
EtOH
C-3 only
C-3 only
1:1.1
CH3OHb
0
10
0
CD2Cl2
CD2Cl2
i-PrOH-d8
i-PrOH-d8
24 h
CH3OHb
<15 min
90 min
<15 min
i-PrOH-d8
i-PrOH-d8
MeOH
6:1
10
a
1.25 equiv of diethylamine was used.
a
b
Reactions were carried out at room temperature. 1.25 equiv
of methanol was used.
C-2 and C-3 products.13 Thus, instead of considering the
overall reaction time, it is necessary look at the individual
pathways in turn.
TABLE 4. Effect of Solven ts on th e Rea ction of 7 w ith
Nu cleop h iles a t Room Tem p er a tu r e
entry
nucleophile
solvent
reaction time (min)
The factors known to favor the C-2 pathway can be
summarized as follows. The C-2 pathway is favored by
lower reaction temperatures (Figure 1). There is a slight
increase in the C-2 product as the concentration of the
nucleophile is increased (Figure 2). From Figure 3 and
Table 2, both the overall rate of reaction and the amount
of C-2 product increase in the presence of a weak acid
such as the ammonium mesylate salt. This suggests that
the C-2 pathway of 7 is acid catalyzed. To evaluate the
effect of protic solvents on the C-2 pathway, R-lactam 7
was reacted with diethylamine in a series of alcohol
solvents at room temperature (Table 5). The results show
that the proportion of C-2 product is very sensitive to the
acidity of the alcohol and increases markedly with an
increase in solvent acidity.
1
2
3
4
5
6
7
MeOH-d4
EtOH-d6
i-PrOH-d8
t-BuOH-d10
Et2NH
MeOH-d4
EtOH-d6
i-PrOH-d8
t-BuOH-d10
CD2Cl2
<10
30
90
240
80
80
Et2NH
Et2NH
CD3CN
i-PrOH-d8
<10
quired 90 min for completion, whereas addition of 10 µL
of methanesulfonic acid caused complete disappearace of
7 in less than 15 min. These results, presented in Table
3, reveal that specific acid catalysis produces a dramatic
rate acceleration on the reaction of 7 with weak nucleo-
philes.
Since acids clearly can catalyze the reaction 7 with
nucleophiles, the influence of protic solvents was next
considered. R-Lactam 7 was dissolved in a series of
deuterated alcohols, and the reaction time was measured
by NMR. The results presented in Table 4, entries 1-4,
show that the rate increases several 100-fold in the order
t-BuOH-d10 < i-PrOH-d8 < EtOH-d6 < MeOH-d4. The
differences between different alcohol solvents could be
due to increasing acidity or increasing polarity since both
properties increase in the opposite order as the reaction
times. This was evaluated by using diethylamine as a
nucleophile and monitoring the reaction times in deu-
terated dichloromethane, acetonitrile, and 2-propanol,
which have dielectric constants17 of 8.9, 36.2, and 18.3,
respectively (Table 4, entries 5-7). The results show that
reaction times are the same for acetonitrile and dichlo-
romethane but much faster in 2-propanol. Thus, solvent
acidity and not solvent polarity appears to play the
dominant role in accelerating the reaction, and the rate
increases with increasing solvent acidity. It is likely that
hydrogen bonding to the R-lactam is capable of activating
it toward nucleophilic attack.
Taken together, these studies demonstrate that nu-
cleophilic attack on the R-lactam ring of 7 is very
sensitive to acid catalysis (specific or general) and protic
solvents (H-bonding). Thus, the R-lactam ring can be
activated by partial or complete proton transfer or by
hydrogen bonding. In the absence of such activation, the
rate of reaction is much slower.
The above data pertain to the overall rate of reaction
of 7 with nucleophiles. In order address the question of
regioselectivity, it is necessary to understand how various
factors affect the individual pathways that lead to the
The factors known to favor the C-3 pathway can be
summarized as follows. The C-3 pathway is favored by
higher reaction temperatures (Figure 1). In alcohol
solvents where the alcohol serves as the nucleophile and
only the C-3 product is formed, the rate of the C-3
reaction is increased by more acidic alcohols (Table 4,
entries 1-4). The rate is also increased dramatically by
specific acid catalysis (Table 3) These results show that
the C-3 process is also acid catalyzed. When stronger
nucleophiles are present in solution, the C-3 process is
favored by protic solvents which are less acidic, e.g., tert-
butyl alcohol (Table 5). It appears that the C-3 process
is less sensitive to solvent acidity than the C-2 pathway
(even though both are faster in more acidic solvents). As
the solvent becomes more acidic going from tert-butyl
alcohol to methanol, the rate of the C-2 reaction increases
more rapidly than that of the C-3 process until in
methanol it is the major pathway. In the less acidic tert-
butyl alcohol, the C-2 process is slowed markedly and
the C-3 process is dominant.
Another major factor influencing the regioselection is
nucleophilicity. To clarify this dependence, R-lactam 7
was reacted with various nucleophiles in dichloromethane
at room temperature and the ratio of regioisomers was
measured. These reactions did not contain any added acid
catalysts nor protic solvents. The results that are col-
lected in Table 6 demonstrate several trends. Poor
nucleophiles such as alcohols or mesylate ion give only
C-3 product. Secondary and bulky primary amines give
only C-3 attack. A less hindered primary amine such as
isopropylamine gives a mixture of C-2 and C-3 products.
Strong alkoxide nucleophiles give mostly C-2 product.
These results reveal that the C-2 process is more de-
pendent on nucleophilicity than is the C-3 pathway.
These data parallel the results found for R-lactam 3.10
Whether or not the nucleophile bears a proton is ir-
(17) Gordon, A. J .; Ford, R. A. The Chemist’s Companion; J ohn Wiley
and Sons: New York, 1972; pp 4-13.
5288 J . Org. Chem., Vol. 67, No. 15, 2002

Reactions of R-Lactams with Nucleophiles
TABLE 6. Regioch em istr y of th e Rea ction of 7 w ith
Va r iou s Nu cleop h ilesa
TABLE 7. Regioch em istr y of th e Rea ction of 7 w ith
Am in e Nu cleop h iles
nucleophile
ROH
CH3SO3
Et2NH
Me2EtCNH2
i-PrNH2
CH3O-b
tBuO-b
equiv
time (h)
product ratio C-2/C-3
nucleophile
Et2NH
reaction conditionsa
product ratio C-2/C-3
1.25
1.25
1.25
1.25
1.25
2.5
24
24
1.5
3
1
1
C-3 only
C-3 only
C-3 only
C-3 only
0.79:1
A
B
A
B
A
B
C-2 only
C-3 only
C-2 only
C-3 only
1/15
-
Et2NH
tert-amylamine
tert-amylamine
t-BuNHCH3
i-PrNH2
4:1
11:1
1/3
2.5
1
a
Key: (A) 1.25 equiv of amine, MeOH, corresponding am-
monium mesylate salt, -18 °C; (B) 1.25 equiv of amine, t-BuOH,
38 °C.
a
Reactions were carried out in CH2Cl2 at room temperature.
Solvent was toluene.
b
relevant in determining the reaction outcome, it is the
nucleophilicity that is the determining property.
phile and some steric concerns have been at the center
of most discussions.
Our work with N-sulfonyloxyamide 2 as the precursor
to R-lactam 3 led us to conclude that two different
pathways were involved in the production of the C-2 and
C-3 products (Scheme 3).13 The C-2 product results from
direct nucleophilic attack on the acyl carbon of 3. The
C-3 product results from N-protonation of the R-lactam,
which activates the C-3 carbon toward nucleophilic attack
so even weak nucleophiles can attack C-3. Protonation
of the R-lactam requires an acid, and thus, we became
interested in the role of acid catalysts in changing the
regioselectivity.
The results of this study show that for R-lactam 7 there
are two different pathways for the formation of the C-2
and C-3 products, and both the C-2 and C-3 pathways
are subject to some type of acid catalysis. Tables 2-4
show that acids of nearly any type increase the overall
rate of conversion to products markedly. Furthermore,
significant rate increases are seen for reactions that give
C-2 products or C-3 products; thus, both pathways are
acid catalyzed.
For example, the reaction of 7 with diethylamine in
dichloromethane at room temperature takes 1.5 h for
completion and gives only the C-3 product (Table 6).
Addition of 1 equiv of diethylammonium mesylate causes
the reaction to be complete in 30 min but still gives the
C-3 product as the major product (Table 2). The same
reaction of diethylamine with 7 in 2-propanol at room
temperature is complete in less than 10 min and gives
only C-3 product. Thus, the C-3 pathway is catalyzed by
weak acids and protic solvents.
A similar analysis can be made for the C-2 pathway.
Reaction of diethylamine with 7 at room temperature in
dichloromethane gives only the C-3 product (Table 6).
Addition of 1 equiv of diethylammonium mesylate gives
14% of the C-2 product. Reaction of 7 with diethylamine
in methanol gave 84% of the C-2 product and shortens
the reaction time to less than 10 min (Table 7). Thus,
the C-2 path is also catalyzed by weak acids and protic
solvents.
Discu ssion
Evidence for the formation of R-lactams from the
reaction of N-sulfonyloxyamides with bases is largely
derived from mechanistic studies12 that showed that a
1,3-elimination was occurring. Moreover, the products
obtained from these reactions were typical of those
obtained from R-lactams. It was logical to deduce that
the 1,3-elimination did, in fact, produce an R-lactam in
situ that was converted to products. The recent sugges-
tion10a that an intermediate other than an R-lactam could
be involved in the reactions of N-mesyloxyamides with
amine bases was the genesis of this study.
It is assumed that 1,3-elimination in 6 would give
R-lactam 7.12 Moreover, R-lactam 7 can be prepared and
16,18
isolated as a pure compound.
Comparison of the
chemistry of 6 with the chemistry of 7 would reveal if
the reacting species are the same. The reaction of 6 and
7 with diethylamine was used for comparison. To main-
tain similar reaction conditions for 7, the equivalents of
amine were adjusted to reflect the different stoichiometry
of the two systems (7 requires one less amine equivalent
than 6). Moreover, a solution of 1 equiv of diethylammo-
nium mesylate and R-lactam 7 was added slowly to a
solution of diethylamine to simulate the slow production
of 7 and the ammonium salt from the 1,3-elimination of
6. These adjustments provided similar, but not identical,
reaction conditions for the two systems. The C-2/C-3
product ratio was used to compare the chemistry of the
two systems.
The response of the C-2/C-3 ratio for 6 and 7 to changes
in temperature, amine concentration, and ammonium
salt concentration are nearly identical (Figures 1-3).
These results show that the chemical properties of the
intermediate produced from 6 and R-lactam 7 are the
same. The results provide convincing evidence that
R-lactam 7 is produced in situ from 6. It is also reasonable
to assume that other N-sulfonyloxyamides react with
bases to produce R-lactams as well.
One difference between these two pathways is that the
C-2 pathway is favored at low temperatures, whereas the
C-3 pathway is favored at higher temperatures. While it
might appear that this results from a kinetic vs thermo-
dynamic partitioning, the C-2 and C-3 products do not
interconvert under the reaction conditions. While there
are obviously different activation barriers for the two
pathways, the product formation steps are irreversible.
Thus, both pathways must be multistep processes involv-
ing intermediates.
In conjunction with the above study, a study of the
regiochemistry of the reaction of 7 with nucleophiles was
initiated. Regioselectivity has been a core problem in the
chemistry of R-lactams and has yet to be resolved. The
source of regioselectivity in nearly all previous work has
focused on the nucleophile and whether it attacks the
C-2 or the C-3 carbon. The protic nature of the nucleo-
(18) Baumgarten, H. E. J . Am. Chem. Soc. 1962, 84, 4975.
J . Org. Chem, Vol. 67, No. 15, 2002 5289

Hoffman et al.
SCHEME 5
F IGURE 4. Energy diagram for the reaction of 7 with
nucleophiles.
little change in the steric environment of the carbonyl
group and C-2 addition should be sterically retarded as
it is in the unprotonated R-lactam.
Nucleophilicity also plays an important role in these
pathways. Poor nucleophiles such as alcohols or mesylate
give only the C-3 product, irrespective of whether acids
are present or not. Better nucleophiles such as unhin-
dered amines are capable of adding to the acyl carbon
and give C-2 products. The nucleophilicity of bulkier
amines is reduced, and they give only C-3 products.
Diethylamine and tert-amylamine fall in a middle ground
and give mixtures of C-2 and C-3 products. Furthermore
the amount of C-2 product does not increase in a first-
order fashion for these amines reacting with 7 as it did
for the reactions of the less hindered 3.12 Thus, it is fair
to say that direct attack on the C-2 carbon is not possible
for 7 because the phenyl and tert-butyl groups block
nucleophilic approach to the acyl carbon. In order for C-2
attack to be successful, the acyl group of 7 must be
activated by acid catalysis.
A reaction mechanism based on these requirements is
pictured in Scheme 5. Two distinct protonated species
are proposed. One is an O-protonated (or hydrogen
bonded) R-lactam 7-OH. The second is an N-protonated
(or hydrogen bonded) R-lactam 7-NH. These protonated
species do not interconvert directly. The N-H proton in
7-NH is nearly orthogonal to the lone pair on oxygen
while the O-H proton in 7-OH is anti and orthogonal to
the lone pair on nitrogen.19 To interconvert they must
pass through the unprotonated R-lactam 7. The product
formation steps are irreversible.
The proposed mechanism shows that O-protonation
7-OH leads to C-2 attack while N-protonation 7-NH leads
to C-3 attack. Two factors led to this choice. First the
effect of temperature requires that the less stable pro-
tonated form gives the C-2 product. Calculations on 1,3,3-
trimethylazridinone have shown that the N-protonated
species is of slightly lower energy than the anti O-
protonated species. This difference should be greater for
7, which has a conjugating phenyl group at C-3.13 Thus,
if 7-OH is of higher energy than 7-NH, then it is the
intermediate that leads to the C-2 product.
Second, protonation on oxygen causes significant flat-
tening of the nitrogen atom in 7-OH (it is pyramidal in
the unprotonated compound).14 This would reduce steric
hindrance to nucleophilic attack at C-2 and should
facilitate C-2 addition relative to the unprotonated R-lac-
tam. Conversely, protonation on nitrogen should cause
This mechanistic assignment differs from the theoreti-
cal study of the reaction of protonated 1,3,3-trimethyl-
14
aziridinone with chloride ion.
In that study, four
transition states were located for chloride addition to C-2
or C-3 of N-protonated and O-protonated species in the
gas phase. Significant H-bonding between the protonated
species and the chloride ion nucleophile appears to play
an important role in the transition state energies amd
structures. It was concluded that C-3 attack on the anti-
O-protonated form had the lowest energy barrier, and
C-2 attack on the N-protonated form was energetically
favored for C-2 addition. In the present case, particularly
where general acid catalysis is observed, there should be
minimal H-bonding with the incoming nucleophile. More-
over, steric effects are probably much more important for
reactions of 7 with amines than for the reaction of 1,3,3-
trimethylaziridinone with chloride ion. Thus, it is not
unreasonable that the mechanisms of the two systems
have different energy profiles.20
An energy diagram consistent with this mechanism is
shown in Figure 4. Assuming that the N-protonated
species is of lower energy than the anti-O-protonated
species, the temperature data suggest that the energy
barrier for O-protonation is lower than the energy barrier
for N-protonation. The lower barrier for O-protonation
means that at lower temperatures, it is possible to
achieve equilibrium between 7 and 7-OH but not between
7 and 7-NH. Thus, the concentration of 7-OH is greater
than the concentration of 7-NH at low temperature.19
The energy barrier (E3) for nucleophilic attack at C-2
of 7-OH is significant because of steric hindrance despite
the activation by protonation. At lower temperature
where 7 and 7-OH are in equilibrium, the Curtin-
Hammett principle applies. Thus, E3, the energy barrier
for acyl attack on 7-OH, must be less than E2, the energy
barrier for protonation of 7 (E3 < E2), because the C-2
process proceeds but N-protonation and hence the C-3
process does not. As the temperature increases, the
protonation equilibrium between 7 and 7-NH can be
established and the concentration of 7-NH increases
because of its greater stability. Protonation on nitrogen
makes C-3 susceptible to attack by even weak nucleo-
philes. The energy barrier for nucleophilic attack on C-3
(19) The syn protonated form of 7-OH is about 2 kcal/mol higher in
energy than the anti form shown in Scheme 5,14 and the proton is
nearly orthogonal to the lone pair on nitrogen as well.
(20) How much greater is not known because the relative energies
of 7-OH and 7-NH are not known. The temperature dependence
suggests, however, that the difference in concentration is large.
5290 J . Org. Chem., Vol. 67, No. 15, 2002

Reactions of R-Lactams with Nucleophiles
(E4) is small and 7-NH proceeds rapidly to the C-3
product. The rapid conversion of 7-NH to product shifts
the protic equilibrium toward 7-NH and the C-3 product
becomes favored at higher temperatures.
appears that even trace amounts of water activate the
acyl carbon by protonation or H-bonding toward nucleo-
philic attack by tert-butoxide.
The data in Table 7 illustrates these considerations.
The use of low temperature, methanol solvent, and the
ammonium salt as an acid catalyst would favor the C-2
process. Thus, for diethylamine and tert-amylamine, only
the C-2 product is observed under these conditions. The
use of higher temperature and the less acidic tert-butyl
alcohol as solvent produces only the C-3 product for these
same amines. Thus the energy barrier E3 for these two
amines is smaller than, but comparable to, E2 as shown
in Figure 4.
The nucleophile also plays an important role in deter-
mining the regiochemistry in reactions of 7. Good nu-
cleophiles lower the energy barrier E3 between 7-OH and
the C-2 product (Figure 4) and make the C-2 process
faster. If the nucleophile is very good, E3 can become
sufficiently low that N-protonation and, hence, the C-3
process cannot compete. Thus, in Table 7, isopropylamine
is a better nucleophile than tert-amylamine (both pri-
mary) by virtue of its smaller size so that even under
conditions favoring the C-3 process, some C-2 product is
found. Poor nucleophiles (bulky or otherwise poor, e.g.,
MsO-) result in an increase in E3. If E3 becomes much
larger than E2, then the C-2 process does not compete
and the C-3 product is favored. Thus, in Table 7,
N-methyl-tert-butylamine is a poorer nucleophile than
diethylamine (both secondary) by virtue of its steric bulk
so that even under conditions favoring the C-2 process,
very little C-2 product is obtained.
The mechanism shown in Scheme 5 shows activation
by protonation of either oxygen or nitrogen. Such activa-
tion does not require complete proton transfer to be
effective. Similar arguments can be made for activation
of the acyl carbon by hydrogen bonding to the oxygen or
activation of the C-3 carbon by hydrogen bonding to
nitrogen. In fact, the rate accelerations seen for am-
monium salts and alcohol solvents suggest that general
acid catalysis or H-bonding rather than complete proton
transfer is the dominant mode of catalysis under the
conditions investigated. Theoretical investigations have
shown that hydrogen bonding to either the oxygen or
nitrogen of 1,3,3-trimethylazridinone result in similar
(but less developed) changes in structure as complete
proton transfer.14
The competing pathways in Scheme 5 are independent
of whether the nucleophile is protic or aprotic, but only
depend on the nucleophilicity of the nucleophile, the
presence of Bronsted acids (general or specific) and/or
protic solvents, and the temperature. This mechanism
is thus a departure from earlier mechanisms in which
the protic nature of the nucleophile and competing C-2
and C-3 attacks were invoked.1,2,7,10
It is important to note that the mechanism in Scheme
5 was developed for R-lactam 7. This R-lactam is steri-
cally stabilized, and thus, direct nucleophilic attack on
the acyl carbon is greatly retarded. This also translates
to a significant energy barrier E3 between 7-OH and the
C-2 product. Furthermore, the phenyl group at C-3
facilitates positive charge development at C-3. Thus, 7
is structurally set up to favor the C-3 process.
It is likely that the same general pathways are present
in the reactions of R-lactams that lack steric stabilization.
In such cases, acyl attack is much faster, i.e., the energy
barrier E3 between the O-protonated (or H-bonded)
R-lactam and its C-2 product is low. In fact, for good
nucleophiles there is the additional pathway of direct
nucleophilic attack on the acyl carbon of the unproto-
nated (or non-H-bonded) R-lactam. This is the case for
R-lactam 3, which undergoes direct nucleophilic attack
at C-2 by tert-butylamine.13 Nevertheless, very bulky
nucleophiles can increase energy barrier E3 to the point
that the C-3 process becomes competitive.12
Groups at C-3 that do not stabilize electron deficiency
at the C-3 carbon should raise the barrier E2 for N-
protonation and thus slow the C-3 pathway. This is
precisely the behavior of R-lactam 13, which gives no C-3
products, even with very bulky amine nucleophiles. Only
C-2 products are found.13
This scenario is consistent with several observations
of the system. At low temperatures, the addition of
ammonium salts or more acidic solvents increases the
concentration of 7-OH and the rate of conversion to the
C-2 product increases. (At low temperatures, protonation
of nitrogen is not competitive.) At higher temperatures,
only protic solvents are needed to give sufficient quanti-
ties of 7-NH for reaction by the C-3 pathway. Greater
amounts of stronger acids increase the concentration of
both 7-OH and 7-NH and some of the 7-OH could form
the C-2 product.
In summary, the mechanism developed in Scheme 5
and the energy diagram in Figure 4 are very useful for
understanding the changes in rates and product distribu-
tions in the reactions of sterically stabilized R-lactams
with nucleophiles. They also can be extrapolated to other
R-lactams so that a more coherent picture of R-lactam
reactivity can be developed.
This scenario also explains the stringent moisture
control necessary in the synthesis of 7. Although not
emphasized in the literature,18 we found that if freshly
dried ether and freshly sublimed potassium tert-butoxide
were not used, the product 7 was contaminated with
significant amounts of tert-butyl amino ester 12. It
Exp er im en ta l Section
Gen er a l Meth od s. Ether was distilled from benzophenone
ketyl. Pentane and dichloromethane were dried and distilled
from P2O5. Acetonitrile was dried with CaH2 and distilled.
Alcohols were dried using magnesium turnings and distilled.
J . Org. Chem, Vol. 67, No. 15, 2002 5291

Hoffman et al.
Potassium tert-butoxide was freshly sublimed. 1H NMR and
13C NMR spectra were recorded at 200 and 50 MHz, respec-
tively. Elemental analyses were performed by M-H-W Labo-
ratories, Phoenix, AZ. Melting points are uncorrected. Thin-
layer chromatography was performed on silica gel 60 F254
plates and visualized by UV irradiation or vanillin in a mixture
of methanol and sulfuric acid. N-Mesyloxyamide 611e and
N-tert-butyl-2-bromophenylacetamide 816 were prepared by
literature procedures.
Gen er a l P r oced u r e for P r ep a r in g Am m on iu m Mesy-
la te Sa lts. Methanesulfonic acid (25 mmol) in 10 mL of ethyl
acetate was added over a period of 20 min to a 0 °C solution
of the amine (25 mmol) in 30 mL of ethyl acetate. The mixture
was stirred for an additional 10 min. The solvent was
evaporated. The products were recrystallized.
δ 171.72, 141.39, 128.62, 127.12, 126.89, 56.61, 50.91, 46.02,
43.38, 29.29, 25.56, 25.22, 24.25; IR 1631 cm-1. Anal. Calcd
for C17H26N2O: C, 74.41; H, 9.55; N, 10.21. Found: C, 74.29;
H, 9.46; N, 10.11.
11c was obtained as a white solid in 9% yield: mp 107-
108 °C; 1H NMR δ 1.36 (s, 9H),1.41-1.58 (m, 6H), 2.32 (t, 4H,
J ) 5.14), 3.70 (s, 1H), 7.26-7.28 (m, 5H); 13C NMR δ 170.99,
136.55, 129.09, 128.34, 127.79, 52.80, 50.55, 18.82, 26.51,
24.40; IR 3415, 1661 cm-1. Anal. Calcd for C17H26N2O: C,
74.41; H, 9.55; N, 10.21. Found: C, 74.28; H, 9.86; N, 10.21.
Rea ction of 6w ith m or p h olin e gave a mixture of 10d and
11d (6.7:1) in 90% crude yield. These were separated by flash
chromatography (hexane/ethyl acetate ) 5/1).
10d was obtained as a white solid in 73% yield: mp 96-97
1
°C; H NMR δ 1.12 (s, 9H), 3.15-3.59 (m, 8H), 4.52 (s, 1H),
7.23-7.32 (m, 5H); 13C NMR δ 172.53, 141.02, 128.92, 127.35,
Dieth yla m m on iu m m esyla te was obtained as an oil in
90% yield after low-temperature recrystallization from ac-
etone: 1H NMR δ 1.38 (t, 6H, J ) 7.28 Hz), 2.77 (s, 3H), 3.04
(m, 4H), 8.95 (bs, 2H).
-1
127.22, 66.36, 56.99, 51.12, 45.96, 29.48; IR 3296, 1650 cm
.
Anal. Calcd for C16H24N2O2: C, 69.53; H 8.75; N, 10.14.
Found: C 69.70; H 8.66; N, 10.10.
ter t-Am yla m m on iu m m esyla te was obtained as a white
solid in 93% yield after recrystallization from ethyl acetate:
mp 103-104 °C; 1H NMR δ 0.98 (t, 3H, J ) 7.56 Hz), 1.36 (s,
6H), 1.72 (q, 2H), 2.79 (s, 3H), 7.50 (bs, 3H).
11d was obtained as a white solid in 11% yield: mp 138-
139 °C; 1H NMR δ 1.35 (s, 9H), 2.41 (m, 4H), 3.68 (s, 1H), 3.70
(m, 4H), 6.97 (bs, 1H), 7.31 (s, 5H); 13C NMR δ 169.80, 135.79,
128.60, 128.43, 128.02, 66.95, 52.06, 50.58, 28.60; IR 3436,
1662 cm -1. Anal. Calcd for C16H24N2O2: C, 69.53; H 8.75; N,
10.14. Found: C, 69.69; H, 8.63; N, 9.96.
Gen er a l P r oced u r e for th e Rea ction of 6 w ith Am in es.
To a 0 °C solution of N-mesyloxyamide 6 (130-250 mg, 0.456-
0.876 mmol) in dichloromethane (10 mL) was added a neat
amine (1.03-1.97 mmol). The reaction mixture was stirred at
room temperature until TLC analysis showed no starting
material to be present (hexane/ethyl acetate ) 4:1). After
rotary evaporation, the residue was treated with 1.0 M
potassium carbonate (10 mL) and extracted with ethyl acetate
(2 × 10 mL). The organic extracts were washed with water (2
× 10 mL) and brine (10 mL) and dried (MgSO4). After rotary
evaporation, the crude products were placed under high
R ea ct ion of 6 w it h 1-a d a m a n t ylm et h yla m in e gave a
mixture of 10e and 11e (5.45:1) in 89% crude yield.
10e was obtained as a crystalline solid in 76% yield after
recrystallization of the crude product (ethyl acetate): mp 153-
1
154 °C; H NMR δ 1.14 (s, 9H), 1.49-1.97 (m, 15H), 2.96 (d,
2H J ) 6.22), 4.34 (s, 1H), 7.26-7.34 (m, 5H), 8.03 (bs, 1H);
13C NMR δ 173.93, 141.99, 129.32, 128.20, 127.74, 62.27, 52.27,
51.31, 40.74, 37.40, 34.31, 29.68, 28.69; IR 1651 cm-1. Anal.
Calcd for C23H34N2O: C, 77.92; H 9.67; N, 7.90. Found: C,
77.68; H 9.49; N, 7.98. 11e was not isolated.
1
vacuum for several hours. H NMR was used to measure the
product ratio by measuring the relative height of the tert-butyl
group of the C-2 product (δ ) 1.1) and the C-3 product (δ )
1.3). The results are collected in Table 1. Individual regioiso-
mers were separated and characterized.
Rea ction of 6 w ith isop r op yla m in e gave a mixture of
10a and 11a (27.6:1) in 88% crude yield. These were separated
by flash chromatography (hexane/ethyl acetate ) 4/1).
10a was obtained as a white solid in 80% yield: mp 66-67
°C; 1H NMR δ 1.048 (s, 9H), 1.10 (s, 3H), 1.14 (s, 3H), 4.19 (s,
1H), 3.95-4.12 (m, 1H), 7.22 (m, 5H), 7.75 (bs, 1H); 13C NMR
δ 172.79, 140.85, 128.95, 127.70, 127.19, 61.71, 52.02, 40.77,
29.28, 22.95, 22.60; IR 3296, 1651 cm-1. Anal. Calcd for
Rea ction of 6 w ith d ieth yla m in e gave a mixture of 10f
and 11f (0.20:1) in 86% crude yield. These were separated by
flash chromatography (hexane/ethyl acetate ) 4/1).
10f was obtained as a white solid in14% yield: mp 47-48
1
°C; H NMR δ 1.01 (t, 3H, J ) 7.14), 1.11 (t, 3H, J ) 7.10),
1.12 (s, 9H), 3.02-3.50 (m, 4H), 7.24-7.29 (m, 5H); 13C NMR
δ 12.68, 14.00, 29.61, 40.75, 41.55, 51.51, 57.35, 127.42, 127.62,
129.05, 141.49, 172.80; IR 3026, 2970, 1682 cm-1. Anal. Calcd
for C16H26N2O: C, 73.24; H, 9.99; N, 10.68. Found: C, 73.50;
H, 9.85; N, 10.71.
11f was obtained as a white solid in 67% yield: mp 65-66
1
°C; H NMR δ 0.99 (t, 6H, J ) 7.10), 1.36 (s, 9H), 2.29-2.47
C
15H24N2O: C, 72.54; H, 9.74; N, 11.28. Found: C, 73.02; H,
(sextet, 2H, J ) 6.96), 2.54-2.72 (sextet, 2H, J ) 7.08), 4.09
(s, 1H), 7.27-7.32 (m, 5H); 13C NMR δ 11.88, 28.79, 43.75,
50.63, 72.35, 127.80, 128.29, 129.41, 136.44, 171.57; IR 3018,
2976, 1691 cm-1. Anal. Calcd for C16H26N2O: C, 73.24; H, 9.99;
N, 10.68. Found: C, 73.30; H, 9.77; N, 10.73.
9.16; N, 11.29.
11a was obtained as a white solid in 3% yield: mp 67-68
1
°C; H NMR δ 1.36 (s, 9H), 1.10 (d, 3H, J ) 1.3), 1.07 (d, 3H,
J ) 1.5), 2.81 (septet, 1H, J ) 6.3), 4.08 (s, 1H), 7.32 (m, 5H),
7.75 (bs, 1H); 13C NMR δ 171.87, 140.76, 128.66, 127.61,
127.04, 66.25, 50.42, 48.42, 28.64, 23.26, 23.02; IR 3404, 1645
cm-1. Anal. Calcd for C15H24N2O: C, 72.54; H, 9.74; N, 11.28.
Found: C 72.42; H 9.57; N, 11.12.
Rea ction of 6 w ith ter t-a m yla m in e gave a mixture of 10g
and 11g (0.18:1) in 91% crude yield. These were separated by
flash chromatography (hexane/ethylethyl acetate ) 4/1)
10g was obtained as a solid in 12% yield: mp 71-72 °C; 1H
NMR δ 0.86 (t, 7.40, 3H), 1.13 (s, 9H), 1.32 (s, 3H), 1.34 (s,
Rea ction of 6 w ith p yr r olid in e gave a mixture of 10b
and 11b (21.1:1) in 93% crude yield.
3H), 1.74 m, 2H), 4.18 (s, 1H), 7.30 (m, 5H), 7.85 (bs, 1H); 13
C
10b was obtained as a crystalline solid in 79% yield after
recrystallization (ethyl acetate) of the crude product: mp 148-
149 °C; 1H NMR δ 1.13 (s, 9H), 1.81 (m, 4H), 3.22 (m, 1H),
3.54 (m, 3H), 4.40 (s, 1H), 7.31 (m, 5H); 13C NMR δ 172.34,
140.82, 128.93, 127.80, 127.46, 58.79, 51.42, 46.30, 46.12,
29.61, 26.08, 24.13; IR 2976, 1636 cm-1. Anal. Calcd. for
NMR δ 8.31, 26.07, 29.13, 34.22, 51.85, 53.03, 62.11, 127.03,
127.52, 128.76, 141.94, 172.68; IR 3056, 1681 cm-1. Anal. Calcd
for C17H28N2O: C, 73.87; H, 10.21; N, 10.13. Found: C, 73.84;
H, 10.05; N, 9.81.
11g was obtained as a solid in 68% yield: mp 69-70 °C; 1H
NMR δ 0.83 (t, 3H, J ) 7.32), 1.03 (s, 3H), 1.08 (s, 3H), 1.38
(m, 2H), 4.12 (s, 1H), 7.28 (m, 5H), 7.86 (bs, 1H); 13C NMR δ
8.31, 25.59, 26.84, 28.57, 34.50, 50.26, 54.17, 61.76, 127.03,
127.46, 128.78, 141.93, 172.97; IR 3415, 3029, 1661 cm-1. Anal.
Calcd for C17H28N2O: C, 73.87; H, 10.21; N, 10.13. Found: C,
73.04; H, 9.97; N, 9.69.
C
16H24N2O: C, 73.81; H, 9.29; N, 10.76. Found: C, 73.71; H,
9.09; N, 10.68. 11b was not isolated.
Rea ction of 6 w ith p ip er id in e gave a mixture of 10c and
11c (7.04:1) in 81% crude yield. These were separated by flash
chromatography (hexane/ethyl acetate ) 5/1).
10c was obtained as a white solid in 66% yield: mp 95-96
°C; 1H NMR δ 1.15 (s, 9H), 1.47-1.53 (m, 6H), 3.31-0.3.35
(m, 2H), 3.55-5.59 (m, 2H), 4.57 (s, 1H), 7.23 (m, 5H); 13C NMR
Rea ction of 6 w ith d iisop r op yla m in e gave a mixture of
10h and 11h (0.16:1) in 83% crude yield.
5292 J . Org. Chem., Vol. 67, No. 15, 2002

Reactions of R-Lactams with Nucleophiles
11h was obtained as a white solid in 68% yield after flash
chromatography (hexane/ethyl acetate ) 7:3): mp 89-90 °C;
1H NMR δ 0.82 (d, 3H, J ) 6.58 Hz), 1.12 (d, 3H, J ) 6.68
Hz), 1.42 (s, 9H), 3.23 (m, 2H), 4.51 (s, 1H), 7.27 (s, 5H), 7.78
(bs, 1H); 13C NMR δ 174.32, 139.42, 130.98, 128.47, 127.48,
64.80, 51.14, 47.06, 29.27, 24.23, 21.56; IR 3315, 2967, 1661
cm-1. Anal. Calcd for C18H30N2O: C, 74.44; H, 10.41; N, 9.65.
Found: C, 74.42; H 9.95; N, 9.42. 10h was not isolated.
Rea ction of 6 w ith d icycloh exyla m in e gave a mixture
of 10i and 11i (0.10:1) in 81% crude yield.
11i was obtained as a white solid in 69% yield after flash
chromatography (hexane/ethyl acetate ) 3/1): mp 109-110
°C; 1H NMR δ 0.99-1.82 (m, 20H), 1.41 (s, 9H), 2.66 (m, 2H),
4.56 (s, 1H), 7.26 (s, 5H), 7.85 (bs, 1H); 13C NMR δ 174.46,
139.99, 130.87, 128.32, 127.40, 65.62, 57.06, 51.16, 35.28,
33.31, 29.32, 27.38; IR 2927, 1666 cm-1. Anal. Calcd for
was stirred for 3 h. After rotary evaporation, the residue was
treated with 1 M potassium carbonate (10 mL) and extracted
with ethyl acetate (2 × 10 mL). The organic extracts were
washed with water (2 × 10 mL) and brine (10 mL) and dried
(MgSO4). After rotary evaporation, the crude product was
placed under high vacuum for several hours. H NMR was
used to measure the ratio of the C-2 and C-3 regioisomers in
the crude product. The results are presented in Figure 1.
1
A second comparison was done using tert-amylamine and
tert-amylammonium mesylate as the amine and salt, respec-
tively, and similar results were obtained.
Effect of Am in e Con cen tr a tion on Rea ction s of 6 a n d
7. -Sulfonoxyamide 6 (25 mg, 0.0876 mmol) in dichloromethane
(5 mL) was added to a solution of diethylamine (0.197, 0.394,
0.788, and 1.183 mmol) in dichloromethane (5 mL) at 0 °C.
The reaction mixture was stirred at 0 °C until TLC showed
no starting material to be present. The reaction was worked
up as above. Changes of the product ratio were observed as a
function of amine concentration. The results are presented in
Figure 2.
C
24H38N2O: C, 77.79; H, 10.34; N, 7.56. Found: C, 77.64; H,
10.12; N, 7.48. 10i was not isolated.
Rea ction of 6 w ith 1-a d a m a n ta n a m in e gave a mixture
of 10j and 11j (0.083:1) in 86% crude yield.
11j was obtained as a white solid in 67% yield after
recrystallization (methanol-water): mp 101-102 °C; 1H NMR
δ 1.38 (s, 9H), 1.58-1.75 (m, 12H), 2.07 (bs, 3H), 4.29 (s, 1H),
7.26-7.29 (m, 5H), 7.97 (bs, 1H); 13C NMR δ 173.04, 142.11,
128.75, 127.16, 59.86, 51.87, 50.25, 42.97, 36.28, 29.29, 28.55;
IR 3414, 1661 cm-1. Anal. Calcd for C22H32N2O: C, 77.60; H,
9.47; N, 8.23; Found: C, 77.73; H, 9.29; N, 8.31. 10j was not
isolated.
In a similar set of experiments, a solution of 7 (17 mg, 0.0898
mmol) and diethylammonium mesylate (0.0898 mmol) in
dichloromethane (3 mL) was added to a solution of diethy-
lamine (0.112, 0.314, 0.718, 1.122 mmol) in dichloromethane
(5 mL) at 0 °C over a period of 1.5 h. The mixture was stirred
at 0 °C for 3 h. The reaction was worked up as above. The
results are presented in Figure 2.
A similar set of experiments using tert-amylamine gave
Rea ction of 6 w ith N-m eth yl-ter t-bu tyla m in e gave only
11k in 89% crude yield. 11k was obtained as a white solid in
78% yield after flash chromatography (hexane/ethyl acetate
) 3/1): mp 88-89 °C; 1H NMR δ 1.17 (s, 9H), 1.42 (s, 9H),
1.99 (s, 3H), 4.58 (s, 1H), 7.15-7.29 (m, 5H), 7.921 (bs, 1H);
13C NMR δ 172.85, 137.75, 130.23, 127.77, 126.90, 66.49, 55.29,
50.18, 32.17, 28.55, 27.17; IR 3435, 1661 cm-1. Anal. Calcd
for C17H28N2O: C, 73.87; H, 10.21; N, 10.13. Found: C, 73.84;
H, 10.05; N, 9.81.
similar results.
Effect of Am m on iu m Sa lt Con cen tr a tion on Rea ction s
of 6 a n d 7. N-Sulfonyloxyamide 6 (40 mg, 0.14 mmol) and
diethylammonium mesylate (0.315, 0.63, 1.26, 1.89, mmol) in
dichloromethane (5 mL) were added to a solution of diethyl-
amine (0.315 mmol) in dichloromethane (5 mL) at 0 °C. The
reaction mixture was stirred at 0 °C until TLC showed no
starting material to be present. The reaction was worked up
as above. Changes of the product ratio were observed as a
function of ammonium mesylate salt concentration. The results
are shown in Figure 3.
In a similar set of experiments, R-lactam 7 (30 mg, 0.158
mmol) and diethylammonium mesylate (0.513, 0.869, 1.58,
2.29 mmol) in dichloromethane (5 mL) was added to a solution
of diethylamine (20 (L, 0.197 mmol) in dichloromethane (5 mL)
at 0 °C over a period of 1.5 h. The mixture was stirred at 0 °C
for 3 h. The reaction was worked up as above. The results are
presented in Figure 3.
N-ter t-Bu tyl-3-p h en yla zir id in on e 7. A stirred suspension
of freshly sublimed potassium tert-butoxide (0.5516 g, 4.913
mmol) in dry ether (40 mL) was added over a period of 1 h to
a well-stirred solution of 8 (1.021 g, 3.779 mmol) in dry ether
(70 mL) at 0 °C under a nitrogen atmosphere. The resulting
mixture was filtered quickly. The filtrate was placed on a
rotary evaporator, and the solvent was removed. The residue
was recrystallized from dry pentane at -18 °C. R-Lactam 7
was obtained as a white crystalline solid in 65% yield: mp
1
32-33 °C; IR 1845 cm -1; H NMR δ 1.37 (s, 9), 3.87 (s, 1H),
A set of experiments using tert-amylamine and tert-amyl-
7.34 (s, 5H). The physical properties of the product are
identical to literature values.16 Pure 7 can be stored for at least
1 month at -18 °C, but only a few days at room temperature.
Effect of Tem p er a tu r e on Rea ction s of 6 a n d 7. Mesyl-
oxyamide 6 (20-40 mg, 0.07-0.14 mmol) in dichloromethane
(5 mL) was added to a solution of diethylamine (0.157-0.315
mmol) in dichloromethane (5 mL) at -18, 0, +23, or +40 °C.
The reaction mixture was stirred until TLC showed no starting
material present. After rotary evaporation, the residue was
treated with 1 M potassium carbonate (10 mL) and extracted
with ethyl acetate (2 × 10 mL). The organic extracts were
washed with water (2 × 10 mL) and brine (10 mL) and dried
(MgSO4). After rotary evaporation, the crude product was
ammonium mesylate gave similar results.
NMR Mon itor in g of Rea ction Tim es. The time required
for the disappearance of the R-lactam was determined by
monitoring 1H NMR peaks at δ 3.87 and 1.37. Different
reaction conditions were evaluated by adding 7 to deuterated
solvent (0.75 mL) containing nucleophiles and/or acid catalysts
to be evaluated. The mixture was well shaken and kept at a
specified temperature (either room temperature or 0 °C). The
tube was periodically placed in the NMR probe and the
spectrum recorded. The reaction time was taken as the time
required for complete disappearance of the starting material.
Effect of Dieth yla m m on iu m Mesyla te R-Lactam 3 (26
mg, 0.14 mmol) was added to an NMR tube containing
diethylamine (18 µL, 0.17 mmol) and diethylammonium me-
sylate (0.14 mmol) in CD2Cl2 at room temperature. A similar
reaction mixture that did not contain diethylammonium
mesylate was used for comparison. A similar set of experi-
ments was carried out using tert-amylamine and tert-amyl-
ammonium mesylate. The results are collected in Table 2.
1
placed under high vacuum for several hours. H NMR was
used to measure the ratio of the C-2 and C-3 regioisomers by
measuring the relative height of the tert-butyl groups at δ )
1.1 (C-2) and 1.3 (C-3) in the crude product. The results are
presented in Figure 1.
In a similar sets of experiments, R-lactam 7 (15-30 mg,
0.0792-0.158 mmol) and diethylammonium mesylate (13.4-
26.7 mg, 0.0792-0.158 mmol) in dichloromethane (3 mL) were
added to a solution of diethylamine (10-20 (L, 0.0991-0.197
mmol) in dichloromethane (5 mL) at -18, 0, +23, and +40
°C. This gave a final molar ratio of 1:1.25:1 of 7/amine/salt,
which was the same as the final ratio used for 6. The mixture
Effect of Sp ecific Acid . R-Lactam 3 (37 mg, 0.195 mmol)
was added to a mixture of methanol (10 µL, 0.24 mmol) and
methanesulfonic acid (4µL, 0.054 mmol) in CD2Cl2 at room
temperature. A similar reaction mixture that did not contain
methanesulfonic acid was run for comparison.
J . Org. Chem, Vol. 67, No. 15, 2002 5293

Hoffman et al.
R-Lactam 3 (35 mg, 0.185 mmol) was added to a solution of
methanesulfonic acid (4 µL. 0.054 mmol) in 2-propanol-d8 at
room temperature. Reaction of 3 in 2-propanol-d8 was used
for comparison. The results are shown in Table 3.
alcohol were carried out at room temperature. The solvent was
evaporated, and the residue was worked up as described above.
The ratio of the C-2/C-3 isomers in the crude product was
measured by NMR. The results are collected in Table 5.
Effect of P r otic Solven ts R-Lactam 3 (20 mg, 0.106 mmol)
was added to methanol-d4, ethanol-d6, 2-propanol-d8, and tert-
butyl alcohol-d10 at room temperature. For comparison, R-lac-
tam 3 (20 mg, 0.106 mmol) was added to solutions of diethyl-
amine (14 µL, 0.132 mmol) in CD2Cl2, acetonitrile-d3, and
2-propanol-d8 at room temperature. The results are collected
in Table 4.
Ack n ow led gm en t. This work was supported by the
National Science Foundation (CHE 9900402). We thank
Prof. Sergei Smirnov for helpful discussions.
Reaction 3 (0.16 mmol) with diethylamine (0.2 mmol) in 3.0
mL solutions of methanol, ethanol, 2-propanol, and tert-butyl
J O020246H
5294 J . Org. Chem., Vol. 67, No. 15, 2002