Amide bond formation via reversible, carboxylic acid-promoted lactone aminolysis

Article abstract of DOI:10.1021/op1001269

A rapid carboxylic acid-promoted lactone aminolysis is reported. A number of carboxylic acids were found to promote this amide bond-forming transformation, with aliphatic acids being the most efficient. This reaction is an equilibrium process (K_eq

Full text of DOI:10.1021/op1001269

Organic Process Research & Development 2010, 14, 1177–1181
Amide Bond Formation via Reversible, Carboxylic Acid-Promoted Lactone Aminolysis
Megan A. Foley and Timothy F. Jamison*
Department of Chemistry, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, U.S.A.
Abstract:
inexpensive reagents and catalysts, (d) minimal solvent usage,
and (e) a nonchromatographic purification. The closest precedent
to this approach involves prolonged heating of the lactone 2
and amine 4 in the presence of 2-hydroxypyridine (5) and
triethylamine, either neat4 or using tert-butylmethyl ether as a
solvent (Scheme 1).5
A rapid carboxylic acid-promoted lactone aminolysis is reported.
A number of carboxylic acids were found to promote this amide
bond-forming transformation, with aliphatic acids being the most
efficient. This reaction is an equilibrium process (Keq ≈ 1.8), and
mechanistic investigations are consistent with mediation of
a kinetically important proton-transfer step by the car-
boxylate, i.e., the conjugate base of the acid employed.
Scheme 1. Direct, aminolytic lactone ring-opening
Introduction
Aliskiren (1) is the first orally available renin inhibitor
approved by the FDA for the treatment of hypertension (Figure
1). Recently our laboratory has become interested in multistep,
continuous flow chemical synthesis.1 We selected 1 as a test
case for the development of a robust continuous manufacturing
process, beginning with an advanced intermediate and culminat-
ing in isolation of a pure drug substance. The near-term goal
of the project is development of a synthetic route to 1 from
lactone 2 (Scheme 1) with the long-term goal of implementing
the reactions and purifications in a continuous system.
Results and Discussion
Optimization of the Lactone Aminolysis. We initially
investigated several refinements of the literature process. For
example, we screened several different catalysts6 and solvents
and found that 5 was the most efficient and that the reaction
was the fastest in the absence of solvent, but nevertheless
required 6 h to reach 73% conversion at 90 °C. Increasing the
reaction temperature to 120 °C and the stoichiometry of 4 from
3 to 5 equiv accelerated the reaction, resulting in 87%
conversion after 90 min. While this finding was promising,
2-hydroxypyridine is of low water solubility, a trait that would
complicate the purification step for a reaction in a continuous
system.
Figure 1. Aliskiren.
Aliskiren possesses a secondary amide moiety (3) that is
commonly formed via ring-opening of the corresponding lactone
(Scheme 1). Preparation of the amide from lactone 2 often
involves a multistep sequence beginning with hydrolytic lactone
opening to the corresponding acid, followed by protection of
the liberated alcohol and amide bond formation, employing a
peptide-coupling reagent.2 Alternatively, direct lactone ami-
nolysis procedures suffer from long reaction times or require
harsh reaction conditions such as, AlMe3.3 With the ultimate
goal of developing a continuous process, we set the following
as criteria for conversion of 2 to 3: (a) one chemical step, (b)
a reaction time of less than one hour, (c) the use of only
In 2001, Novartis disclosed a general lactone-opening
procedure employing a combination of sodium 2-ethylhexanoate
(Na+ salt of 6) and an ammonium chloride salt as the catalytic
system.7 These conditions seemed promising with regard to
* Author to whom correspondence may be sent. E-mail: tfj@mit.edu.
(1) Bedore, M. W.; Zaborenko, N.; Jensen, K. F.; Jamison, T. F. Org.
Process Res. DeV. 2010, 14, 432.
(4) (a) Ru¨eger, H.; Stutz, S.; Go¨schke, R.; Spindler, F.; Maibaum, J.
Tetrahedron Lett. 2000, 41, 10085. (b) Sandham, D. A.; Taylor, R. F.;
Carey, S. J.; Fa¨ssler, A. Tetrahedron Lett. 2000, 41, 10091. (c)
Dondoni, A.; De Lathauwer, G.; Perrone, D. Tetrahedron Lett. 2001,
42, 4819. (d) Herold, P.; Stutz, S.; Spindler, F. WO/02/02508, 2002.
(e) Dong, H.; Zhang, Z.-L.; Huang, J.-H.; Ma, R.; Chen, S.-H.; Li, G.
Tetrahedron Lett. 2005, 46, 6337. (f) Hanessian, S.; Guesne´, S.;
Che´nard, E. Org. Lett. 2010, 12, 1816.
(2) (a) Kempf, D. J.; de Lara, E.; Stein, H. H.; Cohen, J.; Egan, D. A.;
Plattner, J. J. J. Med. Chem. 1990, 33, 371. (b) Nadin, A.; Lo´pez,
J. M. S.; Neduvelil, J. G.; Thomas, S. R. Tetrahedron 2001, 57, 1861.
(c) Yin, T. I.; Yang, G.; Tarassishin, L.; Li, Y.-M. J. Org. Chem.
2004, 69, 7344. (d) Maibaum, J.; Stutz, S.; Go¨schke, R.; Rigollier,
P.; Yamaguchi, Y.; Cumin, F.; Rahuel, J.; Baum, H.-P.; Cohen, N.-
C.; Schnell, C. R.; Fuhrer, W.; Gruetter, M. G.; Schilling, W.; Wood,
J. M. J. Med. Chem. 2007, 50, 4832.
(5) Sedelmeier, G.; Mickel, J. S.; Ru¨eger, H. WO/2006024501, 2006.
(6) Nucleophilic catalysts examined include DMAP, DABCO, phenol,
thiophenol, PPh3, and 4-pyrrolopyridine.
(3) Lindsay, K. B.; Skrydstrup, T. J. Org. Chem. 2006, 71, 4766.
10.1021/op1001269  2010 American Chemical Society
Published on Web 08/20/2010
Vol. 14, No. 5, 2010 / Organic Process Research & Development
•
1177

purification; the acid and the excess amine might be easily
separated from the product using a simple continuous liquid-
liquid extraction.8 However, use of this procedure for the
opening of lactone 2 with amine 4 resulted in only a trace of 3
(Table 1, entry 1). Additionally, as it would likely clog a continuous
flow reactor; NaCl, a byproduct generated from the ammonium
chloride salt and sodium carboxylate, was a highly undesirable
component. A modified procedure where free amine (4) in the
presence of the carboxylic acid (6) was used, only afforded
moderate improvement, with a the rate of product formation
corresponding to a reaction time far longer than our target of one
hour (Table 1, entries 3 and 4). Altering the stoichiometry of 4
with respect to acid 6 was beneficial, but the reaction outcome
still fell short of our established criteria (Table 1, entry 5). Screening
a number of solvents and mixtures of solvents capable of solu-
blizing both the lactone and the amine did not afford a significant
improvement in either reaction rate or yield.
We also confirmed the necessity of the carboxylic acid as
promoter; extremely low conversion was observed in its
absence, even with extended reaction times (Table 3, entry 1).
While 10% catalyst loading gave good conversion after 3 h
(Table 3, entries 4-6), equilibrium could only be achieved in
the targeted 1 h with use of a full equiv of 6. Amounts of acid
greater than 1 equiv did not significantly improve the rate.10
Thus, conditions shown in entry 2, Table 3 satisfied our criteria
for reaction time, number of chemical steps (1), reagent and
promoter costs, and solvent usage (none). There was some
concern that in the absence of solvent the neat reaction mixture
would be too viscous for conventional pumping mechanisms,
thus hindering the development of a continuous system for this
transformation. However, experimental data have indicated that
the reaction mixture was flowable11 at elevated temperature.
Investigations directed toward realization of a continuous
process for this reaction are ongoing.
Table 1. Initial studies employing 2-ethylhexanoic acid
Table 3. Effect of 2-ethylhexanoic acid on reaction rate
entry
equiv 6
time (min)
conversion (%)
1
2
3
4
5
6
7
8
0
50
50
3
84
17
54
77
84
80
87
1
1 (6 Na+)
0.1
50
50
0.1
120
180
50
0.1
promoter
(equiv)
amine
(equiv)
4 HCl (1.5)
4 HCl (1.5)
4 (1.5)
4 (1.5)
4 (3)
time
(h)
17
6.5
50 min
6
6
conversion
(%)
0.5
entry
T (°C)
0.5
90
1
6 Na+ (2.5)
6 Na+ (2.5)
6 (2.5)
25
70
70
70
70
<5
15
5
23
67
2
3
The general scope of the reaction was examined using a
variety of alkyl amines. 2-Ethylhexanoic acid proved to be a
versatile promoter of the lactone-opening reaction, as moderate
to good yields were observed for each case (Table 4). Moreover,
the remaining lactone could be recovered nearly quantitatively
(Table 4, entry 1). We did not observe any conversion when
tert-butyl amine was subjected to the reaction conditions at 46
°C (bp of amine). However, this observation was not surprising
as we had established that the reaction proceeded best as a melt
at elevated temperatures. Increasing the reaction temperature
to 83 °C (mp of 2 and 4) did not afford product; it is unclear
if this is due to complete vaporization of tert-butyl amine or
due to the steric nature of the amine. Closed continuous systems
with no headspace can offer advantages over traditional batch
reactors when volatile reagents are employed,1 and thus this
reaction could potentially proceed in a continuous flow reactor
at elevated temperature.
4
6 (2.5)
6 (1)
5a
a [2]o ) 1.0 M.
In an effort to maximize both rate and yield we turned to an
examination of solvent-free4 or “melt” conditions. Both lactone
2 and amine 4 have relatively low melting points, and therefore,
we chose 100 °C as the starting point of our investigations
(Table 2, entry 1). Increasing the temperature to 120 °C provided
84% conversion in 50 min (Table 2, entry 3). The conversion
did not change upon extended heating (Table 2, entry 4), and
further experiments demonstrated that the reaction was in fact
an equilibrium process with a Keq of approximately 1.8 (see
Supporting Information). Additional amine was only minimally
effective in shifting the equilibrium toward product (see
Supporting Information for complete table), and thus we carried
out the remainder of our experiments using 5 equiv of 4.
Table 4. Amine scopea
Table 2. Effects of temperature and amine stoichiometrya
entry
T (°C)
equiv 4
time (min)
conversion (%)9
100
110
120
120
130
120
120
120
120
5
5
50
50
71
80
84
83b
80c
89
90
93
94
2
3
4
5
6
7
8
9
5
50
5
5
10
15
20
100
120
50
50
50
120
240
entry
amine
product
3
3a
3b
5c
3d
yieldb (%)
1
2
3
4
5
4
83 (13% recovered 2)c
hexylamine
82
79
76
60
benzylamine
neopentyl amined
cyclohexyl amine
a Reactions were run using 1 equivalent of 2 (250 mg, 0.47 mmol) and 1
equivalent of 6. b Over 20 experiments were conducted using these conditions on
scales from 200 mg to 25 g. The conversions ranged from 81-85%. c Byproduct
formation was observed after 30 min. When the reaction was heated to 140 °C, the
byproduct formation was significant. Several additional products were formed at
higher temperatures (not isolated).
a We have not confirmed that entries 2-5 are an equilibrium process. b Isolated
yields average of two runs. c This reaction was performed on scales up to 25 g of
lactone 2. The conversion and purity as determined by HPLC were in good
agreement with the data for small reactions from which the isolated yield was
determined. d Reactions run at 83 °C.
1178
•
Vol. 14, No. 5, 2010 / Organic Process Research & Development

Although 2-ethylhexanoic acid seemed to be an effective
promoter for lactone aminolysis we nonetheless explored other
acids for this purpose. Given the pKa differences between the
amine and the acid, we expected that the predominant acidic
species in the reaction would be the protonated amine. Thus,
we prepared the 2-ethylhexanoate salt of 4 (7) and found that
it performed as predicted: In a reaction of 1 equiv of 7 and 4
equiv of free amine 4, 85% conversion was observed after 50
min. Surprisingly, the ammonium chloride salt of 4 (8) gave
much poorer results under identical reaction conditions (40%
conversion). We suspected that the hexanoate counterion was
important for the solubility of the ammonium species, whereas
the HCl salt was largely insoluble, resulting in a much slower
reaction.
We screened several other carboxylic acids and found further
support for the importance of reaction homogeneity. In cases
where the reaction was homogeneous, the reactions proceeded
with near identical conversion (Table 5; see Supporting
Information for complete table), but reactions with visible
precipitate were slower (Table 5; entries 1, 6-9), some to the
point of being impractical (Table 5; entry 10).12
Mechanistic Investigations. We extended our study to
include benzoic acid derivatives with a range of pKa values
(Table 6).13 A surprising general trend emerged. A decrease in
the acidity of the promoter employed gave an increase in
reactivity; i. e., pKa and rate are positively correlated. These
results suggest that, in addition to solublizing the ammonium
salt, the conjugate base of the acid catalyst may play an
important role in, or prior to, the rate-determining step of the
reaction.
A plot of log (rateX/rateH) versus the σ14 values for the meta-
and para-substituted benzoic acids affords a line with a small
negative slope (F ) -0.432; R2 ) 0.973) (Chart 1).15 The small,
negative value of F suggests that there is a small increase in
positive charge16 on the carboxylate during the rate-determining
step. In the absence of solvent, a kinetically important proton
transfer facilitated by the carboxylate could account for these
observations.
Deuterium-labeling experiments also support the hypothesis
that proton transfer is kinetically important. A small but
measurable normal isotope effect was found for the reaction
shown in Scheme 2 (rateH/rateD ) 1.4).17
Scheme 2. Deuterium-labeling studies
Table 5. Screen of aliphatic carboxylic acidsa
entry
acid
rxn appearance conversion (%)
1
2
3
4
5
6
7
8
9
10
acetic
visible solid
homogeneous
homogeneous
homogeneous
homogeneous
visible solid
visible solid
visible solid
visible solid
visible solid
75
83
82
85
82
51
35
66
10
7
propanoic
hexanoic
pivalic
cyclopentane
trichloroacetic
perfluorooctanoic
camphoric
malic
Deuterium-labeling experiments employing salts 11-14 and
either amine 4 (salts 11 and 13) or amine 9 (salts 12 and 14)
also exhibited a small normal isotope effect of 1.4 (R1 ) H)
and 1.6 (R1 ) NO2) (Scheme 3). The isotope effect for these
substituents is similar in value to that observed for 2-ethylhex-
anoic acid, and we thus conclude that the reaction proceeds
through a similar mechanism with all three catalysts.
tartaric
a Lactone 2 (250 mg, 0.47 mmol) amine 4 (271 mg, 2.33 mmol), and one
equivalent of the acid were heated to 120 °C for 50 min.
Table 6. Effect of benzoic acid substitution on reaction ratea
entry
benzoic acid substituent
pKa (in DMSO)
R
relative rateb (ratex/rateH)
conversion at 50 min (%)
1
2
3
4
5
6
7
8
3-(CH3)2Nc
3-Me
H
2-OMe
3-OMe
2-Me
3-I
3-Cl
3-Br
3-CF3
3-NO2
4-NO2
2-Cl
5.1
4.24
4.2
-0.211
-0.069
0
1.16d
1.14d
1
70d
71d
69d
59d
68d
72d
53d
55d
54d
51d
49d
42d
45
4.09
4.09
3.91
3.86
3.83
3.81
3.79
3.45
3.44
2.94
2.85
2.85
2.84
1.07
0.74
0.94d
1.12
0.69d
0.65d
0.60d
0.61d
0.53d
0.45
0.49
0.51
0.47
0.115
0.352
0.373
0.391
0.46
0.71
0.778
9
10
11
12
13
14
15
16
17
2-Br
2-I
nicotinic
picolinic
42
44
46
27
0.27
a Lactone 2 (250 mg, 0.47 mmol) amine 4 (271 mg, 2.33 mmol) and one equivalent of the acid were heated to 120 °C for 50 min. b The overall order of the reaction and
rate constants were not determined rate constants. The initial reaction rates used to calculate the relative rates in the table were determined by taking the slope of the line
obtained by plotting conversion versus time for the first 10 minutes of the reaction. The scale of the reaction (0.46 mmol) and the size of the reaction vessel (5-mL
microwave vial) were the same in all cases. c We have not ruled out the possibility that the aniline nitrogen participates in the reaction. d Average of two or more reactions
(conversions were within 2-5% of one another).
Vol. 14, No. 5, 2010 / Organic Process Research & Development
•
1179

Scheme 3. Deuterium-labeling studies
pathways for the acid-catalyzed reaction are presented in Figure
2. Path A involves proton transfer to afford activated lactone
A. Attack by amine 4 gives tetrahedral species B, which
undergoes another proton transfer, giving neutral intermediate
C. Protonation of the ether oxygen (D) generates a good leaving
group, facilitating conversion to 3. Alternatively, under the basic
reaction conditions expulsion of an alkoxide followed by proton
transfer to give 3 is also possible.
One can also arrive at intermediate C via attack on a neutral
lactone by the amine to generate the charge-separated interme-
diate G. A carboxylate mediated proton transfer then affords
C in either a stepwise (H) or concerted fashion. While we
cannot distinguish between these two reaction paths, formation
of intermediate C through Path B is consistent with our data
and with the mechanism proposed by Jencks.19 While inter-
mediate C can undergo ring-opening without protonation, the
reaction is significantly faster when carboxylic acid, rather than
sodium carboxylate, is employed (Table 3, entries 2 and 3),
leading us to propose that lactone opening proceeds via
intermediate D.
Chart 1. Plot of relative reaction rate versus σx
Conclusion
We have developed a rapid method for lactone aminolysis
reactions that will be expected to be amenable for use in a
continuous process. These conditions are particularly effective
for certain hindered, electron-poor amines such as 4. Aliphatic
carboxylic acids were found to be the most effective promoters
for this reaction, which we discovered to be an equilibrium
process. While the reaction is acid-catalyzed, the conjugate base
nonetheless plays a critical role in a kinetically important proton-
transfer step. We found that the electronic properties of the
carboxylate have a significant effect on the reaction rate, opening
the possibility of developing an acid promoter or catalyst that
will afford an even faster reaction.
While this reaction is not directly analogous to aminolysis
reactions carried out in water, we think comparison to these
reactions is instructive. Jencks has demonstrated that aminolysis
reactions involving poor leaving groups, such as alkyl alcohols,
proceed through a water-mediated rate-determining proton-
transfer step from nitrogen to oxygen.18 By analogy, in our
system the carboxylate may play the role of water serving to
transfer a proton in a kinetically important step. If this is the
case we would expect to (and do) see a correlation between
benzoic acid derivatives and reaction rate.
The deuterium-labeling studies and the linear correlation
between reaction rate and the Hammet parameters for substi-
tuted benzoic acids present a strong argument for a kinetically
important proton-transfer step involving the carboxylate. Two
Experimental Section
Materials and Methods. All solvents were either anhydrous,
obtained from a commercially available purification system, or
Figure 2. Possible reaction mechanisms.
1180
•
Vol. 14, No. 5, 2010 / Organic Process Research & Development

ACS reagent grade. Reactions were carried out in oven-dried
glassware. Oxygen was not removed from the reaction atmo-
sphere unless specifically noted. Lactone 2 and amine 4 were
obtained from Novartis and used without further purification.
Benzylamine was distilled before use. Analysis was performed
using reverse phase Agilent 1200 series LC\MS with a Zorbax
SB-C18 column using an acetonitrile and an ion pairing reagent
(1 L water, 5.65 g of hexanesulfonic acid sodium salt mono-
hydrate, 2.75 g of sodium dihydrogen phosphate pH 2.3 adjusted
with phosphoric acid) mobile phase (66:36 acetonitrile to ion
pairing reagent). The product eluted in 3.7 min, and the starting
lactone eluted in 14.2 min). The eluent was monitored at 254
and 280 nm. Preparative chromatographic separations were
performed using Silicycle silica gel (230-400 mesh).
General Procedure for Lactone Opening. Lactone 2 (500
mg, 0.93 mmol) was placed in a 5-mL microwave vial (Biotage)
equipped with a stir bar. Amine 4 (4.7 mmol) and 2-ethylhex-
anoic acid 6 were added and the vial was sealed. The vial was
placed in an oil bath (120 °C except where noted) and heated
for 50 min. The vial was then removed from the bath and
unsealed. EtOAc (2 mL) and H2O (2 mL) were added
immediately, and the solution was allowed to stir and cool to
room temperature over 10 min. The mixture was poured into a
separatory funnel, and the layers were separated. The aqueous
phase was extracted twice with EtOAc (2 × 10 mL). The
combined organic extracts were washed with brine, dried over
MgSO4, and concentrated under reduced pressure. The resulting
oily solid was purified via column chromatography (2:1
hexanes/EtOAc to 1:1 hexanes/EtOAc).
General Procedure for Reaction Screening. A 5-mL
microwave vial was charged with a stir bar. The lactone 2 (250
mg, 0.467 mmol), the amine 4 (271 mg, 2.33 mmol, unless
otherwise indicated in Table 1), and the acid catalyst (1 equiv)
were added, and the reaction was placed in a 120 °C oil bath.
Samples were taken at the indicated time points by inserting a
capillary pipet into the reaction vial and removing an aliquot
of the viscous reaction mixture. In the absence of solvent,
accurate addition of an internal or external standard on small
scale was not reproducible. The reactions were very clean,
showing only product and starting materials. The reaction
progress was measured using the relative areas of the starting
lactone and the amide product peaks on the LC/MS traces.
Isolated yield and conversions were in good agreement.
Acknowledgment
This work was supported by the Novartis-MIT Center for
Continuous Manufacturing. We thank the members of this team
for advice, in particular Gerhard Penn, Berthold Schenkel, Oljan
Repic, Thierry Schlama, Mike Girgis, Lukas Padeste, and Felix
Kollmer. We thank Novartis for the generous donation of
lactone 2 and amine 4. We thank Dr. Randy Ewoldt (Hatso-
poulos Microfuidics Laboratory at MIT) and Patrick Heider
(Jensen Laboratory at MIT) for assistance with viscosity
measurements. We also thank Professor Klavs F. Jensen,
Professor Stephen L. Buchwald and their respective group
members, Dr. Jeffery A. Byers and Dr. Matt Bedore (all MIT)
for insightful discussions.
(7) Liu, W.; Xu, D. D.; Repic, O.; Blacklock, T. J. Tetrahedron Lett. 2001,
42, 2439.
(8) Kralj, J G.; Sahoo, H. R.; Jensen, K. F. Lab Chip 2007, 7, 256.
(9) In reactions where 3 was isolated, conversion (LC/MS) and isolated
yield were in good agreement (see Table 4, entry 1). The excess
amine and acid could be recovered as a mixture from the aqueous
phase in near quantitative yield (composition determined by 1H
NMR in D2O).
(10) Increasing the amount of 6 to either 1.5 or 2 equiv did not significantly
improve the reaction rate. Employing an excess of 6 with respect to
amine 4 led to byproduct formation.
(11) The viscosity at 120 °C was determined to be 77 cP.
(12) We screened several sulfonic acids and mineral acids and found that,
even for the acids that afforded soluble ammonium salts, the reaction
rates were much slower, (+)-CSA, (-)-CSA, and (()-CSA gave the
best results but only reached ∼30% conversion after 2 h.
(13) The pKa values given in Table 6 are for the acids in DMSO. The pKa
values of the acids in the reaction medium are not known.
(14) (a) Jaffe´, H. H. Chem. ReV. 1953, 53, 191. (b) McDaniel, D. H.; Brown,
H. C. J. Org. Chem. 1958, 23, 420.
Supporting Information Available
Experimental procedures, complete tables, and spectral data
for new compounds. This material is available free of charge
via the Internet at http://pubs.acs.org.
(15) The ammonium salts formed from 4-substituted benzoic acids (CF3,
OMe, Cl, and I) were not fully soluble in the reaction mixture, and
the results are not included in Table 6.
(16) The small, negative F value can also be thought of as a decrease in
negative charge.
(17) No equilibrium isotope effect was observed. The rates using 2 with
ND-Boc were the same as when the proteo compound was used.
(18) (a) Satterthwait, A. C.; Jencks, W. P. J. Am. Chem. Soc. 1974, 96,
7018. (b) Barnett, R.; Jencks, W. P. J. Am. Chem. Soc. 1968, 90, 4199.
Received for review May 7, 2010.
OP1001269
Vol. 14, No. 5, 2010 / Organic Process Research & Development
•
1181