A practical synthesis of the dual matrix metalloprotease/tumor necrosis factor inhibitor MMP090

Article abstract of DOI:10.1021/op050066k

A practical 9-step synthesis of the dual matrix metalloprotease/ tumor necrosis factor inhibitor MMP090, trans-(R)-[[N-(4-ethoxyphenylsulfonyl)-N-(4- pyridinylmethyl)]amino]-N-hydroxy-4-propoxy cyclohexaneacetamide 10, was accomplished in 12% overall yield from D-4-hydroxyphenylglycine. Highlights include: (1) selective hydrogenation of D-hydroxyphenylglycine to afford predominantly the trans isomer without racemization; (2) virtually complete removal of the undesired cis diastereoisomer by fractional recrystallization of a TBS ether derivative of the functionalized cyclohexylglycine; (3) direct conversion of the TBS ether into the n-propyl ether in the presence of a catalytic amount of bismuth bromide in high yield; and (4) conversion of the carboxylic acid into the hydroxamic acid using an aqueous solution of hydroxylamine.

Full text of DOI:10.1021/op050066k

Organic Process Research & Development 2005, 9, 608−620
A Practical Synthesis of the Dual Matrix Metalloprotease/Tumor Necrosis
Factor Inhibitor MMP090
Joel S. Slade,* James A.Vivelo, David J. Parker, Joginder Bajwa, Hui Liu, Michael Girgis, David T. Parker,†
Oljan Repi cˇ , and Thomas Blacklock
NoVartis Institutes for Biomedical Research, One Health Plaza, East HanoVer, New Jersey 07936, U.S.A.
Abstract:
throughput and tedious product isolation, as well as reaching
a trans/cis selectivity of only about 3:1. The preparation of
the benzyl ester 6 was also a low-throughput process, and
in addition, the product was contaminated with 3-5% of
dicyclohexylammonium bromide. This contaminant interferes
with the oxidation in the next step such that additional quant-
ities of oxidant were required. The reductive alkylation (Step
7 f 8) could not be scaled up in the pilot plant since it was
considered operationally dangerous due to the high reactant
concentrations, the exothermicity of the reaction, as well as
the high viscosity of the reaction mixture. All attempts to
change the conditions shut down the reaction completely.
In addition, approximately 30% of 7 was lost via the
formation of trifluoroacetate esters. The isolated product 8
was a 3:1 mixture of diastereoisomers (trans predominating)
which required chromatography for purification.
A practical 9-step synthesis of the dual matrix metalloprotease/
tumor necrosis factor inhibitor MMP090, trans-(R)-[[N-(4-
ethoxyphenylsulfonyl)-N-(4-pyridinylmethyl)]amino]-N-hydroxy-
4
-propoxy cyclohexaneacetamide 10, was accomplished in 12%
overall yield from -4-hydroxyphenylglycine. Highlights in-
clude: (1) selective hydrogenation of -hydroxyphenylglycine
to afford predominantly the trans isomer without racemization;
2) virtually complete removal of the undesired cis diastereoi-
D
D
(
somer by fractional recrystallization of a TBS ether derivative
of the functionalized cyclohexylglycine; (3) direct conversion
of the TBS ether into the n-propyl ether in the presence of a
catalytic amount of bismuth bromide in high yield; and (4)
conversion of the carboxylic acid into the hydroxamic acid using
an aqueous solution of hydroxylamine.
In addition, we found that 3-7% of 8 had undergone
racemization due to the strongly acidic reaction conditions.
Multiple recrystallizations were required in order to obtain
acceptable material. The actual yield after purification was
only 22% for this step. For the preparation of 9 we found
that an impurity from the alkylation step (either 4-picolyl
chloride or a polymeric derivative) acted as a catalyst poison
thus requiring additional quantities of catalyst after the
reaction had stalled. Finally, we found that the free base of
MMP090 was not a solid, thus making purification of the
penultimate intermediate extremely difficult. These problems
provided severe limitations to the use of this chemistry for
large scale preparation.
1
. Introduction
Matrix metalloprotease (MMP) inhibitors are currently
being developed for the treatment of rheumatoid arthritis and
osteoarthritis, and they have high clinical attractiveness due
to their ability to slow or even halt the progression of arthritis.
Extensive work on the medicinal aspects of these compounds
has been reported. The preparation of some of the com-
pounds currently in development has also been described.
In this paper we report a practical and convergent synthesis
of the clinical candidate MMP090, a dual MMP/tumor
necrosis factor (TNF) inhibitor recently identified by the
Novartis Drug Discovery group.3
Structurally, MMP090 is a highly functionalized, reduced
hydroxyphenylglycine derivative in which the cyclohexyl
oxygen is alkylated with an n-propyl group and the nitrogen
1
2
After considering several alternative routes and comparing
them with the original synthesis, it was decided that the latter
was still the most attractive for the initial scale-up, providing
the following criteria could be met:
3
is alkylated and sulfonylated. The original 12-step synthesis
of MMP090 was designed to provide a route to a wide
variety of structurally related compounds (Scheme 1). In
assessing the viability of this route for scale-up purposes,
several problems became apparent. In the preparation of the
sulfonyl chloride 2, the potential existed for the buildup of
thionyl chloride in the reaction mixture as the scale increases,
(
i) the selectivity of the Raney nickel reduction (3 f 4)
should remain unchanged or improve on larger scale;
ii) conditions would need to be identified for the
O-alkylation of 6 which are safe and efficient;
iii) the chromatographic purification to remove the cis
isomer should be eliminated;
iv) an effective method for the preparation of the
(
(
resulting in the subsequent rapid evolution of HCl and SO
The Raney nickel reduction of 3 was hampered by low
2
.
(
hydroxamic acid could be identified.
*
Corresponding author. Telephone: 862-778-3479. Fax: 973-781-2188.
Scheme 2 summarizes our efforts to improve the synthesis
and to circumvent the major problems. Specifically, by
combining good selectivity in the Raney nickel reduction
with fractional crystallization of an advanced intermediate,
we were able to eliminate the undesired cis isomer from the
final product. In addition, we developed a new catalytic
E-mail: joel.slade@pharma.novartis.com.
†
New address: 400 Technology Square, Cambridge, MA 02139, USA.
(
1) Leung, D.; Abbenante, G.; Fairlie, D. P. J. Med. Chem. 2000, 43, 305.
Hagmann, W. K.; Larkk, M. W.; Becker, J. W. Annu. Rep. Med. Chem.
1
996, 31, 231.
(
(
2) Wojtowicz-Praga, S. Drugs in R&D 1999, 1, 117.
3) Parker, D. T. (Novartis) U.S. Patent 5,770,624, 1998.
6
08
•
Vol. 9, No. 5, 2005 / Organic Process Research & Development
10.1021/op050066k CCC: $30.25 © 2005 American Chemical Society
Published on Web 08/27/2005

Scheme 1. Original 12-step synthesis of MMP090
reaction.3 The maximum trans/cis ratio of 4 found under
-5
procedure for the reductive alkylation of secondary alcohols
which allowed for an efficient preparation of ether 8.
Our efforts, which culminated in the successful realization
of the above strategy for the preparation of MMP090, are
outlined below.
6
these conditions was 3.2/1 (no information with regard to
the trans/cis ratio of the cyclohexylglycine is given in refs 4
and 5).
Although the conversion of the starting material was
>
95%, analysis of the crude reaction mixture by NMR
2. Results and Discussion
revealed that hydrogenolysis of the hydroxyl group had also
taken place. Cleavage of the amino group may also have
5
(R)-R-[N-(4-Ethoxyphenylsulfonyl)amino]-4-hydroxy-
7
,8
occurred under these conditions. In addition, there was
concern regarding the partial racemization of the substrate
during the hydrogenation, since this had been observed under
certain conditions in the case of (R)-phenylglycine.7
Initially, our efforts were directed towards the separation
of the trans/cis mixture of reduction products 4 since it had
cyclohexaneacetic Acid (5). The sodium salt of the sulfonic
acid was O-alkylated using diethyl sulfate in a basic aqueous
medium. The product (isolated as the sodium salt) was
converted to the sulfonyl chloride 2 using a mixture of
thionyl chloride/DMF in toluene. The original process was
modified in that the sulfonyl chloride was added slowly to
a hot (78 °C) solution of the substrate in toluene/DMF. This
addition-controlled procedure allowed for moderation of the
gas evolution as well as the heat of the reaction. Although
crystalline, it was easier to handle compound 2 as a toluene
solution. The overall yield for the two steps was 65%.
We next turned our attention to the reduction of D-
hydroxyphenylglycine 1. The original hydrogenation condi-
tions involved the use of excess Raney Nickel in aqueous
base at elevated temperature (50-70 °C) to effect the
(
4) Blaha, K.; Farag, A. M.; Van Der Helm, D.; Hossain, N. B.; Budesinsky,
M.; Malon, P.; Smolikova, J.; Tichy, M. Collect. Czech. Chem. Commun.
1984, 49, 712.
(5) Edelson, J.; Fissekis, J. D.; Skinner, C. J.; Shive, W. J. Am. Chem. Soc.
1958, 80, 2698.
(
6) Since it proved difficult to isolate 4 from the aqueous solution for analysis,
an NMR method was developed to determine the trans/cis ratio. This
involved the comparative integration of the cyclohexyl ring proton
-CH(OH) (see Experimental Section).
(
7) Tamura, M.; Harada, K. Synth. Commun. 1978, 8, 345.
8) Minnaard, A. J.; Boesten, W. H. J.; Zeegers, H. J. M. Synth. Commun. 1999,
29, 4327.
(
Vol. 9, No. 5, 2005 / Organic Process Research & Development
•
609

Scheme 2. Revised synthesis of MMP090
been demonstrated in the literature that it was possible to
accomplish this on the corresponding Boc benzyl ester 11.
was true regardless of the catalyst system used (Raney nickel,
Rh/alumina, or Pt/C).
9
While this investigation was proceeding, it was discovered
that the trans/cis isomer ratio could be improved to >120/1
in a later step by fractional recrystallization of 6a (see below).
Thus, it was decided to convert the mixture of the diaste-
reoisomers 4 to the sulfonamides 5 and carry out the isomer
separation at the later stage. Compound 5 was prepared by
treating the aqueous solution of 4 with the sulfonyl chloride
2
in a mixture of triethylamine/acetonitrile. The reaction mix-
ture was concentrated and extracted with ethyl acetate to
remove organic impurities, and the product 5 was precipitated
from the aqueous phase by acidification. The trans/cis ratio
at this point was approximately 4/1.
With this result in hand, we focused our attention on
improving the trans/cis ratio to the greatest extent practicable
for the hydrogenation of 3 and also determining the amount
of racemization (if any) during this process.
Thus a variety of functionalized derivatives of 3 were
prepared and subjected to hydrogenation using several
different catalysts and conditions. We found that substitution
at the carboxyl group (i.e., esters) slowed the rate of reaction
and only afforded a 2:1 trans/cis mixture of diastereoisomers
regardless of the substituents present on the nitrogen and
oxygen functionalities. Furthermore, the same was found to
be the case for the N-CBZ and O-alkylated ethers. This
A review of the literature revealed that although less
reactive than the common hydrogenation catalysts (e.g., Pd,
Pt, Rh, and Ru) for the reduction of carbocyclic aromatic
(9) Banfi, A.; Benedini, F.; Sala, A. Synth. Commun. 1990, 20, 3585.
610
•
Vol. 9, No. 5, 2005 / Organic Process Research & Development

Table 1. Step 3 f 4: screening of reaction conditionsa
conversion of starting material (85% at 20 psi vs 95% at
0 psi).
(5) Higher temperatures caused some racemization to
occur.
5
H
2
base
(equiv)
press. temp time conv
b
ref catalyst
(psi)
(°C)
(h) (%) cis/trans
(
6) The source of the Raney nickel affected both the %
-
31 RaNi
35 RaNi
38 RaNi
43 RaNi
44 RaNi
47 RaNi
52 RaNi
58 RaNi
59 RaNi
60 RaNi
NaOH (2)
NaOH (1)
50 40
56 50-70 16
50 55
20
41.5 1/2
1
3
-
-
-
-
-
-
-
-
-
98.4 1/3.2
98.4 1/2.7
96.5 1/3.8
conversion as well as the trans/cis ratio.
K CO
2
3
(1)
15
22
22
20
18
19
18
4
We found that trans/cis ratios as high as 7:1 could be
obtained at temperatures between 70 and 90 °C and hydrogen
pressures of 20-30 psi. However, it was also discovered
that this ratio was dependent on the source of the Raney
nickel catalyst. Different lots from the same supplier gave
ratios that varied from 3.5:1 to 7:1. In fact, we observed
that most of the catalyst types that gave the highest trans/cis
ratios also gave the most variable results and the most
racemization. After extensive testing of 10 different sample
lots of catalyst from one supplier, a lot was chosen which
gave a high trans/cis ratio with little or no racemization.
However, when the full supply of catalyst was received and
tested on a large scale, this same lot gave only a 3:1 trans/
cis ratio of diastereoisomers and in addition caused extensive
racemization (16% L-isomer detected by chiral HPLC).
Therefore, a catalyst type was chosen which had given a
lower ratio but had been the most consistent (3.5:1 trans/
NaOH (0.8) 22 55
NaOH (0.8) 55 55
NaOH (0.8) 22 75
NaOH (0.8) 10 80
NaOH (0.8) 17 90
NaOH (1)
NaOH (1)
>99
1/3.7
1/4.3
1/5
1/6
1/5.5
94
90
93
94
17 90
17 100
48
70
15
61
1/5
-62 RaNi
-63 RaNi
-67 RaNi
-79 RaNi
NaOH (0.8) 27 56-87 21
98.6 1/3.7
NaOH (0.8) 30 92
NaOH (0.8) 50 90
NaOH (0.9) 20 90
15
20
4
14
8
72
72
20
48
22
93
90
83
1/6.1
1/5
1/6.6
65
-39 Rh/Al
-40 Rh/Al
-48 Rh/Al
-64 Rh/Al
-50 Pt/C
O
2 3
O
2 3
O
2 3
O
2 3
K
2
CO
NaOH (0.8) 50 58
aq HCl (1) 52 45
3
(1)
56 55
44
>99
>99
56
1/1
1/1.1
2/1
1/1.6
1/3.2
NaOH (1.0) 50 100
NaOH (0.8) 50 28
40
58
-
-
53 Pt/C
55 Pt/C
aq HCl (1)
NaOH (1.1) 50 28
50 52
5.5
48
16
87c
14
1/2.1
1/3.2
58
1
4
cis, 4-5% racemization). This catalyst performed well on
scale-up in the pilot plant, with the expected results being
obtained in most cases.
a
Note: All reactions were carried out in water. b Determined by NMR. c Large
1
5
amount of impurities formed.
Although we were somewhat perplexed at first with regard
to the cause of the racemization which occurred in two of
the batches, we were able to explain these results once the
conditions under which the racemization had occurred were
examined. In these two pilot plant runs, the L-isomer content
of 4 was 17.7% and 29.4%, respectively.
rings, Raney nickel would be most likely to afford product
mixtures in which the thermodynamically favored isomer
(
i.e., 1,4-trans) would predominate.10,11 Thus, a wide variety
of conditions were screened to optimize the formation of
the trans isomer using this catalyst as well as several others.
Employing a standard loading of a 2:1 w/w ratio of catalyst
and substrate, the following reaction parameters were
studied: amount of base, nature of the base, temperature,
time, and hydrogen pressure. The results are summarized in
Table 1.
Since Raney nickel appeared to provide the best results
in terms of conversion and trans/cis ratio, several different
lots were selected for further study. The results are presented
in Table 2.
Upon examination of the conditions under which these
two reductions were carried out, we found that, in both cases,
the substrate was in contact with the catalyst for an extended
period of time in the absence of a hydrogen atmosphere. This
was due to some operational difficulties which had devel-
oped. Based upon these results, we felt that the high degree
of racemization might be explained in the following manner.
In the first case (17.7% of the trans L-isomer), we learned
that the catalyst and the compound 3 solution had been
charged to the reactor but that the vessel was not pressurized
with hydrogen for several hours. Since Raney nickel was
present for an extended period without the hydrogen
atmosphere (oxygen present), we believe that racemization
occurred through the redox process shown below. The
existence of this intermediate could lead to racemic product
Based upon this large number of experiments, the fol-
lowing information was obtained:
(
1) The reaction temperature must be at least 50 °C to
achieve >90% conversion.
2) Sodium hydroxide (0.8-2.0 equiv based on 3) gave
the most consistent results in terms of the trans/cis ratio of
(
4.
16
upon hydrogenation.
(
3) The highest trans/cis ratios were obtained at higher
Another piece of evidence supporting this hypothesis
comes from the second case mentioned above wherein the
isolated product was found to contain 29.4% of the undesired
L-isomer. In this instance, the hydrogenator was filled up to
12
temperatures and lower hydrogen pressures.
4) Although higher temperatures afforded higher trans/
cis ratios, the lower hydrogen pressures used resulted in lower
(
(
(
10) Augustine, R. L. Catalytic Hydrogenation; Marcel Dekker: New York, 1965.
11) Rh/Al afforded 2:1 trans/cis mixtures under both acidic and basic
O
2 3
(13) The activity grade was equivalent to W-2.
conditions over a wide range of temperatures. Pt/C catalysis resulted in
incomplete conversion as well as affording significant amounts of hydro-
genolysis products (see ref 5). No reaction was observed with either Pd or
Ru catalysts. The trans/cis ratios were monitored by NMR (see ref 6).
12) The highest trans/cis ratio obtained was 7:1 (70 °C, 20-25 psi hydrogen,
(14) The presence of small amounts of the D-cis isomer ceased to be of concern
when it was learned that this impurity could be removed in a later step (see
text).
(15) A total of 15 hydrogenations were carried out with good consistency among
13 of these (see text).
(
2
2 h).
(16) In all cases, less than 1% of the L-cis isomer was formed.
Vol. 9, No. 5, 2005 / Organic Process Research & Development
•
611

Table 2. Step 3 f 4: screening of reaction conditions with Raney Nickel catalysta
4
temp
(°C)
time
(h)
unreduced
(%)
ref
catalyst
Aldrich
R-trans
S-trans
R-cis
S-cis
1
1
9804-2
9807-2
60
60
7
8
2.1
2.4
72.3
72.7
4.4
5.5
21.9
21.2
0.5
0.6
PMC
A-5000
PMC
A-5B00
PMC
A-5300
PMC
A-5300
PMC
1
1
1
1
1
9806-2
9805-2
9808-2
9809-2
9811-2
60
60
60
65
60
11
14
21
11
11
1.2
3.7
2.4
5.1
72.7
68.7
70.9
b
9.3
14.5
9.9
17.2
15.7
18.4
0.9
1.1
0.9
A-5300
PMC
71.2
6.6
21.4
0.8
A-5300
(different lot)
a
Note: The trans/cis ratios were determined by NMR analysis of the solution from the reduction step (compound 3). The analysis for enantiomeric purity was
carried out on compound 4 using an HPLC method developed in our analytical group. All reactions were carried out using 1 equiv of sodium hydroxide and 40 psi
of hydrogen. b The optical purity was not measured. The trans/cis ratio was 6.1/1 for compound 3.
the full capacity of the vessel. Thus, there was little or no
atmosphere of hydrogen above the level of the liquid, and
agitation of the batch was poor. It was also observed that
the rate of the reduction was considerably slower. These
conditions would favor the redox equilibrium shown below
with an intermediate quinine-methide 12, with the result
being racemization. Subsequently, the process conditions
the methine proton on the cylcohexyl carbon bonded to the
hydroxyl group. Fortunately, the chemical shifts for this
proton in the cis and trans isomers were different enough
that one could obtain a rather accurate isomer ratio based
on the integration. However, in the long run, a more
quantitative method (e.g., HPLC) will be required.
Initially, it was decided that it would be advantageous to
isolate 4 as a solid intermediate for the purposes of
purification. In addition, we hoped that further enrichment
of the trans/cis mixture could be effected at the same time.
Thus, an aqueous solution containing 4 was acidified under
a variety of conditions, and the amino acid products were
analyzed by HPLC for content. Acidification of an aqueous
solution of the sodium salt 4 with 0.1 N HCl (0.43 equiv
based on theory) afforded the corresponding acid in an
overall yield of 33% with a trans/cis ratio of 18/1.
At this point, we decided to investigate this reaction
further and to determine more precisely the composition of
the hydrogenation mixture. Thus, a sample of the hydrogena-
tion solution (after filtration of the catalyst) was analyzed
by LC/MS. The results are represented in Figure 1 with
tentative structures assigned based upon the observed mo-
lecular weights (note: these impurities account for 10-15%
of the total mixture).
were modified in that the reaction temperature was lowered
slightly and full hydrogen pressure was applied rapidly after
charging and maintained throughout the reaction. The last
five batches prepared under these conditions all contained
less than 5% of the L-isomer (combined cis and trans).
In an effort to further substantiate the oxidative racem-
ization hypothesis, samples of D-4-hydroxyphenylglycine and
D-phenylglycine were subjected to the identical reaction
conditions in the absence of hydrogen. It was assumed that
if the mechanism for the racemization involved the quinone-
methide-like structure 12, then no racemization would occur
with the D-phenylglycine.
In the event, it was found that both substrates underwent
a small degree of racemization (2-5%) when exposed to
aqueous sodium hydroxide and Raney nickel in the absence
of hydrogen. Thus, the explanation for this phenomenon is
still unclear at this time. However, the possibility cannot be
ruled out that two different mechanisms are operating here.
It was found that the course of the reaction could be
followed most easily using NMR. One could see the
disappearance of the aromatic protons as the reduction
progressed and, at the same time, view the appearance of
With the formation of such a multitude of products, we
decided that perhaps the most expeditious course would be
to react the sodium salt 4 with the sulfonyl chloride 2 and
isolate the sulfonamide 5, which could be purified further if
necessary. The reaction was carried out in the presence of
acetonitrile and triethylamine at 0 °C, and the product 5 was
isolated in an overall yield of 60% (based on 3) with the
following results:
(
i) 98.4% purity by HPLC area normalization (cis and
trans);
(ii) 3.8/1 trans/cis ratio (D-isomers);
(iii) 3.4% trans L-isomer;
(iv) 0.5% cis L-isomer.
612
•
Vol. 9, No. 5, 2005 / Organic Process Research & Development

Figure 1.
We found that this mixture was suitable for use in the
date 14 of the alcohol 13 as the only product. At the time
subsequent steps, and therefore, no further purification of
this compound was effected.
Phenylmethyl trans-R-[N-(4-Ethoxyphenylsulfonyl)-
amino]-4-propoxycyclohexaneacetate (8). As mentioned in
the Introduction, the formation of the n-propyl ether was the
most problematic aspect of the synthesis. Not only was the
compound a secondary alcohol but also, in addition, there
were base sensitive functionalities present in the molecule,
such as the R-amino ester. Using typical Williamson ether
1
7
synthesis conditions (e.g., NaH, n-propyl bromide, or allyl
bromide) lead only to complex mixtures. The palladium
18
catalyzed reaction of 6 with diallyl carbonate afforded the
N-allylated compound 13 as the sole product.
this work was in progress, a publication appeared in which
it was shown that the reductive heterocoupling of a carbonyl
compound with an alkoxy silane could be smoothly effected
with triethylsilane in the presence of bismuth bromide
2
0
(eq 1).
This approach appeared to be applicable to our situation.
The TMS ether of compound 6 was prepared and reacted
with propionaldehyde and triethylsilane in the presence of a
catalytic amount of bismuth bromide under conditions similar
to those described by Komatsu. We were gratified to find
that the desired ether 8 was formed in 70% yield along with
Reacting 6 with allyl 2,2,2-trichloroacetimidate under
Lewis acid catalyzed conditions gave the trichloroacetimi-
19
(
17) For a review, see: Feuer, H.; Hong, J. The Chemistry of the Ether Linkage;
20
John Wiley & Sons: New York, 1969; pp 446, 460.
18) Lakhmiri, R.; Lhoste, P.; Sinou, D. Tetrahedron Lett. 1989, 30, 4669.
19) Wessel, H.-P.; Iverson, T.; Bundle, D. R. J. Chem. Soc., Perkin Trans. 1
(
(
1
985, 2247.
(20) Komatsu, N.; Ishida, J.; Suzuki, H. Tetrahedron Lett. 1997, 38, 7219.
Vol. 9, No. 5, 2005 / Organic Process Research & Development
•
613

1
5% of alcohol 6. The starting alcohol was the result of
period was observed followed by a rise in the temperature
of the mixture. To maintain better control, we decided to
add the propionaldehyde to the mixture of the TBDMS ether,
triethylsilane, and bismuth bromide in the acetonitrile. By
monitoring the rate of addition, the exothermic nature of the
hydrolysis of the TMS ether due to the presence of trace
21
amounts of water in the acetonitrile. The use of the more
stable TBDMS ether 6a afforded the n-propyl ether 8 in 88%
yield. The generality of this approach to ether formation was
investigated further with the following observations and
conclusions.22
Over the course of the reaction, a gray solid precipitate
formed. By elemental analysis this was shown to be bismuth
metal. In a separate experiment, it was demonstrated that
the catalytic activity resided with the filtrate and not with
the metal. When the interaction of triethylsilane and bismuth
bromide was monitored by NMR, it was found that the signal
at δ 0.61 ppm for the methylene protons in the triethylsilane
changed from a multiplet to a simple quartet upon addition
of bismuth bromide. Furthermore, the multiplet for the -SiH
at δ 3.60 ppm disappeared while a new signal appeared at δ
2
4
reaction could be controlled.
Based on the experience from several experiments, we
learned that the water content of the solvent (acetonitrile)
and the reagents had a significant influence on the yield of
the reaction, due to the hydrolysis of the TBDMS ether in
the presence of water and bismuth bromide. Extra precautions
were taken to ensure the anhydrous nature of the reaction.
Thus, using the conditions mentioned above, the product 8
was isolated after an aqueous workup by precipitation from
ethyl acetate/heptane in 85% yield and greater than 99%
purity (HPLC, area normalization). The product was found
to contain 0.24% of the L-trans-isomer of 8, and less than
0.5% of the D-cis-isomer.
10.36 ppm. All of these results are consistent with the
formation of hydrogen bromide, bismuth metal, and trieth-
ylsilyl bromide as the products of this reaction. The hydrogen
bromide goes on to react with the acetonitrile to form
acetimidoyl bromide.23 This adduct was ineffective in
catalyzing O-alkylations. On the other hand, commercial
triethylsilyl bromide when used in catalytic amounts together
Phenylmethyl trans-R-[[N-(4-Ethoxyphenylsulfonyl)-N-
(4-pyridinylmethyl)]amino]-4-propoxy-cyclohexaneace-
tate (8a). In addition to the procedure used in the Research
synthesis, two other sets of reaction conditions were inves-
tigated, i.e., potassium t-amylate in toluene, sodium hydride
in THF. However, neither of these provided useful amounts
of product. Thus, it was decided to optimize the potassium
carbonate/DMF process for scale-up.
with triethylsilane worked as effectively as the BiBr
SiH combination in the reductive etherification of a variety
3 3
/Et -
2
2
of silyl ethers. This served further to confirm our conclu-
sion, based on the NMR experiment, that the triethylsilyl
The main issues which needed to be addressed were the
following:
bromide formed in situ from BiBr
catalyst in the reaction. For our scale-up work, we found it
more convenient to use the BiBr /Et SiH combination to
generate the triethylsilyl bromide due to the difficulties in
handling the latter (hygroscopicity).
3
/Et
3
SiH is the actual
(i) since the yield was low, optimization studies were
needed in order to increase the conversion in the alkylation
step and to improve the workup to provide for a better fit in
the pilot plant (reaction time, concentration, stoichiometry,
etc.);
3
3
One unexpected benefit of using the TBDMS ether 6a
was that it allowed for the virtually complete removal of
the undesired cis isomer. Thus, the reaction was carried out
using TBDMSCl and imidazole in DMF followed by an
aqueous workup and extraction of the product into methyl
tert-butyl ether. The desired product 6a was crystrallized by
the addition of heptane to the partially concentrated solution
(ii) it appeared that the alkylation product (8a) was not a
solid, nor was the HCl salt (isolation issues);
(iii) if 8a could not be isolated, a workup procedure would
be needed in order to ensure that the crude compound would
undergo the debenzylation successfully and that 9 could be
isolated and purified further, if necessary.
(the pure cis isomer was isolated from the mother liquors
by chromatography and was found to be an oil). Starting
with a 3.5/1 trans/cis mixture of 6, 6a was isolated in 68%
yield (87% efficiency based on the amount of trans isomer
present in 6) and then with a 121/1 trans/cis ratio based upon
the HPLC and NMR analysis. At this point, the enantiomeric
purity (D-6a, trans + cis) was 96%.
For the scale-up of the succeeding step (formation of the
n-propyl ether 8), there were two issues which needed to be
addressed. These were the exothermic nature of the reaction
and the water content of the starting materials. On a small
scale, the catalyst (bismuth bromide) was added to the
reaction mixture last. Under these conditions, an induction
Fortunately, we were able to begin our efforts starting
from the procedure which had been used for the first scale-
up in the kilo lab. This involved the use of potassium
(
21) Alkyl TBDMS ethers can be hydrolyzed in wet acetonitrile in the presence
of a catalytic amount of bismuth bromide: Bajwa, J. S.; Vivelo, J.; Slade,
J.; Repi cˇ , O.; Blacklock, T. Tetrahedron Lett. 2000, 41, 6021.
22) Bajwa, J. S.; Jiang, X.; Slade, J.; Prasad, K.; Repi cˇ , O.; Blacklock, T.
Tetrahedron Lett. 2002, 43, 6709.
(
(
(24) As measured in an RC-1 calorimeter, the adiabatic temperature rise was
23) Janz, G. L.; Danyluk, S. S. J. Am. Chem. Soc. 1959, 81, 3850.
18 °C.
614
•
Vol. 9, No. 5, 2005 / Organic Process Research & Development

Table 3. Step 8 f 8a: optimization experimentsa
Table 4. Step 8a f 9: initial optimization experiments
c
HPLC (%)
8a cis-8a Sol
time yield HPLC recryst.
HPLC
recovery (%)
mL of DMF/ time extract. yieldb
ref solvent (h) (%)
(%)
solv.
ref
g of 8
(h) solvent (%)
8
-
-
-
121 HOAc 16
128 EtOAc 16
45
80
86.8 2-propanol
MeOH
98.0 MeOH
56
55
72
96.7
>99
>99
-106
-112
-115
-124c
10/1
10/1
2.6/1
2.1/1
16 EtOAc 87 93.2 0.9 5.9 trace
42 toluene 85 92.2 0.9
68 toluene 80 98.8 1.2
22 toluene 89 95.5 1.0
0
0
0
6.9
trace
3.5
133 HOAc
3
The final series of experiments included the following:
(i) The maximum reaction volume was further reduced
a
Note: All reactions were conducted at room temperature using the
stoichiometry described above. Corrected for solvent content. c Percentages
based on area normalization.
b
by using the extraction solvent (toluene) rather than DMF
to wash the potassium carbonate filter cake. This allowed
for a proportional reduction in the amounts of citric acid
and toluene. Overall, these changes reduced the maximum
volume by 50%.
carbonate along with a catalytic amount of TDA-1.25 It had
been found that, without the TDA-1, the reaction was
considerably slower and that 1-2% racemization occurred.
Under these conditions, the reaction time was reduced by
(ii) The slow filtration to remove the potassium carbonate
was improved by adding Celite to the reaction mixture.
Initially, we had planned to crystallize the benzyl ester
8a as either the free base or the HCl salt. Thus, a sample
was purified by preparative HPLC to provide seed crystals
for the crystallization. However, even using seeding tech-
niques, the free base could not be crystallized probably due
to its high solubility in all the solvents tested with the
exception of water and heptane. In addition, it proved to be
extremely difficult to crystallize the HCl salt, even starting
with a sample of the free base that had been purified by
chromatography. The best that could be obtained was a
solidified foam. Attempts at preparing other salts were
equally unsuccessful.
trans-R-[[N-(4-Ethoxyphenylsulfonyl)-N-(4-pyridinyl-
methyl)]amino]-4-propoxycyclohexaneacetic Acid (9). Since
the isolation of the alkylated sulfonamide (8a) proved to be
difficult, it was decided to convert it directly to the carboxylic
acid by palladium-catalyzed hydrogenolysis. Initially, the
following conditions were employed. The toluene solution
of the 8a was filtered through alumina to remove polar
impurities and to dry the solution. Next, the solvent was
exchanged for acetic acid (final ratio ) 10:1 mL/g 8a). Acetic
acid was used because we had observed that compound 9
was not soluble in most organic solvents and would
precipitate during the hydrogenolysis, making it difficult to
separate from the catalyst. The hydrogenolysis was carried
out at 50 psi using 10% Pd/C, 50% water wet (a 10% loading
based on the theoretical amount of 8a present was used).
The reaction was monitored by HPLC and was found to be
complete in 1-3 h. The product 9 was isolated by filtration,
concentration of the solution, and addition of water. The
crude product was then recrystallized. With this as a starting
point, a series of experiments was carried out in order to
further define the reaction conditions as well as the crystal-
lization and recrystallization procedures. The results are
summarized in Table 4.
5
0% and the amount of racemization was <0.3%. One other
problem identified at this time involved the presence of an
impurity in the crude product which acted as a powerful
catalyst poison in the next step. This was believed to be
picolyl chloride or a homopolymer which was derived from
the excess picolyl chloride hydrochloride after neutralization.
The use of an acidic workup did not completely eliminate
this impurity.
Initially, the reaction was carried out in the following
manner: compound 8, picolyl chloride hydrochloride (1.5
equiv), powdered potassium carbonate (10 equiv), and
TDA-1 (10 mol %) were stirred in DMF at room temperature
overnight. The workup involved filtration to remove the
solids and addition to 10% aqueous citric acid followed by
extraction into ethyl acetate. The extracts were washed with
6
N HCl (to remove the some of the excess picolyl chloride,
DMF, and TDA-1) and water. Under these conditions,
approximately 6% of 8 was present in the crude product. It
is interesting to note that, based on an HPLC analysis of the
acid washes, none of the product 8a was extracted despite
the fact that a basic pyridine ring was present in the molecule.
This procedure was used as the basis for the following
investigation. The results are summarized in Table 3.
In experiment -106, the ethyl acetate solution was washed
with water until the washes were neutral to pH paper. That
the reaction was incomplete and a large volume of waste
was generated were the main concerns. In experiment -112,
the reaction time was extended to ensure complete conversion
of 8. Monitoring was done by HPLC. Toluene was used as
the extraction solvent since ethyl acetate had given poor
phase separations. An alumina treatment was added to
remove colored polar impurities. One other benefit of using
toluene was that less water washes were required. To further
improve the throughput, experiment -115 was carried out at
about 4 times the concentration. This helped further reduce
the volume of waste generated. Extending the reaction time
from 42 to 68 h did not have a significant effect on the yield
or purity of 8a. Increasing the concentration further (experi-
ment -124c) and frequent monitoring indicated that the
reaction was actually complete in only 22 h.
For the first experiment in the table, isolated 8a (gummy
solid) was used in order to become familiar with the process.
The product was isolated by concentrating the solution (after
filtration of the catalyst) and triturating with water. In one
instance we attempted to hydrogenolyze the ethyl acetate
solution of 8a as it was obtained from the previous reaction
(experiment -128). Due to the insolubility of 9 in ethyl
(
25) Tris[2-(2-methoxyethoxy)ethyl]amine (TDA-1) has been shown to be a good
solid-liquid-phase transfer catalyst: Soula, J. J. Org. Chem. 1985, 50, 3717.
Vol. 9, No. 5, 2005 / Organic Process Research & Development
•
615

Table 5. Step 8a f 9: further optimization experiments
This was done to further confirm the stability of the product
in acetic acid since it was our intention to accumulate the
solution from several batches and combine them for the
crystallization.
yielda
(%)
HPLC
(%)
Pd
(ppm)
ref
objective
-
-
162
176
improve filterability of 9
scale-up using catalyst
from pilot plant
80
74
98.3
98.7
71
26
As was mentioned previously, the filtration of the acetic
acid solution of 9 was problematic due to the poor physical
characteristics of the solid. The filterability was greatly
improved by pouring the concentrated acetic acid solution
into a dilute aqueous solution of sodium chloride. The
product obtained was granular in nature and easily filtered
(experiment -162). Experiment -176 involved a use test of
the actual lot of catalyst we intended to employ and also
provided us with an opportunity to scale-up the reaction in
the laboratory prior to implementation in the pilot plant.
Although 9 was not the final drug substance, it was
decided to control the palladium content at this point since
we wished to avoid extensive purification in the last step, if
possible. To ensure low palladium content (<2 ppm in the
final drug substance), the following strategy was adopted.
Filter aid was added before the catalyst filtration, the
suspension was filtered through a wet-packed pad of filter
aid, and the isolated product was further purified by
recrystallization which included an adsorbent.
A series of experiments was carried out in which differ-
ent recrystallization solvents as well as different adsorbents
were used in order to effect the final purification and remove
the last traces of palladium. The results are summarized in
Table 6.
Based on these results, it was decided to use methanol
and the carbon PICA1400 (experiment -31) since the
recovery was reasonably good and the palladium content was
within acceptable limits. Thus, the optimized process in-
volved dissolving the substrate in methanol (25 mL/g)
containing 10% of PICA1400 at reflux. The suspension was
filtered while hot, and the filtrate concentrated to ap-
proximately 50% of the original volume at which point 9
began to crystallize. When this purification method was
scaled up in the laboratory (150 g), the same results were
obtained except that the recovery was better (88%) due to a
longer hold time during the crystallization.
-189
reduce Pd levels
75
>99
20
a
Yields are calculated based on 8.
acetate, the isolation was extremely difficult. However, it
was possible to recrystallize the crude product to high purity.
Obviously, neither of these procedures could be scaled up.
Thus, in the third experiment (-133), acetic acid was used
and the reaction was monitored carefully for completeness.
After 3 h, the catalyst was removed, and the solution was
concentrated to approximately 20% of the initial volume and
then added carefully to water with seeding. This afforded a
good yield of the desired product which after recrystallization
gave 9 in high purity.
One other major concern with the hydrogenolysis step
was the concomitant reduction of the pyridine ring. Depend-
ing on the conditions used, a significant amount of the
piperidine analogue was formed.26 Fortunately, the rate of
debenzylation was quite a bit faster than the rate of ring
reduction.27 In a control experiment, compound 9 was
subjected to the conditions of the hydrogenolysis for 72 h.
Analysis of the reaction mixture by HPLC indicated that
about 75% of the starting material had been converted to
the overreduced side product at this point (an authentic
sample was prepared for comparison). Therefore, the reaction
was carefully monitored for the disappearance of the starting
material. This served to preclude the formation of the
overreduced side product.
With these results in hand, we next sought to further
optimize the reaction conditions in preparation for scale-up.
Specifically, the following items were addressed:
(
i) demonstrate that the solvent exchange from toluene
8a) to acetic acid is feasible with regard to the stability of
the compound and ease of operation;
ii) further develop the isolation of crude 9 (addition of
(
(
trans-R-[[N-(4-Ethoxyphenylsulfonyl)-N-(4-pyridinyl-
methyl)]amino]-N-hydroxy-4-propoxy-cyclohexaneaceta-
mide Hydrochloride (10, MMP090). With an efficient
synthesis of 9 in hand, all that was needed to complete the
synthesis was to form the hydroxamic acid. Most importantly,
we needed to carry out this transformation in such a way as
to prevent epimerization at the chiral center. It was felt that
the best way to accomplish this would be to carry out the
the acetic acid solution to water gave a solid which was very
slow to filter);
(iii) control the amount of palladium in 9 (up to 500 ppm
of palladium was present due to breakthrough during the
filtration stage);
(iv) further refine the conditions for the recrystallization.
Several experiments were carried out in order to address
these issues, and the results are summarized in Table 5.
That the solvent exchange from toluene to acetic acid by
replacement distillation was feasible was demonstrated in
all three of the experiments listed in the table as this was
done in each case. In addition to the results obtained, a study
was carried out in which 9 was held as an acetic acid solution
for 12 days with monitoring by HPLC at periodic intervals.
28
reaction under Schotten-Baumann conditions. Thus, the
acid chloride was prepared by the addition of oxalyl chloride
to a slurry of 9 and a catalytic amount of DMF in THF at 0
°C. The reaction mixture was held for 1 h at this temperature
and then added to a solution of hydroxylamine in THF/
29
water at -10 °C. After 2 h at 0-5 °C, the THF and water
were removed by distillation and replaced by ethyl acetate.
(26) The compound was identified by comparison with an authentic sample
prepared independently.
(28) For an alternative procedure, see: Ando, W.; Tsumaki, H. Synth. Commun.
1983, 1053.
(29) The hydroxylamine was purchased as a 50% aqueous solution.
(27) A reaction carried out for 72 h afforded a mixture containing 75% of the
over-reduced product and only 25% of 9.
616
•
Vol. 9, No. 5, 2005 / Organic Process Research & Development

Table 6. Step 8a f 9: recrystallization of compound 9a
ref
-21
-30
-39
-27
-29
-31
-33
-35
-37
i
solvent
adsorbent
EtOAc/CH
none
3
CN CH
3
CN/MeOH n-BuOAc/ PrOH MeOH
MeOH
MeOH MeOH
MeOH
neutral
alumina
MeOH
none
carbon
PICA
P1400
75
>99
<2
carbon
PICA
P1400
62
chelating chelating carbon carbon
resin
SIR-400
75
resin
SIR-300
59
>99
39
PICA
P1400
75
PICA
GX213
75
recovery (%) 86
75
>99
27
83
>99
30
b
HPLC (%)
Pd (ppm)
>99
15
>99
<2
>99
>99
<2
>99
4
7
a
Note: All reactions were carried out on a 1 g scale. b Area normalization.
Table 7. Step 9 f 10: reaction conditions
Table 8. Step 9 f 10: isolation
HPLC (%)
HPLC (%)
(
COCl) a
2
H
2
NHOHa
yieldb
(%)
yielda
(%)
ref
(equiv)
(equiv)
10
9
ref
solvent (ratio)
10
8
9
1
1
1
1
1
6-2
1.5
10
10
10
20
20
20
96.1
98.3
94.5
97.3
95.9
97.5
2.5
1.0
2.7
1.5
1.8
0.3
78
98
78
90
92
111-1 EtOH/MTBE (1:2)
114-1 EtOAc/EtOH/MTBE (1/1.6/3.2) 98.2
115-4 EtOAc/EtOH/MTBE (1/1.1/2.3) 98.2
97.2
1.7
1.1
1.4
1.9
66
87
90
88
08-1
09-2
13-4
15-4
19-7
1.25
1.25
1.5
1.25
1.25
118-1 EtOH/EtOAc (1/5.7)
97.1
c
d
56
a
Refers to the yield from crude to crystallized 10.
a
b
Based on the amount of 9 used. The product was obtained as a gummy
c
enough for further development. The final conversion was
carried out in a mixture of ethanol/tert-butyl methyl ether
solid or foam after solvent removal. The ethyl acetate solution of 10 was washed
d
with an aqueous Borax solution. Overall yield from 9 after crystallization (see
below).
(the solution was seeded with the correct polymorph to ensure
reproducibility) and afforded a 92% recovery of the desired
form with greater than 99 wt % chemical purity and 99.9%
optical purity.
From our initial work, it appeared that the yield was mainly
dependent upon the quantities of oxalyl chloride and hy-
droxylamine used. Therefore, these two parameters were
studied further. The results are summarized in Table 7.
Although 10 equiv of aq. hydroxylamine appeared to
afford a high conversion in one of the experiments (108-1),
it was decided to develop the process using 20 equiv and
add a Borax wash in order to ensure a low level of 9 in the
drug substance since this impurity was found to be difficult
to remove (experiment 119-7).
Since it proved rather difficult to crystallize the crude
product, the compound was converted to the HCl salt. This
was easily accomplished by the addition of ethanolic HCl
to a solution of the free base. The basic procedure involved
dissolving crude 10 in a suitable solvent or solvent mixture
and adding an antisolvent (50% of the total amount to be
used). This was followed by the addition of ethanolic HCl
In summary, a highly efficient 9-step synthesis of the dual
matrix metalloprotease/tumor necrosis factor MMP090 was
developed and demonstrated on a kilogram scale in 11%
overall yield starting from D-phenylglycine. The most
important features of the chemistry are the following: (1)
the use of Raney nickel to effect the reduction of the aromatic
ring of D-phenylglycine 3 to afford predominantly the trans-
cyclohexanol without racemization; (2) the highly efficient
fractional crystallization of trans-6a to remove the undesired
cis isomer; (3) the bismuth bromide catalyzed reductive
coupling of TBDMS ether 6a with propionaldehyde to give
the n-propyl ether 8 in good yield and purity; (4) the selective
hydrogenolysis of the benzyl ester without reduction of the
pyridine ring; and finally (4) hydroxamic acid formation
under mild conditions without racemization.
(60% of the total amount needed). After seeding, the
Experimental Section
remaining antisolvent and acid were added. Although anti-
solvent and the HCl could be added all at once, this procedure
afforded crystalline 10 with the highest purity and the best
reproducibility. A variety of solvent mixtures were investi-
gated, and the results are summarized in Table 8.
Since no definitive advantage was gained by using the
three-solvent mixture, it was decided in the interest of
simplicity to use the ethanol/ethyl acetate mixture (experi-
ment 118-1). The procedure scaled up well in the pilot plant
where a 74% overall yield from 9 was obtained.
Although the purity of the drug substance at this point
was acceptable, it became necessary to perform a recrystal-
lization in order to obtain the desired polymorph. During
the course of this work, several different polymorphic forms
had been identified; however, only one was considered stable
General. Melting points were determined on a Thomas-
Hoover melting point apparatus and were uncorrected. NMR
spectra were recorded on a Bruker AVANCE DPX-300
1
spectrometer ( H NMR at 300 MHz). Analytical high
performance liquid chromatography (HPLC) was carried out
using a Waters Alliance 2690 Separations Module, a Waters
996 Photodiode Array Detector (MaxPlot), and the following
columns: 4.6 mm × 25 cm Inertsil 5 µ ODS-2 (MetaChem)
and 4.6 × 15 cm Nucleosil 3 µ C18 (MetaChem). Reactions
were carried out under an atmosphere of nitrogen. Residual
water content was determined by Karl Fischer titration. The
Raney nickel catalyst (Grade A-5300) was purchased from
Activated Metals and Chemicals, Inc, Sevierville, TN.
4-Ethoxybenzenesulfonyl Chloride 2. Into a 12-L four-
necked round-bottomed flask equipped with a mechanical
Vol. 9, No. 5, 2005 / Organic Process Research & Development
•
617

stirrer, condenser, nitrogen inlet, addition funnel, heating unit,
and thermometer probe were charged 4-hydroxybenzene-
sulfonic acid sodium salt dehydrate (2.0 kg, 8.62 mol), water
diameter) under 30 psi of nitrogen. At the completion of the
filtration, the filter cake was washed with deionized water
(200 mL), and the filtrates were combined. A portion of this
solution containing 4 was concentrated, and the residue was
analyzed by NMR to determine the trans/cis ratio (5.3:1)
(
1.7 L), and 15% sodium hydroxide solution (3.8 kg). The
solution was warmed to 35 °C, and diethyl sulfate (1.76 kg,
1.46 mol) was added over 10 min. The solution was heated
3
0
1
and the extent of conversion (96.6%). The solution of 4 (1
kg, 0.36 mol based on 3) was placed in a 3-L four-necked
round-bottomed flask equipped with a mechanical stirrer,
condenser, nitrogen inlet, addition funnel, and thermometer
probe, and acetonitrile (660 mL) was added. The solution
was cooled to 0 °C, and triethylamine (73.3 g, 0.72 mol)
was added followed by compound 2 (104 g, 0.47 mol), which
was added dropwise over 1.5 h while maintaining the
temperature at 0 °C. The reaction mixture was held at this
temperature for an additional 2 h, allowed to warm slowly
to room temperature over 1 h, and held at this temperature
for an additional 2 h. The acetonitrile and excess triethy-
lamine were removed by distillation in vacuo to reduce the
volume of the batch by 50%. The pH of the solution was
adjusted to 7.7-8.5 with 1 N sodium hydroxide, and the
solids were removed by suction filtration. The solids were
washed with water (100 mL), and the combined filtrate and
wash were extracted with ethyl acetate (300 mL). The
aqueous portion was treated with 2 N HCl (4.5 mL, 9 mmol)
and ethyl acetate (300 mL). After stirring for 15 min, the
organic layer was discarded, and the pH of the aqueous
portion was adjusted to 2.0 with 2 N HCl (125 mL added
slowly over a 1 h period). The resulting solids were collected
by suction filtration, washed with water (600 mL), and dried
in vacuo at 70 °C for 16 h to give 77.1 g (59%, based on 3)
at an internal temperature of 95-100 °C and held for 16 h.
The reaction mixture was cooled to 23 °C over 2.5 h and
held at this temperature for an additional 1 h. The solid
product was collected by filtration, and the filter cake was
washed with acetone (2 × 750 mL). After drying under
vacuum at 45 °C, 1766 g (87% corrected yield) of 4-ethoxy-
benzenesulfonic acid sodium salt was obtained as an off-
1
white solid: Karl Fischer 4.3% water; H NMR (CD
3
OD) δ
7.75 (d, J ) 9.4 Hz, 2H), 6.94 (d, J ) 9.4 Hz, 2H), 4.05 (q,
J ) 7.5 Hz, 2H), 1.38 (t, J ) 7.5 Hz, 3H). This compound
was used in the following step without further purification.
Into a 12-L four-necked round-bottomed flask equipped
with a mechanical stirrer, nitrogen inlet, addition funnel,
thermometer, and a condenser set for downward distillation
were placed 4-ethoxybenzenesulfonic acid sodium salt (1.2
kg, 5.3 mol), toluene (6 L), and DMF (77.5 g). The
heterogeneous reaction mixture was heated to a 115 °C jacket
temperature, and the toluene/water mixture was removed by
azeotropic distillation (600-700 mL). Once the distillation
was completed, the batch temperature was reduced to 80 °C,
and thionyl chloride (1.9 kg, 16 mol) was added over 5.5 h
while maintaining an internal temperature of 75 °C. Upon
completion of the addition, the mixture was held for an
additional 5 h at 80 °C, cooled to room temperature, and
held for 16 h. The reaction mixture was heated at an internal
temperature of 45-50 °C under reduced pressure (95-100
mmHg), and toluene and excess thionyl chloride (ap-
proximately 3 L) were removed by distillation. After cooling
to room temperature, the concentrate was filtered and the
filter cake was washed with toluene (300 mL). The filtrate
was further concentrated by distillation (45-65 °C, 95-100
mmHg) to afford 4-ethoxybenzenesulfonyl chloride (1.18 kg)
as an orange liquid. The product was found to contain toluene
of 5 as an off-white solid (4/1 trans/cis mixture): mp 80-
1
8
6
(
1
3 °C; H NMR (D
2
O-NaOD) δ 7.58 (2H, d, J ) 8.9 Hz),
.91 (2H, d, J ) 8.9 Hz), 4.06 (2H, q, J ) 7.0 Hz), 3.38
1H, m), 2.87 (1H, d, J ) 7.9 Hz), 1.85-1.65 (3H, m), 1.39-
.15 (5H, m), 1.02 (2H, m), 0.89 (1H, m), 0.65 (1H, m).
Anal. Calcd for C16 S‚H O: C, 51.19; H, 6.71; N,
.73. Found: C, 51.31; H, 6.85; N, 3.66.
H23NO
6
2
3
Phenylmethyl (R)-R-[N-(4-Ethoxyphenylsulfonyl)amino]-
-hydroxycyclohexaneacetate 6. Into a 4-L four-necked
4
round-bottomed flask equipped with a mechanical stirrer,
condenser, nitrogen inlet, addition funnel, and thermometer
probe were placed 5 (50.0 g, 0.137 mol) and DMF (500 mL).
To the solution was added dicyclohexylamine (32.6 g, 0.178
mol) at such a rate that the temperature was maintained below
(
27%) and DMF (<1%) by NMR analysis resulting in a
1
corrected yield of 863.3 g (74%): H NMR (CDCl
3
) δ 7.87
(d, J ) 10.0 Hz, 2H), 6.94 (d, J ) 10.0 Hz, 2H), 4.07 (q, J
)
6.9 Hz, 2H), 1.38 (t, J ) 6.9 Hz, 3H). This mixture was
used in the following step without further purification.
R)-R-[N-[(4-Ethoxyphenyl)sulfonyl]amino]-4-hydroxy-
2
5 °C. Benzyl bromide (31.1 g, 0.178 mol) was introduced
(
while again maintaining the temperature below 25 °C during
the addition. The reaction mixture was stirred at ambient
temperature for 5 h. The resulting solids were filtered with
suction, and the filter cake was washed with DMF (50 mL).
A 2-L four-necked round-bottomed flask equipped with a
mechanical stirrer, nitrogen inlet, and addition funnel was
charged with water (1755 mL), sodium chloride (35.1 g),
and n-heptane (206 mL), and the mixture was stirred for 5
cyclohexanacetic Acid 5. Into an RC1 Mettler Toledo
Reactor Calorimeter fitted with a 1-L MP 10 glass reactor
were added deionized water (300 g) and sodium hydroxide
(
(
12 g, 0.3 mol). To the stirred solution was added d-(-)-
4-hydroxyphenyl)glycine 3 (50.2 g, 0.3 mol), and the
mixture was stirred at 30 °C for 15 min to dissolve the solids.
The solution was treated with Raney nickel catalyst (250
g), and the reactor was purged with nitrogen 3 times and
hydrogen 3 times. The reactor was pressurized to 30 psi with
hydrogen and heated to 62 °C over 20-30 min and held at
this temperature for 15 h with stirring. After cooling to room
temperature and purging of the vessel with nitrogen, the
suspension was filtered through a pressure filter equipped
with a cellulose acetate filter pad (pore size 0.8 µ, 142 mm
(30) The trans/cis ratio was determined by comparing the integrated peak areas
1
at 3.99 ppm (cis) and 3.56 ppm (trans): H NMR (270 MHz, D
(
(
2
O) δ 3.99
m, part of 1H), 3.56 (m, part of 1H), 3.05 (d, J ) 6.0 Hz, 1H), 2.0-0.9
m, 9H). Aromatic protons for the unreduced amino acid: 7.02 (d, J ) 8.2
Hz, 2H), 6.61 (d, J ) 8.2 Hz, 2H). The percent conversion was calculated
from the following equation: conversion % ) [1 - A/2 ÷ A/2 + B] ×
1
00. A ) the total area of the olefin/aromatic hydrogens at 6-8 ppm. B )
area of the peak at 3.05 ppm.
618
•
Vol. 9, No. 5, 2005 / Organic Process Research & Development

min. The solution of the filtrate from above was added to
the water/heptane solution over 30 min. The resulting slurry
was stirred for an additional 30 min and then filtered using
suction. The solid was washed with water (800 mL) and dried
at 60-65 °C under reduced pressure to give 62.0 g (93%,
atmospheric pressure, and acetonitrile (100 mL) was removed
by distillation. The solution was allowed to cool to room
temperature, and triethylsilane (17.5 g, 0.150 mol) was added
over 5 min. In a separate flask, bismuth bromide (3.0 g,
0.0067 mol) was dissolved in acetonitrile (25 mL), and this
solution was added to the solution of 6a and triethylsilane
over 5 min. Once the addition was complete, propionalde-
hyde (8.7 g, 0.150 mol) was introduced at such a rate that
the reaction temperature did not exceed 30 °C. The reaction
mixture was stirred for 30 min and quenched with a 5%
solution of sodium bicarbonate (375 g). Ethyl acetate (320
mL) was added, and the suspension was filtered through
Celite (10 g). The solids were washed with ethyl acetate (50
mL), and the filtrate and wash were combined and allowed
to stand for 10 min to allow the layers to separate. The
organic layer was washed with 1% sodium chloride solution
(300 mL) and water (300 mL) and then concentrated under
reduced pressure to a final volume (400 mL). To the resulting
white suspension was added n-heptane (500 mL), and the
solution was concentrated under reduced pressure to a final
volume (300 mL). This was repeated two more times. To
the suspension were added n-heptane (888 mL) and ethyl
acetate (245 mL), and the mixture was heated to reflux (70
°C) to give a clear solution. The solution was cooled slowly
to 0 °C and held at this temperature for 30 min. The solids
were collected by suction filtration, washed with ethyl
acetate/n-heptane (1/5 v/v, 190 mL), and dried at 45 °C under
corrected) of 6 as a 3.8/1 mixture of trans/cis isomers: mp
1
1
13-116 °C; H NMR (CDCl
3
) δ 7.71 and 6.87 (d of d, J
)
8.3 Hz, 4H), 7.15 (m, 5H), 5.12 (m, 1H), 4.88 (s, 2H),
4.04 (q, J ) 6.60 Hz, 2H), 3.75 (m, 1H), 3.49 (m, 1H), 1.94
1
3
(
m, 2H), 1.44 (t, J ) 6.60 Hz, 3H), 1.14-1.59 (m, 8H); C
NMR (CDCl ) δ 171.7, 171.6, 162.8, 135.1, 131.1, 129.8,
29.0, 128.7, 115.0, 70.5, 67.7, 66.1, 64.3, 60.3, 40.8, 40.5,
3
1
3
5.1, 35.0, 32.3, 27.7, 26.9, 26.7, 26.1, 23.4, 22.2, 15.0; MS
+
m/z 448 (M ). Anal. Calcd for C23
6
H29NO S: C, 61.72; H,
6.53; N, 3.13. Found: C, 61.94; H, 6.68; N, 3.09.
Phenylmethyl trans-4[[Dimethyl(1,1-dimethylethyl)si-
lyl]oxy]-[N-(ethyloxyphenylsulfonyl)-amino]cyclohexa-
neacetate 6a. Into a 1-L three-necked round-bottomed flask
equipped with a mechanical stirrer, thermometer, nitrogen
inlet, and addition funnel were charged 6 (53.0 g, 0.116 mol)
and DMF (400 mL). Imidazole (11.9 g, 0.174 mol) was
introduced followed by a solution of tert-butyldimethylsilyl
chloride (23.6 g, 0.150 mol) in DMF (100 mL) (the solution
of tert-butyldimethylsilyl chloride in DMF was prepared by
warming to 35 °C followed by cooling to 20 °C prior to
addition), which was added over a 10 min period. The
reaction mixture was stirred for 4 h at room temperature and
then diluted with methyl tert-butyl ether (MTBE, 500 mL)
and water (600 mL). The layers were separated, and the
aqueous layer was extracted with MTBE (500 mL). The
organic layers were combined, washed with water (2 × 500
mL), and filtered through Celite. The cake was washed with
MTBE (100 mL), which was combined with the filtrate. The
solution was concentrated under reduced pressure to a final
volume (600 mL). This solution was added to n-heptane (500
mL) with stirring, and the resulting mixture was concentrated
under reduced pressure to a final volume (600 mL). n-
Heptane (250 mL) was added to the resulting suspension,
and the mixture was concentrated once again under reduced
pressure to a final volume (600 mL). n-Heptane (400 mL)
was added, and the suspension was heated to reflux to effect
dissolution. After cooling to room temperature, the solids
were collected by suction filtration, washed with n-heptane
reduced pressure to give 41.5 g (84.8%, corrected) of 8 as a
1
white solid: mp 108-111 °C; H NMR (CDCl
3
) δ 7.71 and
6.87 (d of d, J ) 8.3 Hz, 4H), 7.33 (m, 3H), 7.15 (m, 2H),
5.07 (d, J ) 9.9 Hz, 1H), 4.88 (m, 2H), 4.04 (q, J ) 6.6 Hz,
2H), 3.75 (m, 1H), 3.36 (t, J ) 8.3 Hz, 2H), 3.08 (m, 1H),
2.02 (m, 1H), 1.62-1.05 (m, 12H), 0.89 (t, J ) 8.3 Hz, 3H).
Anal. Calcd for C26H35NO S: C, 63.78; H, 7.21; N, 2.86.
6
Found: C, 63.62; H, 7.23; N, 2.33.
trans-(R)-[[N-(4-Ethoxyphenylsulfonyl)-N-(4-pyridinyl-
methyl)]amino]-4-propoxycyclohexanacetic Acid 9. Into
a 1-L three-necked round-bottomed flask equipped with a
mechanical stirrer, condenser, nitrogen inlet, and thermometer
probe were placed 8 (116.4 g, 0.24 mol), 4-picolyl chloride
hydrochloride (58.8 g, 0.36 mol), powdered potassium
carbonate (325 mesh) (310.1 g, 2.24 mol), and N,N-
dimethylformamide (DMF) (357 mL). The slurry was stirred
in a water bath at 18-22 °C, and tris[2-(2-methoxyethoxy)-
ethyl]amine (TDA-1) (15.2 mL, 0.023 mol) was added in
one portion. The slurry was stirred at room temperature for
24 h, and Celite (46.1 g) was added. After stirring for 10
min, the mixture was filtered using vacuum, and the filter
cake was washed with toluene (3 × 225 mL). The filtrate
and washes were combined and cooled to 5 °C, and 10%
aqueous citric acid (715 mL) was added at such a rate that
the temperature was kept below 20 °C. After stirring for 5
min, the layers were separated, and the aqueous layer was
extracted with toluene (2 × 300 mL). All of the toluene
fractions were combined and cooled below 5 °C. The cold
toluene solution was washed sequentially with the following
cold (5 °C) solutions: 3 N HCl (2 × 207 g), water (490
mL), and 3% aqueous sodium chloride (490 g). This was
followed by washing the toluene solution at room temper-
(100 mL), and dried at 45 °C under reduced pressure to give
4
6
1
5.0 g (68.3%, corrected, based on the chemical purity of
) of 6a with a trans/cis isomer ratio of 121/1: mp 108-
1
09 °C; H NMR (CDCl
3
) δ 7.70 and 6.86 (d of d, J ) 7.5
Hz, 4H), 7.33 (m, 3H), 7.16 (m, 2H), 5.05 (d, J ) 11.9 Hz,
1
3
H), 4.88 (m, 2H), 4.04 (q, J ) 6.0 Hz, 2H), 3.75 (m, 1H),
.43 (m, 1H), 1.83 (m, 2H), 1.56-1.07 (m, 7H), 1.44 (t, J
)
6.0 Hz, 3H), 0.86 (s, 9H), 0.03 (s, 6H). Anal. Calcd for
29 6
C H43NO SSi: C, 62.00; H, 7.72; N, 2.49. Found: C, 61.93;
H, 8.02; N, 1.99.
Phenylmethyl (R)-trans-R-[N-(4-Ethoxyphenylsulfony-
l)amino]-4-propoxycyclohexaneacetate 8. Into a 1-L four-
necked round-bottomed flask equipped with a mechanical
stirrer, condenser, nitrogen inlet, addition funnel, and ther-
mometer probe was placed 6a (56.2 g, 0.100 mol) in dry
acetonitrile (600 mL). The mixture was heated to 82 °C under
Vol. 9, No. 5, 2005 / Organic Process Research & Development
•
619

ature with water (2 × 490 mL) and 25% aqueous sodium
chloride solution (204 mL). The toluene solution was filtered
through aluminum oxide (activated, acidic, Brockmann I,
34 2 6
Anal. Calcd for C25H N O S: C, 61.20; H, 6.98; N, 5.71;
S, 6.54. Found: C, 61.19; H, 6.92; N, 5.69; S, 6.48.
trans-(R)-[[N-(4-Ethoxyphenylsulfonyl)-N-(4-pyridinyl-
methyl)]amino]-N-hydroxy-4-propoxy-cyclohexaneaceta-
mide Hydrochloride 10 (MMP090). Into a 1-L four-necked
round-bottomed flask equipped with a mechanical stirrer,
condenser, nitrogen inlet, addition funnel, and thermometer
probe was charged 9 (73.5 g, 0.15 mol), THF (750 mL),
and DMF (1.2 mL). The slurry was cooled in an ice bath to
0-5 °C, and oxalyl chloride (23.9 g, 0.18 mol) was added
at such a rate that the temperature did not exceed 5 °C. The
time of the addition was 20 min. The homogeneous mixture
was stirred at 0-5 °C for 1 h and then added via cannula to
a cold (-15 °C), stirred solution of 50% aqueous hydroxy-
lamine (198 g, 3.0 mol) in THF (2.25 L). The addition was
carried out at such a rate that the temperature did not exceed
121.3 g) using gravity. The alumina was washed with toluene
(
3 × 50 mL), and the combined fractions were concentrated
under reduced pressure to a final volume (275 mL). The
toluene solution containing the benzyl ester of 9 (109.5 g,
79% yield, determined by HPLC assay) was used as is in
the following step. A sample of the solution was concentrated
1
to an oil for charactization purposes: H NMR (CDCl
3
) δ
8
4
1
.48 (d, J ) 3.8 Hz, 2H), 7.6 and 6.8 (d of d, J ) 10.0 Hz,
H), 7.34 (m, 3H), 7.26 (m, 2H), 7.24 (m, 2H), 4.9 (d, J )
1.3 Hz, 1H), 4.78 (d, J ) 11.3 Hz, 1H), 4.63 (m, 2H), 4.32
(d, J ) 11.3 Hz, 1H), 4.0 (m, 2H), 3.3 (m, 2H), 3.06 (m,
1
4
H), 1.93 (m, 2H), 1.75 (m, 1H), 1.54 (m, 3H), 1.44 (m,
H), 1.03 (m, 3H), 0.85 (t, J ) 7.5 Hz, 3H), 0.75 (m, 1H);
-
5 °C. After the addition was complete, the reaction mixture
+
MS m/z 581 (M ).
was allowed to warm to 0-5 °C and held at this temperature
for 2 h. Ethyl acetate (750 mL) was added, and the lower
aqueous layer was separated and discarded. The organic layer
was concentrated under reduced pressure to a final volume
A toluene solution of the benzyl ester of 9 (475.5 g, 0.5
mol) was diluted with glacial acetic acid (1.4 L). The solution
was concentrated under reduced pressure at 50 °C to a final
volume (475 mL). The concentrate was diluted with glacial
acetic acid (2.2 L), and a portion (1.4 L) was charged to a
(600 mL), and ethyl acetate (750 mL) was added. The
solution was concentrated again in the same manner to a
final volume (600 mL), and ethyl acetate (500 mL) was
added. The ethyl acetate solution was washed with a 2%
sodium tetraborate decahydrate solution (3 × 150 mL), water
2-L Parr bottle that had been inerted with hydrogen and
which contained 10% Pd/C, 61.5% water wet (18.1 g). The
mixture was hydrogenated under an atmosphere of 40 psi
of hydrogen until the uptake slowed appreciably (ap-
proximately 60 min). The reaction mixture was treated with
Celite (18.1 g) and stirred for 15 min. The suspension was
filtered through a wet-packed pad of filter aid (42 g), and
the filter cake was washed with glacial acetic acid (2 × 100
mL). The filtrate and washes were combined and concen-
trated under reduced pressure at 50 °C to a final volume
(150 mL), and 5% sodium chloride (150 mL) and concen-
trated under reduced pressure to a final volume (650 mL).
Ethyl acetate (750 mL) was added, and the solution was
concentrated to a final volume (550 mL) under reduced
pressure. The process was repeated using ethyl acetate (500
mL) until a final volume (650 mL) was reached. After
cooling to room temperature, absolute ethanol (130 mL) was
added, and the solution was filtered using filter aid. The
filtrate was heated to 55 °C, and 4.2 N ethanolic hydrogen
chloride (23 mL, 96.6 mmol) was added. After seeding, the
mixture was held at 55 °C for 20 min, and 4.2 N ethanolic
hydrogen chloride (16 mL, 67.2 mmol) was added. Seed
crystals were again added followed by ethyl acetate (115
mL). After stirring for 20 min, the mixture was cooled to
20 °C over 2 h and held at this temperature for an additional
2 h. The solids were isolated by suction filtration, washed
with ethyl acetate (2 × 150 mL), and dried under reduced
pressure at 50 °C overnight to give 60.3 g (74%) of MMP090
(293 mL). This solution was added slowly to a stirred
solution of sodium chloride (40.3 g) in water (2.05 L), which
contained seed crystals of the desired product (9). When the
addition was complete, the slurry was stirred for 10 min and
then cooled to 2 °C and stirred at this temperature for 1 h.
The solids were collected by suction filtration, washed with
cold water (1.08 L), and dried at 55 °C under reduced
pressure to give 69.6 g (54.5%, overall yield from 8) of 9 as
an off-white solid. The crude product (69.0 g) was dissolved
in methanol (1.7 L) containing activated carbon (6.9 g) at
reflux. After 1 h, the suspension was filtered through Celite
3
as an off-white solid: mp 140 °C (lit. 131 °C); enantiomeric
(34.3 g), and the filter cake was washed with hot methanol
1
purity (HPLC assay) 100%; H NMR (MeOD) δ 8.74 and
(60 °C). The filtrate and wash were combined and concen-
8
.02 (d of d, J ) 7.5 Hz, 4H), 7.78 and 7.05 (d of d, J ) 9.0
Hz, 4H), 5.17 (d, 2H), 4.13 (q, J ) 7.50 Hz, 2H), 3.90 (d,
H), 3.40 (t, J ) 6.0 Hz), 3.11 (m, 1H), 2.04 (m, 1H), 1.86
m, 1H), 1.75 (m, 1H), 1.52 (m, 4H), 1.42 (t, J ) 7.50 Hz,
H), 1.11 (m, 3H), 0.89 (t, J ) 7.5 Hz, 3H), 0.50 (m, 1H).
trated under atmospheric pressure until the product began
to crystallize (approximately 1 L of distillate was collected).
The slurry was cooled to -10 °C and held for 1 h. The solids
were collected by suction filtration, washed with cold
methanol (-10 °C), and dried at 55 °C under reduced
pressure to give 61.5 g (48.2%, overall yield from 8) of 9 as
1
(
3
1
Acknowledgment
a white solid: mp 206-208 °C; H NMR (DMSO, d
1
7
6
) δ
2.91 (br s, 1H), 8.48 and 7.37 (d of d, J ) 5.0 Hz, 4H),
.78 and 7.03 (d of d, J ) 10.0 Hz, 4H), 4.62 (m, 2H), 4.10
We wish to thank Karl Gunderson for recording all NMR
spectra.
(q, J ) 6.2 Hz, 2H), 3.96 (d, J ) 11.2 Hz, 1H), 3.25 (t, J )
7
1
.50 Hz, 2H), 3.02 (m, 1H), 1.88 (m, 1H), 1.72 (m, 1H),
Received for review May 4, 2005.
OP050066K
.54 (m, 1H), 1.37 (m, 7H), 0.93 (m, 2H), 0.78 (t, J ) 7.50
+
Hz, 3H), 0.75 (m, 1H), 0.52 (m, 1H); MS m/z 491 (M ).
620
•
Vol. 9, No. 5, 2005 / Organic Process Research & Development