A convergent kilogram-scale synthesis of the PPARα Agonist LY518674: Discovery of a novel acid-mediated triazolone synthesis

Article abstract of DOI:10.1021/op700040v

The first kilogram-scale synthesis of the PPARα agonist LY518674 (1) is described. The de novo convergent synthetic approach involved coupling of two rapidly assembled components, triazolone formation via a novel acid-promoted cyclization reaction, and final step saponification, delivering the compound in 32.5% overall yield via eight total steps with a six-step longest linear sequence. A regioselective alkylation on the dianion of 4-hydroxyphenylbutyric acid allowed the direct preparation of one of the convergent coupling partners, carboxylic acid 12, and an unusual solvent effect enabled the installation of a urea group on a protected hydrazine, permitting the regiospecific preparation of the other coupling partner, semicarbazide mesylate 17. Sulfonic acids were found to effect the desired triazolone ring formation, affording 25 from the coupled precursor acyl semicarbazide 23. Following saponification of 25 to 1, a wide solubility differential between ethyl acetate extracts of 1 and solutions of 1 in anhydrous ethyl acetate was harnessed in the final crystallization step to deliver the final compound in high yield and purity. The novel acid-mediated triazolone formation was further evaluated on a range of additional substrates, showing the new methodology to be largely complementary to existing base-mediated triazolone syntheses.

Full text of DOI:10.1021/op700040v

Organic Process Research & Development 2007, 11, 431−440
A Convergent Kilogram-Scale Synthesis of the PPARr Agonist LY518674:
Discovery of a Novel Acid-Mediated Triazolone Synthesis
Timothy M. Braden, D. Scott Coffey, Christopher W. Doecke, Michael E. LeTourneau, Michael J. Martinelli,†
Christopher L. Meyer, Richard D. Miller, Joseph M. Pawlak, Steven W. Pedersen, Christopher R. Schmid,*
Bruce W. Shaw, Michael A. Staszak, and Jeffrey T. Vicenzi
Chemical Product Research and DeVelopment, Eli Lilly and Company, Lilly Corporate Center, Indianapolis, Indiana
46285-4813, U.S.A.
Abstract:
The PPARR receptor has been associated with regulation
of serum triglyceride levels; modulation of this receptor also
triggers production of high-density lipoprotein (HDL) cho-
lesterol, elevated levels of which have been associated with
diminished risk of CHD.3 Consequently, administration of
a highly selective and potent PPARR agonist could provide
therapeutic benefit to patients suffering from either dyslipi-
demia or increased risk of CHD by virtue of low HDL levels.
The co-morbidity of both these maladies in the so-called
metabolic syndrome X, associated with increasing obesity
in the first world patient population, further supports
development of new therapies in this area.4
The first kilogram-scale synthesis of the PPARr agonist
LY518674 (1) is described. The de novo convergent synthetic
approach involved coupling of two rapidly assembled compo-
nents, triazolone formation via a novel acid-promoted cycliza-
tion reaction, and final step saponification, delivering the
compound in 32.5% overall yield via eight total steps with a
six-step longest linear sequence. A regioselective alkylation on
the dianion of 4-hydroxyphenylbutyric acid allowed the direct
preparation of one of the convergent coupling partners, car-
boxylic acid 12, and an unusual solvent effect enabled the
installation of a urea group on a protected hydrazine, permitting
the regiospecific preparation of the other coupling partner,
semicarbazide mesylate 17. Sulfonic acids were found to effect
the desired triazolone ring formation, affording 25 from the
coupled precursor acyl semicarbazide 23. Following saponifica-
tion of 25 to 1, a wide solubility differential between ethyl
acetate extracts of 1 and solutions of 1 in anhydrous ethyl
acetate was harnessed in the final crystallization step to deliver
the final compound in high yield and purity. The novel acid-
mediated triazolone formation was further evaluated on a range
of additional substrates, showing the new methodology to be
largely complementary to existing base-mediated triazolone
syntheses.
LY518674 (1), a 2,4-dihydro-3H-1,2,4-triazol-3-one (tria-
zolone)-based PPARR agonist, was recently disclosed as part
of an effort to identify potent, selective compounds targeting
this receptor isoform.5 The compound showed excellent in
vitro binding affinity and selectivity for the PPARR receptor
and proved highly potent in the human apoA-I transgenic
mouse in vivo model; accordingly, the compound was
selected for further preclinical and clinical evaluation.
Kilogram quantities of 1 were required to support these
studies, and we detail here our successful preparation of 1
on kilogram scale, using a highly convergent synthetic
approach that employs a novel acid-catalyzed triazolone ring
formation.
Introduction
The peroxisome proliferator activated receptors (PPARs)
comprise a family of nuclear hormone receptors which play
a central role in the metabolic regulation of dietary carbo-
hydrates and fats in mammals.1 By virtue of their integral
involvement in these processes, the PPARs have become
targets of significant interest for drug researchers, as suc-
cessful modulation of these receptors via small-molecule-
based therapies offers the potential to treat diseases ranging
from dyslipidemia and coronary heart disease (CHD) to
diabetes. Three receptor isoforms, PPARR, PPARδ, and
PPARγ, have been identified to date, and their respective
functions in metabolic processes have been established.2
Results and Discussion
The first synthesis of 1, accomplished by Mantlo and co-
workers, is shown in Scheme 1. Thus acyl hydrazide 2 was
(3) (a) Staels, B.; Zuwerx, J. Curr. Pharm. Des. 1997, 3, 1-14. (b) Staels, B.;
Dallongeville, J.; Auwerx, J.; Schoonjans, D.; Leitersdorf, E.; Furchart, J.
C. Circulation 1998, 98, 2088-2093.
* To whom correspondence should be addressed. Telephone: (317)-276-5155.
E-mail: crs@lilly.com.
(4) (a) Timar, O.; Sestier, F.; Levy, E. Can. J. Cardiol. 2000, 16, 779-789.
(b) Lopez-Candales, A. J. Med. 2001, 32, 283-300. (c) Jeppesen, J. Metab.
Syndr. Relat. Disord. 2003, 1, 33-53.
(5) Xu, Y.; Mayhugh, D.; Saeed, A.; Wang, X.; Thompson, R. C.; Dominianni,
S. J.; Kauffman, R. F.; Singh, J.; Bean, J. S.; Bensch, W. R.; Barr, R. J.;
Osborne, J.; Montrose-Rafizadeh, C.; Zink, R. W.; Yumibe, N. P.; Huang,
N.; Luffer-Atlas, D.; Rungta, D.; Maise, D. E.; Mantlo, N. B. J. Med. Chem.
2003, 46, 5121-5124.
† Present address: Chemistry Research and Discovery, Amgen, Inc., One
Amgen Drive, Thousand Oaks, CA 91320.
(1) (a) Reifel-Miller, A. Circulation 2006, 67, 574-578. (b) Lemberger, T.;
Desvergne, B.; Wahli, W. Annu. ReV. Cell DeV. Biol. 1996, 12, 335-363.
(2) Wilson, T.; Brown, P.; Sternbach, D.; Henke, B. J. Med. Chem. 2000, 43,
527-550.
10.1021/op700040v CCC: $37.00 © 2007 American Chemical Society
Published on Web 04/06/2007
Vol. 11, No. 3, 2007 / Organic Process Research & Development
•
431

Scheme 1. Mantlo synthesis of LY518674 (1)a
Scheme 2. Retrosynthesis for a convergent approach to 1
a Reagents and conditions taken from Ref 5: (a) 2,4-dimethoxybenzyl
isocyanate, THF, rt; (b) KOH, MeOH, reflux; (c) MeOH, H2SO4, rt, then
4-methylbenzyl bromide, K2CO3, DMF, 45 °C; (d) HBr, HOAc, rt.
A search of the literature revealed essentially no examples
of direct condensations to provide the desired N2-alkylated
N4-unsubstituted triazolones from acyl semicarbazides with
the substitution pattern shown. The reported base-promoted
triazolone-forming reactions (using hydroxide as base) have
all employed acyl semicarbazide precursors containing latent
N4-triazolone substitution.5,6 A single reference employing
a combination of triflic acid and hexamethyldisilazide
provided the only encouraging precedent.7
To determine if such a cyclization to the desired triazolone
was indeed possible, gram quantities of a suitable acyl
semicarbazide were required. Having ready access to acyl
hydrazine 2, we undertook a synthesis of acyl semicarbazide
6 (Scheme 3). Hydrazone formation on the terminal nitrogen
of 2 using 4-methylbenzaldehyde furnished the crystalline
7 (92%), which was subjected to several sets of conditions
to effect reduction. Hydrogenation using catalytic platinum
reacted with 2,4-dimethoxybenzyl isocyanate to form the acyl
semicarbazide 3, which was then subjected to potassium
hydroxide (KOH) in methanol (MeOH), affording the N4-
benzylated N2-unsubstituted triazolone 4. Re-esterification
and alkylation at N2 with 4-methylbenzyl bromide provided
the dialkylated triazolone 5. Acid-mediated deprotection of
the 2,4-dimethoxybenzyl group (HBr/HOAc) revealed the
N4 position while also accomplishing ester saponification to
afford 1.5
The published synthesis of 1 was attractive as a potentially
scaleable approach to the moleculesthe compound was
assembled in a straightforward mannersand it was felt that
any issues arising on scale-up could likely be addressed with
alternative conditions and minor modifications to the inter-
mediates. However, the selection of 1 as the sole compound
of interest permitted us to consider additional synthetic
approaches unique to the target. We were intrigued in
particular with the retrosynthetic possibility depicted in
Scheme 2, where 1 would be derived from the acyl
semicarbazide shown. Formation of the desired triazolone
core, a condensation reaction, could potentially be promoted
by either acid or base; dehydrative conditions to facilitate
the desired reaction could also be evaluated. The acyl
semicarbazide itself would be derived from convergent
assembly of a suitably activated carboxylic acid and the
benzylated semicarbazide 14. We decided to evaluate this
route aggressively, while keeping the original synthesis as a
fall-back approach should acceptable cyclization conditions
not be identified.
(6) (a) Kubota, S.; Uda, M. Chem. Pharm. Bull. 1973, 21, 1342-1350. (b)
Madding, G. D.; Smith, D. W.; Sheldon, R. I.; Lee, B. J. Heterocycl. Chem.
1985, 22, 1121-1126. (c) Mazzzone, G.; Puglisi, G.; Corsaro, A.; Panico,
A.; Bonina, F.; Amico-Roxas, M.; Caruso, A.; Trombadore, S. Eur. J. Med.
Chem. 1986, 21, 277-284. (d) Kane, J. M. Synthesis 1987, 912-914. (e)
Kane, J. M.; Baron, B. M.; Dudley, M. W.; Sorensen, S. M.; Staeger, M.
A.; Miller, F. P. J. Med. Chem. 1990, 33, 2772-2777. (f) Dobosz, M.;
Pitucha, M.; Dybala, I.; Koziol, A. E. Collect. Czech. Chem. Commun. 2003,
68, 792-800. (g) Roman, G.; Manciulea, I.; Adreleanu, R.; Dumitrescu, L.
Heterocycl. Commun. 2005, 11, 311-316. (h) Wujec, M.; Pitucha, M.;
Dobosz, M. Heterocycles 2006, 68, 779-785. For an alternative approach
to N4-unsubstituted triazolones, see: (i) Chang, L. L.; Ashton, W. T.;
Flanagan, K. L.; Strelitz, R. A.; MacCoss, M.; Greenlee, W. J.; Chang, R.
S. L.; Lotti, V. J.; Faust, K. A.; Chen, T.-B.; Bunting, P.; Zingaro, G. J.;
Kivlighn, S. D.; Siegl, P. K. S. J. Med. Chem. 1993, 36, 2558-2568.
(7) Rigo, B.; Valligny, D.; Taisne, S.; Couturier, D. Synth. Commun. 1988, 18,
167-180.
432
•
Vol. 11, No. 3, 2007 / Organic Process Research & Development

Scheme 3. Preparation of cyclization precursor 6a
a Reagents and Conditions. (a) 4-methylbenzyaldehyde (1 equiv), ethyl acetate,
rt, then hexanes, 92%; (b) 40 psig H2, PtO2, THF, rt, 16 h, 97%; (c) TMS-
NCO (1.5 equiv), IPA, rt, 16 h, 81%.
Figure 1. Rationales for the observed rate increase in the
reaction of 8 with TMS-NCO in IPA.
oxide in tetrahydrofuran (THF) served to deliver the desired
product 8 (97%).
Scheme 4. Acid- and base-promoted cyclization attempts
on 6a
We uncovered an intriguing solvent-related effect during
installation of the urea moiety on the basic nitrogen of acyl
hydrazine 8 to afford 6. Reaction of the compound with
trimethylsilyl isocyanate (TMS-NCO) in dichloromethane
(DCM), at rt gave no reaction. Heating the reaction in 1,2-
dichloroethane8 also produced only recovered starting mate-
rial. We were surprised to find that the reaction proceeded
satisfactorily in ethanol (EtOH) and isopropanol (IPA). IPA
was selected as the solvent, and a slight excess (1.5 equiv)
of TMS-NCO was employed to compensate for a basal rate
of reaction between reagent and solvent. Compound 6 was
thus prepared in 81% yield.
Two speculative explanations could be offered for this
unusual effect (Figure 1). The solvent’s hydroxyl group could
be involved in a 6-membered transition state, facilitating
proton exchanges between substrate and reagent. Alterna-
tively, simple silyl group exchange with solvent hydroxyl
proton could serve to slowly release isocyanic acid (HNCO),
which would be expected to react more readily with 8 than
TMS-NCO.
With quantities of 6 in hand, we explored a range of acid-
and base-promoted cyclization conditions (Scheme 4). Treat-
ment with KOH in MeOH (50 °C, 16 h) gave a complex
mixture by HPLC and NMR. Treatment with K2CO3 in
MeOH returned 8 through removal of the urea functionalitys
a competing side reaction that would later be observed under
several sets of conditions with several substrates. Non-
nucleophilic bases such as lithium hexamethyldisilazide
(LiHMDS, THF, rt to reflux), gave a complex mixture by
HPLC, but encouragingly showed the desired 9 as the main
component (∼40%).
a Reagents and conditions: (a) K2CO3, MeOH, reflux, 30 min; (b) LiHMDS,
THF, rt to reflux, 4 h, ca. 40%; (c) TsOH, toluene, reflux, 4 h, ca. 55%.
Acid-promoted cyclization fared better. In particular,
reaction of 6 with p-toluenesulfonic acid (TsOH) in toluene
at reflux gave desired product 1 directly (TsOH additionally
effecting de-esterification of the tert-butyl ester group) in a
reaction that also afforded small quantities of 10, the
saponified version of 8. The results from this preliminary
triazolone-forming experiment provided us with enough
encouragement to pursue the approach as a means of secur-
ing 1.
(8) Bolton, R. E.; Moody, C. J.; Rees, C. W.; Tojo, G. Tetrahedron Lett. 1987,
28, 3163-3166.
Vol. 11, No. 3, 2007 / Organic Process Research & Development
•
433

Scheme 5. Design considerations for the synthesis of 1
of the ethyl ester instead of the tert-butyl ester on the
alkylating reagent was desirable for two reasons. First, we
had observed that alkylation of phenols was significantly
faster with ethyl bromoisobutyrate than with tert-butyl
bromoisobutyrate; additionally, the competing elimination
reaction of the reagent to form the corresponding methacry-
late was less pronounced. Second, although the tert-butyl
ester had been conveniently deprotected concomitantly with
triazolone formation in the one-pot conversions of 5 to 1
and 6 to 1, a switch to the ethyl ester would install a discreet
saponification step at the end of the synthesis, adding a useful
potential purification point to the proposed route. These
considerations outweighed any concerns over introduction
of an additional step in the synthesis.
Research on the first leg of the synthesis began with the
use of common conditions (potassium carbonate, dimethyl-
formamide) to effect the alkylation of 11 with ethyl bro-
moisobutyrate; the reactions proceeded poorly, and additional
charges of base and alkylating reagent were routinely
required to drive them to completion. It was felt that the
dianion of 11 might have limited solubility in the alkylation
medium, and so other conditions were explored. A switch
to EtOH as solvent proved more successful, and use of
potassium tert-butoxide in EtOH gave even cleaner, higher-
yielding reactions. We ultimately settled on the use of sodium
ethoxide (NaOEt) in EtOH as the base and solvent choice.
The alkylation proceeded at reflux over several hours and
required about 2 equiv of ethyl bromoisobutyrate. In practice
2 equiv of NaOEt were initially added to form the dianion
of 11. A second charge of NaOEt (1 equiv) was added later
to drive the reaction to completion, reforming phenoxide that
had been consumed in the competing elimination reaction
noted earlier. Upon completion of reaction, the mixture was
quenched with aqueous phosphoric acid, the ethanol was
removed by evaporation, and tert-butyl methyl ether (MTBE)
was added. We were pleased to observe that the desired 12
was fully extracted into the aqueous layer by adjusting the
pH to 8 with NaOH. Following layer separation, reacidifi-
cation, and extraction of the product into fresh MTBE, drying
and concentration afforded 12 as a crystalline solid. While
other workers have recently noted difficulties with polym-
erization of ethyl methacrylate formed during alkylations
with bromoisobutyric esters, we did not encounter these
issues under our conditions.11
TsOH was replaced with methanesulfonic acid (MsOH)
when a quick screen of additional acids showed that it gave
a slightly better reaction profile. The MsOH-mediated
cyclization of 6 to provide 1 was performed in toluene at
reflux over about 6 h. The undesired byproduct 10, produced
in 5-10% yield (1:10 ratio of about 10:1), could be removed
from the reaction by extensive washes with aqueous acid (1
M HCl, 8-10 washes (!)) following workup. Concentration
of the organics followed by warming the product residue in
ethyl acetate then afforded crystalline 1 in about 55% overall
yield from 6. Using essentially the conditions outlined above,
we prepared 112 g of 1 from acyl hydrazide 2 to support
initial studies on the compound.
With a viable end-game strategy for the compound in
place, we turned to implementation of the convergent
approach to 1 outlined in Scheme 2. We envisioned the
carboxylic acid of Scheme 2 arising from 4-hydroxyphenyl-
butyric acid (11), itself readily accessed in one step from
commercially available 4-methoxyphenylbutyric acid via our
recently reported pyridinium hydrochloride melt demethy-
lation.9 We then sought to embed two design features into
the synthesis that would likely prove useful on scale (Scheme
5). First, we wanted to monoalkylate 11 directly, avoiding a
protection/deprotection sequence on the carboxylic acid.10
Second, we wanted to replace tert-butyl bromoisobutyrate
with ethyl bromoisobutyrate as the alkylating reagent. Direct
monoalkylation of 11 with ethyl bromoisobutyrate would
afford 12, which presumably could be isolated and purified
from the organic reaction mixture by acid/base partition. Use
Some alkylation reactions gave noticeable levels of an
impurity identified as diacid 13; this was produced by
saponification of 12 from sodium hydroxide found in some
commercial solutions of NaOEt in EtOH (eq 1). Addition
of ethyl acetate to the dianion solution consumed this
adventitious hydroxide, producing EtOH and sodium acetate,
and suppressing formation of 13.
The final conditions developed for the alkylationsNaOEt
(2 equiv then 1 equiv), ethyl acetate (1 equiv), EtOH, and
ethyl bromoisobutyrate (2.25 equiv)safforded 12 in yields
of 85% with 95-98% potency following the extractive
workup noted above (eq 2). Implementation on 22-L scale
(9) Schmid, C. R.; Beck, C. A.; Cronin, J. S.; Staszak, M. A. Org. Process
Res. DeV. 2004, 8, 670-673. For an alternative approach, see: Delhaye,
L.; Diker, K.; Donck, T.; Merschaert, A. Green Chem. 2006, 8, 181-182.
(10) The methyl ester of 11 has been alkylated with tert-butyl bromoisobutyrate,
see ref 5.
(11) Guo, J.; Erickson, G. A.; Fitzgerald, R. N.; Matsuoka, R. T.; Rafferty, S.
W.; Sharp, M. J.; Sickles, B. R.; Wisowaty, J. C. J. Org. Chem. 2006, 71,
8302-8305.
434
•
Vol. 11, No. 3, 2007 / Organic Process Research & Development

Scheme 6. Attempted direct approach to semicarbazide 17a
was smooth, with slightly higher yields (92-95%) of slightly
lower-potency material (93-95%) being obtained.
Interestingly, however, on 22-L scale we observed crystal
formation on quench of the alkylation reaction with phos-
phoric acid and ethanol removal; cooling of the resulting
slurry to 0 °C and filtration delivered 12 directly from the
reaction mixture. A heptane wash dissolved entrained organic
contaminants and delivered 12 of excellent purity (97-99%)
in very good yields (86-95%). This variant was employed
in the remaining 22-L runs to secure 12; overall, 8.43 kg of
12 was prepared, 3.57 kg using the initial extractive workup,
and 4.86 kg using the direct crystallization procedure.
For the second leg of the convergent synthesis, it was
envisioned that the coupling partner benzyl semicarbazide
14 could be obtained in one of two ways. Most directly, the
compound might be obtained via urea group installation on
4-methylbenzylhydrazine (15), as semicarbazide formation
is known to proceed primarily on the interior nitrogen in
similar substrates;12 purification would then be required to
separate 14 from the regioisomeric semicarbazide 16.
Compound 15 would be derived from 4-methylbenzaldehyde
via reduction of its corresponding hydrazone. Alternatively,
a monoprotected hydrazine could be combined with 4-me-
thylbenzaldehyde and the resulting hydrazone reduced in situ
to afford 15 regiospecifically protected on its terminal
nitrogen; installation of the urea group followed by depro-
tection would then deliver 14. While this latter approach
substituted a chemical step in place of a purification, it did
ensure a fully regiospecific delivery of 14; the direct approach
would depend critically on a means for purification of the
presumed mixture of 14 and 16 to obtain the desired
semicarbazide.
a Reagents and conditions: (a) NH2NH2 (excess), 50 psig H2, 10% Pd/C,
EtOH, rt; (b) TMS-NCO (1.5 equiv), IPA, rt, 16 h; (c) MsOH (1.0 equiv),
IPA; (d) IPA, reflux, 30 min.
(50 psig H2, 10% Pd/C, EtOH, rt).13 Reaction of 15 with
TMS-NCO resulted in an 85:15 mixture of 14 and 16,
respectively. Isolation of pure 14 from these mixtures proved
surprisingly problematic. Direct recrystallization of the
isomeric mixture from IPA or EtOH gave solutions at reflux,
but the crystals deposited on cooling showed no significant
enrichment in 14. Addition of MsOH to the mixture followed
by recrystallization from EtOH on small scale gave 14 as
its methanesulfonate (mesylate) salt 17, but in modest yield
(54%) with only a minor upgrade in purity (90%) over
regioisomer 18. Moreover, heating 17 at reflux in IPA for
more than a few minutes resulted in the contamination of
17 with significant amounts of the mesylate salt of 4-me-
thylbenzylhydrazine (19) via loss of the urea group. With
the need to deliver 1 against a timeline and purification
barriers still to overcome, we opted to deprioritize this direct
approach to 14.
The regiospecific approach to secure 14 proved more
successful (Scheme 7). We selected the commercially
available tert-butyl carbazate as the monoprotected hydrazine
source, with the rationale that 14 should be compatible with
the acid-mediated Boc deprotection conditions. Reaction of
the carbazate with 4-methylbenzaldehyde under the hydro-
genation conditions previously employed for the conversion
of 7 to 8 (50 psig H2, PtO2, THF) provided the Boc-protected
4-methylbenzylhydrazine 20. A further screen of platinum-
based catalysts identified 5% Pt/C as a more ideal choice
Pursuit of the direct approach (Scheme 6) began with
formation of 4-methylbenzylhydrazine 15 from 4-methyl-
benzaldehyde via reduction of the corresponding hydrazone,
prepared from 4-methylbenzaldehyde and excess hydrazine
(12) LeBoucq, J. J. Pharm. Chim. 1930, 11, 200-207.
(13) Kline, G. B.; Cox, S. H. J. Org. Chem. 1961, 26, 1854-1856.
Vol. 11, No. 3, 2007 / Organic Process Research & Development
•
435

Scheme 7. Regiospecific approach to semicarbazide 17a
Scheme 8. Coupling and triazolone formationa
a Reagents and conditions: (a) 50 psig H2, 5% Pt/C, THF, 50 °C, 8 h, 85%;
(b) TMS-NCO (1.5 equiv), IPA, rt, 16 h, 86%; (c) MsOH (1.1 equiv), DCM,
reflux, 16 h, 95%.
than PtO2, affording the compound in better quality as
evidenced by 1H NMR comparison. Filtration and concentra-
tion delivered 20 as an oil that crystallized slowly on
standing.
Reaction of 20 with TMS-NCO (1.5 equiv, IPA, rt, 16
h) afforded the desired Boc-protected semicarbazide 21,
which gratifyingly crystallized directly from the reaction
mixture as an easily filtered solid in very good yield (86%).
Boc group removal was smoothly effected by reaction with
1.1 equiv of MsOH in DCM at reflux. Again, we were
pleased to observe that mesylate 17 crystallized directly from
the reaction mixture; 19 was not formed under these
conditions. Cooling of the reaction mixture and filtration
delivered 17 in excellent yield (95-98%) and purity.
Minor scale-up modifications to this leg of the convergent
synthesis centered primarily around the hydrogenation step
used to deliver 20. In practice, the carbazate and aldehyde
were combined in IPA (replacing THF), and then 50% water
wet 5% Pt/C was added as catalyst. During optimization on
small scale, hydrogenations were observed to stall on
occasion, necessitating additional charges of catalyst. Inter-
estingly, we also observed formation of some p-xylene by
HPLC (∼5%), presumably derived from hydrogenolysis of
20san unexpected result from a platinum-based catalyst.
Scale-up in 10-gal autoclave equipment (IPA, 50 psig H2,
50 °C, 25 wt % load of 5% Pt/C, 4 h) gave reactions that
typically stalled at high but incomplete conversion; a second
equivalent charge of catalyst and continued hydrogenation
was used to drive the reaction to completion. Filtration and
concentration afforded 20 delivered as a solution in IPA
ready for direct use in the next step. By this method 9.61 kg
of material was prepared in five runs with an average weight
yield of 86.0% and average potency of 92.0%.
a Reagents and conditions: (a) oxalyl chloride (1.15 equiv), DMF (0.05 equiv),
ethyl acetate, rt, 30 min; (b) pyridine (2.3 equiv), ethyl acetate, 0 °C; (c) CSA
(1.1 equiv), ethyl acetate, reflux, 6 h; (d) Amberlyst 15, ethyl acetate, reflux, 1
h; MTBE (recrystallization), 50-55% overall from 12.
from the reaction mixture was similarly uneventful. The
compound was obtained in an average yield of 95.4% over
seven runs with an average potency of 97.8%. In this manner
7.25 kg of 17 was obtained.
With the successful preparation of both partners for the
convergent synthesis, we turned our attention to the coupling
and cyclization reactions (Scheme 8). The acid chloride 22
was easily formed from 12 by reaction with oxalyl chloride
and a catalytic amount of DMF in ethyl acetate; reaction
completion was conveniently determined by HPLC analysis
on the methyl ester of 12. Oxalyl chloride was chosen over
other acid chloride-forming reagents by virtue of its low
boiling point, affording the opportunity to remove any excess
reagent via concentration or distillation of solvent.14 Coupling
with 17 using 2.3 equiv of triethylamine proceeded smoothly
to afford cyclization precursor 23. An acidic aqueous wash
removed triethylammonium mesylate and triethylammonium
hydrochloride, leaving a solution of 23 for cyclization.
During optimization, the stoichiometry of acid chloride
formation was found to be significant, as residual oxalyl
chloride would consume 17 to form the undesired byproduct
Installation of the urea group on 20 was conducted on
22-L scale essentially as developed; crystalline 21 precipi-
tated directly from the reaction mixture, and heptane was
added to crystallize additional material. The compound was
obtained in an average yield of 80.1% across seven runs,
affording 9.17 kg of 21. Optimization revealed that only 1.3
equiv of TMS-NCO was needed to compensate for the basal
rate of reaction between TMS-NCO and isopropanol noted
earlier.
(14) The commonly employed procedure using catalytic DMF with stoichiometric
thionyl chloride (SOCl2) to convert carboxylic acids to acid chlorides results
in formation of dimethylcarbamoyl chloride, a confirmed animal carcinogen,
via an oxidative mechanism. See: Levin, D. Org. Process Res. DeV. 1997,
1, 182. We are unaware of any similar concerns with the DMF/oxalyl
chloride combination used here.
Similarly, Boc deprotection to reveal 17 scaled without
incident to 22-L equipment; direct filtration of this material
436
•
Vol. 11, No. 3, 2007 / Organic Process Research & Development

24. In practice this problem was solved by using a combina-
tion of stoichiometry and a subsequent concentration of
solvent to remove residual oxalyl chloride. With these
controls in place, the coupling proceeded in good yield and
purity, with levels of 24 reduced below 1%. Although 23
could be crystallized, filtrations, even on gram scale, were
unacceptably long, and so the decision was made to hold
the compound as a solution in ethyl acetate for the cyclization
reaction. Laboratory yields in excess of 90% were obtained
using this procedure.
not appreciably soluble in the aqueous phase of the saponi-
fication mixture, as analysis of the reaction mixture as it
progressed showed no signs of 25 in the aqueous phase, even
in those reactions not proceeding to completion. Following
completion of saponification, the aqueous layer containing
1 was separated and neutralized to pH 7. Ethyl acetate was
then added and the aqueous phase acidified to pH 1-2,
affording the compound as a solution in ethyl acetate in high
yield following layer separation. This two-stage neutraliza-
tion, pausing at pH 7, was put in place to avoid precipitation
of final compound in the form of a white gum that proved
difficult to dissolve. On scale-up, however, it was found that
slow neutralization of the aqueous phase with good agitation
gave precipitated 1 that was much more easily dissolved in
ethyl acetate; accordingly, the two-stage neutralization was
abandoned. In practice the exact yield was not obtained from
this step as it was combined with the final recrystallization
in a single operation.
We set out to optimize the new cyclization conditions on
23 to provide triazolone 25, examining solvent, acid, and
temperature to see if formation of byproduct 26 could be
minimized. The experiments revealed subtle but synthetically
useful results. Camphorsulfonic acid (CSA) gave more
favorable ratios of 25 to 26 than either methanesulfonic acid
or TsOH; higher temperatures produced more of 26, and ethyl
acetate gave better ratios of 25 to 26 than did toluene at
comparable temperatures. Reactions ran to completion with
substoichiometric amounts of CSA (0.5 equiv), but were
prohibitively long. Other acids either decomposed 23 or
failed to produce any desired product. The best conditions
identified (1 equiv CSA, ethyl acetate, reflux, 6 h) produced
a ratio of 25:26 of about 15:1 and afforded the desired
triazolone 25 in 55-60% yield. Attempts to convert 26 back
to 23 (and thus on to 25) via addition of TMS-NCO to the
reaction were unsuccessful. The extensive and tedious
aqueous acid washes used previously on smaller scale to
remove byproduct 10 from 1 were eliminated when it was
found that stirring the completed cyclization reaction with
Amberlyst 15 sulfonic acid resin for 1 h at reflux could
effectively remove 26 through simple filtration. Drying of
the filtered ethyl acetate layer and concentration afforded
crude product, which was recrystallized from tert-butyl
methyl ether (MTBE) to give crystalline 25 in 50-55%
overall yield from 12 with potencies in the range of 95-
98%.
On scale-up, the coupling and cyclization steps were
combined without isolation of the intermediate 23. In fact,
additional research demonstrated that aqueous washes to
remove triethylammonium methanesulfonate byproduct from
the coupling step could be eliminated; CSA could simply
be added to the completed coupling reaction to effect the
desired cyclization. The overall yield of the cyclization
reaction and the ratio of product 25 to main byproduct 26
were only mildly diminished by the presence of methane-
sulfonic acid in the cyclization. Accordingly, the washes were
omitted. The reactions scaled to 22 L essentially without
incident, providing 4.25 kg of 25 (52.0% average yield over
two steps) in five runs with an average potency of 96.0%
after recrystallization from MTBE.
A screen of suitable crystallization solvents for 1 identified
ethyl acetate as the ideal choice. This was puzzling as the
compound had extracted quite readily into ethyl acetate from
the acidic aqueous solution obtained in the saponification
step. This led to concerns that solutions of 1 extracted into
ethyl acetate might in fact be supersaturated and that 1 might
crystallize without warning from the solution. These concerns
were allayed when the extract solution was further evaluated
1
by H NMR. The extract was in fact a mixture of ethyl
acetate, water (5 wt %), and ethanol (2 wt %). Samples of
the solution held for extended times showed no propensity
to deposit 1 up to concentrations of about 100 mg/mL at 23
°C. The compound also showed much higher solubility in
water-saturated ethyl acetate (38.1 mg/mL, 23 °C) than it
did in anhydrous ethyl acetate (3.7 mg/mL, 23 °C). The
profound effect of small amounts of water in reactions,
workups, and isolations has been previously documented.15
This solubility difference of 1 between the extract solution
and anhydrous ethyl acetate was parlayed into a process to
deliver crystalline 1 of pharmaceutically acceptable purity.
Thus, the ethyl acetate extract from the saponification was
distilled at atmospheric pressure, collecting the lower-boiling
azeotropic mixture of water, ethanol, and ethyl acetate. Back
addition of anhydrous ethyl acetate was then performed until
the distillate temperature exceeded 76 °C, at which point
the ethyl acetate was essentially free of water and ethanol.
The solution was reduced to about 5 volumes by further
distillation and was then cooled and seeded at about 65 °C
to induce crystallization. Following cooling to ice-bath
temperature, the compound was filtered and rinsed with cold
ethyl acetate. The final step crystallization thus combined
solubility differences derived from compositional changes
in the solvent mixture with solubility differences derived
from temperature changes to deliver crystalline 1. The overall
process is shown in eq 3. Three lots of 1 totaling 3.58 kg
were prepared in 95.2% yield over both saponification and
crystallization from 25 in 99.2% potency vs a fully charac-
terized reference standard.
Optimization of the final-step saponification/recrystalli-
zation sequence to afford 1 revealed some interesting
features. Two-phase saponification of 25 (toluene, aq NaOH,
rt, 4 h), required two full equivalents of base to proceed to
completion, indicating that the triazolone NH proton was
acidic enough to undergo full deprotonation and consume
one equivalent of base. Curiously, however, this anion was
(15) Anderson, N. G. Practical Process Research & DeVelopment; Associated
Press: New York, 2000; Chapter 6.
Vol. 11, No. 3, 2007 / Organic Process Research & Development
•
437

Scheme 9. Preparation of cyclization precursor substrates
30a-fa
a Reagents and conditions: (a) aldehyde (1.1 equiv), 50 psig H2, Pt/C, THF,
50 °C; (b) R2NCO (1.1 equiv), IPA or DCM, rt, 2-8 h; (c) MsOH (1.1 equiv),
DCM, reflux, 8 h; (d) 22 (1.1 equiv), pyridine (2.2 equiv), ethyl acetate, 0 °C,
2 h; (e) n-PrNCO (1 equiv), THF, rt, 30 min.
The preparation of 1 from 4-methoxyphenylbutyric acid
and 4-methylbenzaldehyde was thus accomplished in eight
total steps, with a six-step longest linear sequence (from
4-methylbenzaldehyde), to deliver 3.58 kg of LY518674 in
32.5% overall yield.
Table 1
Having successfully prepared LY518674 using the new
triazolone ring-formation conditions, we wanted to briefly
examine the scope and limitations of this new methodology.
We chose to examine triazolone formation using our condi-
tions for compounds of the type previously prepared by
Mantlo and co-workers.5 As shown in Scheme 9, cyclization
substrates 30a-e were assembled in convergent fashion using
22 and the requisite substituted semicarbazide mesylates,
which were themselves prepared from the corresponding
aldehydes and isocyanates employing the regiospecific
preparation sequence used to secure 17. Substrate 30f was
prepared from acyl hydrazide 2 by reaction with n-propyl-
isocyanate.5 The cyclization substrates prepared spanned the
range of substitution possibilities for the corresponding
triazolones: N2-monoalkylated, N4-monoalkylated, and N2,N4-
dialkylated. Subjection of the substrates to the cyclization
conditions (CSA, ethyl acetate, reflux) for 30a-e afforded
the corresponding triazolones 31a-e captured in Table 1.
Several interesting findings emerged from this brief study.
The dialkylated substrates underwent cyclization much faster
and in higher yield than did those producing the N2-
monoalkylated triazolones, the slower rates of cyclization
for 30d and 30e presumably allowing side reactions of the
substrate to compete with the main reaction. Perhaps most
interestingly, compound 30f failed to undergo cyclization at
all, instead affording a quantitative yield of the CSA
hydrazone adduct 32 within about 30 min (eq 4).
precursor
R1
R2 time (h) yield (%) product
30a
30b
30c
30d
30e
30f
3,4-dimethylbenzyl Me
2-napthylmethyl
benzyl
3-methoxybenzyl
3,5-difluorobenzyl
H
1
94
98
96
47
39
100
31a
31b
31c
31d
31e
32
Et
2.5
2
16
17
n-Pr
H
H
n-Pr 0.5
promoted cyclizations of dialkylated substrates such as
30a-c have not been reported in the literature, so no
comparison could be made.
Conclusions
The synthesis of the PPARR agonist LY518674 was
accomplished on kilogram scale in a highly convergent
fashion using a novel acid-promoted triazolone-forming
methodology. The compound was prepared in 32.5% overall
yield from 4-methoxyphenylbutyric acid and 4-methylben-
zaldehyde, requiring eight total steps, and employing a six-
step longest linear sequence from the latter compound. A
selective alkylation on a phenoxide carboxylate allowed the
direct preparation of one of the convergent coupling partners,
and an unusual solvent effect enabled the installation of a
urea group on a hydrazine, permitting the preparation of the
On the basis of these limited studies, the acid-promoted
cyclization chemistry appeared to serve as a complementary
methodology to the established hydroxide-promoted approach
in that at least one substrate (6) undergoing successful
cyclization under acid conditions failed using the hydroxide-
promoted approach, and 30f, a successful substrate for
hydroxide-promoted cyclization,5 failed using CSA. Base-
438
•
Vol. 11, No. 3, 2007 / Organic Process Research & Development

other coupling partner. A combination of solvent composi-
tional and temperature changes was employed to provide an
effective crystallization of the final compound. Finally, the
scope and limitations of this new triazolone forming meth-
odology were briefly explored, and found to be complemen-
tary to existing base-promoted methodology.
as before, and hydrogenation was continued for 4 h. The
contents were cooled to rt and then filtered, and the autoclave
and filter cake were rinsed with IPA (15 L). The resulting
solution was concentrated to a weight of 5.33 kg. Gravimetric
assay gave a 31.7 wt % solution of product (1.9 kg, 8.04
1
mol, 85%) with an HPLC potency of 90.2%. H NMR
(DMSO-d6, 500 MHz): δ 8.20 (br s, 1 H), 7.13 (dd, 4 H),
4.61 (br s), 3.79 (s, 2 H), 2.25 (s, 3 H), 1.37 (s, 9 H).
2-(Aminocarbonyl)-2-[(4-methylphenyl)methyl]hydra-
zinecarboxylic Acid, 1,1-Dimethylethyl Ester (21). To a
22-L four-neck flask equipped with overhead stirring ap-
paratus, cooling bath, thermometer probe, and a 2-L addition
funnel was charged solution 20 in IPA (4988 g of a 28.6 wt
% solution, 1583 g, 6.70 mol). Additional IPA (3130 mL)
was added to dilute the solution. Trimethylsilyl isocyanate
(1179 mL, 8.71 mol, 1.3 equiv) was charged to the 2-L
addition funnel and added dropwise to the stirring solution
over 45 min while maintaining the 15-25 °C, and the
mixture stirred for 16 h. The reaction mixture was treated
with heptane (7800 mL), cooled to 5-10 °C, stirred for 0.5
h, and then filtered; the filter cake washed with heptane (2
× 1000 mL). The material was dried in vacuo at 30-35 °C
to a constant weight to afford the title compound as a white
Experimental Section
4-(2-Ethoxy-1,1-dimethyl-2-oxoethoxy)benzenebutano-
ic Acid (12). A 12-L three-neck round-bottom flask equipped
with an overhead, air-driven stirrer apparatus, thermometer/
thermocouple, condenser, nitrogen inlet, and heating mantle
was charged with ethyl acetate (450 mL) and sodium
ethoxide (21 wt % solution in EtOH, 3318 mL, 8.87 mol, 2
equiv). The resulting mixture was heated to reflux under a
nitrogen atmosphere and maintained at reflux for 30 min.
The mixture was then allowed to cool to slightly below
reflux, and 4-hydroxyphenylbutyric acid9 was added (800
g, 4.43 mol). The flask contents were reheated to reflux for
30 min, and ethyl 2-bromoisobutyrate was added (1954 mL,
13.32 mol, 3.0 equiv). After 1 h at reflux, the flask was
equipped with a 2-L addition funnel, which was charged with
additional sodium ethoxide (21 wt % solution in EtOH, 1660
mL, 4.43 mol, 1 equiv). The sodium ethoxide solution was
added dropwise to the refluxing reaction over 1 h. After an
additional 30 min at reflux, HPLC analysis showed the
reaction was complete. The flask contents were cooled to
5-10 °C and transferred to a 22-L bottom outlet flask. While
stirring, the mixture was quenched with 0.5 M phosphoric
acid (6000 mL), transferred to a 20-L evaporatory flask, and
concentrated in vacuo to remove EtOH. The resulting
aqueous slurry was held at 5-10 °C overnight. The slurry
was filtered and washed with water (2000 mL). The filter
cake was removed from the funnel and placed in a 20-L
evaporatory flask equipped with an overhead stirring ap-
paratus. Water (2000 mL) was added, and the resulting slurry
was stirred at ambient temperature for 30 min. The solids
were filtered and dried on the filter for 30 min and then
washed with heptane (2 × 2500 mL). The resulting solids
were dried in vacuo at 45 °C to afford the title compound
as an off-white solid (1247 g, 95.5%). 1H NMR (500 MHz,
CDCl3): δ 1.25 (t, 3 H), 1.57 (s, 6 H), 1.92 (m, 2 H), 2.35
(t, 2 H), 2.60 (t, 2 H), 4.23 (q, 2 H), 6.76 (d, 2 H), 7.03 (d,
2 H). 13C NMR (CDCl3): δ 14.1, 25.4, 26.3, 33.3, 34.2, 61.4,
119.4, 129.1, 129.4, 134.9, 153.7, 174.4, 179.7.
1
solid (1423 g, 84%). H NMR (DMSO-d6, 300 MHz): δ
1.35 (s, 9 H), 2.26 (s, 1 H), 4.45 (br s, 2 H), 6.06 (s, 2 H),
7.10 (s, 4 H), 8.88 (br s, 1 H). Anal. Calcd for C14H21N3O3:
C, 60.20; H, 7.58; N, 15.04. Found: C, 59.65; H, 7.34; N,
14.87.
1-[(4-Methylphenyl)methyl]hydrazinecarboxamide,
Monomethanesulfonate (17). To a 22-L 4- neck flask
equipped with overhead stirring apparatus, warming/cooling
bath, thermometer probe, condenser, and 500-mL addition
funnel was charged 21 (1100 g, 3.94 mol) and dichlo-
romethane (12 L). The mixture was warmed to 30-35 °C
to dissolve all solids, and then was cooled to 25-30 °C.
MsOH (398 g, 4.14 mol) was added dropwise over 30 min.
The water bath was replaced with a heating mantle and the
reaction solution heated at reflux for 20 h. The reaction
mixture was diluted with heptane (4 L), cooled to 10-20
°C, stirred for 30 min, and then filtered. The filter cake was
washed with heptane (2 × 1000 mL) and dried in vacuo at
35-45 °C to a constant weight to afford the title compound
as a white solid (1070 g, 98.6%). 1H NMR (DMSO-d6, 300
MHz): δ 1.35 (s, 9 H), 2.26 (s, 1 H), 4.45 (br s, 2 H), 6.06
(s, 2 H), 7.10 (s, 4 H), 8.88 (br s, 1 H). Anal. Calcd for
C10H17N3O4S: C, 43.62; H, 6.22; N, 15.26. Found: C, 43.33;
H, 6.21; N, 14.97.
2-[4-[3-[2,5-Dihydro-1-[(4-methylphenyl)methyl]-5-oxo-
1H-1,2,4-triazol-3-yl]propyl]phenoxy]-2-methylpropano-
ic Acid, Ethyl Ester (25). To 22-L four-neck flask equipped
with a thermometer/thermocouple, addition funnel, overhead
stirring apparatus, nitrogen inlet, and cooling bath was
charged 12 (1250 g, 4.247 mol), ethyl acetate (11,250 mL),
and DMF (16.4 mL, 5 mol %), and agitation was begun in
order to dissolve all solids. Oxalyl chloride (426 mL, 4.88
mol, 1.15 equiv) was charged to the addition funnel and then
added dropwise to the stirring reaction mixture over 45 min,
maintaining the temperature below 30 °C. After 30 min,
2-[(4-Methylphenyl)methyl]hydrazinecarboxylic Acid,
1,1-Dimethylethyl Ester (20). To a 10-gal stainless steel
autoclave was charged tert-butyl carbazate (1.25 kg, 9.46
mol) and IPA (2 L). Additional IPA (10 L) was added and
the resulting mixture heated to 35 °C. 4-Methylbenzaldehyde
(1.15 kg, 9.57 mol) was then added, followed by an IPA
rinse (0.3 L). The temperature was then raised to 50 °C and
the contents stirred for 1 h. In a separate vessel, 5% Pt/C
(0.3 kg) and water (0.3 kg) were combined, and IPA (4 L)
was added. The resulting slurry was added to the autoclave,
followed by rinses with IPA (2 × 1 L). The resulting mixture
was hydrogenated at 50 °C and 50 psig H2 for 4 h. Analysis
at that time showed an 83% conversion. An additional charge
of catalyst (0.3 kg) was prepared and added to the autoclave
Vol. 11, No. 3, 2007 / Organic Process Research & Development
•
439

HPLC analysis indicated the reaction was complete. The
reaction mixture was transferred to a 20-L evaporatory flask
and concentrated in vacuo to afford crude acid chloride 22.
A 22-L four-neck flask, equipped with a thermometer/
thermocouple, addition funnel, nitrogen inlet, overhead
stirring apparatus, and cooling bath, was charged with 17
(1169 g, 4.247 mol, 1 equiv), ethyl acetate (8750 mL), and
pyridine (790 mL, 9.77 mol, 2.3 equiv). The contents of the
flask were cooled to 0-5 °C. The acid chloride 22 was
dissolved in ethyl acetate (1000 mL), transferred to the
addition funnel, and added dropwise to the mixture contain-
ing 17 over 20 min, maintaining the pot temperature below
25 °C. The resulting mixture containing 23 was stirred at
ambient temperature for 1 h, then treated with CSA (1973
g, 8.494 mol, 2 equiv). The contents were heated to reflux
and allowed to stir overnight. HPLC analysis after 16.5 h
showed the reaction was complete. The reaction was cooled
to 20 °C and transferred into a 22-L bottom outlet flask
containing 1 N HCl (7500 mL). After stirring, the layers
were separated, and the organic layer was washed with sat’d
sodium carbonate solution (7500 mL) and water (1000 mL)
and then was dried (MgSO4).
ic Acid (LY518674, 1). To a 22-L four-neck flask equipped
with overhead stirring apparatus and thermometer probe was
charged compound 25 (800 g, 1.828 mol) and toluene (4000
mL) followed by 1 N NaOH (4023 mL, 4.023 mol, 2.2
equiv). The resulting biphasic mixture was stirred at ambient
temperature for 5 h. HPLC analysis at that point indicated
the reaction was complete. The layers were separated, and
the aqueous layer was transferred to a 22-L three-neck round-
bottom flask equipped with a thermometer, overhead stirring
apparatus, and addition funnel and was acidified to pH 2 by
dropwise addition of concentrated HCl (337 mL). The
resulting slurry was extracted with ethyl acetate (8000 mL),
and the layers were separated. The organic extracts were
transferred to a 22-L three-necked round-bottom flask
equipped with a distillation head and overhead stirring
apparatus. The mixture was concentrated by distillation of
the ethyl acetate to approximately 4000 mL. Ethyl acetate
(3600 mL) was added to the reaction vessel, and distillation
was continued until 3600 mL more of distillate were
recovered. The mixture was allowed to cool slowly to 60-
65 °C, at which point seed crystals (0.8 g) of the product 1
were added. The flask contents were allowed to cool slowly
until crystallization had initiated (55-57 °C), then the
mixture was cooled to 0-5 °C and stirred for 1 h. The
product was filtered, washed with cold ethyl acetate, and
dried in vacuo at 55 °C to afford the title compound as a
The ethyl acetate solution was transferred to a 22-L four-
neck flask equipped with a condenser, nitrogen inlet,
thermometer/thermocouple, and a heating mantle. The flask
was charged with Amberlyst-15 resin (1975 g), and the
mixture was heated to reflux for 1 h. After cooling to 20
°C, the material was gravity filtered to remove the resin and
the resin washed with ethyl acetate (2 × 2000 mL). The
filtrate was transferred to a 20-L evaporatory flask and
concentrated in vacuo to afford a tan solid. The crude
material was treated directly with MTBE (5000 mL) and
warmed to 45-50 °C until all solids were dissolved. The
solution was allowed to cool slowly while rotating in the
evaporatory flask at low rpm, inducing crystallization. The
material was then cooled to 0-5 °C and held for 1 h, filtered,
rinsed with cold MTBE (1500 mL), and dried in vacuo at
45 °C to a constant weight. The title compound was obtained
1
white solid (713.6 g, 95.3%). H NMR (DMSO-d6, 500
MHz): δ 1.46 (s, 6 H), 1.79 (m, 2 H), 2.25 (s, 3H), 2.35 (t,
2 H), 2.47 (t, 2 H), 4.70 (s, 2 H), 6.71 (d, 2H), 7.03 (d, 2H),
7.10 (m, 4 H). 13C NMR (DMSO-d6): δ 20.6, 25.0, 25.5,
27.7, 47.0, 78.3, 118.6, 127.4, 127.7, 128.9, 128.9, 134.4,
134.5, 136.4, 145.9, 153.4, 154.3, 175.0.
Acknowledgment
We thank Bill Diseroad, John Howell, Jim Marshall,
Cindy Steadham, and Kurt Willis for technical assistance.
1
Supporting Information Available
as a white solid (1027 g, 55.2%). H NMR (500 MHz,
Experimental procedures and characterization for com-
pounds 30a-f, 31a-e, and 32. This material is available
free of charge via the Internet at http://pubs.acs.org.
DMSO-d6): δ 1.14 (t, 3 H), 1.47 (s, 6 H), 1.76, (m, 2 H),
2.26 (t, 2 H), 2.49 (t, 2 H), 3.55 (s, 3 H), 4.15 (q, 2 H), 6.69
(d, 2 H), 7.05 (d, 2 H). 13C NMR (CDCl3): δ 14.0, 21.0,
25.3, 26.1, 28.1, 34.1, 48.3, 61.2, 79.0, 119.3, 127.8, 129.0,
129.3, 133.2, 134.8, 137.5, 147.2, 153.6, 155.8, 174.3.
2-[4-[3-[2,5-Dihydro-1-[(4-methylphenyl)methyl]-5-oxo-
1H-1,2,4-triazol-3-yl]propyl]phenoxy]-2-methylpropano-
Received for review February 15, 2007.
OP700040V
440
•
Vol. 11, No. 3, 2007 / Organic Process Research & Development