Initial Scale-Up and Process Improvements for the Preparation of a Lead Antibacterial Macrolone Compound

Article abstract of DOI:10.1021/op100199t

Macrolones are a novel class of potent antimicrobial agents that consist of a macrolide scaffold to which a quinolone unit is tethered by various linkers to the 4 -O-position of the cladinose sugar. In this paper is described a modified 13-step route to a

Full text of DOI:10.1021/op100199t

Organic Process Research & Development 2010, 14, 1393–1401
Initial Scale-Up and Process Improvements for the Preparation of a Lead
Antibacterial Macrolone Compound
ˇ
ˇ
Vlado Stimac,* Maja Matanovic´ Skugor, Ivana Palej Jakopovic´, Adrijana Vinter, Marina Ilijasˇ, Sulejman Alihodzˇic´, and
Stjepan Mutak‡
GlaxoSmithKline Research Centre, Zagreb Ltd., Prilaz baruna FilipoVic´a 29, HR-10000 Zagreb, Croatia
Abstract:
procedure was successfully applied using a laboratory 5 L
reactor to prepare an initial 350 g of compound 15. The material
was isolated as the free base, by precipitation in the form of an
amorphous solid with purity of 97.7% area by HPLC-UV. The
process does not include final crystallization or salt preparation.
Herein we describe the optimized synthesis and present
improved reaction conditions that significantly contributed to
the scalability, overall yield and final purity of the product 15.
Macrolones are a novel class of potent antimicrobial agents that
consist of a macrolide scaffold to which a quinolone unit is tethered
by various linkers to the 4′′-O-position of the cladinose sugar. In
this paper is described a modified 13-step route to a lead
compound in the series. Critical reaction steps in the medicinal
chemistry route were modified for an initial scale-up process, and
as a result, a synthetic procedure suitable for preparation of
multihundred gram quantities of the final product, with 98%
purity, has been developed. The new procedure does not require
any purification by column chromatography for any of the reaction
steps. The overall yield was increased from 5-8% in the medicinal
chemistry route to 27% in the improved procedure.
Results and Discussion
Synthesis of the Quinolone Building Block 6: Steps 1-5.
The synthesis of the quinolone moiety was recently described,4
and only minor modifications in the in the workup, isolation
and purification of the final product were required to furnish
2.0 kg of 6 in 66% overall yield with purity of >99.5% area.
The major changes involved using compounds 2 and 3, Scheme
1, without purification. Compound 4 was isolated by crystal-
lization from EtOH rather than n-hexane, and compound 5 was
crystallized from water rather than THF. In step 5 THF was
replaced with EtOH and compound 6 isolated directly from the
reaction mixture by adjusting the pH of the solution.
Preparation of iodoquinolone 6 was carried out in five linear
steps, starting from 2-fluoro-5-iodo-benzoic acid (1) on 500-700
g scales.
Reagents and conditions used in the final optimized synthetic
procedure:
Step 1: compound 1, oxalyl chloride (1.2 equiv), DMF (0.1
equiv), toluene (5 vol), 20 °C, 1.5-2 h, partially concentrated,
2 used in the next step as toluene solution.
Step 2: compound 2, TEA (1.3 equiv), dimethylaminoacry-
late (1.05 equiv), toluene (6 vol), 90 °C, 2 h, insoluble salts
filtered off, concentrated, EtOH added, toluene azeotropically
removed, 3 used in the next step as a suspension in EtOH.
Step 3: suspension of 3 in EtOH (3 vol), 2 M ethylamine in
THF (1.14 equiv), rt, 1.5 h, THF evaporated, 4 crystallized from
EtOH. Yield ) 67-70% starting from 1. HPLC-MS: 96.9%
area, HPLC-UV: 98.2% area.
Step 4: compound 4, K2CO3 (2 equiv), DMF (3 vol), 120
°C, 1 h, water (3 vol) added, crystals filtered off, aliquot dried.
Yield ) 97%. HPLC-MS: 99.7% area, HPLC-UV: 99.9%
area.
Introduction
Recently our team discovered a new class of antimicrobial
compounds which consist of a macrolide scaffold and a
quinolone unit, covalently connected by a suitable linker1-3
which were named “macrolones”. The recently prepared
analogue 15, Figure 1, is a key lead for the series, and the
combination of a 10-atom bis-ether chain linking the quinolone
moiety to an azithromycin core fulfils the medicinal chemistry
target product profile.4
The initial medicinal chemistry route for the preparation of
compound 15 consisted of thirteen reaction steps including
several chromatographic purifications, with an overall yield of
5-8%. A more efficient route which avoided expensive and
time-consuming chromatography was required to prepare suf-
ficient material to support preclinical in ViVo safety assessment
studies. Given the limited time available it was decided to adapt
the existing medicinal chemistry route for scale-up rather than
design a completely new approach. This modified synthetic
* Corresponding author. Telephone: +385 1 6051032. Fax: +385 1 6051019.
E-mail address: vlado.stimac1@zg.t-com.hr.
‡ Present address: Hercegovacˇka 99, HR-10000 Zagreb, Croatia.
ˇ
ˇ
(1) Hutinec, A.; Ðerek, M.; Lazarevski, G.; Sunjic´, V.; Cipcˇic´ Paljetak,
H.; Alihodzˇic´, S.; Erakovic´ Haber, V.; Dumic´, M.; Mutak, S. Bioorg.
Med. Chem. Lett. 2010, 20, 3244–3249.
ˇ
(2) Fajdetic´, A.; Cipcˇic´ Paljetak, H.; Lazarevski, G.; Hutinec, A.;
ˇ
ˇ
Alihodzˇic´, S.; Ðerek, M.; Stimac, V.; Andreotti, D.; Sunjic´, V.; Berge,
J. M.; Mutak, S.; Dumic´, M.; Lociuro, S.; Holmes, D. J.; Marsˇic´, N.;
Erakovic´ Haber, V.; Spaventi, R. Bioorg. Med. Chem. 2010, 18, 6559–
6568.
Step 5: compound 5, NaOH (3 equiv), EtOH/H2O 1:1 (8
vol), 75 °C, 35-45 min, cooled to 50 °C, 6 M aq HCl (1.3
vol) added (pH ) 6), cooled to rt, filtered off, dried. Yield )
96-100%. HPLC-MS 99.5% area, HPLC-UV 99.7% area.
By using these optimized conditions the initial small-scale
procedure was significantly improved. In summary, the number
ˇ
(3) Kapic´, S.; Cipcˇic´ Paljetak, H.; Alihodzˇic´, S.; Antolovic´, R.; Erakovic´
Haber, V.; Jarvest, R. L.; Holmes, D. J.; Broskey, J. P.; Hunt, E.
Bioorg. Med. Chem. 2010, 18, 6569–6577.
ˇ
ˇ
(4) Matanovic´ Skugor, M.; Stimac, V.; Palej Jakopovic´, I.; Lugaric´, Ð.;
ˇ
Cipcˇic´ Paljetak, H.; Filic´, D.; Modric´, M.; Ðilovic´, I.; Gembarovski,
D.; Mutak, S.; Erakovic´ Haber, V.; Holmes, D. J.; Ivezic´ Schoenfeld,
Z.; Alihodzˇic´, S. J. Bioorg. Med. Chem 2010, 18, 6547–6558.
10.1021/op100199t  2010 American Chemical Society
Published on Web 09/28/2010
Vol. 14, No. 6, 2010 / Organic Process Research & Development
•
1393

Figure 1. Outline of the synthetic route to compound 15 and key building blocks.
Scheme 1. Synthesis of 1-ethyl-6-iodo-4-oxo-1,4-dihydro-3-quinolinecarboxylic acid (6)
of organic solvents used in the route decreased from 8 to 4,
and evaporation of solutions to dryness was avoided. Intermedi-
ates 4 and 5 were isolated by a crystallization from ecologically
acceptable solvents and were used in the next steps without
drying. Only the final compound needed to be dried. Overall
yield on 6 was increased from 37% to 66%.
Synthesis of the Linker-Quinolone Building Block 12:
Steps 6-10. An outline of the improved route to compound
12 is given in Scheme 2.
Step 6: Small-Scale Procedure and Reaction Conditions. A
synthesis of the intermediate 8, step 6, started from com-
mercially available benzyloxyethanol (7) which was alkylated
with propargyl bromide (2 equiv) in THF using NaH (1.2 equiv)
as base at room temperature. After removal of the solvent, the
product was isolated by EtOAc/H2O extraction and evaporation
of the resultant organic layer to dryness.
Step 6: Optimization of the Reaction Conditions and
ImproVed Procedure. During examination of the reaction’s
parameters it was found that if toluene was used as the reaction
solvent, the quantity of propargyl bromide could be reduced to
1.5 equiv, and the reaction was complete after 1 h compared to
24 h in THF. The purity, assessed by HPLC-UV, was increased
from 71% to 83%. Three separate runs were performed starting
from 440, 600, and 700 mL of compound 7 all using a modified
isolation protocol of azeotropical removal of toluene, from the
crude reaction solution, by addition of EtOH. The resulting
solution of compound 8 in EtOH was used in the next step
without further manipulation, calculating the yield of 8 as 100%.
Step 7: Small-Scale Reaction Conditions and Scale-Up
Limitations. The next step was a Sonogashira coupling5 of
quinolone 6 and the linker 8. The issues in this step were mainly
related to usage of large amounts of the catalysts (CuI 15 mol
% and Pd(PPh3)2Cl2 10 mol %) and the solvents (TEA 12 vol,
MeCN 12 vol), as well as isolation of the product 9 by column
chromatography.
Step 7: Modifications of Reaction Conditions and Isolation
Procedure. The isolation of compound 9 was modified by
diluting the reaction mixture with water (40 vol), washing with
i-Pr2O at pH 12, and treatment of the aqueous layer with
activated charcoal at 60 °C. Acidification of the aqueous layer
to pH 8.5 resulted in a precipitation of the product that was
removed by a filtration. Although the yields were acceptable
(80-82%) and the purity excellent (>98% area as determined
by HPLC-UV), large amounts of solvents and reagents (12
vol MeCN, 12 vol TEA, 40 vol H2O, 30 vol i-Pr2O) were
wasted and limited the amount of starting compound 6 in the
5-L reactor to just 50 g. Attempts to avoid washing with i-Pr2O
and to precipitate the product from a mixture MeCN/TEA/H2O
resulted in resinous gum.
Several solvent systems were investigated, and a particular
effort was made to find an ecologically acceptable solution.6
Initial reactions on 1-g scale were performed in toluene, acetone,
i-PrOAc, and absolute EtOH; in all systems the reaction was
complete within 1 h. Simultaneously, the catalyst loading was
reduced to 5 mol % each without any detrimental effect on the
(5) Sonogashira, K. J. Organomet. Chem. 2002, 653, 46–49.
(6) Alfonsi, K.; Colberg, J.; Dunn, P. J.; Fevig, T.; Jennings, S.; Johnson,
T. A.; Kleine, H. P.; Knight, C.; Nagy, M. A.; Perry, D. A.; Stefaniak,
M. Green Chem. 2008, 10, 31–36.
1394
•
Vol. 14, No. 6, 2010 / Organic Process Research & Development

Scheme 2. Synthetic route and modified reaction conditions for a preparation of intermediate 12
Scheme 3. Stepwise hydrogenation of compound 9
rate of reaction. Toluene was the preferred choice as it was
used as the solvent in a previous step and would avoid the
necessity of solvent exchange. However, isolation of compound
9 as a tractable solid proved impossible.
Using EtOH as the solvent, on a 5-g scale as described
above, the reaction mixture was filtered to remove the catalysts,
and the majority of the TEA was evaporated under vacuum.
The resultant ethanolic solution of 6 was diluted with water
and acidified by addition of 6 M HCl. At pH 7.5 precipitation
of the product occurred, but the initial solid residue soon
transformed to a sticky gum.
Step 7: Final Modified Procedure. Repeating the above
procedure on a 10-g scale, but using only 1 mol % of the
catalysts, incurred an increase in reaction time from 1 to 3 h.
However, it was found that precipitation from a biphasic mixture
EtOH/H2O/toluene, 1:1:0.8 at pH 8.3-8.5, gave the product 9
in excellent purity as a free-flowing powder that was much
easier to filter and dry than the previously isolated gums.
By using this procedure six large-scale batches (300-500 g)
were performed, affording the compound 9 in high yield (81-86%)
and purity (97.7% area by HPLC-MS, 99.2% area by HPLC-UV).
Approximately 2.3 kg of the compound was produced.
NMR analysis of the solid residue was performed in order
to investigate possible formation of the TEA salt of the product
9 at relatively high pH 8.4. For the analysis, the material after
precipitation and filtration at pH 8.4 was used and compared
with an aliquot of the material isolated by precipitation at lower
pH 5.0. In the 1H NMR spectra of compound 9 after precipita-
tions at pH 8.4 (wet and dried) and pH 5.0 (dried), no TEA
signals were observed, confirming that compound 9 is isolated
as a free acid. The only differences observed were higher levels
of the residual solvents H2O and toluene in the wet product.
Step 8: Small-Scale Reaction Conditions and Scale-Up
Limitations. The small-scale (0.6-10 g) reactions of hydrogena-
tion of compound 9 were performed at 5-20 bar pressure at rt
overnight in a mixture MeOH/DCM, 3:1 (20 vol), in the
presence of 10% Pd/C as a catalyst. DCM was added to improve
the solubility of 9. In all batches the reaction stalled at
intermediate 9a, Scheme 3, requiring additional amounts of fresh
catalyst to complete the conversion.
After filtration of the catalyst and evaporation of the solvent,
in the first experiments compound 10 was isolated by column
chromatography. Crystallization of crude residue from EtOH
under controlled conditions resulted in compound 10 in
35-78% yields and with 83-93% purity, assessed by
HPLC-MS and HPLC-UV.
Step 8: Modifications of Reaction Conditions, Isolation, and
Final ImproVed Procedure. The first step in the optimisation
was to replace MeOH as the solvent with EtOH in order to
enable direct crystallization of the product 10. The goal was
also to reduce the amount of DCM which is required because
of low solubility of starting compound 9 in MeOH. Initial
reactions were performed in EtOH/DCM ) 4:1 (10 vol) at
5-10 bar pressure, and monitoring of the progress revealed a
profile similar to that of small-scale reactions. After removal
of the catalyst by filtration, the following three different isolation
methods of compound 10 were examined:
Procedure P1. The solution was concentrated, and the
product was dissolved in DCM, extracted to water at pH 10,
and then precipitated by acidification of the aqueous layer to
pH 6.
Procedure P2. After evaporation of DCM into the
resulting EtOH solution, water was added, the solution
was concentrated to a smaller volume and the precipitation
occurred.
Procedure P3. The solution was concentrated, and the
product crystallized from EtOH by heating to 60 °C and slowly
cooling to 5 °C.
Vol. 14, No. 6, 2010 / Organic Process Research & Development
•
1395

Scheme 4. Undesired ether bond hydrogenolysis and formation of byproduct 9b at elevated temperature
Table 1. Conditions and results of hydrogenation of compound 9
batch
9/g
Pd cat./g
Tmin - Tmax/°C
H2/bar
isolation procedure
yield/%
HPLC-MS/%; HPLC-UV/%
1
2
3
173
200
1683
60.0 + 9.0
29.1 + 3.0
258.0 + 96.2
20-33
11-27
8-20
5-10
5
3-5
P1
P2
P3
49.7
61.3
82.4
67.1/85.8
85.6/96.2
94.1/97.6
Isolation procedure P3 proved to be the best, affording the
highest yield and final purity of 10, as will be shown below.
After initial laboratory reactions the hydrogenation was
repeated in a 47-L reactor starting from 173 g of 9. The first
batch showed a highly exothermic reaction as the temperature
of the mixture rose from 20 to 33 °C while cooled by water.
HPLC-MS analysis revealed that this inadequate temperature
control caused ether bond hydrogenolysis resulting in the
formation of 33% of 6-propylquinolone 9b as the only byprod-
uct, Scheme 4.
Although the byproduct 9b does not take an active part in
the next step and can be tolerated in crude product, the goal
was to reduce its formation in order to increase the yield of the
product 10. Therefore, additional batches were carried out by
controlled addition of hydrogen gas while maintaining the
temperature below 20 °C. By using this procedure the amount
of 9b was reduced to 5.9% area, determined by HPLC-MS,
and 2.4% by HPLC-UV, batch 3, Table 1.
As mentioned previously and as can be seen in Table 1,
isolation procedure P3 gave the highest yields and purities.
Step 9: Small-Scale Reaction Conditions and Scale-Up
Limitations. Addition of the final three-atom portion of the linker
was achieved by a Michael reaction of compound 9 with
acrylonitrile. Initially, acrylonitrile was used as solvent for the
reaction with DBU as base at 80 °C for 24 h. This combination
required a tedious isolation procedure due to extensive formation
of polymers and afforded desired cyano derivative 11 in low
yield and purity (66% by HPLC-UV). Furthermore, in two
runs exothermic polymerisation of acrylonitrile resulted in a
runaway reaction.
Step 9: Modifications of Reaction Conditions and Isolation,
and Final ImproVed Procedure. A definite improvement was
achieved when 40% aq NaOH was used as base and methyl
isobutyl ketone (MiBK) as solvent, allowing a lower molar ratio
of acrylonitrile to be used. The reaction was completed after
2-3 h at 10-20 °C, and cyano derivative 11 was isolated in
70-80% yields. The optimised isolation protocol included
extraction of 11 into the aqueous phase at pH 11 and
precipitation from aqueous solution at pH 6.0-6.5, affording
the desired product in >90% purities, as determined by HPLC.
Further improvements of the reaction conditions included
complete elimination of the organic solvent and reducing the
quantity of NaOH used.
Five batches of compound 11 were successfully prepared
on 96-380 g scales, affording the product in high yields and
purities. The reactions were carried iout n 10% aq NaOH at
room temperature and were completed in only 50 min. The
product was isolated by diluting with water and adding conc.
HCl until precipitation at pH 6.3 started. Compound 11 was
used in the next step as a wet powder, and aliquots weredried
in order to calculate the yield. The yields were 92-97% and
purities 92-95% by HPLC-MS and 94-95% by HPLC-UV
methods.
Step 10: Small-Scale Reaction Conditions and Scale-Up
Limitations. In the first experiments a hydrolysis of the nitrile
11 was attempted in ∼60% sulfuric acid, Scheme 2. These
conditions afforded quinolone diacid 12 usually contaminated
with 20-30% of alcohol 10 formed by retro-Michael reaction.
The resulting alcohol 10 is a competitor to 2′-O-acetylazithro-
mycin 13 in the esterification step (step 12), and thus formation
of this byproduct needs to be suppressed. In the original
synthesis compound 12 was isolated by extraction with DCM
at pH 3.5 followed by purification by column chromatography.
Step 10: Modifications of Reaction Conditions and Isolation,
and Final ImproVed Procedure. Virtually complete suppression
of the retro-Michael reaction could be achieved by careful
temperature control of the reaction. This resulted in initial
formation of the intermediate amide 11a, Scheme 5, at low
temperature and high concentration of acid followed by hy-
drolysis to the desired product 12 at a higher temperature and
lower concentration of acid.
According to this protocol the nitrile 11 was added to conc.
H2SO4 portionwise with stirring over 1 h, maintaining the
temperature at <5 °C. After addition was complete, the solution
was then stirred at rt for 18 h, monitoring by HPLC-MS to
confirm completion of reaction. The reaction was treated
dropwise with water over a 5-h period, keeping the temperature
of the mixture below 8 °C until ∼0.3 mol/L concentration of
the acid was reached. The temperature was then raised to 70
°C and the mixture stirred overnight. The approach of diluting
1396
•
Vol. 14, No. 6, 2010 / Organic Process Research & Development

Scheme 5. Two-step hydrolysis of nitrile 11 to diacid 12
the reaction mixture with water, adjustment to pH 9.3-9.5, and
filtration of the resultant crystalline product. Starting from 700 g
of azithromycin, 609 g of compound 13 was produced in a
single batch. The yield was 88% and the purity 98.0% area,
assessed by HPLC-MS.
Synthesis of the Final Compound 15: Steps 12 and 13.
To complete site-selective acylation of compound 13 at
C4′′-OH by dicarboxylic acid 12, the combination of EDAC
× HCl/DMAP was used.9 After careful comparison of the
relative acidity of the two carboxylic groups in diacid 1210 it
was concluded that a less acidic carboxylic group attached to
the quinolone ring would be a much poorer substrate for
esterification than the aliphatic acid at the chain terminus. This
selectivity is explained by preferred proton transfer from the
more acidic carboxylic group to the N-atom of carbodiimide,
followed by an addition of carboxylate anion to form the O-acyl
urea active intermediate. Several acylation procedures were
investigated, e.g. mixed anhydride of diacid 12 with pivaloyl
chloride and DCC, but the most suitable reagent was EDAC
× HCl/DMAP. Formation of up to 5% of macrolide byproduct
was regularly observed with various molar ratios of acylating
agents.
Based on this finding and other exploratory experiments, the
final protocol for modified preparation of 14 includes the use
of 2.5 equiv of EDAC × HCl and 3 equiv of DMAP. DMF
was replaced as the reaction solvent by DCM, which simplified
the workup procedure and ensured complete dissolution of
reagents and product at low temperatures. In spite of the longer
reaction time, the large-scale reactions were conducted and
maintained at low temperature, which resulted in reduced levels
of byproduct. Furthermore by increasing the concentration of
the reactants the reaction rate was accelerated and the level of
starting material 13 was kept to <1% in the final product, as
monitored by HPLC-MS. The purification method in the final
step 13, described below, has allowed almost complete elimina-
tion of this byproduct.
The workup included washing the crude reaction mixture
with aq NaHCO3 and water to remove the EDAC and the
majority of DMAP. Evaporation of the solvent to dryness gave
the desired product 14. Since the required amount of compound
14 had been produced in this single batch, no further optimi-
sation of the isolation procedure was carried out at this stage.
Additional experimentation on a 1-g scale showed us that DCM
can be easily exchanged with MeOH, and compound 14 was
forwarded to the next step as MeOH solution. A total of 1178 g
of compound 14 was produced in the form of a syrup with
HPLC-MS purity of 91% area. It transpired that intermediate
14, thus formed, can be used in the next step without any
detrimental effect on the course of the next reaction.
the acid with water was adopted, rather than the dropwise
addition of acid to water, because of problems with the high
viscosity of the 11a/H2SO4 solution. Isolation included further
dilution with water and adjustment of pH to 0.1-0.5 by slow
addition of 40% NaOH over 2.5 h, keeping the temperature of
the mixture below 8 °C. Dicarboxylic acid 12 was collected
by suction filtration in ∼90% yield. HPLC-MS analysis of
dried compound 12 revealed less than 5% of retro-Michael
product, quinolonic alcohol 10.
Two batches were carried out starting from 224 and 300 g
of compound 11 affording 215 and 281 g of 12, respectively,
with 88.1% purity, assessed by HPLC-UV.
Step 11: Small-Scale Reaction Conditions and Scale-Up
Limitations. The second key intermediate in the convergent
route to 15, the 2′-O-acetyl derivative 13, was prepared from
the parent azithromycin with slight excess of acetic anhydride
(1.1 equiv) in the presence of NaHCO3 (4.5 equiv) in DCM at
rt, Scheme 6. Upon completion of the reaction, the mixture was
diluted with water, and the product was isolated by extraction
with DCM and evaporation to dryness.
By using this procedure compound 13 was always ac-
companied with 2′,4′′-O,O-diacetyl byproduct 13a, Figure 2,
in up to 5%, as determined by HPLC-MS.
Exploratory experimentation suggested that reducing the
amount of acetic anhydride to 1.0 equiv suppressed formation
of the byproduct but also resulted in incomplete conversion of
starting azithromycin. Replacement of DCM with environmen-
tally more acceptable EtOAc and reducing the amount of
NaHCO3 gave a similar product profile.
Step 11: Modifications of Reaction Conditions and Isolation
Procedure. Eventually it was discovered that formation of 13a
could be avoided by using i-PrOH as the solvent. Since the
reaction rate of the desired acylation at the 2′-OH group is
enhanced due to intramolecular catalysis of vicinal dimethy-
lamino group on the C-3′ position,7,8 the secondary hydroxyl
in the solvent is only a competitor to the less reactive 4′′-OH
group. After applying these modified conditions (i-PrOH 3 vol,
NaHCO3 1.5 equiv, Ac2O 1.1 equiv, 0 °C to rt, 20 h)
HPLC-MS analysis revealed less than 0.1% of diacetyl
byproduct 13a. Isolation of the product 13 included dilution of
Step 13: Modified Extraction Method and Isolation of Solid
Amorphous Product 15. Methanolysis of 2′-O-acetyl protecting
group was performed at 55 °C overnight in two batches starting
from 698 and 480 g of crude 14. After the completion of the
reaction, in addition to the product 15, TLC and HPLC indicated
(9) Tanikawa, T.; Asaka, T.; Kashimura, M.; Suzuki, K.; Sugiyama, H.;
Sato, M.; Kameo, K.; Morimoto, S.; Nishida, A. J. Med. Chem. 2003,
46, 2706–2715.
(7) Wessjohann, L. A.; Zhu, M. AdV. Synth. Catal. 2008, 350, 107–112.
(8) Hoffmann, H. M. R.; Schrake, O. Tetrahedron: Asymmetry 1998, 9,
1051–1057.
(10) Perrin, D. D.; Dempsey, B.; Serjeant, E. P. pKa Prediction for Organic
Acids and Bases; Chapman and Hall: London, 1981.
Vol. 14, No. 6, 2010 / Organic Process Research & Development
•
1397

Scheme 6. Modified reaction conditions for a preparation of the target compound 15
traces of azithromycin, the quinolone diacid 12, DMAP, and
two macrolide byproduct. Extensive exploration of purification
and isolation procedures, in order to avoid column chromatog-
raphy, resulted in a modified extraction method.
process represents a solid base for further development of a
large-scale process. The number of different organic solvents
used in the process is reduced, as well as the amount of catalysts.
Isolation of intermediates 2, 3, 8 and 14 is avoided and other
intermediates are obtained by crystallization/precipitation and
filtration, preferably from water or EtOH. In all these steps, it
is worth to point out that the most significant modification is
probably the isolation of compound 9 in Sonogashira reaction
step. It was found that it can be isolated by precipitation from
a mixture of EtOH/H2O/toluene at a defined pH value, thereby
affording the product in high purity and yield. The most
important improvement in the quality of the final product 15
in the modified procedure was achieved by performing extrac-
tions at ascending pH values resulting in 97-99% purity, thus
avoiding column chromatography. Overall yield from 1 to 15
was increased from 5%, obtained in the laboratory-scale
procedure, to 27%.
After concentration of the MeOH solution, the reaction
mixture was partitioned between i-PrOAc and water. Gradient
extraction at various pH values resulted in the following: at
pH 4.0 i-PrOAc extracted quinolone diacid 12 and DMAP; at
pH 4.8, 5.0, and 5.3 DCM extracted macrolide byproduct;
at pH 5.8 DCM extracted pure product 15 (probably as
hydrochloride salt), while unreacted azithromycin remained in
the aqueous layer; at pH 8.8 DCM extracted the product 15 as
a free base. The organic layer at pH 8.8 was concentrated almost
to dryness and DCM exchanged for i-PrOAc. The solid product
precipitated after slow addition of the i-PrOAc solution to a
large volume of i-Pr2O. i-PrOAc was chosen since it showed
encouraging results in initial crystallization experiments. A total
of 682 g of the final product 15 was prepared in the form of an
amorphous powder with purity of 96.4% area by HPLC-MS
and 97.7% area by HPLC-UV, and in 65% yield starting from
compound 13. All impurities were not more than 0.5% area
each, as determined by HPLC, and their structures have not
been disclosed in this paper.
Experimental Section
All commercial reagents (Merck, Sigma-Aldrich) were used
as provided unless otherwise indicated, and all solvents are of
the high purity unless otherwise noted.
HPLC-UV and HPLC-MS analyses were carried out on
a system comprising an Agilent 1100 HPLC system equipped
with diode array detector (Agilent Technologies, Waldbronn,
Germany) and a Micromass ZQ 2000 single quadrupole mass
spectrometer (Waters, Milford, USA), operating in electrospray
ionization (ESI) positive mode. 1D and 2D NMR spectra (1H,
APT, COSY, HSQC, HMBC) were recorded at 25 °C in
DMSO-d6 with TMS as the internal standard on Bruker Avance
DRX500 spectrometer using QNI probe and Bruker Avance
Conclusions
The initial laboratory 13-step procedure for the preparation
of compound 15 was significantly improved, and the new
1
DPX300 spectrometer using a dual H/13C probe. DSC was
performed on a Mettler Toledo DSC 822e calorimeter equipped
with a refrigerated cooling system. The sample was heated in
a pin-holed aluminium pan at heating rate 10 °C/min from 25
to 300 °C. A nitrogen purge at 50 mL/min was maintained over
the sample. Reaction flow was monitored by thin layer
chromatography (TLC) on Merck Kieselgel 60 (230 -400
Figure 2. 2′,4′′-O,O-Diacetyl byproduct formed in step 11.
1398
•
Vol. 14, No. 6, 2010 / Organic Process Research & Development

mesh) using specifically solvent systems indicated in the
protocol. I2, UV-light (254 nm), and H2SO4, followed by heating
to >120 °C, were used for detection.
Equipment. The reactions were performed in a 5-L jacketed
glass reactor, equipped with a warming-cooling circulation
bath, thermostat, electromechanical stirrer, thermometer, distil-
lation condenser, reflux condenser, addition funnel, and Bu¨chner
funnel. The whole apparatus can be attached to a vacuum station
or be filled with nitrogen.
138.6, 136.76, 136.3, 136.2, 134.0, 133.8, 117.8, 117.6, 99.4,
87.5, 58.8, 44.5, 15.7, 13.7.
1-Ethyl-6-iodo-4-oxo-1,4-dihydroquinoline-3-carboxylic Acid
Ethyl Ester (5). Compound 4 (616.5 g, 1.58 mol) was dissolved
in DMF (2.0 L) at 22 °C. K2CO3 (431.0 g, 2 equiv) was added,
and the suspension was stirred at 125 °C for 1 h (TLC in EtOAc/
n-hexane ) 1:1 revealed no starting material). The mixture was
cooled to 25 °C, water (2.0 L) was added, and the suspension
was stirred for 15 min. The precipitate was filtered off through
a Bu¨chner funnel, and the reactor and the cake were washed
with water (2 × 400 mL). An aliquot (1.281 g) of the wet cake
(707.7 g wet) was dried in vacuum oven at 70 °C until constant
weight, and 1.029 g of dry title compound 5 was obtained. The
707.7 g of wet cake contained 568.5 g of dry title compound 5
(yield ) 97%). HPLC-MS: 99.7%, HPLC-UV: 99.9%, mp
2-Fluoro-5-iodo-benzoyl Chloride (2). A 5-L reactor was
filled with N2 gas. Toluene (3.5 L) and DMF (42 mL) were
introduced, and then 2-fluoro-5-iodo-benzoic acid (1) (700 g,
2.63 mol) was added to the mixture. Oxalyl chloride (269.0
mL, 1.2 equiv) was introduced dropwise during 1 h at 22 °C.
The reaction mixture was stirred at rt for an additional 1 h and
monitored by TLC in a solvent system, EtOAc/n-hexane ) 1:1.
The excess of oxalyl chloride was removed by evaporation
under reduced pressure at 50 °C to distill ∼250 mL of toluene/
oxalyl chloride mixture. The resulting solution of the title
compound 2 was cooled to 20-25 °C and directly forwarded
into the next step.
3-Dimethylamino-2-(2-fluoro-5-iodo-benzoyl)acrylic Acid Eth-
yl Ester (3). Into a toluene solution of compound 2 (2.63 mol)
from previous step cooled at 25 °C, was added a mixture of
ethyl 3-(dimethylamino)acrylate (414.3 g, 1.1 equiv) and TEA
(475 mL, 1.3 equiv) in toluene (700 mL). The resulting solution
was stirred at 90 °C for 2 h and the reaction monitored by TLC
in a solvent system EtOAc/n-hexane ) 1:1. The mixture was
cooled to 22 °C, and insoluble salts were filtered off through a
Bu¨chner funnel. The cake was washed with toluene (400 mL).
The mother liquor was concentrated under reduced pressure at
50 °C almost to dryness. To the thick solution was added EtOH
(1.2 L) and the mixture evaporated further to remove residual
toluene. To the resulting suspension was added EtOH (2.5 L)
(total volume of 3.0 L), and the suspension of the title compound
3 was used in next step.
1
(DSC): 173-177 °C. H NMR (300 MHz, DMSO) δ: 8.71
(s, 1H), 8.50 (d, 1H), 8.05 (dd, 1H), 7.64 (d, 1H), 4.39 (q, 2H),
4.23 (q, 2H), 1.36 (t, 3H), 1.29 (t, 3H). 13C NMR (75 MHz,
DMSO) δ: 171.3, 164.3, 149.2, 140.6, 138.0, 134.8, 129.9,
119.6, 110.5, 90.0, 59.7, 47.9, 14.2, 14.1.
1-Ethyl-6-iodo-4-oxo-1,4-dihydro-quinoline-3-carboxylic Acid
(6). Compound 5 (589.2 g, 1.59 mol) was suspended in water
(2.5 L) and EtOH (2.3 L) at 22 °C. NaOH (190.0 g, 3 equiv)
was added over 5 min; the mixture was stirred at 78 °C for 35
min. The solution was cooled to 60 °C, and the pH was adjusted
from 13.7 to 6.0 using 6 M HCl (820 mL). The resulting
suspension was cooled to 5 °C and filtered through a Bu¨chner
funnel. The reactor and the cake were washed with a mixture
H2O/EtOH ) 4:1 (500 mL). The wet cake (775.6 g) was dried
in a vacuum oven at 80 °C until constant weight, yielding 528.1
g of dry title compound 6 (yield ) 97%). HPLC-MS: 99.5%.
HPLC-UV: 99.7%. mp (DSC): 228-232 °C. 1H NMR (300
MHz, DMSO) δ: 8.48 (s, 1H), 8.44 (d, 1H), 8.11 (dd, 1H),
7.87 (d, 1H), 4.22 (q, 2H), 3.64 (m, 1H), 1.28 (t, 3H), 1.23 (m,
2H), 1.11 (m, 2H). 13C NMR (75 MHz, DMSO) δ: 171.4,
164.1, 148.6, 140.6, 139.9, 134.4, 129.2, 120.0, 110.3, 90.3,
59.8, 34.63, 14.2, 7.5.
({[2-(2-Propyn-1-yloxy)ethyl]oxy}methyl)benzene (8). Ben-
zyloxyethanol (7) (344 mL, 2.42 mol) was added to toluene
(2.4 L) in a 5-L reactor under N2 at 10 °C. NaH (w ) 60% in
mineral oil, 116 g, 1.2 equiv) was added portionwise over 1 h
while keeping the temperature between 10-20 °C. The mixture
was stirred at 10 °C for 45 min. Propargyl bromide (w ) 80%
in toluene, 404 mL, 1.5 equiv) was added dropwise over 1 h
while keeping the temperature below 20 °C. The mixture was
stirred at 25 °C for 1 h until TLC in EtOAc/n-hexane ) 1:4
showed no starting 7. Water (1.8 L) was added dropwise over
30 min while keeping the temperature below 20 °C. The mixture
was stirred for an additional 10 min, the layers were separated,
and the organic layer was washed with 10% aq NaHCO3 (450
mL). The organic layer was concentrated under reduced pressure
to a smaller volume (∼500 mL) and abs. EtOH (2 L) added.
The solution was concentrated to a smaller volume (∼1 L),
fresh abs. EtOH (2 L) was added, and the solution of the title
compound 8 was used in the next step. HPLC-UV: 82.5%.
1H NMR (300 MHz, DMSO-d6) δ: 7.33-7.35 (m, 5H), 4.57
(s, 2H), 4.21 (d, 2H), 3.70-3.74 (m, 2H), 3.63-3.67 (m, 2H),
3-Ethylamino-2-(2-fluoro-5-iodo-benzoyl)acrylic Acid Ethyl
Ester (4). To the suspension of compound 3 in EtOH (3 L,
2.63 mol) at 10 °C cold was added a solution (4 °C) of ethyl
amine in THF (2 M, 1.60 L, 1.22 equiv), and the mixture was
stirred at 22 °C for 1.5 h. TLC in the solvent system EtOAc/
cyclohexane ) 1:1 revealed no starting 3 but one product. THF
was evaporated under reduced pressure (450 mbar, temp. of
the solution 48 °C) to ∼3 L of solution. The solution was slowly
cooled to 10 °C (crystallization occurred at 35 °C), and the
resulting suspension was stirred for 15 min. The crystals were
filtered off through a Bu¨chner funnel, and the cake was washed
with cold EtOH (2 × 300 mL). A quantity of 741.2 g of wet
cake was obtained. An aliquot (1.564 g) of the wet cake was
dried in a vacuum oven at 50 °C for 4 h, and 1.456 g of dry
title compound 4 was obtained. The 741.2 g of wet cake
contained 690.0 g of dry title compound 4 (yield ) 67% starting
from compound 1). HPLC-MS: 96.9%. HPLC-UV: 98.2%.
1
mp (DSC): 118-122 °C. H NMR (300 MHz, DMSO) δ:
7.77-7.71 (m, 1H), 7.60 (bs, 1H), 7.55 (dd, 1H), 7.02 (dd, 1H),
3.89 (q, 2H), 3.50 (q, 2H), 1.21 (t, 3H), 0.94 (t, 3H). 13C NMR
(75 MHz, DMSO) δ: 186.9, 166.0, 159.8, 159.5, 156.3, 138.7,
Vol. 14, No. 6, 2010 / Organic Process Research & Development
•
1399

2,43 (t, 1H). 13C NMR (75 MHz, DMSO) δ: 138.4, 128.5,
128.4, 127.8, 127.6, 73.3, 69.2, 61.1, 58.4, 19.7.
added, and the suspension was evaporated to dryness. The crude
residue was suspended in EtOH (1370 mL) and the suspension
charged to a 5-L glass reactor. The suspension was heated to
60 °C and stirred for 30 min until all solid material dissolved.
The solution was cooled to 5 °C and the resulting suspension
stirred at that temperature for 2 h. The suspension was filtered
through a Bu¨chner funnel and the cake washed with EtOH/
H2O ) 1:1 mixture (300 mL). Wet cake (1561 g) was dried in
a vacuum oven at 40 °C until constant weight, yielding 1093 g
of the title compound 10 (yield ) 82.4%). HPLC-MS: 94.1%;
1-Ethyl-4-oxo-6-[3-({2-[(phenylmethyl)oxy]ethyl}oxy)-1-pro-
pyn-1-yl]-1,4-dihydro-3-quinolinecarboxylic Acid (9). Into the
solution of compound 8 (w ≈ 82%, 501.5 g, 1.36 equiv) in
abs. EtOH (2.5 L) from the previous step were added TEA (800
mL, 4 equiv), CuI (2.77 g, 1 mol %), 1-ethyl-6-iodo-4-oxo-
1,4-dihydro-quinoline-3-carboxylic acid (6) (500.0 g, 1.46 mol),
and Pd(PPh3)2Cl2 (10.2 g, 1 mol %). The suspension was heated
to 57 °C and stirred for 80 min. TLC of the solution in DCM/
MeOH/NH4OH ) 90:15:1.5 revealed complete conversion of
starting 6. The solution was concentrated under reduced pressure
to 2.5 L volume and cooled to 10 °C. Water (2.5 L) and toluene
(2.0 L) were added, and the pH was adjusted from 9.4 to 8.3
by dropwise addition of 6 M aq HCl (220 mL). The resulting
suspension was stirred at 10 °C for 20 min. The precipitate
was filtered off through a Bu¨chner funnel, and the reactor and
the cake were washed with toluene (100 mL), then water (500
mL) and EtOH (600 mL), respectively. An aliquot (2.076 g)
of wet cake (674.7 g) was dried in a vacuum oven at 65 °C for
5 h, yielding 1.517 g of dry title compound 9. The 674.7 g of
wet cake contained 493.0 g of dry title compound 9 (yield )
1
HPLC-UV: 97.6%; weight loss (105 °C): 0.65%. H NMR
(300 MHz, DMSO-d6) δ: 9.02 (s, 1H), 8.19 (s, 1H), 7.98 (d,
1H), 7.84 (d, 1H), 4.62 (q, 2H), 3.51 (t, 2H), 3.43-3.39 (m,
4H), 2.86 (t, 2H), 1.95-1.85 (m, 2H), 1.43 (t, 3H). 13C NMR
(75 MHz, DMSO-d6) δ: 177.6, 166.3, 148.7, 140.5, 137.5,
135.1, 125.6, 124.7, 118.2, 108.0, 72.2, 69.4, 60.4, 49.1, 31.2,
30.8, 14.7.
6-[3-({2-[(2-Cyanoethyl)oxy]ethyl}oxy)propyl]-1-ethyl-4-
oxo-1,4-dihydro-3-quinolinecarboxylic Acid (11). Water (2.6 L)
was added into the reactor followed by the addition of crude
NaOH (280 g, 7 equiv). After all NaOH dissolved, compound
10 (320.4 g, 1 mol) was added and the solution cooled to 0 °C.
Acrylonitrile (330 mL, 5 equiv) was added dropwise through
an addition funnel over 20 min while maintaining temperature
at 0 °C. The temperature was alowed to rise to 20 °C over 30
min (a suspension was formed) and the mixture stirred for 50
min until TLC in DCM/MeOH/NH4OH ) 90:15:1.5 revealed
no starting 10. Water (500 mL) and EtOH (300 mL) were
added, and the solution was cooled to 0 °C. HCl (conc., 605
mL) was added dropwise through an addition funnel over 50
min to adjust the pH to 6.3 while keeping the temperature at
0-10 °C. The resulting suspension was stirred at 0 °C for 2 h.
The precipitate was filtered off through a Bu¨chner funnel and
the cake washed with a water (700 mL)/EtOH (300 mL)
mixture, yielding 413.2 g of wet title compound 11. An aliquot
(2.935 g) was dried in a vacuum oven at 50 °C until constant
weight, yielding 2.582 g of dry title compound 11. The 413.2
g of wet cake contained 362.1 g of the title compound 11 (yield
) 97%). HPLC-MS: 94.7%; HPLC-UV: 94.9%; mp (DSC):
66-74 °C, peak at 71.2 °C. 1H NMR (300 MHz, DMSO-d6)
δ: 8.57 (s, 1H), 8.09 (d, 1H), 7.68 (d, 1H), 7.59 (dd, 1H), 4.34
(q, 2H), 3.62 (t, 2H), 3.57-3.59 (m, 2H), 3.49-3.51 (m, 2H),
3.42 (t, 2H), 2.76 (t, 2H), 2.75 (t, 2H), 1.34 (t, 3H). 13C NMR
(75 MHz, DMSO-d6) δ: 175.4, 167.2, 146.7, 137.1, 136.9,
132.6, 128.1, 125.1, 119.2, 116.5, 110.1, 69.5, 69.3, 65.2, 47.3,
31.0, 30.7, 14.4.
1
83.5%). HPLC-MS: 97.7%; HPLC-UV: 99.2%. H NMR
(300 MHz, CDCl3) δ: 8.78 (s, 1H), 8.57 (d, 1H), 7.83 (dd,
1H), 7.57 (d, 1H), 7.38-7.27 (m, 5H), 4.60 (s, 2H), 4.48 (s,2H),
4.39 (q, 2H), 3.82-3.80 (m, 2H), 3.72-3.69 (m, 2H), 1.59 (t,
3H). 13C NMR (75 MHz, CDCl3) δ: 177.7, 166.7, 148.1, 138.5,
136.6, 130.6, 128.4, 126.5, 121.0, 116.5, 109.3, 87.8, 84.3, 73.4,
69.5, 59.1, 49.8, 14.6.
1-Ethyl-6-{3-[(2-hydroxyethyl)oxy]propyl}-4-oxo-1,4-dihy-
dro-3-quinolinecarboxylic Acid (10). An 8.0-L glass reactor was
filled with DCM (4.14 L) and compound 9 (1904.4 g wet,
1683.3 g calculated on dry substance). The suspension was
stirred at 20-25 °C for 10 min to dissolve the material. The
resulting solution was charged to a 47-L hydrogenator. The glass
reactor was washed with EtOH (5.0 L) that was charged to the
hydrogenator. The glass reactor was filled with EtOH (5.0 L)
to which was added 10% Pd/C (497.2 g wet, 258.0 g dry). The
resulting suspension was charged to the hydrogenator. The glass
reactor was washed with EtOH (6.5 L) that was charged to the
hydrogenator. The reaction mixture was cooled to 8 °C and
stirred at 75% of the value of max rpm. The hydrogen pressure
was set to 3 bar and kept at that value. The temperature of the
mixture was raised to 19.9 °C over 15 min and then started to
decrease. Hydrogen pressure was increased to 5 bar, and the
mixture was stirred at 15-20 °C for 2 h. Hydrogen consumption
was significantly decreased during that period. After 2 h
HPLC-MS showed 28.9% of the product 10, 63.9% of the
intermediate 9a, and 5.3% of a byproduct. Additional amounts
of 10% Pd/C (200.0 g wet, 96.2 g dry) were added, and the
hydrogenation was continued at 5 bar pressure and 15-20 °C
for an additional 4 h until TLC in DCM/MeOH/NH4OH ) 90:
15:1.5 revealed complete conversion to the product 10. Hy-
drogen was evacuated and the reactor filled with N2. The
reaction mixture was filtered through a sintered glass funnel,
and the catalyst was washed with 2 × 2.5 L DCM and 1.0 L
EtOH, respectively. The mother liquors were concentrated on
a rotavap to a smaller volume (∼0.5 L). EtOH (1370 mL) was
6-[3-({2-[(2-Carboxyethyl)oxy]ethyl}oxy)propyl]-1-ethyl-4-
oxo-1,4-dihydro-3-quinolinecarboxylic Acid (12). To the reactor
was added conc. H2SO4 (1.0 L), and the mixture was cooled to
0 °C. Wet compound 11 (300 g dry calculated, 0.81 mol) was
added portionwise over 30 min, while keeping the temperature
at 0-5 °C. The suspension was stirred at 0 °C for 1 h and then
at rt overnight. The solution was cooled to 0 °C, and water
(1.6 L) was added dropwise over 5 h while keeping the
temperature at 4-8 °C. The solution was stirred at 20 °C for
1 h and at 70 °C overnight. The solution was cooled to 0 °C,
and water (1.0 L) was added portionwise over 30 min. NaOH
(40% aq 1.10 L) was added portionwise at 4-8 °C over 2.5 h
1400
•
Vol. 14, No. 6, 2010 / Organic Process Research & Development

to adjust the pH to 0.1. The resulting suspension was stirred at
4 °C for 45 min and filtered through a Bu¨chner funnel. The
cake was washed with water (1.0 L) and then with a water (700
mL)/EtOH (300 mL) mixture. The cake was dried in a vacuum
oven at 70 °C until constant weight, yielding 280.72 g of the
title compound 12 (yield ) 88.5%). HPLC-MS: 80.1%;
HPLC-UV: 88.1%. 1H NMR (300 MHz, DMSO-d6) δ: 9.02
(s, 1H), 8.18 (d, 1H), 7.98 (d, 1H), 7.85 (dd, 1H), 4.60 (q, 2H),
3.62 (m, 2H), 3.47-3.53 (ov, 4H), 3.41 (t, 2H), 2.85 (t, 2H),
2.45 (t, 2H), 1.87 (m, 2H), 1.43 (t, 3H). 13C NMR (75 MHz,
DMSO-d6) δ: 177.4, 172.5, 166.1, 148.4, 140.2, 137.3, 134.9,
125.4, 124.5, 118.0, 107.4, 69.5, 69.3, 69.2, 66.2, 48.9, 34.7,
30.9, 30.5, 14.5.
2′-O-Acetyl-9a-methyl-9-dihydro-9a-aza-9a-homoerythromy-
cin (13). Into a 5-L glass reactor was introduced N2 gas, and
then 2-propanol (2.1 L), azithromycin (700.0 g, 0.875 mol),
and NaHCO3 (112.0 g, 1.33 mol) were added. The mixture was
cooled to 0-5 °C, and acetic anhydride (99.4 mL, 1.05 mol)
was slowly added over 15 min through an addition funnel. The
reaction mixture was warmed to rt and stirred for 2 h. To the
reaction mixture was added water (1.4 L), and the pH was
adjusted to 9.3-9.5 using 10% aq NaOH (300 mL). Additional
amounts of water (700 mL) were added, and the mixture was
cooled to 5-10 °C. The resulting suspension was stirred for
2 h and then filtered. The cake was washed with a water/2-
propanol mixture (500 mL, 1:1) and then with water (2 × 350
mL). The wet cake (929 g) was dried in a vacuum oven at
45-50 °C and then at 75-80 °C until constant weight to
provide 609.9 g of the title compound 13 (yield ) 88.3%).
HPLC-MS: 98.0%; mp (DSC): 180-198 °C.
4′′-O-(3-{2-[3-(3-Carboxy-1-ethyl-4-oxo-1,4-dihydro-quino-
line-6-yl)propoxy]ethoxy}propionyl)-9-deoxo-9a-methyl-9a-aza-
9a-homoerythromycin A (15). The reactor was filled with N2,
and DCM (dry, 1.0 L) was added. Compound 12 (280.0 g, 88%,
1.3 equiv) was added and the mixture cooled to 0 °C. EDAC
× HCl (241.4 g, 2.6 equiv) was added and the mixture stirred
for 5 min. Compound 13 (384.6 g, 0.486 mol) and DMAP
(178.2 g, 3 equiv) were added, and the mixture was stirred at
3 °C overnight. NaHCO3 (sat. aq solution, 1.4 L) was added,
and the layers were separated. The organic layer was concen-
trated under reduced pressure at 30 °C to an oily residue. The
residue was dissolved in i-PrOAc (2.0 L) and the solution
concentrated (∼ 600 mL of the solvent was collected in a
receiving flask). Water (1.0 L) was added, the pH was adjusted
from 9.5 to 6.7 by using AcOH (60 mL), and the layers were
separated. The organic layer was concentrated under reduced
pressure, resulting in 698 g of oily residue of the compound
14. The residue (theor. 566 g of 14) was dissolved in MeOH
(5.0 L) and the solution stirred at 55 °C overnight. MeOH was
evaporated and the oily residue (620 g) dissolved in i-PrOAc
(1.2 L). Water (1.5 L) was added, and the pH was adjusted
from 7.3 to 4.0 by using 6 M HCl (140 mL). The aqueous layer
was separated and washed with i-PrOAc (3 × 500 mL). The
aqueous layer was separated and washed with DCM (3 × 500
mL) at pH 4.8, 5.0, and 5.3 (40% aq NaOH), respectively. The
product was extracted from the aqueous layer with DCM (3 ×
500 mL) at pH 5.8. Onto the DCM layer at pH 5.8 was added
water (0.5 L), and the pH was adjusted to 8.8 using 10% aq
NaOH. The DCM layer was separated and concentrated under
reduced pressure to a smaller volume (∼0.5 L). i-PrOAc (0.5
L) was added, and the solution was concentrated under reduced
pressure to a smaller volume. Fresh i-PrOAc (0.4 L) was added
and the solution added dropwise over 3.5 h to a stirring i-Pr2O
(5.0 L). The resulting suspension was filtered through a Bu¨chner
funnel, and the cake was washed with i-Pr2O (1.0 L) and dried
in a vacuum oven at 40 °C, yielding 340.8 g of the title
compound 15 (yield ) 62.4%). HPLC-MS: 96.4%; HPLC-
UV: 97.7%. During the filtration a precipitation occurred in the
mother liquor. The residue was filtered off through a Bu¨chner
funnel, washed with i-Pr2O (100 mL), and dried. The second
crop yielded 16.7 g of the title compound 15 (yield ) 3.0%).
HPLC-MS: 99.0%; HPLC-UV: 99.1%. 1H NMR (300 MHz,
DMSO-d6) δ: 9.03 (s, 1H), 8.18(d, 1H), 7.98(d, 1H), 7.83 (dd,
1H), 4.91 (d, 1H), 4.73 (dd, 1H), 4.59 (t, 2H), 4.55 (d, 1H),
4.43 (d, 1H), 4.33 (m, 1H), 4.17(dd, 1H), 3.66 (m, 1H), 3.64
(m, 2H), 3.50 (ov, 2H), 3.47 (ov, 1H), 3.45 (ov, 1H), 3.43 (ov,
2H), 3.38(t, 2H), 3.22 (s, 3H), 3.05 (dd, 1H), 2.81(t, 2H), 2.67
(ov, 1H), 2.67 (ov, 1H), 2.59(m, 2H), 2.40 (m, 1H), 2.35 (dd,
1H), 2.31 (d, 1H), 2.21 (s, 3H), 2.18 (s, 3H), 2.11 (t, 1H), 1.88
(ov, 1H), 1.85 (ov, 2H), 1.85 (ov, 1H), 1.78 (m, 1H), 1.66 (dd,
1H), 1.59 (m, 1H), 1.51(d, 1H), 1.42 (t, 3H), 1.37 (m, 1H),
1.27 (dd, 1H), 1.12 (s, 3H), 1.1 (s, 3H), 1.09 (ov, 1H), 1.08 (d,
3H), 1.07 (dd, 3H), 1.03 (d, 3H), 1.01 (s, 3H), 0.96 (d, 3H),
0.94 (d, 3H), 0.84 (d, 3H), 0.79 (t, 3H). 13C NMR (75 MHz,
DMSO-d6) δ: 177.4, 177.0, 170.9, 166.1, 148.4, 140.2, 137.3,
134.9, 125.5, 124.5, 118.0, 107.4, 102.0, 94.3, 82.6, 78.0, 77.3,
76.3, 74.9, 73.5, 72.4, 72.0, 70.5, 69.3, 69.6, 69.3, 68.6, 66.7,
66.0, 64.8, 62.1, 61.3, 48.8, 44.5, 41.6, 40.2, 35.6, 34.9, 34.2,
31.1, 31.0, 30.6, 27.3, 25.9, 22.0, 21.6, 20.8, 20.5, 17.6, 14.5,
10.8, 8.8, 6.6, 4.9.
Acknowledgment
We thank Biserka Metelko and Gorjana Lazarevski for NMR
ˇ
support, Darko Filic´ for DSC analysis, and Vitomir Sunjic´ and
John Berge for the discussion and critical reading of the
manuscript. A special thanks goes to Ður{ica Lugaric´ for her
help and technical expertise.
Received for review July 21, 2010.
OP100199T
Vol. 14, No. 6, 2010 / Organic Process Research & Development
•
1401