Synthesis of erythromycin derivatives via the olefin cross-metathesis reaction

Article abstract of DOI:10.1021/jo049737n

Olefin cross metathesis (CM) was applied to the synthesis of 6-O-substituted erythromycin derivatives. The reactions were catalyzed by transition metal alkylidene complexes, particularly bis(tricyclohexylphosphine)benzylidine ruthenium (IV) dichloride (Grubbs' first-generation catalyst). This approach allowed for the elaboration of the 6-O-allyl group of highly functionalized macrolides at various stages of the synthetic sequence, affording 6-O-3-aryl-propenyl products with excellent E-selectivity. Little or no self-dimerization of the reacting components was found in the crude mixtures. Preliminary kinetic data accounts for the observed cross-selectivity based on substrate reactivity and steric factors.

Full text of DOI:10.1021/jo049737n

Syn th esis of Er yth r om ycin Der iva tives via th e Olefin
Cr oss-Meta th esis Rea ction
†
‡
Margaret C. Hsu,* Adam J . J unia, Anthony R. Haight, and Weijiang Zhang
Process Chemistry Research and Development, Global Pharmaceutical R&D, Abbott Laboratories,
1
401 Sheridan Road, North Chicago, Illinois 60064
margaret.hsu@abbott.com
Received February 13, 2004
Olefin cross metathesis (CM) was applied to the synthesis of 6-O-substituted erythromycin
derivatives. The reactions were catalyzed by transition metal alkylidene complexes, particularly
bis(tricyclohexylphosphine)benzylidine ruthenium (IV) dichloride (Grubbs’ first-generation catalyst).
This approach allowed for the elaboration of the 6-O-allyl group of highly functionalized macrolides
at various stages of the synthetic sequence, affording 6-O-3-aryl-propenyl products with excellent
E-selectivity. Little or no self-dimerization of the reacting components was found in the crude
mixtures. Preliminary kinetic data accounts for the observed cross-selectivity based on substrate
reactivity and steric factors.
In tr od u ction
ABT-773 (1) is a semi-synthetic ketolide anti-infective
candidate previously selected for clinical development at
1
Abbott Laboratories. Beginning with erythromycin A (2),
a sequence of chemical reactions are applied to effect the
following overall transformations: (a) elaboration at the
6
-hydroxyl group, (b) 11,12-cyclic carbamate formation
from the corresponding diol, and (c) cladinose hydrolysis
2
and oxidation at the 3-position. Of these three transfor-
mations, the construction of the 6-O-3-(3′-quinolyl)-
propenyl side chain poses the greatest opportunity for
exploring various synthetic approaches. Herein, we report
the application of olefin cross metathesis (CM) to eryth-
romycin derivatives.
The application of olefin metathesis in organic synthe-
3
ses has gained popularity in recent years due, in large
part, to the commercial availability of well-defined cata-
†
Current address: Department of Chemistry, 907 Washington
Street, Northwestern University, Evanston, IL 60202.
‡
Current address: Theravance Inc., 901 Gateway Blvd., South San
Francisco, CA 94080.
(1) (a) Or, Y. S.; Ma, Z.; Clark, R. F.; Chu, D. T.; Plattner, J . J .;
Griesgraber, G. U.S. Patent 5,866,549, 1999. (b) Or, Y. S.; Ma, Z.; Clark,
R. F.; Chu, D. T.; Plattner, J . J .; Griesgraber, G. U.S. Patent 6,028,-
1
81, 2000. (c) Or, Y. S.; Ma, Z.; Clark, R. F.; Chu, D. T.; Plattner, J . J .;
Griesgraber, G. U.S. Patent 6,075,133, 2000.
2) Stoner, E. J .; Peterson, M. J .; Ku, Y.; Cink, R. D.; Cooper, A. J .;
(
Deshpande, M. N.; Grieme, T.; Haight, A. R.; Hill, D. R.; Hsu, M. C.;
King, S. A.; Leanna, M. R.; Lee, E. C.; McLaughlin, M. A.; Morton, H.
E.; Napier, J . J .; Plata, D. J .; Raje, P. S.; Rasmussen, M.; Riley, D.;
Tien, J . J .; Wittenberger, S. J . U.S. Patent 6,437,106 B1, 2002.
(3) For recent reviews on olefin metathesis, see: (a) Handbook of
Metathesis; Grubbs, R. H., Ed.; Wiley-VCH: Weinheim, 2003. (b)
Trnka, T. M.; Grubbs, R. H. Acc. Chem. Res. 2001, 34, 18-29. (c)
Furstner, A. Angew. Chem., Int. Ed. 2000, 39, 3012-3043. (d) Roy,
R.; Das, S. K. Chem. Commun. 2000, 519-529. (e) Philips, A. J .; Abell,
A. D. Aldrichimica Acta 1999, 32, 75-90. (f) Armstrong, S. K. J . Chem.
Soc., Perkin Trans. 1 1998, 371-388. (g) Grubbs, R. H.; Chang, S.
Tetrahedron 1998, 54, 4413-4450. (h) Ivin, K. J . J . Mol. Catal. A:
Chem. 1998, 133, 1-16. (i) Randall, M. L.; Snapper, M. L. J . Mol. Catal.
A: Chem. 1998, 133, 29-40. (j) Schuster, M.; Blechert, S. Angew.
Chem., Int. Ed. Engl. 1997, 36, 2036-2056.
lysts such as the Grubbs’ (3)4 and the Schrock’s (4)5
catalysts. A cursory overview of the chemical literature
reveals that C-C double bond formation via olefin
(4) Schwab, P.; Grubbs, R. H.; Ziller, J . W. J . Am. Chem. Soc. 1996,
118, 100-110.
1
0.1021/jo049737n CCC: $27.50 © 2004 American Chemical Society
Published on Web 05/01/2004
J . Org. Chem. 2004, 69, 3907-3911
3907

Hsu et al.
SCHEME 1
SCHEME 2
metathesis has been utilized extensively for intramo-
lecular ring-closing (RCM) and for ring-opening reactions
(
including ring-opening metathesis polymerization), but
6
few examples of olefin CM reactions are known. Un-
doubtedly, the intermolecular reactions are complicated
by competition between the desired CM and the undes-
ired self metathesis to produce homodimers (Scheme 1).
Nonetheless, these literature precedents suggest a po-
tential for expanding the scope of olefin CM to complex
substrates, especially applicable as a late-stage tool in
designing a total synthesis. Inspired by the availability
dependent on the catalyst loading when 5, 10, and 25
mol % catalysts were compared (entries 2, 6, and 7). At
the highest loading of 25 mol % 3, 75% product yield was
obtained after refluxing for 65 h (entry 8).
2
of 6-O-allyl erythromycin derivatives, our strategy was
to perform an intermolecular CM reaction between 6-O-
allyl protected erythronolides (5) and vinylquinoline (6)
to generate the alkene product 7 with the E-configuration
The ratio of the two CM substrates was examined next.
The typical conditions employ an extra equivalent of 6
relative to 8. However, it was observed that the yield
after 20 h was only slightly lower when the same molar
amounts of 8 and 6 were employed, but with the tradeoff
of increased amounts of impurities (entries 7 and 9).
When 3.0 equiv of 8 and 1.0 equiv of 6 were subjected to
the reaction conditions with 10 mol % catalyst, an
impressively high yield of 64% was obtained after 20 h,
and 79% yield after 65 h (entries 10 and 11). On the other
hand, when the reaction was overwhelmed with 5.0 equiv
of 6, the opposite trend was observed in which a low yield
of 23% was obtained after 20 h (entry 12). The kinetics
of these three reactions were plotted in Figure 1. The
graph indicates that, although it is theoretically possible
for these three reactions to reach comparable yields, the
kinetics of product appearance has a significant depen-
dence on the stoichiometry of the two olefinic components.
(
eq 1).
Resu lts a n d Discu ssion
Table 1 summarizes the results of olefin CM reactions
between 8 and 6 in methylene chloride with 3 as the
catalyst. A general method was adopted to generate data
for comparison based on the reaction of 1.0 equiv of 8
and 2.0 equiv of 6 in the presence of 10 mol % 3 for 20 h.
It was speculated that an excess of one of the olefin
components would drive the conversion forward to yield
9
(see Scheme 2). Compound 6 was chosen to be the
excess reagent because it is more readily available than
the highly derivatized macrolide, and any unreacted
portion or its homodimer would not hinder product
isolation.
At ambient temperature, the conversion was unaccept-
ably low under the typical reaction conditions (entry 1).
Longer reaction times at reflux temperature significantly
improved the yield, reaching 71% after 7 days (entries
2
-5). Not surprisingly, the reaction progress was also
(5) (a) Fox, H. H.; Yap, K. B.; Robbins, J .; Cai, S.; Schrock, R. R.
Inorg. Chem. 1992, 31, 2287-2289. (b) Schrock, R. R.; Murdzek, J . S.;
Bazan, G. C.; Robbins, J .; DeMare, M.; O’Regan, M. J . Am. Chem. Soc.
1
990, 112, 3875-3886.
6) For recent reviews on olefin cross metathesis, see: (a) Vol. 2,
(
In general, the olefin CM reaction between 8 and 6
produced very little undesired byproducts. Analyzed by
Chapter 2.8 of ref 3a. (b) Conner, S. J .; Blechert, S. Angew. Chem.,
Int. Ed. 2003, 42, 1900-1923.
3
908 J . Org. Chem., Vol. 69, No. 11, 2004

Synthesis of Erythromycin Derivatives
TABLE 1. Olefin CM betw een 8 a n d 6 in th e P r esen ce of 3
macrolide
8
(equiv)
3-vinyl-quinoline
catalyst
3
(mol %)
impurities (% PA)b
6
temp
(°C)
reaction
time (h)
product 9
entry
(equiv)
(%)a
10
11
12
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
3
3
1
2
2
2
2
2
2
2
2
1
1
1
5
10
10
10
10
10
5
25
25
25
10
10
10
ambient
40
40
40
40
40
40
40
40
20
20
65
91
168
20
20
65
20
20
65
20
5
35
59
65
71
27
49
75
45
64
79
23
2.1
2.8
1.5
1.1
1.1
ND
6.5
5.4
10.8
1.0
ND
1.4
ND
ND
1.1
1.4
1.8
ND
0.8
2.0
0.6
ND
ND
0.6
ND
ND
ND
1.3
ND
ND
1.1
1.5
2.6
13.7
25.0
ND
1
1
1
0
1
2
40
40
40
a
Yield based on the limiting substrate, by HPLC potency relative to a standard solution of 9. b Impurities with >0.5% peak area (PA)
based on HPLC integration; not adjusted for UV response factor at 235 nm. ND ) <0.5% PA.
SCHEME 3
F IGURE 1. Kinetics of product formation is dependent on
the stoichiometry of reactants. The plot label indicates the
stoichiometry of 6 relative to 8 in the reaction mixtures (e.g.,
means 5 equiv of 6 and 1 equiv of 8).
SCHEME 4
5
HPLC, the peak integration consistently furnished >95:5
ratio in favor of the desired E-configuration in the newly
formed disubstituted olefin,7 and >90% of the macrolide
was typically accountable as either starting material 8
or product 9. The anticipated vinylquinoline homodimer
,8
1
1 and macrolide homodimer 12 were observed in small
quantities when an excess of 6 was present, indicating a
slow depletion of 8 to nonproductive pathways. When 8
was present in equal amounts or greater, dimer 12 grew
significantly (entries 9-11). These results suggest that
macrolide dimerization occurs at a much slower rate than
productive CM and is therefore suppressed when 6 is
present in excess. The slow dimerization of 6 is consistent
with the literature report that stilbene formation via
dimerization of styrene is a slow process in the presence
of the Schrock’s catalyst (4).5b
(
PA) by HPLC). These data allowed us to conclude that
the product formation was an irreversible process.
The small amounts of homodimers 11 and 12 found in
these CM reactions led us to investigate their individual
reactivity under the reaction conditions. After refluxing
in methylene chloride for 20 h in the presence of Grubbs’
catalyst 3, macrolide 8 was converted to a complex
mixture consisting of recovered 8 (39.6% PA), cinnamyl
ether 10 (11.9% PA), and homodimer 12 (43.6% PA).
Styrene was also identified by HPLC (1.0% PA), indicat-
ing that 3 was initiated as proposed in Scheme 3. When
Along with the desired product, a cinnamyl impurity
1
0 was also generated in these CM reactions. Based on
the mechanism, 10 can result from 13, which can be
derived either directly from 3 and 8 or via 15 and the
free styrene released from the catalyst (Scheme 3).
The stability of the desired product 9 was challenged
to determine if the reactions were susceptible to second-
ary metathesis reactions. In the presence of 1.0 equiv of
styrene under reaction conditions, 9 was recovered
without any detectable conversion to 10. In this case,
some styrene dimerized to give stilbene (∼20% peak area
1
0 equiv of 1-pentene was included at the beginning of
the reaction (Scheme 4), a large amount of the coupled
product 16 (87.1% PA, 1.8:1 E/Z selectivity) was de-
tected.7 Being an unhindered alkene, 1-pentene dimer-
izes quickly to form 4-octene. It can be rationalized that
due to substrate sterics, 8 is not as reactive as unhin-
dered terminal alkenes in intermolecular metathesis
reactions.
,8
1
0
(9) (a) Engelhardt, F. C.; Schmitt, M. J .; Taylor, R. E. Org. Lett.
(
7) The E/Z selectivity does not change over the course of the
2001, 3, 2209-2212. (b) BouzBouz, S.; Cossy, J . Org. Lett. 2001, 3,
1451-1454. (c) Crowe, W. E.; Zhang, Z. J . J . Am. Chem. Soc. 1993,
115, 10998-10999.
(10) 4-Octene was identified by GC, but its yield and E/Z selectivity
were not determined.
reaction.
(
8) The rationale behind olefin selectivity is not completely under-
9a,b 9c
stood. Literature precedence suggests chelation
both modulate the olefin geometry.
and electronics
J . Org. Chem, Vol. 69, No. 11, 2004 3909

Hsu et al.
SCHEME 5
When 6 was refluxed with 10 mol % 3 in methylene
chloride (in the absence of 8), most of the starting olefin
was recovered after 20 h. A very small amount of
homodimer 11 was detected by HPLC (<2.0% PA). With
1
0 equivalents of 1-pentene added (Scheme 5), the CM
F IGURE 2. Kinetic study in 1,2-dichloroethane vs methylene
chloride.
product (17, 94.6% PA, 19:1 E/Z selectivity) was formed
along with 4-octene.7
,8,10
These data suggest that the
quinoline-carbene species formed is electronically unfa-
vorable to continue the catalytic cycle in the absence of
a suitable coupling partner such as 8 or 1-pentene. In
fact, the authentic sample of 11, required for character-
ization purposes, was generated from reacting 6 with 9
mol % of Grubbs’ second-generation catalyst11 in refluxing
methylene chloride for 4 days. In conclusion, these
experiments showed that the differences in reactivity
between 8 and 6 provide the key opportunity to pursue
the CM pathway as the preferred one, where self-
dimerization pathways are suppressed by high steric
demands of 8 and the poor relative reactivity of 6.12
Two other solvents commonly used with the Grubbs’
catalyst were tested in the reaction between 8 and 6.
Toluene was investigated briefly and was found to be
ineffective in dissolving 6, resulting in very low conver-
sion to 9. On the other hand, 1,2-dichloroethane consis-
tently provided results more inferior than did methylene
chloride. This discrepancy led us to perform a kinetics
study (Figure 2). When 1.0 equiv of 8 and 2.0 equiv of 6
were stirred with 10 mol % 3, 1,2-dichloroethane initially
created a more reactive system than methylene chloride;
however, the catalyst activity stalled within 6 h. In
methylene chloride, the catalyst remained active for over
SCHEME 6
2
0 h, shown by the continual increase in product over
1
3
time (Table 1, entries 2-5).
Other attempts to optimize the reaction were qualita-
tively examined. In their mechanistic and reactivity
studies, Dias et. al have reported significant reaction rate
1
4
enhancement by addition of CuCl to reaction mixtures.
With our substrates, a faster turnover was indeed
observed initially, but the stability of the catalyst was
poor, resulting in lower overall conversions to 9 with the
added CuCl. It was speculated that the basicity of
nitrogen atoms in 8 or 6 might interfere with the catalyst
activity. Whereas the intramolecular RCM of amino
1
5
diene-hydrochloride salts was successful, the HCl salt
of 8 reacted poorly to give only 8% conversion to the
desired product. Contrary to literature examples of en-
Several other catalytic systems were investigated.
hanced catalyst turnovers under an ethylene atmo-
1
7
sphere,16 no desired product was found with 8 and 6
Other ruthenium-based catalysts (alkylidene, homobi-
metallic,18 allenylidene, and Nolan’s ) tend to give
19
20
under similar conditions.
(
11) Scholl, M.; Ding, S.; Lee, C. W.; Grubbs, R. H. Org. Lett. 1999,
, 953-956.
12) In the course of preparing this manuscript, reference 3a and
(16) Harrity, J . P. A.; La, D. S.; Cefalo, D. R.; Visser, M. S.; Hoveyda,
A. H. J . Am. Chem. Soc. 1998, 120, 2343-2351.
1
(
(17) Schwab, P.; France, M. B.; Ziller, J . W.; Grubbs, R. H. Angew.
Chem., Int. Ed. Engl. 1995, 34, 2039-2041.
the following article were published: Chatterjee, A. K.; Choi, T.;
Sanders, D. P.; Grubbs, R. H. J . Am. Chem. Soc. 2003, 125, 11360-
1370.
(18) Dias, E. L.; Grubbs, R. H. Organometallics 1998, 17, 2758-
2767.
1
(
13) Reactions in chloroform resulted in unsatisfactory results with
several new impurities.
14) Dias, E. L.; Nguyen, S. T.; Grubbs, R. H. J . Am. Chem. Soc.
997, 119, 3887-3897.
15) Fu, G. C.; Nguyen, S. T.; Grubbs, R. H. J . Am. Chem. Soc.
(19) (a) Furstner, A.; Ackermann, L. Chem. Commun. 1999, 95-
96. (b) Picquet, M.; Bruneau, C.; Dixneuf, P. H. Chem. Commun. 1998,
2249-2250. (c) Furstner, A.; Picquet, M.; Bruneau, C.; Dixneuf, P. H.
Chem. Commun. 1998, 1315-1316.
(20) Huang, J .; Stevens, E. D.; Nolan, S. P.; Petersen, J . L. J . Am.
Chem. Soc. 1999, 121, 2674-2678.
(
1
1
(
993, 115, 9856-9857.
3
910 J . Org. Chem., Vol. 69, No. 11, 2004

Synthesis of Erythromycin Derivatives
lower conversions than 3. Grubbs’ second-generation
catalyst11 was also inferior to 3 as a facilitator of the
olefin CM reaction. Instead, its higher reactivity pro-
moted homodimerization of the substrates. Nugent’s
tungsten catalyst21 and Schrock’s molybdenum catalyst5
decomposed in the presence of our substrates, and no
metathesis product was detected in either case.
The scope of the olefin CM involving a 6-O-allyl-
protected macrolide was further demonstrated by treat-
ment of 8 and 3 (catalytic) with 2.0 equiv of styrene in
2 2
refluxing CH Cl (Scheme 2) to generate 10. The desired
E-configuration of the newly formed olefin was identified
3
by the large J coupling between the two olefinic protons
,22
(
15.8 Hz)7 Under similar conditions, 18 also reacted
2
2
with styrene to give 19 (see Scheme 6). 3-Chlorostyrene
and 4-methylstyrene also coupled with 8 (3.0 equiv) in
the presence of 10 mol % 3 to yield products 20 and 21
after 20 h in 58 and 78% yield, respectively.2
2,23
Three other macrolide substrates were tested in olefin
CM reactions with 6. Utilizing the optimal conditions
developed for 9, although not practical, excess macrolide
2
2 (3.0 equiv) underwent olefin CM reaction with 6 (1.0
equiv) in refluxing methylene chloride (20 h) to give 23
in 86% yield based on the amount of the limiting reagent
2
2,23
charged.
6
Under the same conditions, 24 furnished
4% yield of 25.22,23 Finally, 26 also reacted with 6 to yield
ABT-773 (1, 39% yield) as the ultimate step in the
synthesis of 1.2
2,23
Con clu sion s
Exp er im en ta l Section
Olefin CM has been demonstrated for the construction
of the 6-O-propenylquinoline side chain of ABT-773 (1)
and intermediates. In most cases, reactions of 8 with 6
resulted in product 9 with >95:5 E-selectivity in modest
to good yields. These CM reactions were generally clean
with small amounts of homodimer. However, these
reactions were complicated by the formation of the
cinnamyl impurity 10.
Gen er a l P r oced u r e for Olefin CM. To a 10-mL
round-bottom flask was charged 8 (0.1 mmol), 6 (0.2
mmol), and CH Cl (5.0 mL) followed by 3 (10 mol %).
2 2
The solution was heated to reflux under either nitrogen
or argon atmosphere. At the conclusion of the reaction,
the mixture was diluted and assayed for product against
an authentic sample of product.
The data suggest that the large size of the macrolide
substrates hindered formation of homodimer 12 under
the reaction conditions. In the case of vinylquinoline 6,
the electronic nature of this substrate precluded dimer-
ization to occur. These two advantages combined to favor
the CM pathway. As of this report, we have not identified
a catalyst more effective than 3 for the elaboration of 6-O-
allyl erythronolides.
Ack n ow led gm en t. We thank Dr. Casey Chun Zhou
of the Structural Chemistry Department for her as-
sistance with NMR studies of product mixtures, and Dr.
Steve A. King of the Process Chemistry Research and
Development Department for his helpful discussion.
Su p p or tin g In for m a tion Ava ila ble: Includes the general
methods section, plus synthesis and characterization data for
compounds 6, 10-11, 16-17, and 19-21. This material is
available free of charge via the Internet at http://pubs.acs.org.
(21) Nugent, W. A.; Feldman, J .; Calabrese, J . C. J . Am. Chem.
Soc. 1995, 117, 8992-8998.
(22) Greater than 95:5 E/Z selectivity.
(23) Unoptimized yields.
J O049737N
J . Org. Chem, Vol. 69, No. 11, 2004 3911