Synthesis, stereochemical assignment and biological activity of a novel series of C-4″ modified aza-macrolides

Article abstract of DOI:10.1016/S0960-894X(03)00358-5

Modification of the cladinose C-4″ position via manipulation of the corresponding keto derivatives afforded two stereochemically pure series of compounds. The synthesis and structure determination of these compounds is described within. The in vitro and i

Full text of DOI:10.1016/S0960-894X(03)00358-5

Bioorganic & Medicinal Chemistry Letters 13 (2003) 1955–1958
Synthesis, Stereochemical Assignment and Biological Activity of a
Novel Series of C-400 Modified Aza-Macrolides
Brian S. Bronk,* Michael A. Letavic,* Camilla D. Bertsche, David M. George,
Shigeru F. Hayashi, Barbara J. Kamicker, Nicole L. Kolosko, Laura J. Norcia, Margaret
A. Rushing, Sheryl L. Santoro and Bingwei V. Yang
Pfizer Global Research and Development, Groton Laboratories, Eastern Point Road, Groton, CT 06340, USA
Received 25 February 2003; revised 9 April 2003; accepted 10 April 2003
Abstract—Modification of the cladinose C-400 position via manipulation of the corresponding keto derivatives afforded two ste-
reochemically pure series of compounds. The synthesis and structure determination of these compounds is described within. The in
vitro and in vivo biological activity of this novel series of C-400 modified macrolides is also described.
# 2003 Elsevier Science Ltd. All rights reserved.
One of the leading causes of mortality and morbidity in
North American bovine and swine livestock operations
is respiratory disease. While multiple factors can lead to
disease, outbreaks are frequently associated with ani-
mals that have experienced some form of environmental
stress. It is estimated that the global economic losses
related to bovine respiratory disease (BRD) and swine
respiratory disease (SRD) is in excess of five billion
dollars.1 A wide variety of antibacterial agents are
approved for treating livestock respiratory disease,
including the 16-membered semisynthetic macrolide til-
micosin (Micotil1, 1), which is one of the most popular
therapeutic agents for BRD based in part on an exten-
ded duration of action.1 Tilmicosin, however, is not
useful for the treatment of SRD.
During our efforts to identify compounds with the
above-mentioned properties, we became interested in
other classes of semi-synthetic macrolides, such as the
15-membered aza-macrolides 2 and 3 previously repor-
ted by Pfizer2 and Merck,3 respectively. Our interest in
these chemotypes was primarily due to their favorable
tissue distribution, half-life and broad-spectrum anti-
microbial activity.
Our initial goal in this area was to design an anti-
bacterial agent that would have improved in vitro
potency and spectrum when compared to current
therapies such as tilmicosin. Another goal was to dis-
cover novel chemotypes that would be safe, long-acting
and effective for the treatment of respiratory diseases
caused by Gram negative pathogens in both cattle and
swine.
*Corresponding author. Tel.: +1-860-441-1712; fax: +1-860-715-8157;
e-mail: brian_s_bronk@groton.pfizer.com; tel.: +1-860-441-1803; fax:
+1-860-715-2467; e-mail: michael_a_letavic@groton.pfizer.com.
0960-894X/03/$ - see front matter # 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0960-894X(03)00358-5

1956
B. S. Bronk et al. / Bioorg. Med. Chem. Lett. 13 (2003) 1955–1958
More specifically, we felt that modification of the C-400
position of the cladinose sugar on these templates could
further enhance the potency of our analogues versus the
Gram negative pathogens associated with bovine respira-
tory disease. While there have been previous reports on
the modification of this position for both 2 and 3, includ-
ing a recent report from this group4 describing the triami-
lide class of C-400 modified aminoalcohols, we were intent
on expanding the scope of accessible compounds as well
as identifying a synthetic strategy that would provide both
efficiency and high levels of stereoselectivity to our analo-
gue program. Our efforts in this area have resulted in the
preparation and characterization of a novel series of C-400
tertiary alcohols that are potent antibacterial agents.
Conversion of the ketone intermediate 5 to the corres-
ponding epoxide was also desirable and, as outlined in
Scheme 3 (condition 1), utilization of the stabilized ylide
derived from Me3S(O)I under standard conditions fol-
lowed by deprotection afforded the desired epoxide 7 in
4
good overall yield.
Initial efforts to complete the analogous transformation
with the unstabilized trimethylsulfonium ylide (NaH/
DMSO, Me3SI, THF, 0 ꢀC) were met with failure,
instead affording multiple products attributed to
decomposition of the starting material. The overall effi-
ciency of the reaction was greatly improved by employ-
ing KHMDS as the base and running the reaction in
DME at low temperature, although the yield of desired
product was variable due to the formation of the corre-
sponding thiomethyl ether 8. Formation of 8 likely
proceeds through attack of the ylide as anticipated, fol-
lowed by demethylation of the zwitterionic intermediate
(Scheme 4). It is postulated that the iodide counter ion
introduced via the sulfonium salt is involved in the
demethylation reaction. This hypothesis is supported by
the fact that the yield of epoxide 9 improved and for-
mation of the thiomethyl ether was suppressed when
Me3SBF4 was used as the source of the sulfur ylide
(Scheme 3, condition 2).
Chemistry
Working from templates 2, we identified the C-400 ketones
5 as our key synthetic intermediates. Ketones 5 are readily
prepared in two steps from the corresponding alcohols. As
described previously,4 treatment of 2a (R1, R2=H) with
one equivalent of benzyl chloroformate, followed by
oxidation of the C-400 alcohol of 4a with EDC/PTFA
proceeded smoothly to give ketone 5a in 88% overall
yield. Similar results were obtained when preparing the
corresponding analogue 5b (R2=Me). Although some-
what unstable in solution, the solid state stability of
intermediates 5 was suitable for extended storage and
this sequence was routinely conducted on scale in excess
of one hundred grams (Scheme 1).
1
Comparison of both the H NMR spectra and HPLC
retention times of 7 and 9 indicated that the major pro-
duct obtained from each epoxidation reaction was
unique. In addition, HPLC analysis of the crude reac-
tion mixtures indicated that epoxidation under condi-
tion 1 afforded ca. 10% of 9, while condition 2 afforded
less than 5% of 7. Although significant literature pre-
Initial attempts to effect the addition of organometallic
reagents to ketones 5 afforded very low yields of the
desired compounds. Optimization efforts included pre-
treatment of the macrolide with TMSCl in THF and
modification of the reactivity of the organometallic
reagent (RMgX or RLi) via the addition of additives
(ZnCl2, MgBr2, CeCl3) in combination with solvent
screening (THF, Et2O, DME). The most significant
improvements in the efficiency of the reaction were rea-
lized employing excess Grignard reagents with DME as
the solvent.5 Subsequent removal of the C-20 benzyl
carbonate with methanol proceeded smoothly to afford
the desired products 6 in moderate to good yields as a
single isomer at C-400 (Scheme 2).6
Scheme 2. (a) RMgX, DME, ꢁ78 ꢀC; (b) MeOH.
Scheme 1. (a) CBZ-Cl, CH2Cl2, 0 ꢀC; (b) 1-(3-dimethylaminopropyl)-
3-ethylcarbodiimide hydrochloride, pyridinium trifluoroacetate, 0–
23ꢀC.
Scheme 3. (1) (a) Me3S(O)I, NaH, tetrahydrofuran, 100%; (b)
MeOH.-or- (2) (a) Me3SBF4, KHMDS, toluene, tetrahydrofuran,
58%; (b) MeOH.

B. S. Bronk et al. / Bioorg. Med. Chem. Lett. 13 (2003) 1955–1958
1957
After numerous attempts, we were successful in obtain-
ing a single crystal X-ray structure of tertiary alcohol 6a
(R=Me, R2=H), obtained from the addition of
MeMgBr to ketone 5a. As illustrated in Figure 1, attack
of the Grignard reagent affords a single diastereomer
that places the alcohol in the axial (unnatural) con-
figuration. The high degree of selectivity observed for
this reaction is attributed to chelation of the Grignard
reagent to the neighboring methoxy group and sub-
sequent delivery of the Grignard reagent to the same
face of the cladinose ring structure.
Scheme 4. (a) Me3SI, KHMDS, toluene, tetrahydrofuran, ꢁ78 ꢀC; (b)
Conversion of 7 to 6a using lithium aluminum hydride
confirmed the stereochemical assignment of 7.7 The ste-
reochemistry of the epoxide 9 was determined by reac-
tion of the epoxide with n-propyl amine and subsequent
solution of an X-ray structure of the product as descri-
bed previously.4 These results confirm our assumption
that the stabilized sulfoxonium ylide reacted with the
ketone 5a to form epoxide 7, the result of equatorial
attack on the ketone and that the unstabilized ylide
reacted with the ketone to form epoxide 9, which is the
result of axial attack on the ketone.
MeOH.
A wide variety of substituents at the C-400 position were
prepared via addition of the appropriate Grignard
reagent to the C-400 carbonyl compounds 5a and 5b.
Tables 1 and 2 show the in vitro activity of these com-
pounds against two important livestock pathogens,
Pasteurella multocida and Escherichia coli.8 The tables
also show the activity of many of the analogues in an in
vivo mouse P. multocida infection model.9
Figure 1. The configuration of 6a in the solid state. Carbon atoms are
in green, oxygen atoms in red and nitrogen atoms in blue.
cedence exists for assigning the stereochemical out-
come of these two reactions, our goal was to unam-
biguously assign the stereochemistry of each of the
products. In addition, the stereochemical relationship
between the epoxides and the products obtained via
direct Grignard addition to the ketone was also
unknown at this time.
Small C-400 substituents generally gave the best activity
versus P. multocida (e.g., 6a–e, 6g–j), however slightly
larger substituents such as n-butyl (6f), pyridyl (6m–n)
or phenyl (6o) show weaker activity against this patho-
gen.
A similar structure–activity relationship was
observed against E. coli. Potent in vivo activity was
observed with the methyl substituted analogue 6a and
Table 1. MIC values against P. multocida and E. coli, and ED50 values in a mouse P. multocida infection model for compounds 6, R2=H and
compounds 7 and 9
Compd
R2
P. mult. MIC (mg/mL)
E. coli MIC (mg/mL)
Mouse ED50 (mg/kg)
6a
6b
6c
6d
6e
6f
6g
6h
6i
Me–
Et–
H2C=CH–
n-Pr–
<0.1
0.2
0.2
0.78
1.56
1.56
1.56
0.78
12.5
1.56/3.13
3.13
1.56
0.78/1.56
3.13
1.56
3.13
3.13
6.35/12.5
3.13
0.78
28
61
66
81
0.2
H2C=CHCH2–
n-Bu–
HCC–
H3COCH2CC–
(H3C).2NCH2CC–
H3CCC–
2-Thiophenyl–
2-Furanyl–
2-Pyridyl–
3-Pyridyl–
Phenyl
2-(NMe)-indole–
H3CO(CH2)3–
HOCH2CH2–
H3CSCH2–
—
<0.1
0.78
<0.1
0.2
58
n.d.
24
20
22
21
0.2
6j
6k
6l
0.1/0.2
0.39/0.78
0.2
0.78
0.39
1.56/3.13
0.2
0.1/0.2
0.39
0.1
<0.1
0.1
>80
72
6m
6n
6o
6p
6q
6r
6s
7
>40
>40
n.d.
>40
49
n.d.
44
31
6.25
3.13
1.56
3.13
9
—
n.d.

1958
B. S. Bronk et al. / Bioorg. Med. Chem. Lett. 13 (2003) 1955–1958
Table 2. MIC values against P. multocida and E. coli, and ED50 values in a mouse P. multocida infection model for compounds 6, R2=CH3
Compd
R2
P. mult. MIC (mg/mL)
E. coli MIC (mg/mL)
Mouse ED50 (mg/kg)
6t
6u
6v
6w
6x
6y
6z
Me–
0.05/0.1
0.05
0.1
0.05/0.1
0.39
0.05
0.39/0.78
0.39
0.39
0.39/0.78
6.25
0.39
26
29
40
33
>40
n.d.
>80
42
H3COCH2CC–
(H3C).2NCH2CC–
HCC–
2-(NMe)-indole–
2-(NMe)-imidazole–
H3CO(CH2)3–
H3CCC–
0.1
0.2
0.39
1.56
6aa
2. Bright, G. M.; Nagel, A. A.; Bordner, J.; Desai, K. A.;
Dibrino, J. N.; Nowakowski, J.; Vincent, L.; Watrous, R. M.;
Sciavolino, F. C.; English, A. R.; Retsema, J. A.; Anderson,
M. R.; Brennan, L. A.; Borovoy, R. J.; Cimochowski, C. R.;
Faiella, J. A.; Girard, A. E.; Girard, D.; Herbert, C.; Manou-
sos, M.; Mason, R. J. Antibiot. 1988, 41, 1029.
3. (a) Wilkening, R. R.; Ratcliffe, R. W.; Doss, G. A.; Barti-
zal, K. F.; Graham, A. C.; Herbert, C. M. Bioorg. Med. Chem.
Lett. 1993, 3, 1287. (b) Shankaran, K.; Wilkening, R. R.;
Blizzard, T. A.; Ratcliffe, R. W.; Heck, J. V.; Graham, A. C.;
Herbert, C. M. Bioorg. Med. Chem. Lett. 1994, 4, 1111.
4. Letavic, M. A.; Bronk, B. S.; Bertsche, C. D.; Casavant,
J. M.; Cheng, H.; Daniel, K. L.; George, D. M.; Hayashi,
S. F.; Kamicker, B. J.; Kolosko, N. L.; Norcia, L. J. L.;
Oberton, V. D.; Rushing, M. A.; Santoro, S. L. Bioorg. Med.
Chem. Lett. 2002, 12, 2771.
the alkynes (6g–j). Other substituents including the aryl
and heteroaryl substituted tertiary alcohols showed
weak in vivo activity. For reference, tilmicosin gave an
MIC of 0.78 mg/mL against P. multocida and 50 mg/mL
against E. coli in these same assays. Many of the new
compounds reported here are more potent than tilmi-
cosin against P. multocida and all of the compounds are
significantly more potent than tilmicosin against E. coli.
Epoxides 7 and 9 were also tested. The stereochemistry
of the epoxide had little effect on potency (Table 1).
An analogous set of the N-Me azalides were also pre-
pared and characterized. The trends in activity for
representative analogues were similar to the trends
observed for the N-H compounds. Smaller substituents
were best and the larger or heteroaromatic analogues
were not as potent. In general, the in vitro activity was
comparable or slightly better than the N-H compounds
described previously.
5. The Grignard reagents were typically employed in 8–10-
fold excess. The Grignard acetylides were formed in situ via
treatment of the corresponding alkyne with MeMgBr.
6. Compound 6q was prepared via reduction of 6h via hydro-
genation. Compound 6r was prepared via ozonolysis of 6e and
subsequent reduction with NaBH4.
7. Epoxides 7 and 9 were treated with excess LiAlH4 in THF
and HPLC retention times of the crude product were com-
pared to authentic 6a. The crude material also contained
epoxide starting material and small amounts of products
attributed to reduction of the ring lactone.
In conclusion, we have unambiguously defined the ste-
reochemical outcome of the addition of nucleophiles to
the C-400 ketones of this class of macrolides and we have
utilized this discovery to identify a novel class of C-400
modified azalide antibiotics with good in vitro potency
against P. multocida and E. coli, two significant veterinary
pathogens responsible for BRD. Several members of this
class of azalide antibiotics have potent in vivo activity in
mice and warrant further in vivo studies in livestock.
These studies will be published in due course.
8. The E. coli strain 51A0150 (poultry lung origin) and P.
multocida strain 59A0067 (turkey origin) were used in this
assay to test the antibacterial activity. Both strains were grown
on Brain Heart Infusion (BHI) plates overnight. Several colo-
nies were suspended into saline and adjusted to
OD625nm=0.09 (0.5 McFarland unit). The inoculum solution
was made by preparing a 1:100 dilution of 0.5 McFarland sal-
ine suspension using BHI broth and 100 mL of this suspension
was added to 100 mL of BHI broth containing various con-
centrations of test antibiotics. The test antibiotic solution was
serially diluted two fold by automatic pipette in a 96-well
microtiter format. After inoculation with both strains (final
density was approximately 5ꢂ105 cfu/mL), the microtiter
plates were incubated at 37 ꢀC for 18 h. The minimum inhibi-
tory concentration (MIC) was determined as the lowest con-
centration of the test compound in which the absorbency at
600 nm is less than or to equal 0.025.
Acknowledgements
The authors thank Jon Bordner for solving the X-ray
structure of 6a and Frank DiCapua for providing Fig-
ure 1. The authors also thank Nick Vamvakides for
biology technical assistance.
9. The experimental parameters for the murine model are as
follows. Twenty gram female CF-1 mice were infected intrana-
sally with 50 microliters of a suspension of 2ꢂ104 P. multocida
serotype 5A. Compounds were administered subcutaneously 0.5
h post-infection at doses of 5 to 80 mg/kg. The effective dose for
50% of the aminals (i.e., the ED50) was calculated based on the
number of surviving mice four days after infection.
References and Notes
1. USDA. 2000. Part I–III: Beef Feedlot Study, 1999. USDA:
APHIS:VS, CEAH, National Animal Health Monitoring Sys-
tem (NAHMS). Fort Collins, CO. #N327.0500, #N335.1000
and #N336.1200.