A. M. Thompson et al. / Bioorg. Med. Chem. Lett. 25 (2015) 3804–3809
3805
The more complex chemical syntheses of the reversed linker
analogues (26–42) are described in Scheme 2. Thus, alkylation of
both 4-trifluoromethoxy- and 4-iodo-phenol with the known17
iodide 46 (conveniently obtained from commercial 2-methy
lene-1,3-propanediol via alcohol 4518), followed by hydrobora-
tion19 of the terminal double bond in each product, led to the inter-
mediate alcohols 49 and 50 (Scheme 2A). Alkylation of
2-bromo-4-nitroimidazole with their derived iodides 51 and 52,
acid-catalysed desilylation, and NaH-assisted ring closure of the
resulting alcohols (55 and 56) then gave the target compound 26
and iodo analogue 57, respectively. The latter was further elabo-
rated (via Suzuki couplings) to the biaryl derivatives 31-33, while
26 was separated into its two enantiomers via preparative chiral
HPLC (ChiralPak IA, 27% EtOH/hexane). Benzyl ether analogues of
these were derived from the key alcohol 61 (Scheme 2B), obtained
by alkylation of 2-bromo-4-nitroimidazole with the known20
bis-silyl ether protected iodide 58, followed by desilylation and
ring closure of the resulting diol 60 using excess NaH (3.5 equiv).
Reaction of alcohol 61 with 5-bromo-2-fluoropyridine
(NaH/DMF), followed by Suzuki couplings (in the presence of
DMF to improve solubility), also led to the 2-pyridine derivatives
34–36. Finally, a higher-yielding third route (Scheme 2C) was
developed to access the 3-pyridine derivatives 37–39. Mitsunobu
reaction of the known20 bis-silyl protected triol 64 with
6-bromo-3-pyridinol and desilylation of the product gave the diol
66. Selective monosilylation21 of this diol gave alcohol 67, which
was then converted into iodide 68 and further elaborated via sim-
ilar chemistry to give the desired products. Unfortunately, various
attempts to transform the key alcohol 61 into phenyl ether 26 (via
Mitsunobu reactions or mesylation, followed by reaction with the
sodium salt of the required phenol) all gave predominantly elimi-
nation to the terminal alkene 72, necessitating the alternative syn-
thetic approaches described. All final compounds were
characterised by 1H NMR, MS, melting point, and combustion
analysis.
OCF3
OCF3
O
N
N
O2N
O2N
Y
1
: PA-824
N
N
N
O
O
O
6
O2N
2
: X=H, Y=CH
X
N
3: X=H, Y=N
O
5
: X=Me, Y=CH
O
N
O2N
6
N
4
O
OCF3
: CGI-17341
Figure 1. Structures of antituberculosis leads.
significantly improve in vitro potency. We hypothesised that sim-
ilar substitution in the first benzene ring (proximal to the ether
linkage) of biaryl compounds such as 2, or related aza analogues,
might also lead to a decreased rate of metabolism (e.g., via steric
or electronic means) and/or a greater level of in vivo activity. We
have previously reported wide variations in both microsomal sta-
bility and in vivo efficacy across a range of biaryl derivatives sub-
stituted in the terminal benzene ring.4,5 Fluorine and fluorinated
substituents (e.g., CF3, OCF3) are well known to influence the rate
and extent of drug metabolism, sometimes even at sites distal to
the position of metabolic attack.13,14 In looking towards these
objectives, another important final strategy to examine was rever-
sal of the 6-oxymethylene side chain linkage, based on the better
microsomal stability of 6.10 The findings from these studies are
now described here.
The synthetic methods employed for the proximal ring substi-
tuted biaryl analogues (10–25) are outlined in Scheme 1. The
required 4-halobenzyl ether precursors (I) were prepared by
NaH-promoted alkylation of the known15 chiral alcohol 43 using
appropriately-substituted benzyl bromides (or benzyl iodide in
the case of 18). The benzyl halides were readily available, either
commercially, or via standard methods (e.g., bromination of
known16 or commercial alcohols with HBr/AcOH, or iodination
using I2/PPh3/imidazole). The halobenzyl ethers (I) were typically
Suzuki-coupled directly with arylboronic acids to give the final
products but in one example (14) it was necessary to first trans-
form the 4-bromide into the pinacol boronate ester derivative
(44) and then reverse couple this with the corresponding chloropy-
ridine. The novel N-oxide derivative of pyridine 9 (25) was also
Tables 1 and 3 summarise the structures and in vitro antituber-
cular potencies of the 33 new PA-824 analogues studied, together
with relevant microsomal stability and in vivo efficacy data for
selected compounds (some published data were also included for
comparison). Briefly, minimum inhibitory concentrations (MICs,
for
a growth inhibition of >90%) were determined against
Mycobacterium tuberculosis (M. tb, strain H37Rv) under both aero-
bic and hypoxic conditions, using either an 8 day
microplate-based assay with an Alamar blue readout or an
11 day assay (involving bacteria pre-adapted to low oxygen condi-
tions) with a luminescence readout, respectively.22,23 Screening
obtained in moderate yield via
3-chloroperbenzoic acid.
a buffered oxidation with
O
B
Br (I)
2'
F
O
OH
O
ii
i
iii
N
O
N
O2N
N
6'
O2N
X
O2N
N
O
N
(X=2'-F)
I
O
44
N
O
43
iv
compounds 10-13, 15-24 of Table 1
compound 14 of Table 1
compound 25 of Table 1
OCF3
N
v
O
N
O2N
9
N
O
Scheme 1. Reagents and conditions: (i) ArCH2Br (or ArCH2I), NaH, DMF, 0–20 °C, 2–4 h (73–93%); (ii) ArB(OH)2, toluene, EtOH, 2 M Na2CO3, Pd(dppf)Cl2 under N2, 90 °C, 0.5–
6 h (34–93%); (iii) bis(pinacolato)diboron, KOAc, DMSO, Pd(dppf)Cl2 under N2, 89 °C, 5 h (66%); (iv) 2-Cl-5-CF3pyridine, toluene, EtOH, 2 M Na2CO3, Pd(dppf)Cl2 under N2,
90 °C, 2 h (83%); (v) m-CPBA, Na2HPO4, CH2Cl2, 20 °C, 32 h (38%).