1316
T. P. Tran et al. / Bioorg. Med. Chem. Lett. 17 (2007) 1312–1320
and Gram-positive organisms and their minimum inhib-
itory concentrations (MICs, lg/mL) were determined
using standard microtitration techniques.15 The Gram-
negative strains reported in this paper are Escherichia
coli MC4100 (wild-type) and E. coli Tol C (membrane
efflux-pump deficient); the Gram-positive panel includes
Enterococcus faecalis, Staphylococcus aureus, and Strep-
tococcus pyogenes. Additionally, all compounds were
tested for their ability to inhibit the target enzyme
DNA gyrase, isolated and purified from E. coli
H560.16 The assay measures the concentration of drug
needed to cleave DNA, as visualized by agarose gel elec-
trophoresis and staining with ethidium bromide (cipro-
floxacin control).
eightfold improvement in antibacterial activity against
various strains over 2c. This atypical trend also has been
often observed within the fluoroquinolone SAR.19 In
particular, the 8-methyl analogue (2i) offers the best over-
all antibacterial and enzymatic activity profile of any
compound in this group. These four R8 substituents rep-
resent both withdrawing and donating functional groups.
Other R8 substituents were examined (2k and 2l), but
those substitutions result in a significant loss of activity.
Thus, the antibacterial and enzymatic activity is not
due to the electronics at this position, but likely to be
directed by steric factor imposed by the substituent at R8.
Having established the optimum substituents for R1 and
R8, we set out to examine the relative importance of R5
and R6 (Table 2). Comparing compounds with identical
substitutions elsewhere, it is apparent that fluorine sub-
stitution at R6 provides the best antibacterial and gyrase
activity in this series. For example, compound 2n is
more active than its unsubstituted analogue, 2m, while
2p displays significantly better antibacterial and enzy-
matic activity than its chloro-analogue, compound 2q.
The preference for the R6 fluoro-substituent is also
shared by many of the fluoroquinolone antibiotics.
In general, the 3-aminoquinazolinediones displayed
excellent activity against both Gram-negative and
Gram-positive organisms. The best compounds of this
series, 2r and 2s, although having only modest activity
against the Gram-negative organisms when compared
to ciprofloxacin, display far superior activity against
Gram-positive bacteria. These two compounds also
exhibit better E. coli gyrase inhibition than ciprofloxacin.
In addition, most of 3-aminoquinazolinediones tested are
multiple-fold more potent against the E. coli Tol C strain
than the E. coli wild-type strain, clearly demonstrating
that they, unlike ciprofloxacin, are substrates for efflux.
The data in Table 2 also indicate that there is opportu-
nity for improvement of activity at R5. While the
5-amino (2o) and 5-difluoromethyl analogues (2t) are
significantly less active than their unsubstituted compar-
ators, 2h and 2r, respectively, the 5-methyl analogue (2s)
provides antibacterial and gyrase activity comparable to
2r. This limited data set suggests that while some R5
substituents can be detrimental to antibacterial activity,
there exists some potential at R5 for either maintaining
or improving potency.
The quinazolinedione core offers multiple opportunities
for SAR exploration. In general, subtle modification to
the substitution at any position on the quinazolinedione
ring has a significant impact on the antibacterial and
enzymatic activity. This paper examines the SAR trends
specifically at R1, R3, R5, R6, and R8. Because of the
number of sites investigated, the SAR effect of each po-
sition was determined by varying the substitution at that
site and evaluating the effects of change with compara-
ble substitutions at all other positions.
As can be ascertained from the data shown in Tables 1
and 2, substitution at each site of the 3-aminoquinazol-
inedione core can independently influence the SAR of
the series. Moreover, combining the optimal substitu-
tion at each site affords compounds with the best anti-
bacterial and enzymatic activity of the series (2r and
2s). Thus, there is an additive effect among the substitu-
tions on the quinazolinedione core. Undeniably, the
functionalized pyrrolidines at R7 also play a central role
in the antibacterial SAR of the 3-aminoquinazolinedi-
ones, as clearly observed when comparing compounds
2p to 2g and, more profoundly, 2r to 2i. In fact, much
of our effort was geared toward optimizing the SAR at
R7 for this series. Extensive details of these findings will
be disclosed in future publications.
The 3-aminoquinazolinediones show an explicit prefer-
ence for the cyclopropyl moiety at R1, as shown in Table
1. For example, comparing compound clusters with
R8 = H or Cl, it is clear that R1 = cyclopropyl is supe-
`
`
rior (2c vis-a-vis 2a and 2b; 2g vis-a-vis 2d through 2f).
While 2c has only marginal activity against the three
Gram-positive organisms, it is more potent against E. co-
li and its enzymatic activity is more than ten times better
than 2a and 2b. This trend is reinforced by the superior
overall activity of compound 2g over compounds 2d–
2f.17 Despite their structural similarities, this is where
the 3-aminoquinazolinediones diverge considerably
from their fluoroquinolone counterparts. Whereas com-
pounds 2a and 2f display no significant activity, the eth-
yl- and 2,4-difluorophenyl components at R1 confer
outstanding activity in the fluoroquinolone series.18
Lastly, we also probed the effects of 3-amino substitu-
tion on antibacterial and gyrase activity. As shown in
`
Table 3, both mono-alkylation (2v vis-a-vis 2u) and acyl-
ation (2w vis-a-vis 2g) of the 3-amino moiety result in
`
Table 1 also shows that substitutions at R8 can offer sig-
nificant improvement in potency. The chloro- (2g), fluo-
ro- (2h), methyl- (2i), and methoxy- (2j) analogues all
display significantly greater activity than their unsubsti-
tuted analogue (2c). It is interesting to note that while
completely inactive compounds, along with a corre-
sponding reduction in DNA gyrase activity. Hence,
the unsubstituted 3-amino moiety has proven to be opti-
mal for both antibacterial and enzymatic activity.
`
2c has better enzymatic activity than 2j (1.6 vis-a-vis
In summary, we have prepared and evaluated a series of
3-aminoquinazolinediones as antibacterial agents. The
4.3 lg/mL, respectively), 2j actually displays a two- to