Synthesis, biological activity, and QSAR studies of antimicrobial agents containing biguanide isosteres

Article abstract of DOI:10.1071/CH03146

Analogues of chlorhexidine and chemically related antimicrobial compounds were synthesized, based on a model in which the bisbiguanide moieties were replaced by conformationally restricted cyclic isosteres. This model was tested by measuring the antimicro

Full text of DOI:10.1071/CH03146

Full Paper
CSIRO PUBLISHING
www.publish.csiro.au/journals/ajc
Aust. J. Chem. 2004, 57, 77–85
Synthesis, Biological Activity, and QSAR Studies of Antimicrobial
Agents Containing Biguanide Isosteres
Gregory T. Wernert,A David A. Winkler,A,D George Holan,B and Gina NicolettiC
A CSIRO Division of Molecular Science, Private Bag 10, Clayton South VIC 3169, Australia.
B Starpharma, Baker Heart Research Institute, Melbourne VIC 3004, Australia.
C Department of Biotechnology and Environmental Biology, Royal Melbourne Institute of Technology,
Bundoora VIC 3083, Australia.
D Author to whom correspondence should be addressed (e-mail: dave.winkler@csiro.au).
Analogues of chlorhexidine and chemically related antimicrobial compounds were synthesized, based on a model
in which the bisbiguanide moieties were replaced by conformationally restricted cyclic isosteres. This model was
tested by measuring the antimicrobial activities of the compounds. Quantitative structure–activity relationship
(QSAR) studies showed a parabolic dependence of antimicrobial activity on the lipophilicity of the compounds.
The basicity of the functional groups in the molecules was also very important, as uncharged molecules were not
able to disrupt the microbial phospholipid bilayer and cause an antimicrobial effect. We compared our QSAR results
to those reported in other studies of antimicrobials of diverse structure. We found very similar QSAR models for
all compounds studies with a log P (octanol/water partition constant) optimum at 5.5 (neutral log P value). The
form of the QSAR equations were similar, suggesting a common mode of action for these agents.
Manuscript received: 6 June 2003.
Final version: 8 September 2003.
Introduction
and have been used in agriculture for over a decade.[1] They
show activity against several economically significant fungi.
Chlorhexidine 1, a bisbiguanide, is a widely used topical
antimicrobial compound exhibiting a broad spectrum
of activities (Diagram 1).[2] Azarole, a (pyrrolylimino)-
cyclohexadiene, shows potent antitubercular and antibacte-
rial activity.[3] Bispyridinamines, which may be considered
a type of ‘masked guanide’, have been developed as anti-
microbial agents against dental plaque.[4]
The antimicrobial mode of action of many of these
compounds is not well understood. Some guanidinium com-
pounds that show antimycotic activity inhibit ꢀ14-reductase
and ꢀ8-ꢀ7-isomerase enzymes in the ergosterol biosyn-
thetic pathway.[5] Octenidine, a bispyridinamine, inhibits
extracellular polysaccharide-producing enzymes of some
microbes.[6] Although chlorhexidine was patented 50 years
ago, surprisingly little published work exists.The mechanism
by which chlorhexidine kills bacteria is still ill-defined.[7]
Pathogenic bacteria and fungi are of great economic signifi-
cance in health care and agriculture. Bacterial and fungal
infections acquired by patients undergoing hospital treat-
ment (nosocomial infections) represent severe health care
problems, and the rising incidence of multiply resistant
strains is of major concern. Crop losses and damage due
to attack by fungi and insects amount to billions of dollars
per year and create large markets for effective agrochemical
agents.
There are some classes of chemical agents that exhibit
a relatively broad spectrum of activity against bacteria,
fungi, and insects. Alkyl and aryl guanides, biguanides, bis-
biguanides, and other lipophilic cationic compounds, such
as bispyridinamines and (pyrrolylimino)-cyclohexadienes,
show this type of activity, albeit through different modes
of action in different species. The broad-spectrum fungi-
cides Dodine, Guazatine, and Iminoctadine are guanidines
Cl
NH
NH
H
N
H
N
H
N
N
H
N
H
N
H
NH
NH
Cl
1
Diagram 1.
© CSIRO 2004
10.1071/CH03146
0004-9425/04/010077

78
G. T. Wernert et al.
Chlorhexidine may possibly act by several mechanisms:[8–10]
• in bisbiguanides, the length of the linker between the two
biguanide moieties is important;
• differences in the basicities of these conformationally
restricted isosteres would not dramatically affect their
antimicrobial activities.
• adsorption to the surface of bacteria;
• damage to permeability barriers, facilitating entry of the
bactericide to the cytoplasm and leakage of ions from
the cell;
• precipitation of the cytoplasm and prevention of repair of
The validity of these assumptions was tested experimen-
tally by observing the antimicrobial activities of analogues
designed on the basis of these assumptions.
We devised potential antimicrobial agents based on
chlorhexidine and other antimicrobial agents, where the
biguanide or guanide moieties were replaced by poten-
tial bioisosteres or bioanalogues.[22] Our conformationally
restricted analogues were guanylpyrimidines and amino-
pyrimidines, in which part of the guanide or biguanide is
incorporated into a planar pyrimidine ring.
the cell wall and membrane;
• potential interaction with cellar ATPases.
There are relatively few reports describing structure–
activity relationships in this or similar types of compounds.
Warner and coworkers reported a QSAR study of biguanides,
carbamimidates and bisbiguanides exhibiting activity against
Streptococcus mutans.[11] Lindholm synthesized antibacte-
rial pyridylguanidines and screened them against several
bacteria and yeasts, carrying out comprehensive QSAR anal-
yses on these data.[12] The QSAR analyses found a parabolic
relationship between the logarithm of the antimicrobial activ-
ity and log P in each single chemical class. In two recent
reviews Denyer[8,9] noted that the study of biocide mecha-
nisms of action offers an, as yet, largely untapped initiator
of novel development directions and called for more QSAR
studies of these classes of bioactive agents.
Chemistry
The synthesis of the bisguanidino compounds 1–7,
Table 1, has recently been reported.[23] The alkanebisamino-
pyrimidines 10 and 11 are known structures.[24,25] Other
compounds reported here were made by modification of the
general synthesis or adaptation of an existing method for the
synthesis of bisguanidinoalkanes.[23]
Compounds 8 and 9 were prepared from the diamine and
2-chloropyrimidine by a method[26] based on that using tri-
ethylamine as the base in dioxan (Scheme 1, Method A).
We have an ongoing interest in the application of QSAR
to bioactive agent design.[13–15] We aim to design more
potent, efficacious analogues of these compounds for use as
agents against Staphylococcus aureus, particularly the resis-
tant strain MRSA, and against dental plaque.[16,17] This paper
describes the synthesis and antimicrobial activity of ana-
logues of chlorhexidine 1 and related compounds that contain
guanide or biguanide isosteres. These compounds exhibit
activity against a range of bacteria and yeasts. We also report
a QSAR analysis of these compounds, a comparison with
other QSAR studies of antimicrobial agents, and a discus-
sion of the implications of the QSAR models for the mode of
action of these agents.
4-Chlorphenoxy-4-pyrimidine 13 was prepared by
a
regioselective synthesis[27] in which none of the bis 2,4-
substituted isomer was detected. Treatment of 13 with 1,6-
diaminohexane in dioxane gave compound 10 (Scheme 1,
Method B). Treatment of 4,6-dichloro-2-methylthiopyrim-
idine with 4-chloroaniline in acetic acid and concentrated
hydrochloric acid as catalyst afforded 14, which was oxidized
to the sulfone 15. Catalytic hydrogenolysis with hydrogen
in the presence of an acid scavenger yielded 16 (Scheme 1,
Method C). Compound 11 was prepared by the nucleophilic
displacement of the sulfone group from the intermediate 16
by 1,6-diaminohexane in a similar way to that reported.[23]
The 2,4-bisanilinopyrimidine 12 was prepared following a
literature method.[28]
Design Rationale
It is clear that many topical antimicrobial agents (dis-
infectants) have one or more basic structural features (often
guanides, biguanides, or tertiary amines) that are positively
charged under physiological conditions and linked to a
lipophilic chain.[4,11,18] The basicity (and tautomerism) of
the biguanide and related moieties, intramolecular hydro-
gen bonding, and lipophilicity are potential modulators of
antimicrobial activity. Many antimicrobial agents are also
conformationally flexible molecules. Guanide and biguanide
moieties can adopt several tautomeric forms, as illustrated in
the structural work of Fabrizzi et al.,[19] and Pinkerton and
Schwarzenbach,[20] and described in the theoretical paper
by Jordan and Gready.[21] To design topographical mimics
of chlorhexidine it was necessary to make several assump-
tions about the basicity, tautomeric state, geometry, and
conformational preferences of the active compounds such
as chlorhexidine:
Antimicrobial Activities
The structures and antimicrobial activities of the twelve com-
pounds studied are reported in Table 1. The activities of
the most active compounds 6, 7, and 12 approach that of
chlorhexidine 1. It is clear that all compounds except 8, 9, and
10showedsubstantialantimicrobialactivityagainstallstrains
except Pseudomonas. This microbe is known to contain hun-
dreds of different proteins in the cell wall of the bacterium
that pump material out of the cell, allowing P. aeruginosa
to resist the effects of many antibiotics. P. auriginosa also
excretes an exopolysaccharide biofilm that protects it from
antimicrobial agents better than other bacteria.[29]
QSAR Analyses
• the biguanide moiety adopts an essentially planar confor-
mation in its biologically active form[19,20] which can be
mimicked by conformationally restricted cyclic analogues;
The physicochemical properties and molecular descriptors
used in the QSAR study are reported in Table 2. The multiple

Studies of Antimicrobial Agents Containing Biguanide Isosteres
79
Table 1. Antimicrobial activities (MIC) of bisguanides and bisbiguanides [µM]
ꢁ
4
4'
Y
Y
Y
Y
Aꢀ
5
5'
NH
Z
NH
R
R
6'
6
Compound
R
Y
N
Z
A
S. aureus MRSA E. coli P. aeruginosa C. albicans
1
2
chlorhexidine
8
8
16
63
32
Cl−
Cl−
155
620
1240
>1200
310
NH
4,6-dimethyl
H
N
N
H
NH
3
4
5
6
7
5-chloro
5-chloro
5-chloro
5-chloro
5-chloro
N
N
N
N
N
640
600
33
640
600
33
1300
1200
66
>1300
>1200
>1000
>1000
>1000
>1300
>1200
130
NH
H
N
N
H
NH
Cl−
Br−
Cl−
Cl−
NH
H
N
N
H
NH
NH
H
N
N
H
NH
NH
17
17
33
66
H
N
N
H
NH
NH
16
16
63
63
H
N
N
H
NH
8
H
N
(CH2)6
(CH2)10
(CH2)6
(CH2)6
N
Cl− >3000
Br− >3000
Cl− >1000
>3000 >3000
>3000 >3000
>1000 >1000
>3000
>3000
>2000
>1000
>1500
>3000
>3000
>1000
500
9
H
N
10
11
12
4-chlorophenoxy
4-chloroanilino
5-chloro
N
N
Cl−
Cl−
61
12
61
120
CH
12 >1000
25
N
Method A
Method C
Cl
N
Cl
N
N
N
N
H
Cl
(a), (b)
Cl
N
.2HX
Cl
ꢁ NH (CH ) NH
2
NH(CH2)nNH
2
2 n
(a)
ꢁ
N
N
N
N
N
N
Cl
Cl
n = 6, X = Cl
8
9
NH2
SCH3
SCH3
n = 10, X = Br
14
Method B
H
H
Cl
Cl
N
N
O
Cl
(c)
(b)
(a)
N
N
N
ꢁ
N
N
N
N
Cl
Cl
SO2CH3
SO2CH3
Cl
Cl
OH
13
15
16
N
N
N
N
N
N
N
NH
.2HCl
NH
NH
NH
N
(d), (e)
(b), (c)
.2HCl
O
HN
O
NH
Cl
Cl
Cl
Cl
10
11
Scheme 1. Method A: (a) Et3N, dioxane; (b) HX, MeOH. Method B: (a) NaOH, acetone, H2O; (b) Et3N, dioxane; (c) HCl, MeOH. Method C:
(a) HOAc, HCl; (b) H2O2, HOAc; (c) H2, Pd/C, MgO; (d) NH2(CH2)6NH2·DMSO; (e) HCl, MeOH.

80
G. T. Wernert et al.
Table 2. SAR data for bisguanides and bisbiguanides [µM]
Compound
log P
IN+
Ibis
S. aureus
−log MIC
MRSA
−log MIC
E. coli
−log MIC
C. albicans
−log MIC
1
2
3
4
5
6
7
8
4.78
3.79
1.73
2.81
4.97
4.97
6.08
2.05
4.21
8.30
7.54
5.40
1
1
1
1
1
1
1
0
0
0
1
1
1
0
0
0
0
0
0
0
0
0
0
0
5.10
3.81
3.19
3.22
4.48
4.78
4.80
1.83
1.91
2.42
4.21
4.92
5.10
3.21
3.19
3.22
4.48
4.78
4.80
1.83
1.91
2.42
4.21
4.92
4.80
2.91
2.89
2.92
4.18
4.48
4.20
1.83
1.91
2.42
3.91
2.52
4.50
3.51
2.30
2.33
3.88
4.18
4.20
1.83
1.91
2.42
3.61
4.62
9
10
11
12
regression analyses yielded parabolic relationships between
lipophilicity and the antimicrobial activity for all species
except P. aeruginosa for which there was insufficient bio-
logical activity for a QSAR model. It was clear from Tables 1
and 2 that only those compounds with basic nitrogen atoms
that were likely to be positively charged at the pH of the
screen media show significant activity. Compounds 8, 9, and
10, with weakly basic anilino nitrogen atoms in the linker
chain, were inactive at the highest concentrations tested. An
indicator variable for charged molecules was employed. The
bisbiguanide-containing compound chlorhexidine 1 showed
enhanced activity and this moiety was accounted for by
another indicator variable, as previous studies have done.
This was applied cautiously as the indicator variable was of
marginal statistical significant (t > 1.3, P > 0.75) especially
given only one example of this type of compound. However,
the QSAR models suffered substantially when the variable
was omitted.
−log MIC [µM] {C. albicans} = 1.601( 0.334) log P
− 0.155( 0.036)(log P)2 + 0.153( 0.748)
n 7, s 0.33, r2 0.86, log P(opt) 5.2
where Ibis is an indicator variable for bisbiguanides, n is the
number of compounds in the analysis, s is the standard error,
r2 is the squared correlation coefficient, F is the F-statistic,
and log P (opt) is the value of log P at which the activity is
maximum, derived from the regression equation.
(b) Inactive compounds included at four times their
highest tested concentration:
−log MIC [µM] {S. aureus} = 1.009( 0.314) log P
− 0.071( 0.025)(log P)2 + 1.882( 0.335)IN+
− 0.640( 0.804)
n 12, s 0.45, r2 0.89, F 22.6, log P(opt) 7.1
QSAR equations were derived for two scenarios: (a) all
active compounds included in the model and compounds
inactive at the highest concentration tested excluded; and
(b) all compounds included, with inactive compounds given
a value four times that of the highest concentration at
which they were tested. The following structure–activity
relationships were derived.
−log MIC [µM] {MRSA} = 1.034( 0.370) log P
− 0.072( 0.029)(log P)2 + 1.816( 0.395)IN+
− 0.736( 0.946)
n 12, s 0.53, r2 0.86, F 15.8, log P(opt) 7.2
−log MIC [µM] {E. coli} = 0.895( 0.378) log P
− 0.061( 0.030)(log P)2 + 1.473( 0.396)IN+
− 0.408( 0.961)
(a) Inactive compounds excluded:
−log MIC∗ [µM] {S. aureus} = 1.177( 0.357) log P
n 11 (12 excluded), s 0.53, r2 0.82, F 10.5, log P(opt) 7.3
− 0.098( 0.038)(log P)2 + 0.604( 0.396)Ibis
− 1.107( 0.772)
−log MIC [µM] {E. coli} = 0.826( 0.330) log P
n 9, s 0.36, r2 0.85, log P(opt) 6.0
− 0.056( 0.026)(log P)2 + 1.569( 0.320)IN+
− 0.327( 0.897)
−log MIC [µM] {MRSA} = 1.048( 0.518) log P
n 12 (12 not ionized), s 0.51, r2 0.83, F 12.9, log P(opt) 7.4
− 0.081( 0.056)(log P)2 + 0.725( 0.575)Ibis
− 1.224( 1.121)
−log MIC [µM] {C. albicans} = 1.252( 0.387) log P
n 9, s 0.52, r2 0.74, log P(opt) 6.4
− 0.091( 0.031)(log P)2 + 1.131( 0.413)IN+
− 1.160( 0.990)
−log MIC [µM] {E. coli} = 1.829( 1.028) log P
− 0.148( 0.098)(log P)2 + 0.835( 0.657)Ibis
− 1.392( 2.535)
n 12, s 0.56, r2 0.79, F 10.1, log P (opt) 6.9
n 7, s 0.57, r2 0.76, log P(opt) 6.1
where IN+ is an indicator variable for charged nitrogen atoms
∗Minimum inhibitory concentration.
and other terms are as defined above.

Studies of Antimicrobial Agents Containing Biguanide Isosteres
81
Discussion
(a)
It is clear from Table 1 that several of the topograph-
ical analogues of chlorhexidine show substantial broad-
spectrum antimicrobial activity. In particular, the activities
of 6 and 7 approach to that of chlorhexidine, except against
P. aeruginosa. The conformational restriction introduced by
the pyrimidine ring, and the changes in basicity on incorpora-
tion into a ring appear to have a relatively small influence on
activity as long as the nitrogen atom is protonated at physio-
logical pH. Compounds 8–10 do not meet this criterion and
are inactive in all microbial strains tested.
Ca2ꢁ
Ca2ꢁ
Ca2ꢁ
Ca2ꢁ
Ca2ꢁ
Ca2ꢁ
(b)
However, as the QSAR equations show, the overriding
dependence of antimicrobial activity is on the lipophilicity
of the compounds not the length of the central linker chain
(defining distances between positively charged groups) or
any other structural feature. The dependence on lipophilicity
(log P) is parabolic in each case, resulting in optimal val-
ues of log P at which activity is greatest. As compound 12
shows, it is possible to have high antimicrobial activity with
compact molecules provided the log P value is near the opti-
mum and the compound is charged. However, the smaller
molecular dimensions of this compound and its high anti-
microbial activity were useful in eliminating the molecular
size as a significant structural variable in the QSAR.The indi-
cator variable for the presence of the biguanide moiety in the
molecule was significant, as others have found in QSAR stud-
ies. However, this must be viewed cautiously in our models
as this feature only occurred once in the dataset.
Ca2ꢁ
Ca2ꢁ
Ca2ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
Ca2ꢁ
Ca2ꢁ
Ca2ꢁ
(c)
Ca2ꢁ
Ca2ꢁ
Ca2ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
The QSAR models that include the poorly active/inactive
compounds in particular are consistent with the mechanism
that antimicrobial compounds of this type interact with and
disrupt phospholipid bilayers in microbes. In this model
proposed by Ikeda et al.[30] for oligomerized biguanide com-
pounds similar to those we studied, positively charged groups
on the biocide interact with the acidic phospholipid (anionic)
head groups, which, while being present in the membrane in
relatively low concentrations, fulfill vital roles in stabilizing
the membrane. The calcium counterions normally present at
the surface of the cell membrane are displaced by the cationic
moieties of the antimicrobial agent. The hydrophobic linker
regions can insert into the hydrophobic interior of the bilayer,
helping stabilize the bond between the charged head group of
the phospholipids and the delocalized positive charge on the
biocide. The linker is not required to span the phospholipid
membrane, consistent with a lack of linker length dependence
in the QSAR models. Phospholipid phase separation occurs
causing ion leakage from the cytoplasm and cell death. This
mechanism is illustrated in Fig. 1.[31]
Ca2ꢁ
Ca2ꢁ
Ca2ꢁ
(d)
Ca2ꢁ
Ca2ꢁ
Ca2ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
Ca2ꢁ
Ca2ꢁ
Ca2ꢁ
ꢁ
ꢁ
ꢁ
ꢁ
ꢂ biocide
ꢂ neutral
ꢂ proteins
ꢂ acidic
phospholipids
phospholipids
The indicator variable denoting charged/uncharged
molecules shows that there is a very large loss in activity
when the molecule is not charged (large positive coefficient
in the QSAR model). Compound 12 has a log P value very
similar to that of chlorhexidine but a weaker basicity than the
guanide and biguanide moieties. Calculations on the ionic
speciation of this compound as a function of pH show that it
has the lowest conjugate acid pKa and is the compound whose
ionization state is most sensitive to the effect of pH under
Fig. 1. (a) Bacterial cytoplasmic membrane conforming to fluid
mosaic model; stabilized by calcium ions and phospholipid mixture and
distribution. (b) Initial wave of biocide displaces surface cations, binds
to acidic phospholipids, causing change in packing. (c) Biocide induces
a phospholipid phase separation, affects concentrate in the area of inte-
gral proteins; causes increase in membrane permeability, K+ efflux,
loss of enzyme function, i.e. bacteriostatic level. (d) Destabilized zones
aggregate into favourable hexagonal phase by building excess of biocide
(electrostatic and hydrophobic); complete loss of membrane function,
i.e. bactericidal level.

82
G. T. Wernert et al.
the conditions of testing. The compound showed inhibitory
activity below 10 µM concentration against S. aureus and C.
albicans, activity of the same order as chlorhexidine. How-
ever it showed a narrower spectrum of activity as it was
inactive against the Gram-negative E. coli and P. aeruginosa.
The weaker basicity may be a factor in its low activity against
E. coli, if this bacterium has a different cytoplasmic pH than
other microbes, or a different ratio of acid and neural phos-
pholipids in its cytoplasmic membrane. It is interesting that
the models containing all twelve compounds show a slightly
higher log P optimum than the models based on just active
compounds.
Gasparrini and coworkers[32] found a log P optimum of 5.2–
5.4 for their acyl-l-carnitine alkyl ester antimicrobials. A
very similar log P optimum was reported byTsubouchie et al.
for biguanide antiseptics in a recent publication.[33] We calcu-
latedlog P valuesforthecompoundsinthispaper, andcarried
out a QSAR analysis of the antimicrobial activity using their
reported activities. We obtained a parabolic dependence of
antimicrobial activity on the lipophilicity with QSAR equa-
tions and log P values for optimum activity very similar to
those from other studies, including ours. These QSAR equa-
tions and optimum log P values for current, and previous
studies, are summarized in Table 3.
The close correspondence of the QSAR equations, and
very similar values for the optimum log P value in all
these studies, across a wide range of structures, suggests
that a common mechanism of action exists for all of these
chemical classes as discussed in the previous section. How-
ever, the mechanistic analyses are complicated by the fact
that some compounds have several mechanisms of toxicity
towards microbes. For example, the antimicrobial biocides
dinitrophenol, tribrominated salicylanilide, polymyxin, and
chlorhexidine all collapse the proton membrane potentials at
inhibitory concentrations. Other agents, including biguanides
also inactivate or inhibit ATPase. However, it appears to
be the cytoplasmic membrane disruption, and not ATPase
inactivation or disruption of electron transport, which is the
mainlethaleventinchlorhexidineaction.[10,34,35] Inaddition,
adherence is an important pathogenic mechanism in Candida
Global or Consensus QSAR Analyses
Our QSAR analyses yielded relationships between anti-
microbial activity and log P that are similar for each species.
Interestingly, the equations we derived for active com-
pounds are essentially the same as those derived by several
other studies of antimicrobial activities of other structural
types. Lindholm[12] studied QSAR in pyridylguanidines and
obtained an optimum log P value of 5.4–5.5 for antimicrobial
agents, consistent with our findings.This value is also similar
to that derived by Warner et al.,[11] whose QSAR analyses,
like ours, used an indicator variable to account for the higher
activity of bisbiguanides compared with biguanides. This
may indicate that biguanides can interact more effectively
with the negatively charged phospholipid head groups due
to the larger size of the cationic biguanide moiety. Recently,
Table 3. QSAR equations from current and previous studies
Organism QSAR equation
Compound
log P0
1
2
3
4
5
6
7
8
9
carnitine esters[32]
carnitine esters[32]
biguanides[11]
Gram+
−0.17 log P2 + 1.81 log P + 0.58
5.2
5.4
5.8
5.5
5.7
6.6
6.5
6.2
5.1
4.2
4.9
6.0
6.0
5.2
…
Yeasts
−0.15 log P2 + 1.57 log P + 0.95
S. mutans
S. mutans
S. mutans
S. mutans
S. mutans
S. mutans
S. aureus
P. aerug.
E. coli
−0.17 log P2 + 1.94 log P − 0.29
carbamimidates[11]
3 + 4[11]
−0.19 log P2 + 2.13 log P − 0.60
−0.17 log P2 + 1.97 log P − 0.32
bisbiguanides[11]
3 + 4 + 6[11]
−0.08 log P2 + 1.04 log P + 2.17
−0.12 log P2 + 1.54 log P + 0.56
3 + 4 + 6[11]
−0.13 log P2 + 1.56 log P + 0.34Ibis + 0.48
−0.18 log P2 + 1.83 log P − 0.57σp − 5.73
−0.11 log P2 + 0.93 log P − 0.35σp − 4.27
−0.20 log P2 + 1.95 log P − 6.05
2-pyridylguanidines[12]
10 2-pyridylguanidines[12]
11 2-pyridylguanidines[12]
12 2-pyridylguanidines[12]
13 2-pyridylguanidines[12]
14 bispyridylguanidines[12]
15 bispyridylguanidines[12]
16 bispyridylguanidines[12]
17 bispyridylguanidines[12]
18 bispyridylguanidines[12]
19 biguanides[33]
P. vulgaris
−0.17 log P2 + 1.40 log P − 0.28π − 5.03
C. albicans −0.11 log P2 + 1.33 log P − 7.00
S. aureus
P. aerug.
E. coli
−0.08 log P2 + 0.83 log P − 2.53σm − 3.26
+0.28 log P − 3.27
−0.12 log P2 + 1.12 log P − 3.60
+0.23 log P + 2.34σm − 3.23
4.7
…
P. vulgaris
C. albicans −0.07 log P2 + 0.80 log P − 2.93
5.7
5.5
5.7
5.7
5.2
5.2
6.0
6.4
6.1
5.2
S. aureus
MRSA
−0.26 log P2 + 2.92 log P − 3.61
20 biguanides[33]
−0.30 log P2 + 3.45 log P − 5.02
21 biguanides[33]
E. coli
−0.21 log P2 + 2.42 log P − 1.82
22 biguanides[33]
P. aerug.
B. cepacia
S. aureus
MRSA
−0.21 log P2 + 2.21 log P − 1.10
23 biguanides[33]
−0.54 log P2 + 5.67 log P − 9.99
24 this work
25 this work
26 this work
27 this work
−0.10 log P2 + 1.18 log P + 0.60Ibis + 1.11
−0.08 log P2 + 1.05 log P + 0.73Ibis + 1.22
−0.15 log P2 + 1.83 log P + 0.84Ibis − 1.39
E. coli
C. albicans −0.16 log P2 + 1.60 log P + 0.15

Studies of Antimicrobial Agents Containing Biguanide Isosteres
83
infections and interference with this process may represent
a major component of the mode of action of antifungal
drugs.[36]
2,2ꢀ-N,Nꢀ-(hexane-1,6-diyl)bis(2-aminopyrimidinium)
Dihydrochloride 8
A
solution of 2-chloropyrimidine (900 mg, 8 mmol) and 1,6-
diaminohexane (690 mg, 6 mmol) in dioxane (12 mL) containing tri-
ethylamine (1.6 mL, 1.16 g, 11 mmol) was heated under reflux for 8 h.
The cooled mixture was evaporated in vacuo, mixed as a slurry in water
and the solid collected and well washed with water. Crystallization from
ethanol afforded the free base as an off-white powder (430 mg, 40%),
mp 178–179.5◦C. δH (CDCl3) 1.28–1.5 (4H, m, 2 × CH2), 1.5–1.70
(4H, m, 2 × CH2), 3.39 (2H, q, J 6, 2 × NHCH2), 5.06–5.36 (2H, br s,
2 × NHCH2), 6.49 (2H, t, J 5, pyrimid H-5), 8.24 (4H, d, J 5, pyrimid
H-6). MS (CI) m/z 273 (MH+, 100%).
A solution of the free base (400 mg, 1.5 mmol) in warm ethanol
(14 mL) was treated with an excess of concentrated hydrochloric acid
(0.6 mL). The solution was treated with ethyl ether until precipita-
tion was complete. The precipitate was collected and as a solution in
methanol (25 mL), treated with activated carbon followed by addition
of excess ethyl ether until precipitation was complete. The solid was
collected to afford the dihydrochloride 8 as off white crystals (335 mg,
65%), mp 233–234◦C. δH ([D4]MeOH) 1.53–1.75 (4H, m, 2 × CH2),
1.75–1.98 (4H, m, 2 × CH2), 3.69 (4H, t, 2 × NHCH2), 4.97–5.20 (8H,
br s, 4 × NH and NH+ and H2O), 7.12 (2H, t, J 6, pyrimid H-5), 8.61–
8.86 (4H, bs, pyrimid H-6). νmax (KBr disc)/cm−1 3600–2200, 1625,
1460, 1330, 1200, 1100, 1080, 1050, 1030, 975, 790, 770. MS(CI)
m/z 273 (MH+ – 2HCl, 100).
Conclusions
Theresultsofthisstudyprovideinsightintothephysicochem-
ical requirements for antimicrobial activity. Although most
of these agents contain linear chains, the molecular length is
not an important factor in the structure–activity relationships
as compound 12 and several of the most active biguanides
reported by Tsubouchi et al.[33] show. Rather, there is a
consistent parabolic dependence of activity on the lipophilic-
ity of the compounds. The optimum log P of 5.5 for anti-
microbial activity is similar for all compound classes studied
and for most species. The activity may also be increased by
the presence of a biguanide group (as in chlorhexidine). It is
essential for antimicrobial activity that the molecule contains
a basic functional group that can become positively charged
in the microbe environment. Our findings are consistent with
current theories of microbe phospholipid biolayer interaction
and destabilization.
2,2ꢀ-N,Nꢀ-(decane-1,6-diyl)bis(2-aminopyrimidinium)
Dihydrochloride 9
This suggests considerable scope for use of topographi-
cal mimics of the biguanide moiety in antimicrobial agents
of this type. It appears that antimicrobial agents require a
log P value near 5.5 and a nitrogen atom capable of carry-
ing at least a partial positive charge at physiological pH.[37]
Conformational restriction of the biguanide analogue within
the cyclic analogues does not seriously affect antimicrobial
activity.
A solution of 2-chloropyrimidine (1.8 g, 8 mmol) and 1,10 diaminode-
cane (2.06 g, 12 mmol) in dioxane (25 mL) containing triethylamine
(3.2 mL, 22 mmol) was treated in an identical manner to 8 to yield,
after recrystallization from ethanol, 9 as colourless crystals (1.19,
45%), mp 105–106◦C. δH (CDCl3) 1.10–1.35 (12H, br s, 6 × CH2),
1.35–1.64 (4H, m, 2 × CH2), 3.35 (4H, q, J 6, 2 × NHCH2), 5.2–5.6
(2H, br s, 2 × NHCH2), 6.49 (2H, t, J 5, pyrimid H-5), 8.26 (2H, d,
J 5, pyrimid H-6). MS(CI) m/z 329 (MH+, 100).
A solution of the free base (490 mg, 1.5 mmol) in ethanol (6 mL)
was treated with concentrated hydrobromic acid (0.7 mL, 6.2 mmol) by
an identical procedure to 8 to afford the dihydrobromide 9 as colourless
crystals (606 mg, 82%), mp 138–139◦C. δH (CDCl3) 1.0–1.5 (12H, br s,
6 × CH2), 1.5–1.71 (4H, m, 2 × CH2), 3.58 (4H, q, J 6, 2 × NHCH2),
6.50–6.84 (2H, m, J 5, pyrimid H-5), 8.13–6.32 (2H, m, J 5, pyrimid
H-6), 8.51–8.86 (2H, m, 2 × NH). νmax (KBr disc)/cm−1 3150, 3050,
2920, 2850, 1640, 1535, 1470, 1420, 1320, 1210, 1080, 1000, 780.
MS(CI) m/z 329 (MH+ – 2HBr, 100), 108(100).
Experimental
General
Antimicrobial design was based on a model structure for chlorhexidine
derived from the crystal structure of biguanide[38] using the molecule
building capabilities of the Sybyl modelling package.[39]
Chemical Synthesis
Melting points were measured (uncorrected) on an Electrothermal appa-
ratus. 1H NMR spectra were measured on a Varian EM 360 instrument
or a Bruker WM 250 spectrometer with trimethylsilane as the inter-
nal standard. All spectra quoted in the text were measured at 250 MHz
unless otherwise specified. 13C NMR spectra were measured on a
Bruker WM 250 MHz instrument at 62.9 MHz. Infrared spectra were
measured on a Perkin–Elmer 783 instrument as KBr discs unless other-
wise reported in the text. Mass spectra were obtained with a Finnigan
3300 instrument or a JEOL JMS-DX 303 instrument as chemical ioniza-
tion spectra with methane as the reagent gas. High-resolution chemical
ionization mass spectra were recorded on a Micron SS 77OF instru-
ment using methane as the reagent gas. Microanalyses were undertaken
at the National Analytical Laboratory, Melbourne. Elemental analy-
ses are within 0.4% of the calculated values. Dioxane was passed
through a column of aluminium oxide before use. 2-Chloropyrimidine
(Janssen) was purified by recrystallization from petroleum spirit before
use. DMSO was dried by stirring over calcium hydride for 48 h. and then
distilling under vacuum. HPLC was carried out using aWaters 501 pump
and a Du Pont Zorbax C18 analytical column and acetonitrile/water
(45 : 55) as the eluting solvent system. Column Chromatography was
carried out by using aluminium oxide (Merck, Art 1077, 90 Activ. neu-
tral 0.063–0.2 mm). Petroleum Spirit was distilled before use and refers
to the 60–80◦C fraction.
2-Chloro-4-(4-chlorophenoxy)pyrimidine 13
4-Chlorophenol (15.42 g, 0.12 mol) was dissolved in a solution of
sodium hydroxide (4.8 g, 0.12 mol) in water (30 mL). Acetone (90 mL)
was added and the stirred mixture maintained at ice-bath temperature.To
this was added, dropwise, a solution of 2,4-dichloroprymidine (17.88 g,
0.12 mol) in acetone (40 mL). The stirred mixture was maintained at
ice-bath temperature for 4 h. during which time a precipitate formed.
The mixture was extracted with ethyl ether (1 × 450 mL) and the ether
layer washed with 2.5 M NaOH (3 × 30 mL), water (3 × 30 mL), dried
(Na2SO4), and the ether evaporated in vacuo to yield a moist solid. Crys-
tallization from light petroleum afforded 13 (17.40 g, 60%) as colourless
flakes, mp 111–113◦C. δH (CDCl3) 6.83 (2H, d, J 5.5, pyrimid H-5),
7.08–7.13 (2H, m, phenyl), 7.36–7.42 (2H, m, phenyl). 8.45 (2H, d,
J 5.5, pyrimid H-6). νmax (KBr disc)/cm−1 1600, 1560, 1480, 1425,
1335, 1310, 1220, 1190, 1160, 1100, 1085, 1015, 980, 950, 845, 810,
810, 760, 720, 650. MS(CI) m/z 241 (MH+, 100), 243 (62). HRMS
calcd for C10H7Cl2N2O, 240.9935; found 240.9910.
2,2ꢀ-N,Nꢀ-(hexane-1,6-diyl)bis[4-(4-chlorophenoxy)2-
aminopyrimidinium] Dihydrochloride 10
A solution of 13 (3.62 g, 15 mmol) and 1,6-diaminohexane (1.31 g,
11.3 mmol) in dioxane (36 mL) containing triethylamine (2.91 mL,

84
G. T. Wernert et al.
21 mmol) was heated under reflux for 10 h. The cooled mixture was
evaporated in vacuo, mixed with water, and the solid collected and
washed well with water. Crystallization from ethanol (activated car-
bon) afforded the free base as an off-white powder (1.55 g, 39%), mp
158–159◦C. δH (CDCl3) 1.05–1.4 (4H, br s, 2 × CH2). 1.4–1.74 (4H, br
s 2 × CH2), 3.1–3.5 (4H, br s, 2 × NHCH2), 4.8–5.7 (2H, br s, 2 × NH),
6.08 (2H, d, J 5, pyrimid H-5), 7.09 (4H, d, J 8, phenyl), 7.37 (4H, d,
J 8, phenyl), 8.10 (2H, d, J 8, pyrimid H-6). MS(CI) m/z 525 (MH+,
100), 527(70).
The free base (787 mg, 1.5 mmol) as a suspension in warm methanol
(5 mL) was treated with concentrated hydrochloric acid (0.6 mL,
7 mmol) and the resulting solution treated with activated charcoal. Ethyl
ether was added until precipitation commenced. Crystallization at ice-
bath temperature afforded 10 as an off-white powder (653 mg, 73%),
mp 214.5–216◦C. δH ([D6]DMSO) 0.8–1.6 (8H, m, 4 × CH2), 2.9–3.15
(2H, br s, NHCH2), 3.15–3.45 (2H, br s, NHCH2), 6.45–6.71 (2H, br d,
pyrimid H-5), 7.30 (4H, d, J 8, phenyl), 7.42–7.62 (4H, br d, phenyl),
8.10–8.45 (2H, m, pyrimid H-6), 8.7–9.1 (2H, bm, 2 × NH). νmax (KBr
disc)/cm−1 3600–2200, 1630, 1550, 1480, 1440, 1420, 1380, 1340,
1320, 1290, 1270, 1210, 1100, 1080, 1010, 970, 860, 800, 700, 610.
MS(CI) m/z 525 (MH+ – 2HCl, 100), 527(70).
7.44 (1H, s, J 6.0, pyrimid H-5), 7.44 (2H, d, J 9.0, phenyl), 7.71 (2H,
d, J 9.0, phenyl), 8.45 (1H, s, pyrimid H-6), 10.40 (1H, s, NH). νmax
(KBrdisc)/cm−1 3330, 1620, 1580, 1500, 1400, 1375, 1350, 1290, 1210,
1125, 1010, 990, 970, 935, 825, 780, 720. MS(CI) m/z 284 (MH+, 100),
286(35). HRMS: calcd for C11H10ClN3SO2 284.0260, found 284.0270.
2,2ꢀ-N,Nꢀ-(hexane-1,6-diyl)bis[4-(4-chloroanilino)2-
aminopyrimidinium Dihydrochloride 11
A solution of 1,6-diaminohexane (244 mg, 2.1 mmol) and 16 (57 mg,
2 mmol) in DMSO (3 mL) was heated to 100–105◦C for 5 h.The mixture
was maintained at room temperature overnight and water was added
dropwise to the stirred solution until precipitation was complete. The
precipitate was well washed with water (decanting) and dried in vacuo to
yield a sticky white solid. The solid was purified by column chromato-
graphy on alumina (60 g) eluting with methanol to yield the free base
as white crystals (160 mg, 31%), mp 90–93◦C. δH (60 MHz; CDCl3)
1.2–1.8 (8H, br s, 4 × CH2), 3.15–3.55 (4H, br m, 2 × CH2, 2 × NH),
5.2–5.6 (2H, br s, 2 × NH), 6.02 (2H, d, J 6, pyrimid H-5), 7.04–7.64
(8H, br s, phenyl), 8.02 (2H, d, J 6, pyrimid H-6). MS(CI) m/z 523
(MH+), 525 (70), 527 (14), 489 (14). HRMS: calcd for C26H29Cl2N8
523.1892, found 523.1861.
A warm solution of the free base (170 mg, 0.325 mmol) in methanol
(6 mL) was treated dropwise with concentrated hydrochloric acid
(0.1 mL). The solution was filtered and the filtrate treated with ethyl
ether (approx. 25 mL) until precipitation commenced. The mixture was
cooled in the refrigerator overnight and the precipitate collected to
afford the dihydrochloride 11 as off-white beads (185 mg, 95%), mp
234–236.5◦C. δH ([D6]DMSO) 1.18–1.51 (4H, br s, 4 × CH2), 1.51–
1.71 (4H, br s, 4 × CH2), 3.1–3.8 (8H, m, 2 × CH2, H2O), 6.42 (2H, d,
J 6.5, pyrimid H-5), 7.18–7.49 (4H, br s, phenyl), 7.49–7.84 (4H, br s,
phenyl), 7.88 (2H, d, pyrimid H-6). 8.4–8.7 (2H, bs, 2 × NH), 11.0–11.4
(2H, bs, 2 × NH), 12.2–12.8 (2H, bs, 2 × NH+). νmax (KBr disc)/cm−1
3550–2300, 1645, 1550, 1520, 1485, 1450, 1380, 1220, 1080, 1010, 830,
780. MS(CI) m/z 523 (MH+ – 2HCl, 100), 525 (70), 527 (14) 489 (14).
6-Chloro-2-methylthio-N-(4-chlorophenyl)-4-pyrimidineamine 14
A mixture of 4,6-dichloro-2-methylthiopyrimidine (21.60 g, 0.11 mol)
and 4-chloroaniline (12.80 g, 0.10 mol) in glacial acetic acid was treated
with concentrated hydrochloric acid (5 mL). The stirred mixture was
heated to 100◦C for 3 h. and during this time a thick, white precip-
itate formed. After cooling the reaction, the solid was collected and
washed well with glacial acetic acid.The solid was suspended in ethanol
(200 mL) and made alkaline (pH 8–9) with ammonium hydroxide. Suffi-
cient water was added to completely precipitate the crude compound that
was collected, which was washed well with water. Crystallization from
a mixture of ethanol and water (1 : 1) afforded 14 as colourless crystals
(18.94 g, 66%), mp 157–159◦C. δH (CDCl3) 2.52 (3H, s, CH3S), 6.31
(1H, s, pyrimid H-5), 6.8–7.0 (1H, bs, NH), 7.29 (2H, d, J 9.0, phenyl),
7.35 (2H, d, J 9.0, phenyl). 13C NMR (250 MHz; CDCl3) 14.17, 98.49,
123.92, 129.58, 130.70, 135.94, 160.02, 161.07. νmax (KBr disc)/cm−1
3280, 3190, 3140, 3070, 3000, 2920, 1610, 1565, 1535, 1480, 1400,
1350, 1285, 1230, 1200, 1120, 1090, 1010, 970, 860, 820, 730, 690, 660.
MS(CI) m/z 286 (MH+, 100), 288 (67), 290 (14), 250 (36), 177 (44).
Bis-N-(4-chlorophenyl)-2,4-pyrimidineamine hydrochloride 12
This compound was prepared by the method of Gosh.[28] The crude
salt was crystallized twice from ethanol to afford 12 as fluffy needles
(2.30 g, 35%), mp 225–227◦C (lit[39] mp 225◦C). The literature melting
point was incorrectly reported as that for the free base. δH ([D6]DMSO)
2.9–4.2 (2H, bs, 2H2O), 6.58 (1H, d, J 7, pyrimid H-5), 7.39–7.54 (6H,
m, phenyl), 7.64 (2H, d, J 8.5, phenyl), 8.03 (1H, d, J 7, pyrimid H-6),
10.90 (1H, s, NH), 11.32 (1H, s, NH), NH+ not seen. δC (250 MHz;
[D6]DMSO) 99.78, 123.46, 124.02, 128.59, 128.76, 128.89, 135.61,
136.33, 143.50, 152.47, 160.92. νmax (KBr disc)/cm−1 3400, 3200,
3120, 3090, 1660, 1605, 1585, 1550, 1525, 1490, 1450, 1380, 1210,
1100, 1010, 820, 780, 740. MS(CI) m/z 331(MH+-HCl, 100). HRMS:
calcd for C16H14Cl3N4·HCl 331.0501, found 331.0517.
6-Chloro-2-methylsulfonyl-N-(4-chlorophenyl)-
4-pyrimidineamine 15
To a stirred suspension of 14 (3.43 g, 0.012 mol) in glacial acetic acid
(12 mL) was added hydrogen peroxide (30% w/v)(4.0 mL, 0.035 mol)
dropwise. The mixture was stirred at room temperature for 3 d, after
which time, an additional aliquot of hydrogen peroxide (30% w/v,
4.0 mL) was added. The mixture was stirred at room temperature for
an additional 24 h. Water (30–40 mL) was added and the resulting
precipitate that formed was collected, washed with water, and dis-
solved in ethyl acetate (50–60 mL). The solution was washed with 1M
sodium carbonate (3 ×), water (3 ×), dried (Na2SO4), and the solvent
removed to yield a white solid. Crystallization from ethanol afforded
the sulfone 15 as colourless needles (1.84 g, 54%), mp 173–174◦C. δH
(CDCl3/[D6]DMSO) 3.29 (3H, s, CH3SO2) 6.92 (1H, s, pyrimid H-5),
7.35(2H, d, J 8.8, phenyl), 7.60(2H, d, J 8.8, phenyl), 10.53(1H, s, NH).
νmax (mull)/cm−1 3390, 1610, 1565, 1300, 1230, 1140, 1100, 1010, 970,
960, 825, 760. MS(CI) m/z 318 (MH+, 100), 320 (67), 322 (14).
Microbiology
Stock solutions of chlorhexidine diacetate (Sigma) and test compounds
were prepared in water or DMSO at 10 000 µg mL−1. Compounds were
screened against Staphylococcus aureus (NCTC 4163), a clinical isolate
of methicillin resistant S. aureus (MRSA), Escherichia coli (NCTC
8196), Pseudomonas aeruginosa (NCTC 6749), and Candida albicans
(RMIT, QAP 1987). Susceptibility testing was by macro broth dilution
(1 mL volumes) in log2 dilution from 512 to 0.25 µg mL−1 in Oxoid
Tryptone Soya Broth, inoculum density of 2–5 × 106 cfu mL−1 and
incubation at 35◦C to 24 h. MIC were expressed as − log10MIC [µM]
and are the results of at least two replicates.
2-Methylsulfonyl-N-(4-chlorophenyl)-4-pyrimidineamine 16
A mixture of 15 (1.91 g, 6 mmol) in methanol was treated with magne-
sium oxide (2.40 g, 40 mmol) and Pd/charcoal catalyst (10%) (800 mg).
The vigorously stirred mixture was hydrogenated at atmospheric pres-
sure and temperature for 6.5 h. when HPLC indicated that no starting
material remained. The mixture was filtered through a celite pad and
the filtrate evaporated in vacuo to yield a white powder. Crystallization
from methanol afforded the pyrimidinesulfone 16 as a white powder
(0.750 g, 44%), mp 190–191◦C. δH ([D6]DMSO) 3.31 (3H, s, CH3SO2),
QSAR Analyses
Multiple regression analyses were run on a personal computer using
the statistical analysis package Statview (ver. 4.02). The log P
(octanol/water) values were calculated by the CLogP (ver. 1.0)
program[40] although this was missing a parameter for the aryl guanide
moiety. We found this could be overcome by using a urea moiety in place
oftheguanidebycomparisonswithexperimentallogP values.ThelogP

Studies of Antimicrobial Agents Containing Biguanide Isosteres
85
values were also checked against those provided by the HINT! program
of Abraham and Kellogg.[41] These programs calculated log P values
for chlorhexidine very close to the experimental log P value of 4.87.[37]
Partition coefficients were not corrected for ionization due to pKa as
these authors have shown the average difference between the log P of
biguanide free bases and their salts was 2.35 0.07 log units. An indi-
cator variable was used to discriminate between bisbiguanides and other
compounds, as this was found to be important in QSAR studies of related
compounds.[11] We omitted compounds with unquantified, low activity.
We also attempted statistical analyses with the inactive compounds set
to values four times their reported ‘greater than’ MIC values to explore
the importance of basicity of the molecules on activity. Due to the small
number of data points care was taken to ensure that a minimum of para-
meters were screened in the QSAR analyses in order to reduce the risk of
chance correlations.[42] The ionic speciation calculations and conjugate
acid pK calculations were performed using the University of Georgia
SPARC Calculator.[43] Calculated pK values were consistent to those
of similar compounds (where available) from Perrin’s compilations.[44]
[15] D. A. Winkler, Briefings Bioinform. 2002, 3, 73.
[16] Y. Murata, E. Miyamoto, M. Ueda, J. Pharm. Sci. 1991, 80, 26.
[17] R. A. Coburn, M. T. Clark, R. T. Evans, R. J. Genco, J. Med.
Chem. 1987, 30, 205.
[18] J. Anné, E. De Clercq, H. Eyssen, O. Dann, Antimicrob. Agents
Chemother. 1980, 18, 231.
[19] L. Fabbrizzi, M. Cicheloni, P. Paoletti, G. Schwarzenbach, J. Am.
Chem. Soc. 1977, 99, 5574.
[20] A. A. Pinkerton, D. Schwarzenbach, J. Chem. Soc., Dalton
Trans. 1978, 8, 989.
[21] M. J. Jordan, J. E. Gready, J. Comput. Chem. 1989, 10, 186.
[22] T. Fujita, Biosci. Biotech. Biochem. 1996, 60, 557.
[23] B. C. Elmes, Aust. J. Chem. 1996, 49, 573.
[24] G. Menichi, J. Het. Chem. 1984, 21, 209.
[25] D. Galanakis, C. A. Davis, B. D. R. Herrero, C. R. Ganellin,
P. M. Dunn, D. H. Jenkinson, J. Med. Chem. 1995, 38, 595.
[26] J. H. Lister, J. Chem. Soc. 1960, 3394.
[27] (a) T. Jojima, H. Takeshiba, T. Kinito, Y. Kondo, Annu. Rep.
Sankyo Res. Lab. 1980, 32, 114.
(b) T. Jojima, H. Takeshiba, T. Kinito, Y. Kondo, Annu. Rep.
Sankyo Res. Lab. 1966, 9, 423.
References
[1] Pesticide Manual, 10th Edition (Ed. C. Tomlin) 1994 (Royal
Society of Chemistry: London).
[2] (a) G. E. Davies, J. Francis, A. R. Martin, F. L. Rose, G. Swain,
Br. J. Pharmacol. 1954, 9, 192. (b) F. L. Rose, G. Swain,
J. Chem. Soc. 1956, 4422.
[3] R. E. Johnson, A. E. Soria, J. R. O’Connor, R. A. Dobson,
J. Med. Chem. 1981, 24, 1314.
[4] D. M. Bailey, C. G. De Grazia, S. J. Hoff, P. L. Schulenberg,
J. R. O’Connor, D. A. Paris, A. McKenzie Slee, J. Med. Chem.
1984, 27, 1457.
[28] D. Gosh, J. Med. Chem. 1966, 9, 423.
[29] E. P. Greenberg, Nature 2000, 406, 947. doi:10.1038/35023203
[30] T. Ikeda, A. Ledwith, C. H. Bamford, R. A. Hann, Biochim. Bio-
phys. Acta 1984, 769, 57. doi:10.1016/0005-2736(84)90009-9
[31] Figure courtesy of Avecia Inc (http://www.avecia.com).
[32] F. Gasparrini, M. O. Tinti, M. Santaniello, N. Scafetta, G. Gallo,
L. Critelli, P. Foresta, in New Antimicrobial Agents: Synthesis
and Structure–Activity Relationship, in QSAR and Molecular
Modelling: Concepts, ComputationalTools and BiologicalAppli-
cations (Eds. F. Sanz, J. Giraldi, F. Manaut) 1995, p. 193 (Prous
Scientific: Barcelona).
[5] M. L. Arnold, A. D. Duriatti, M. Jung, R. B. Katz,
J. W. Liebeschuetz, Pestic. Sci. 1995, 44, 341.
[6] M. A. Ghannoum, N. M. Moussa, P. Whittaker, I. Swairjo,
K. H. Abu-Elteen, Chemother. 1992, 38, 46.
[33] H. Tsubouchie, K. Ohgura, K. Yasumura, H. Ishikawa,
M. Kikuchi, Bioorg. Med. Chem. Lett. 1997, 7, 1721.
doi:10.1016/S0960-894X(97)00297-7
[7] F. C. Sheppard, D. J. Mason, S. F. Bloomfield, V. A. Gant,
FEMS Microbiol. Lett. 1997, 154, 283. doi:10.1016/S0378-
1097(97)00343-1
[8] S. P. Denyer, Int. Biodeterior. Biodegrad. 1995, 36, 227.
doi:10.1016/0964-8305(96)00015-7
[9] S. P. Denyer, G. S. A. B. Stewart, Int. Biodeterior. Biodegrad.
1998, 41, 261. doi:10.1016/S0964-8305(98)00023-7
[10] T. E. Gunda,A. M. P. Koskinen, Eur. J. Med. Chem. 1991, 26, 449.
[11] V. D. Warner, D. M. Lynch, K. H. Kim, G. L. Grunewald,
J. Med. Chem. 1979, 22, 359.
[12] S. Lindholm, Ann. Acad. Sci. Fenn., Ser. A2: Chem. 1990, 231, 1.
[13] D. A. Winkler, G. Holan, in Rational Approaches to Structure,
Activity and Ecotoxicology of Agrochemicals (Eds. T. Fujita,
W. Draber) 1992, p. 123 (CRC Press: Boca Raton, FL).
[14] F. R. Burden, B. Rosewarne, D. A. Winkler, Chemome. Intell.
Lab. Syst. 1997, 38, 127. doi:10.1016/S0169-7439(97)00052-X
[34] T. Kuyyakanond, L. B. Quesnel, FEMS Microbiol. Lett. 1992,
79, 211. doi:10.1016/0378-1097(92)90211-6
[35] J. A. Chawner, P. Gilbert, J. Appl. Bacteriol. 1989, 66, 253.
[36] V. Vuddhakul, J. Antimicrob. Chemother. 1988, 21, 755.
[37] V. D. Warner, D. M. Lynch, R. S. Ajemian, J. Pharm. Sci. 1976,
65, 1070.
[38] S. R. Ernst, F. W. Cagle, Acta Crystallogr. 1977, B33, 235.
[39] Tripos Associates (http://www.tripos.com).
[40] Biobyte Corporation (http://www.biobyte.com).
[41] G. E. Kellogg, S. F. Semus, D. J. Abraham, J. Comput. Aided
Mol. Des. 1991, 5, 545.
[42] J. G. Topliss, R. P. Edwards, J. Med. Chem. 1979, 22, 1238.
[43] University of Georgia SPARC calculator (http://ibmlc2.chem.uga.
edu/sparc/index.cfm).
[44] D. D. Perrin, Dissociation Constants of Organic Bases in
Aqueous Solution, 1972 (Butterworths: London).