Functional foldamers that target bacterial membranes: The effect of charge, amphiphilicity and conformation

Article abstract of DOI:10.1016/j.bmc.2016.07.017

By varying the molecular charge, shape and amphiphilicity of a series of conformationally distinct diarylureas it is possible to control the levels of phospholipid membrane lysis using membranes composed of bacterial lipid extracts. From the data obtained, it appears as though the lysis activity observed is not due to charge, conformation or amphiphilicity in isolation, but that surface aggregation, H-bonding and other factors may also play a part. The work provides evidence that this class of foldamer possesses potential for optimisation into new antibacterial agents.

Full text of DOI:10.1016/j.bmc.2016.07.017

Bioorganic & Medicinal Chemistry 24 (2016) 4241–4245
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Functional foldamers that target bacterial membranes: The effect
of charge, amphiphilicity and conformation
⇑
Yogita Patil-Sen, Sarah R. Dennison, Timothy J. Snape
School of Pharmacy and Biomedical Sciences, University of Central Lancashire, Preston PR1 2HE, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
By varying the molecular charge, shape and amphiphilicity of a series of conformationally distinct diary-
lureas it is possible to control the levels of phospholipid membrane lysis using membranes composed of
bacterial lipid extracts. From the data obtained, it appears as though the lysis activity observed is not due
to charge, conformation or amphiphilicity in isolation, but that surface aggregation, H-bonding and other
factors may also play a part. The work provides evidence that this class of foldamer possesses potential
for optimisation into new antibacterial agents.
Received 26 January 2016
Revised 13 June 2016
Accepted 9 July 2016
Available online 11 July 2016
Keywords:
Foldamer
Ó 2016 Elsevier Ltd. All rights reserved.
Antibacterial
Membrane
Diarylureas
Conformation
1. Introduction
area and discuss the significance of these molecular properties on
the ability of the compounds to lyse membranes composed of
There can no longer be any doubt that new antibacterial agents
are needed, drugs which not only have potent activity against
resistant strains of bacteria, but which are less susceptible to
developing resistance at a later date. The crisis associated with
antimicrobial resistance has generated major world-wide opportu-
nities for science and technology to lead the way, and one area that
could deliver some of the answers is the field dedicated to folda-
mer research.1–5 A foldamer can be defined as ‘a discrete oligomer
that folds into a conformationally ordered state in solution’, and
contemporary research has shown that a number of foldamer con-
structs (in particular, antimicrobial peptides (AMPs)) can interact
with, and disrupt, bacterial cell membranes thus making these
agents valid candidates for future therapeutics, particularly if
selectivity over host cells can be achieved.6–14
In light of this need, our research group has recently developed
a range of AMP-influenced mimetics which are based on a folda-
mer scaffold, under the presumption that control over antimicro-
bial properties could be obtained by fine-tuning the molecules’
charge, amphiphilicity and conformation.8,10–13,15–19 Previous
efforts have looked at the influence of foldamer length and confor-
mation on membrane interaction,15–19 but as yet, the effects of
charge and thereby amphiphilicity have not been studied against
bacterial membranes by us. Herein, we outline recent efforts in this
lipids extracted from Staphylococcus aureus (S. aureus) and Escher-
ichia coli (E. coli), with the aim of developing a new class of antibac-
terial agent.
2. Results and discussion
It is well established that N-unsubstituted diarylureas exhibit a
distinct difference in conformation when compared to their fully
N-substituted counterpart, both in solution and the solid state
(Fig. 1). For instance, upon full methylation, N,N0-diphenylurea
changes from the trans,trans-conformation to the cis,cis-conforma-
tion, as shown.20–23
In such a case the conformation can be determined by either 1H
NMR (solution state conformation), as evidenced by a diagnostic
upfield shift in the aromatic signals, or by obtaining the X-ray crys-
tal structures to determine the conformation in the solid state.24
The consequence of achieving such conformational control by
simple N-substitution has been studied in a number of applica-
tions, not least: for facilitating conformational communication
via stereogenic axes;25 for controlling oligourea helicity;26,27 for
designing promising anticancer and anti-bacterial agents;15–17,28
as a molecular splint;29 for carrying out a so-called ‘impossible’
macrocyclisation;30 and for the development of fluorescent
sensors.31
Another way to potentially exploit this conformational switch is
to prepare and evaluate compounds, which as a result, differ in
⇑
Corresponding author. Tel.: +44 (0)1772 895805.
E-mail address: tjsnape@uclan.ac.uk (T.J. Snape).
http://dx.doi.org/10.1016/j.bmc.2016.07.017
0968-0896/Ó 2016 Elsevier Ltd. All rights reserved.

4242
Y. Patil-Sen et al. / Bioorg. Med. Chem. 24 (2016) 4241–4245
and amphiphilicity, against theses membranes, suggesting that
once optimised, this class of molecule could be developed as
pathogen-selective antimicrobial agents.
3. Comparisons: Structure and lysis
Against S. aureus the trans,trans-compounds 1, 2 and 3 are more
membrane lytic than the cis,cis-isomers 4, 5 and 6 at almost all
concentrations studied, with 3 being the most potent compound
overall. Presumably, this is the case because this amine is the most
ionised at the pH used (pH 7.4, Fig. 2), but that conformation or H-
bonding must play a role too, since these are the main structural
differences between compounds 3 and the less active analogue,
6. That said, compound 7, a cis,cis-diarylurea-NH-amide, which is
less completely protonated at pH 7.4, is better still by several fold
than all the trans,trans-compounds suggesting that more compli-
cated factors are at play.
Interestingly, in the context of pathogen-selectivity, against
E. coli the pattern is different, with 1 being the best of the trans,
trans-analogues (despite being neutral at pH 7.4), although the
overall levels of lysis are slightly lower than the same compounds
against S. aureus. Conversely, in all cases, the cis,cis-compounds (5
and 6) tend to be more active than the trans,trans-compounds (2
and 3), with compound 6 being the best diarylurea overall by sev-
eral fold, and comparable to compound 7.
From the data in Tables 1 and 2 it appears as though lysis does
correlate with the pKa of the nitrogen which is protonated for com-
pounds 1–6, such that the glycine derivatives 3 and 6 (99.9% pos-
itively charged) are consistently the best lytic compounds against
both membrane types. The membrane of S. aureus is mainly com-
posed of the negatively charged dimyristoylphosphatidylglycerol
(DMPG) lipid, whilst the membrane of E. coli is mainly the nega-
tively charged DMPG and the zwitterionic (neutral) dimyris-
toylphosphatidylethanolamine (DMPE) lipid.32 As such, with the
membrane’s overall negative charge, it is expected that the posi-
tively charged compounds 3 and 6 would be the best at interacting
with the membranes and ultimately lysing them at a critical con-
centration (albeit relatively high with these low molecular weight,
un-optimised compounds). In addition, hydrophobic features to
penetrate the lipid layer of the membrane are important confirm-
ing that amphiphilic molecules as a whole are required.8
Interestingly, compound 4 is poor against both strains of bacte-
ria, as would be expected for a neutral, weakly amphiphilic com-
pound with no H-bond donor capability, but 1 sits right in the
middle of 2 and 3 in terms of is lytic ability against both E. coli
and S. aureus. Presumably, this is due, in-part at least, to its capac-
ity to H-bond both to itself and aggregate at the membrane surface,
thus disrupting the membranes’ electrical balance, and subse-
Figure 1. Conformational change induced in diphenylureas upon N-methylation.
their molecular dimensions and functionalities, such that activity
can be studied and apportioned to the individual properties of
interest in both the trans and cis forms, in this case their ability
to lyse bacterial membranes; it is assumed that the pKa of the com-
pounds being prepared, and thus their protonation state in the
assay media, would be the same for both conformations.
In order to be able to test each property individually (conforma-
tion, charge and thereby amphiphilicity) a series of compounds
were designed and prepared which exist in two discrete and stable
conformations depending upon their N-methylation status, as out-
lined in Scheme 1 and Figure 2.
The compounds in Figure 2 were chosen to enable a direct com-
parison to be made between the level of membrane lysis achieved
between compounds in the same conformation state, but differing
protonation levels (series 1–3 and 4–7) vs. compounds in the same
protonation state, but differing conformation (1 vs 4, 2 vs 5 and 3
vs 6). Compound 7 was a hybrid-type structure with both N–Me
and N–H functionality and a protonation site.
Tables 1 and 2 show the concentration-dependent lysis of com-
pounds 1–7 against membrane extracts from S. aureus and E. coli,
respectively, over a treatment period of 1 h, in a calcein-release
assay. The relatively weak maximum levels of lysis obtained, even
at the highest concentrations, suggests that the compound-mem-
brane interactions are not optimised against the membranes being
studied. These results were confirmed in minimum inhibitory con-
centration (MIC) studies against cultures of both bacterial strains
(Table 3), whereby relatively high values were observed. However,
it should be noted that these are small, individual monomer mole-
cules interacting with relatively large phospholipid membranes
which are usually disrupted by large aggregated oligomers. Herein
attempts have been made to identify key features for membrane
interaction and disruption which will be taken forward into larger
oligomers in future work.
Nonetheless, modest levels of membrane lysis are observed at
the lowest concentration studied (Tables 1 and 2), where the high-
est levels are given by 7 (ꢀ40%) against S. aureus and 7 (ꢀ24%)
against E. coli. Importantly, the variety of lysis levels obtained
against both strains is indicative of some selectivity being observed
with the different compounds, which differ in their charge, shape
NH2
NCO
O
O
CH2Cl2
N
N
N
H
N
H
R
R
NaH, MeI (2 eq.)
R=H or NO2
1, R=H
NaH,
R
1a, R=NO2
Pd/C, H2
MeI (2 eq.)
4, R=H
2, R=NH2
i) Boc-Gly-OSu
ii) TFA
4a
, R=NO2
5, R=NH2
6, R=NHC(O)CH2NH2
3, R=NHC(O)CH2NH2
Pd/C, H2
i) Boc-Gly-OSu
ii) TFA
O
Cl
N
CH2Cl2, Et3N
7
, R=NHC(O)C5H4N
Scheme 1. Synthesis of the test compounds.

Y. Patil-Sen et al. / Bioorg. Med. Chem. 24 (2016) 4241–4245
4243
1 3
-
Testing the influence of charge at pH 7.4
-
-diarylureas
trans trans
O
O
O
O
NH2
N
H
N
H
N
H
N
H
NH2
N
H
N
H
N
H
1
2
3
(0%)
(0.16%)
(99.9%)
4-7 Testing the influence of charge at pH 7.4
cis-cis-diarylureas
O
O
O
O
N
O
N
N
N
N
N
N
N
O
NH2
NH2
N
H
N
H
4
5
6
(99.9%)
(0%)
(0.16%)
7 (0.7%)
N
Testing the influence of charge
Figure 2. Compounds synthesised to test the influence of conformation, charge and amphiphilicity on membrane lysis. The percentages in parentheses are the approximate
calculated values for protonation of the amine or pyridine nitrogen at pH 7.4 using the Henderson–Hasselbalch equation (see Supporting information).
Table 1
Percentage release for compounds 1–7 against calcein-loaded lipid vesicles prepared from the total phospholipid content extracted from the membrane of S. aureus. The values
shown are the average and standard deviations of five experiments
Concentration (
lM)
1
2
3
4
5
6
7
0
0
0
0
0
0
0
0
93.73
187.5
375
750
1500
3000
7.18 0.49
9.20 1.32
13.84 0.16
15.34 0.07
18.34 0.23
25.50 0.99
4.58 0.27
8.34 0.19
10.09 0.25
12.67 0.41
13.84 0.52
15.47 0.30
6.91 0.06
11.23 0.57
17.81 0.50
27.24 1.95
31.25 0.40
44.09 0.15
2.59 1.03
5.71 0.13
12.79 0.28
14.16 0.41
15.14 0.48
24.10 0.35
4.45 0.05
7.13 0.05
12.02 0.65
15.29 0.99
21.23 0.53
27.00 0.36
4.45 2.49
10.00 0.18
11.80 0.37
13.21 0.20
15.27 0.99
17.84 0.60
39.98 0.18
48.25 0.11
49.58 0.32
51.53 0.72
53.10 1.07
54.62 1.88
Table 2
Percentage release for compounds 1–7 against calcein-loaded lipid vesicles prepared from the total phospholipid content extracted from the membrane of E. coli. The values
shown are the average and standard deviations of five experiments
Concentration (
lM)
1
2
3
4
5
6
7
0
0
0
0
0
0
0
0
93.73
187.5
375
750
1500
3000
3.17 0.26
6.27 3.19
6.47 1.46
8.53 0.22
12.02 0.14
14.85 1.73
0.15 0.79
0.95 0.58
1.24 0.69
1.73 0.56
1.78 0.47
2.17 0.70
1.10 2.23
6.06 0.68
9.17 2.04
16.63 4.19
23.46 3.42
29.17 0.53
0.82 0.62
2.05 0.65
4.25 1.35
6.44 1.96
7.28 0.87
9.02 0.65
2.30 0.31
5.72 0.20
6.75 0.40
7.44 0.62
8.03 0.43
9.65 0.30
8.45 1.23
27.22 2.42
29.32 0.17
29.98 0.32
30.97 0.42
35.55 1.12
24.05 0.16
24.81 0.05
26.28 0.62
29.08 2.60
33.61 0.12
35.31 0.09
Table 3
Minimum inhibitory concentration (MIC) studies against both S. aureus and E. coli cultures
Bacteria
1
2
3
4
5
6
7
S. aureus
E. coli
2.0 mM
>3.0 mM
2.5 mM
>3.0 mM
1.5 mM
2.0 mM
2.5 mM
>3.0 mM
2.0 mM
>3.0 mM
2.5 mM
2.0 mM
1.0 mM
2.0 mM
quently inducing leakage of cytoplasmic material leading to cell
rupture, whereas 4 is unable to do so.
outer membranes of bacteria are able to exclude this external
hydrophobic probe, forcing it to remain in the aqueous environ-
ment. However, in the presence of certain membrane-permeabilis-
ing agents, the membrane becomes compromised and thus
sensitive to external factors. As a consequence, this sensitisation
allows entry of the hydrophobic probe into the hydrophobic envi-
ronment of the membrane. Thus, by virtue of its fluorescence in
hydrophobic environments, measuring NPN uptake gives an indica-
tion of any changes to the permeability of the outer membrane.
For the purposes of proof-of-concept, the membrane-
permeabilising NPN assay was performed at the MIC concentration
4. Membrane permeabilisation
To confirm that the compounds are acting through a membrane
disruptive mechanism, as a consequence of membrane insertion,
rather than through endocytosis, a membrane permeabilisation
assay was conducted using 1-N-phenylnaphthylamine (NPN) as
the fluorometric probe.33 NPN fluoresces strongly in phospholipid
environments but only weakly in aqueous environments, and intact

4244
Y. Patil-Sen et al. / Bioorg. Med. Chem. 24 (2016) 4241–4245
Table 4
Percentage uptake of NPN by S. aureus and E. coli in the presence of the test compounds at 2 mM
Time
1 h
Bacteria
1
2
3
4
5
6
7
S. aureus
E. coli
46.01 5.57
45.29 4.49
38.91 3.20
33.93 3.21
69.56 7.31
47.09 3.95
43.82 5.79
40.75 1.88
46.26 4.81
40.18 2.83
40.60 5.49
42.47 5.51
50.59 1.31
49.57 1.55
Overnight
S. aureus
E. coli
49.40 4.82
30.02 7.76
19.29 5.60
16.34 4.30
69.56 9.02
50.93 4.70
44.34 3.02
38.67 5.20
53.57 7.47
48.79 11.32
36.81 5.03
52.87 9.19
57.33 5.05
52.34 5.39
(2 mM), since this is the concentration at which the dose-depen-
dent curves begin to plateau and would represent the maximum
levels of insertion. Table 4 shows the percent lysis compared to
polymyxin B, an established membrane-permeable antibacterial
peptide, as control.34 Although uptake of NPN into the membrane
is moderate in the presence of the test compounds (up to 70%,
Table 4), compared to polymyxin B as control, the MIC activity is
overall weak. Nevertheless, the fact that membrane insertion is
confirmed suggests that potential exists to optimise the lytic prop-
erties of this compound class, especially if oligomeric constructs
can be developed to increase the effective concentration of the
amphiphilic monomers exposed to the membrane.
that surface aggregation, H-bonding and other factors may also
be at play, nevertheless, that membrane permeability takes place
has been proven in the NPN assay.
Acknowledgements
The authors would like to thank the EPSRC UK National Mass
Spectrometry Facility at Swansea University for acquiring the accu-
rate mass data, and the University of Central Lancashire for
support.
Supplementary data
5. Integy moment
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmc.2016.07.017.
The change in shape associated with N,N0-dimethylation of
diarylureas (Fig. 1) causes aromatic stacking to occur such that
the net molecular dimensions are reduced and thus the distance
between the centre of the lipophilic aromatic sections and the
hydrophilic charge is reduced in the cis,cis-diarylureas compared
to the trans,trans. As a result, the more folded cis,cis-conformers
would have different integy moments (the integy moment is a
measure of amphiphilicity which expresses the unbalance between
the centre of mass of a molecule and the barycenter of its hydro-
philic or hydrophobic regions)35 compared to the extended trans,
trans-conformers, thereby providing a physical property for which
to correlate against any observed membrane lysis. Unfortunately,
attempts to correlate the levels of lysis of S. aureus and E. coli mem-
branes with the integy moments (IW) of the test compounds, did
not yield any noticeable trends, despite precedent of such a corre-
lation being known with synthetic peptide mimics,8 suggesting
that the lysis activity observed is not due to charge, conformation
or amphiphilicity in isolation, but that surface aggregation, H-
bonding and other factors may also play a part. Although in isola-
tion the change in conformation must affect the integy moment of
compounds 1–3 versus 4–7, presumably the removal of two hydro-
gen-bond donors (N–H) upon N-methylation to give series 4–7,
changes the integy moments and complicates any correlation, as
does the difference in protonation states of the nitrogen atoms
under the physiologically relevant pH of 7.4 used.
References and notes
1. Gellman, S. H. Acc. Chem. Res. 1998, 31, 173.
2. Guichard, G.; Huc, I. Chem. Commun. 2011, 5933.
3. Huc, I. Eur. J. Org. Chem. 2004, 17.
4. Hill, D. J.; Mio, M. J.; Prince, R. B.; Hughes, T. S.; Moore, J. S. Chem. Rev. 2001, 101,
3893.
5. Saraogi, I.; Hamilton, A. D. Chem. Soc. Rev. 2009, 38, 1726.
6. Bechinger, B.; Lohner, K. Biochim. Biophys. Acta 2006, 1758, 1529.
7. Dennison, S. R.; Wallace, J.; Harris, F.; Phoenix, D. A. Protein Pept. Lett. 2005, 12,
31.
8. Thaker, H. D.; Cankaya, A.; Scott, R. W.; Tew, G. N. ACS Med. Chem. Lett. 2013, 4,
481.
9. Strömstedt, A. A.; Ringstad, L.; Schmidtchen, A.; Malmsten, M. Curr. Opin.
Colloid Interface Sci. 2010, 15, 467.
10. Gabriel, G. J.; Tew, G. N. Org. Biomol. Chem. 2008, 6, 417.
11. Som, A.; Vemparala, S.; Ivanov, I.; Tew, G. N. Biopolymers 2008, 90, 83.
12. Tew, G. N.; Scott, R. W.; Klein, M. L.; Degrado, W. F. Acc. Chem. Res. 2010, 43, 30.
13. Choi, S.; Isaacs, A.; Clements, D.; Liu, D.; Kim, H.; Scott, R. W.; Winkler, J. D.;
DeGrado, W. F. Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 6968.
14. Czaplewski, L.; Bax, R.; Clokie, M.; Dawson, M.; Fairhead, H.; Fischetti, V. A.;
Foster, S.; Gilmore, B. F.; Hancock, R. E. W.; Harper, D.; Henderson, I. R.; Hilpert,
K.; Jones, B. V.; Kadioglu, A.; Knowles, D.; Olafsdottir, S.; Payne, D.; Projan, S.;
Shaunak, S.; Silverman, J.; Thomas, C. M.; Trust, T. J.; Warn, P.; Rex, J. H. Lancet
Infect. Dis. 2016, 16, 239.
15. Dennison, S. R.; Akbar, Z.; Phoenix, D. A.; Snape, T. J. Soft Matter 2012, 8, 3258.
16. Dennison, S. R.; Phoenix, D. A.; Snape, T. J. Bioorg. Med. Chem. Lett. 2013, 23,
2518.
17. Dennison, S. R.; Snape, T. J.; Phoenix, D. A. Eur. Biophys. J. Biophys. Lett. 2012, 41,
687.
18. Fahs, S.; Rowther, F. B.; Dennison, S. R.; Patil-Sen, Y.; Warr, T.; Snape, T. J.
Bioorg. Med. Chem. Lett. 2014, 24, 3430.
19. Hester, T. J.; Dennison, S. R.; Baker, M. J.; Snape, T. J. Org. Biomol. Chem. 2015, 13,
8067.
6. Conclusion
20. Lepore, G.; Migdal, S.; Blagdon, D. E.; Goodman, M. J. Org. Chem. 1973, 38, 2590.
21. Lepore, U.; Lepore, G. C.; Ganis, P.; Germain, G.; Goodman, M. J. Org. Chem.
1976, 41, 2134.
22. Kagechika, H.; Azumaya, I.; Tanatani, A.; Fukutomi, R.; Yamaguchi, K.; Shudo, K.
Abstr. Pap. Am. Chem. Soc. 1998, 216, U494.
23. Tanatani, A.; Kagechika, H.; Azumaya, I.; Fukutomi, R.; Ito, Y.; Yamaguchi, K.;
Shudo, K. Tetrahedron Lett. 1997, 38, 4425.
24. Itai, A.; Toriumi, Y.; Saito, S.; Kagechika, H.; Shudo, K. J. Am. Chem. Soc. 1992,
114, 10649.
25. Clayden, J.; Vallverdu, L.; Helliwell, M. Org. Biomol. Chem. 2006, 4, 2106.
26. Kudo, M.; Hanashima, T.; Muranaka, A.; Sato, H.; Uchiyama, M.; Azumaya, I.;
Hirano, T.; Kagechika, H.; Tanatani, A. J. Org. Chem. 2009, 74, 8154.
27. Clayden, J.; Lemiegre, L.; Morris, G. A.; Pickworth, M.; Snape, T. J.; Jones, L. H. J.
Am. Chem. Soc. 2008, 130, 15193.
28. Snape, T. J.; Karakoula, K.; Rowther, F.; Warr, T. RSC Adv. 2012, 2, 7557.
29. Yamaguchi, K.; Matsumura, G.; Kagechika, H.; Azumaya, I.; Ito, Y.; Itai, A.;
Shudo, K. J. Am. Chem. Soc. 1991, 113, 5474.
We have shown that by varying the molecular charge, shape,
and thus amphiphilicity, of a series of diarylureas, it is possible
to control the level of lysis of membranes composed of bacterial
lipid extracts. The molecular structures have not been optimised
and thus future work will set out to combine the best features
required for membrane lysis (high pKa amine(s) or permanent pos-
itive charge(s) and lipophilic group(s), sufficiently separated along
the molecular axis giving rise to significant integy moments) into
new compounds and begin to extend the structures to oligomeric
scaffolds in the hope of designing new antibacterial agents. From
the data obtained so far, it appears as though the lysis activity
observed is not due to charge or amphiphilicity in isolation, but

Y. Patil-Sen et al. / Bioorg. Med. Chem. 24 (2016) 4241–4245
4245
30. Campbell, F.; Wilson, A. J. Tetrahedron Lett. 2009, 50, 2236.
31. Hirano, T.; Osaki, T.; Fujii, S.; Komatsu, D.; Azumaya, I.; Tanatani, A.; Kagechika,
H. Tetrahedron Lett. 2009, 50, 488.
33. Helander, I. M.; Mattila-Sandholm, T. J. Appl. Microbiol. 2000, 88, 213.
34. Falagas, M. E.; Rafailidis, P. I.; Matthaiou, D. K. Drug Resist. Updates 2010, 13,
132.
32. Lohner, K.; Prenner, E. J. Biochim. Biophys. Acta 1999, 1462, 141.
35. http://www.moldiscovery.com/software/vsplus/.