Anion complexation and transport by isophthalamide and dipicolinamide derivatives: DNA plasmid transformation in E. coli

Article abstract of DOI:10.1021/ja304816e

Tris-arenes based on either isophthalic acid or 2,6-dipicolinic acid have been known for more than a decade to bind anions. Recent studies have also demonstrated their ability to transport various ions through membranes. In this report, we demonstrate two

Full text of DOI:10.1021/ja304816e

Communication
pubs.acs.org/JACS
Anion Complexation and Transport by Isophthalamide and
Dipicolinamide Derivatives: DNA Plasmid Transformation in E. coli
Jason L. Atkins,†,‡,§ Mohit B. Patel,†,‡,§ Megan M. Daschbach,† Joseph W. Meisel,†,§
and George W. Gokel*,†,‡,§
‡
§
†Department of Chemistry & Biochemistry, Department of Biology, and Center for Nanoscience, University of MissouriSt.
Louis, St. Louis, Missouri 63121, United States
The approximate thickness of a tris-arene is 3.4 Å. Although
ABSTRACT: Tris-arenes based on either isophthalic acid
or 2,6-dipicolinic acid have been known for more than a
decade to bind anions. Recent studies have also
demonstrated their ability to transport various ions
through membranes. In this report, we demonstrate two
important properties of these simple diamides. First, they
transport plasmid DNA into Escherichia coli about 2-fold
over controls, where the ampicillin resistance gene is
expressed in the bacteria. These studies were done with
plasmid DNA (∼2.6 kilobase (kb)) in JM109 E. coli cells.
Second, known methods do not typically transport large
plasmids (>15 kb). We demonstrate here that trans-
formation of large pVIB plasmids (i.e., >20 kb) were
enhanced over water controls by ∼10-fold. These results
are in striking contrast to the normal decrease in
transformation with increasing plasmid size.
this is similar to the thickness of the aromatic bases in DNA,
the phosphate groups in the backbone are typically separated
by about twice that distance. In the structure reported for B-
DNA, the phosphate spacing varies but it is generally between 6
and 7 Å (as measured from the crystal structure available in the
Protein Data Bank as 1BNA).12 We show here that the simple
tris-arenes mediate the transport of DNA plasmids into E. coli
and that they are effective transformation agents for plasmids
larger than those that can be transported by currently available
chemical agents.
Eight simple tris-arenes were prepared for the present study.
They include isophthalanilides and dipicolinanilides that are
substituted in aniline’s (para) 4,4′-positions. The substituents
are methoxy, hydrogen (no substituent), chloro, and cyano.
The structures are shown in Figure 1. Compounds 2, 6, and 8
rabtree and co-workers reported that such simple tris-
C
arenes as N1,N3-dimesitylisophthalamide complex anions
by dint of their amide NH bonds.1 During the decade since that
report, numerous structural variants have been reported that
bind and/or transport an array of anions such as F−, Cl−, Br−,
and CH3COO−.2 Since these reports, isophthalic acid dianilides
and numerous structural variants have been studied by Gale
and co-workers3 and by others.4 The receptor has also been
used as a module in the construction of ion-pair hosts.5 Aniline
derivatives have been replaced by a range of other residues
including aliphatic, aromatic, and heterocyclic amines. A very
recent report details a supramolecular interaction between
anthracene-substituted pyridinium amides and DNA.6
Figure 1. Structures of tris-arenes 1−8.
were reported in an earlier study from our laboratory.10 The
substituents reported in this work represent a range of
electronic effects from releasing (1, 5) to withdrawing (4, 8).
The tris-arenes were prepared by treating either isophthalic
(1−4) or dipicolinic acid (5−8) with SOCl2, followed by
reaction with the appropriate aniline derivative. The products
were typically crystalline14 and isolated as stable solids. Solid-
state structures have been reported by Malone et al.13 for 2 and
6. The solid-state structures of relatives having hydroxyl,14
methyl,15 n-butyl,16 and nitro17 substituents have all been
reported.
Our interest in anion binding, specifically chloride anion
binding, resulted from work to develop a chloride anion
transporter that functions in phospholipid bilayer membranes.7
We called these compounds synthetic anion transporters
(SATs) and showed that they are selective for Cl− over K+
by more than 10-fold.8 A comparison of these SATs with the
tris-arenes showed that their binding affinities for Cl− were
similar and anion dependent.9 In more recent work, we showed
that at least one tris-arene showed channel-like behavior in
phospholipid bilayers10 and that they are generally good
phosphate-complexing agents.11
The ability of the 4,4′-dinitrodipicolinanilide to exhibit
channel-like behavior and the ability to bind phosphate
combined to suggest the possibility of binding and/or
transporting DNA. We envisioned that a stack of tris-arenes
within the membrane might foster transport of a DNA plasmid.
Studies were undertaken in E. coli to determine the abilities
of 1−8 to mediate the transfer of a DNA plasmid into a
microbe. The overall experiment was done as follows. A
Received: May 21, 2012
Published: August 3, 2012
© 2012 American Chemical Society
13546
dx.doi.org/10.1021/ja304816e | J. Am. Chem. Soc. 2012, 134, 13546−13549

Journal of the American Chemical Society
Communication
plasmid of specified size range was used that codes for
ampicillin resistance. The bacteria were grown on a medium
that is rich in ampicillin so that no bacteria will grow in the
absence of gene insertion (transformation). After exposure to
the plasmid and compounds, the bacteria are permitted to grow
and colonies are counted. The number of colonies obtained
after exposure to the plasmid and carrier is compared with the
number of colonies observed when only the plasmid is present
and when both plasmid and tris-arene are present.
revealed by an enhancement of colony-forming units (CFUs)
counted after 18−24 h on each plate. Each cell that acquires the
ampicillin resistance gene will grow into a colony, hence
“colony forming unit”. The colonies were counted on the plate
under microscopic examination.
The vertical bars in the graph of Figure 2 show trans-
formation efficiencies. Both DMSO and H2O controls gave
CFU counts of ∼32 with essentially no variation over replicates.
The results for transformation by the four isophthalanilides
were as follows: 1, R = OCH3, CFU ≈ 74; 2, R = H, CFU ≈
65; 3, R = Cl, CFU ≈ 27; and 4, R = CN, CFU ≈ 0. The results
observed in these experiments proved to be both encouraging
and surprising. First, a modest enhancement in transformation
was apparent for methoxy-substituted 1 and for unsubstituted
2. The enhancements were ∼2.3-fold and ∼2-fold, respectively.
The bacteria in contact with 4-chloro-substituted 3 showed a
decrease in transformation (only 80% of control values), a
reduction that is outside of experimental error. Moreover,
compound 4 whose para-substituent is cyano, showed
complete suppression of transformation.
Initial studies were done using Promega18 JM109 competent
cells. These cells are commercially prepared so as to be
amenable to plasmid transformation experiments. Two
plasmids were selected for study. The first was a 2.6 kilobase
(kb) (∼1.7 × 106 Da) plasmid that is expected to readily enter
competent E. coli cells. The second was a >20 kb (>1.3 × 107
Da) plasmid that was expected to be prohibitively large. Note
that the molecular weights in both cases were calculated by
using an average base pair weight of 660 Da.19 The smaller
plasmid was studied as a control and the larger plasmid was
expected to pass through the boundary membranes of the
microbe reluctantly if at all. Typically, a plasmid size of 13−15
kb is considered to be too large to traverse the membrane
barrier either of bacterial or eukaryotic cells.20
Both plasmids examined contained an ampicillin resistance
gene.21 Solutions of the synthetic isophthalic acid dianilides, 1−
4, were dissolved in dimethylsulfoxide (DMSO). Aliquots of
the resulting solution were added to standard RNase- and
DNase-free microcentrifuge tubes containing plasmid DNA
dissolved in nuclease-free water. Mixtures were allowed to react
at ambient temperature for 10 min. A suspension of the
competent cells (as provided by Promega), plasmid, and
dianilide or control chemical were mixed as described in the
experimental section. In each case, an E. coli culture was
exposed to a 2 μL aqueous solution containing the plasmid to
be tested [0.1 μg·mL−1] and isophthalic acid dianilide derivative
1−4 [2 μL] and plated on ampicillin-containing LB-agar media.
Each experiment was repeated a minimum of three times, and
the error bars shown on the graph of Figure 2 indicate the
reproducibility for compounds 1−4. The controls were
sampling solvents: either Milli-Q (18 MΩ) H2O or 2 μL of
DMSO added to Milli-Q H2O. Successful transformation was
A similar study was conducted with the ∼2.6 kb plasmid and
E. coli using the 2,6-dipicolinic acid analogues (5−8) of 1−4
under otherwise identical conditions. The results are shown in
the graph of Figure 3. The results generally parallel those
Figure 3. Transformation efficiencies plot for E. coli and 5−8. CFU
counted after E. coli was exposed to a 2 μL aqueous solution
containing 0.1 μg·mL−1 plasmid and 2 μL of dipicolinic acid
derivatives (5−8) and plated on ampicillin.
observed with the corresponding isophthalic acid derivatives.
The relative transformation efficiencies were as follows: 5, R =
OCH3, CFU ≈ 88; 6, R = H, CFU ≈ 44; 7, R = Cl, CFU ≈ 21;
and 8, R = CN, CFU ≈ 34. Although there is some difference in
magnitude comparing 1 to 5, 2 to 6, and 3 to 7, the major
difference is that the cyano-substituted dipicolinic acid tris-
arene 8 showed transformation similar to controls whereas 4
was not only inactive, it apparently suppressed activity.
Two conclusions can be reached from the data shown in
Figures 2 and 3. First, both isophthalanilides and dipicolina-
nilides function as transformation agents effectively in E. coli.
The enhancement over controls is modest, but real and
reproducible. Second, there is a general correlation between the
electronic effect of the para substituent and transformation
efficiencies for methoxy, chloro, and hydrogen. Indeed, plots of
1, 2, and 3 or 5, 6, and 7 vs Hammett22 σp gave straight lines
with r2 values of 0.86 and 0.98, respectively (data not shown).
The relationship obviously breaks down in both cases for 4 and
8, which are cyano-substituted. A plausible explanation for the
poor transformation efficiencies of tris-arenes having electron-
Figure 2. Transformation efficiencies plot for E. coli and 1−4. CFU
counted after E. coli was exposed to a 2 μL aqueous solution
containing 0.1 μg·mL−1 plasmid and 2 μL of isophthalic acid derivative
(1−4) and plated on ampicillin. Figure 1 represents the colony counts
resulting from a 100 μL aliquot, taken from a 1 mL solution. The
results were verified by using the entire 1 mL solution in a separate
experiment.
13547
dx.doi.org/10.1021/ja304816e | J. Am. Chem. Soc. 2012, 134, 13546−13549

Journal of the American Chemical Society
Communication
withdrawing groups is posed below. Notwithstanding, methoxy-
substituted 1 (isophthalic acid) and 5 (dipicolinic acid) show
transformation enhancements over control of ∼2.3-fold and
∼2.8-fold. These increases expand the versatility of the large
family of tris-arenes, the results are clearly of both practical and
of mechanistic interest, but the extent by which transformation
has been improved is modest.
When similar experiments were undertaken using “large”
(>20 kb) plasmids, the results obtained were far more dramatic.
Transformation of such large plasmids currently presents a
significant challenge in biology generally and in genetics in
particular. Plasmids >20 kb are inherently valuable because of
the number of genes they encode. The dianilides described
above, namely 1, 2, 4, and 5 were chosen for study because they
were successful with the smaller plasmid. We thus reproduced
the experiments described above using a >20 kb, rather than a
2.6 kb, plasmid. We did so being fully aware that the efficacy of
typical chemical transformation agents is limited to 13−15 kb
DNA.20 The results of the experiments conducted with E. coli,
large plasmids, and compounds 1, 2, 4, and 5 are shown in
Figure 4.
Figure 5. Schematic illustration of the postulated tris-arene:DNA
complexation model.
passive diffusiontransported through the bilayer. In support
of this hypothesis, we have varied the stoichiometry of
DNA:tris-arene over a wide range (data not shown). Optimal
transformation, as reflected in the highest colony counts, was
achieved with a ratio of 1:1 between receptor and DNA
phosphate. Of course, endocytosis may play a role.23
The complex must be strong enough to support transport
but unless the plasmid is released from the complexing agent
within the cell, no genetic modification can take place and the
phenotype will remain unexpressed. If binding to the phosphate
occurs, it will presumably be strongest when the aniline
substituent is most electron-withdrawing. The acidity of the
corresponding aniline will correlate with the amide acidities and
binding strengths. The pKA values for protonated anilines
increase for para substituents in the order NO2 < CN < Cl < H
< OCH3. The transformations shown in the graphs of Figures 3
and 4 show that rates decreased in the order OCH3 > H > Cl >
CN. Note that the decrease observed in the dipicolinamides
was similar to that of the isophthalamides except that Cl− was
less active than CN, which was equal to controls.
An issue that could affect the observation of colony growth is
the cytotoxicity of the tris-arenes. No difference in toxicity was
observed between any of the tris-arenes and DMSO alone in
Kirby−Bauer tests.24 Even so, minimum inhibitory concen-
trations (MICs, E. coli) were determined for compounds 1−8.
Note that the higher the MIC value, the lower is the
compound’s toxicity. In all cases, the MIC values were >1
mM. The maximum concentration of any tris-arenes used in
any experiment reported here was 40 μM. The concentration of
tris-arenes exposed to JM109 E. coli cells during most of the
experiment is only 2 μM.
Figure 4. Transformation efficiencies plot. CFU after E. coli was
exposed to a 2 μL aqueous solution containing 0.1 μg·L−1 plasmid
(pVIB) and 2 μL of isophthalic acid and dipicolinamide derivative and
plated on ampicillin. Bar: 1, R = OCH3, CFU ≈ 43; 2, R = H CFU ≈
33; 5, R = OCH3, CFU ≈ 48; 6, R = H, CFU ≈ 25. Control: 2 μM
DMSO, CFU ≈ 4; 2 mL of Milli-Q (18 MΩ) H2O, CFU ≈ 3.
The transformation experiment, conducted in the presence
of DMSO but in the absence of any dianilide, resulted in the
observation of ∼4 CFUs. In contrast, isophthalanilides
substituted by OCH3 (1), unsubstituted, (2), dipicolinanilide
substituted by OCH3 (5) and the unsubstituted dipicolinanilide
(6) resulted in 43, 33, 48, and 25 CFUs, respectively. The range
of enhancements is 1, 10-fold; 2, 8-fold; 5, 12-fold; and 6, 6-
fold. When these observations are coupled with the fact that the
plasmids tested were larger than are typically transported
through bilayers by known reagents, the results are both
significant and remarkable.
The transformation efficacy appears to follow a Hammett
relationship, diminishing roughly in the order of the substituent
sigma parameters.22 In all of the cases studied thus far, 4-
methoxy-substituted tris-arenes were better transformation
agents than were those having electron-withdrawing groups
on the arenes. Our current hypothesis is that the DNA
phosphate residues are bound by individual tris-arenes. This is
shown schematically in Figure 5. The array of tris-arenes thus
envelop the plasmid in a hydrophobic coat that ispossibly by
The work reported here is not only a new application of the
broad family of tris-arenes, it represents a breakthrough in the
size restriction normally experienced in chemical vector
transformation or transfection. Additional studies are underway
to determine the broad scope of this application and to better
understand the mechanism by which these compounds
function.
AUTHOR INFORMATION
Corresponding Author
gokelg@umsl.edu
■
Notes
The authors declare no competing financial interest.
13548
dx.doi.org/10.1021/ja304816e | J. Am. Chem. Soc. 2012, 134, 13546−13549

Journal of the American Chemical Society
Communication
ACKNOWLEDGMENTS
■
We thank the National Science Foundation (CHE-0957535)
for a grant that supported a portion of this work, and Profs.
Wendy Olivas and John-Stephen Taylor for valuable
discussions and technical assistance.
REFERENCES
■
(1) Kavallieratos, K.; Bertao, C. M.; Crabtree, R. H. J. Org. Chem.
1999, 64, 1675−1683.
(2) (a) Sessler, J. L.; Gale, P.; Cho, W.-S. Anion Receptor Chemistry;
Royal Society of Chemistry: Cambridge, 2006; 413 pp. (b) Bowman-
James, K.; Bianchi, A.; García-Espana, E., Eds. Anion Coordination
Chemistry; Wiley VCH: New York, 2012; 574 pp.
(3) (a) Gale, P. A.; , Coord. Chem. Rev., 2000, 199, 181. (b) Sessler, J.
L.; Camiolo, S.; Gale, P. A. Coord. Chem. Rev. 2003, 240, 17. (c) Kubik,
S.; Reyheller, C.; Stuwe, S. J. Inclusion Phenom. Macrocyc. Chem. 2005,
52, 137.
(4) (a) Sakai, N.; Brennan, K. C.; Weiss, L. A.; Matile, S. J. Am. Chem.
Soc. 1997, 119, 8276. (b) Otto, S.; Osifchin, M.; Regen, S. L. J. Am.
Chem. Soc. 1999, 121, 7276. (c) Koulov, A. V.; Lambert, T. N.; Shukla,
R.; Jain, M.; Boon, J. M.; Smith, B. D.; Li, H.; Sheppard, D. N.; Joos, J.-
B.; Clare, J. P.; Davis, A. P. Angew. Chem., Int. Ed. 2003, 42, 4931.
(d) Ferdani, R.; Li, R.; Pajewski, R.; Pajewska, J.; Winter, R. K.; Gokel,
G. W. Org. Biomol. Chem. 2007, 5, 2423. (e) Santacroce, P. V.; Davis, J.
T.; Light, M. E.; Gale, P. A.; Iglesias-Sanchez, J. C.; Prados, P.;
Quesada, R. J. Am. Chem. Soc. 2007, 129, 1886. (f) McNally, B. A.;
O’Neil, E. J.; Nguyen, A.; Smith, B. D. J. Am. Chem. Soc. 2008, 130,
17274. (g) Berezin, S. K.; Davis, J. T. J. Am. Chem. Soc. 2009, 131,
2458.
(5) Koulov, A. V.; Mahoney, J. M.; Smith, B. D. Org. Biomol. Chem.
2003, 1, 27.
(6) Ghosh, K.; Sarkar, A. R.; Ghorai, A.; Ghosh, U. New J. Chem.
2012, 36, 1231−1245.
(7) Schlesinger, P. H.; Ferdani, R.; Liu, J.; Pajewska, J.; Pajewski, R.;
Saito, M.; Shabany, H.; Gokel, G. W. J. Am. Chem. Soc. 2002, 124,
1848−1849.
(8) Schlesinger, P. H.; Ferdani, R.; Pajewski, R.; Pajewska, J.; Gokel,
G. W. Chem. Commun. 2002, 840−841.
(9) Pajewski, R.; Ferdani, R.; Schlesinger, P. H.; Gokel, G. W. Chem.
Commun. 2004, 160−161.
(10) Yamnitz, C. R.; Negin, S.; Carasel, I. A.; Winter, R. K.; Gokel, G.
W. Chem. Commun. 2010, 2838−2840.
(11) Carasel, I. A.; Yamnitz, C. R.; Winter, R. K.; Gokel, G. W. J. Org.
Chem. 2010, 75, 8112−6.
(12) http://www.rcsb.org/pdb/explore.do?structureId=1BNA
(13) Malone, J. F.; Murray, C. M.; Dolan, G. M.; Docherty, R.;
Lavery, A. J. Chem. Mater. 1997, 9, 2983−2989.
(14) Malone, J. F.; Murray, C. M.; Nieuwenhuyzen, M.; Stewart, G.;
Docherty, R.; Lavery, A. J. Chem. Mater. 1997, 9, 334−338.
(15) (a) Nimmanpipug, P.; Tashiro, K.; Maeda, Y.; Rangsiman, O. J.
Phys. Chem. B. 2002, 106, 6842−6848. (b) Qi, J. Y.; Yang, Q. Y.; Lam,
K. H.; Zhou, Z. Y.; Chan, A. S. C. Acta Crystallogr. 2003, E59, o415−
o416.
(16) Light, M. E.; Evans, L. S.; Gale, P. A.; , Unpublished deposition
[NIQBOF] to Cambridge Crystallographic Structural Database, 2007.
(17) Yin, Z.; Li, Z.; Yang, W.-Z.; He, J.; Cheng, J.-P. Struct. Chem.
2005, 16, 641−647.
(18) http://www.promega.com/
(19) McMaster, G. K.; Carmichael, G. G. Proc. Natl. Acad. Sci.U.S.A.
1977, 74, 4832−4838.
(20) Zhdanov, R. I.; Podobed, O. V.; Vlassov, V. V. Bioelectrochem-
istry 2002, 58, 53−64.
(21) Sutcliffe, J. G. Proc. Natl. Acad. Sci. U.S.A. 1978, 75, 3737−3741.
(22) Hansch, C.; Leo, A.; Taft, R. W. Chem. Rev. 1991, 91, 165−195.
(23) Shigeyuki, K.; Watara, H.; Kousaku, M. Bioeng. Bugs 2010, 1,
395−403.
(24) Bauer, A. W.; Kirby, W. M. M.; Sherris, J. C.; Turck, M. Am. J.
Clin. Pathol. 1966, 36, 493−96.
13549
dx.doi.org/10.1021/ja304816e | J. Am. Chem. Soc. 2012, 134, 13546−13549