2
J.L. Collins III et al. / Bioorganic & Medicinal Chemistry Letters xxx (2017) xxx–xxx
resulted in the development of synthetic receptors that transport
metal ions,20,21 can carry out enantioselective transport of amino
acids22 and principles of dynamic combinatorial chemistry have
uncovered new transporters.23
of 3 hosts (1–3) and 3 functional handles (see below) we have a
small tool set at our disposal as we look for answers to the drug
transport problem.
As reported phosphonic acid 124,25 could effect transport across
a bulk liquid membrane without a pH gradient, whereas lower rim
functionalized carboxylates calix[4]arene 226 and calix[6]arene 327
responded favorably to an inverse acid gradient (Fig. 1). We
hypothesize that calixarenes provide ion-ion interactions with
choline that overcome two-phase extraction, but are lenient
enough to release a payload into a welcoming receiving phase.
Whereas resorcinarene cavitands were unable to accomplish this
essential latter requirement. While we have suspended our trans-
port efforts with carboxylates, we have confidence that they can
serve as a backup should phosphonic acid 1 present problems in
more relevant cell based assays. While pursuing this, we found
complementary use for the calix[6]arene hexacarboxylic acid
towards binding of Pb, Sr and Ba with a new calix[6]arene octahe-
dral geometry occurring.28–30
Discussion
Trimethylammonium dopamine 6 was produced in fair yields
upon exhaustive methylation of dopamine hydrochloride
(Scheme 1, for full details see ESI). Under our reaction conditions
no phenolic ethers were detected, our recrystallization procedure
easily removes excess potassium carbonate, but does not afford
high yields at this time. Crude NMR during the reaction indicates
clean conversion, but some product is likely lost when separating
the solid trimethylammoniums from carbonate. An ammonium
dicarboxylate handle was installed in two straightforward steps
starting with addition of two t-butyl acetates to the free amine
of 5. Subsequent removal of t-butyl groups using TFA afforded 7
as a TFA salt.
Recently we reported that calixarenes are promising receptors
for liquid membrane transport of choline-fluorophore conju-
gates.15 Our results indicated that the presence of ionizable, pre-
organized functional groups such as those on a calixarene scaffold
provide effective transport of choline-fluorophore conjugates.
Lower rim carboxylic acid 2, 3 and upper rim phosphonic acid 1
groups were sufficient to transport payloads appended with a
trimethylammonium handle such as that found in choline (O-ethy-
lene trimethylammonium). Tetraphosphonic acid calix[4]arene
124,25 was capable of transporting choline conjugates without a
complementary, inverse pH gradient and became our sole focus.15
Our second study examined the nature of the payload and we
found many drug and drug like entities were efficiently trans-
ported through a liquid membrane – usually at rates far superior
compared to controls lacking a calixarene transporter.31 Some lim-
its obviously emerged, but also exciting results showing that sero-
tonin and dopamine with a smaller ammonium handle were also
transported. Concurrently, we serendipitously uncovered a third
useful handle in the form of an ammonium dicarboxylate. With
three potential handles at our disposal we wanted to complete
our initial work to directly compare the effect of handle on trans-
port efficiency. Serotonin and dopamine were chosen to compare
endogenous ammonium, and chemically introduced trimethylam-
monium and ammonium dicarboxylate handles side by side.
Simultaneously we explored the role of HEPES buffer on these
events as we previously noted a surprising advantage while using
it for one payload.31
Following the same two protocols for preparation of dopamine
derivatives 6 and 7, analogs of serotonin were prepared (Scheme 2).
Exhaustive methylation of serotonin hydrochloride 8 with methyl
iodide gave readily isolated trimethylammonium serotonin 9 in
fair, but unoptimized yield. Ammonium dicarboxylate 10 was pre-
pared in two steps starting with addition of two t-butyl acetates to
the free amine of 8. Subsequent removal of t-butyl groups using
TFA afforded 10 as a TFA salt. We will refer to these compounds
as ammonium dicarboxylates from this point forward, when dis-
solved in water - this is a more accurate representation of their
likely protonation state.
With this matrix of 2 neurotransmitters with three handle, we
screened them against tetraphosphonic acid calix[4]arene recep-
tor 1 using a 3-phase U-tube apparatus. Screens were conducted
in both water and 10 mM HEPES Buffer (pH 7.4). A detailed
description of the apparatus as well as representative calibration
curves are found in the ESI, we graph and discuss the results
herein.
As we reported, dopamine
5
had
a transport flux of
1.18 Â 10À4 0.02 Â 10À4
l
moles cmÀ2 minÀ1 in water with virtu-
ally no transport in the absence of host (Fig. 2 and Table 1).31
Transport was enhanced 4.6 times when HEPES buffered source
and receiving phases were used. The enhanced transport in HEPES
was a surprise to us. The exact mechanism of enhancement is
unclear at this time, but one small effect might be on the protona-
tion/deprotonation of the host at the interfaces due to buffering.
Increased salt concentration also could play a role. We then exam-
ined the effect of changing the charged ammonium handle of 5 to
the larger trimethylammonium handle of 6. This handle was the
basis for our first two reports on this subject. A trimethylammo-
nium handle had considerable reach in its ability to transport a
variety of fluorophores, drug-like and drug molecules when com-
bined with hosts 1–3.
These results we believe complete our optimization of handle
and host and afford us some variety of introducing a handle onto
a new membrane resistant drug-candidate. With the versatility
Comparing 5 vs. 6 in water we note that transport is 1.6 times
more efficient for 6. When switching to HEPES, 6 is transported
1.7 times more than 5. We then conducted the same experiments
with an ammonium dicarboxylate handle 7. Comparing 5 vs. 7 we
see a drastic decrease in transport (0.17 water, 0.15 Hepes). In the
cases of 5 and 6 a comparison of the control experiment with no
host to an experiment with host present is virtually meaningless,
in water and HEPES both guests 5 and 6 had little or no detectable
transport after 72 h under control conditions; the presence of host
1 was the required ingredient for transport. For guest 7 however,
the change in the nature of the substrate resulted in non-zero con-
trol transport, in these cases while host mediated transport is
much lower than for 5 and 6, we find that host 1 enhances guest
7 transport (2.5 times in water, 5.1 times in HEPES).
Fig. 1. Structure of upper phophonic acid calix[4]arenen 1 and lower carboxylic
acid calix[4]arene 2 and calix[6]arene 3.
Collins, James L.
