Chiral selective transmembrane transport of amino acids through artificial channels

Article abstract of DOI:10.1021/ja312704e

Peptide-appended pillar[n]arene (n = 5, 6) derivatives have been synthesized. ¹H NMR and IR studies revealed that the molecules adopt a tubular conformation in solution and lipid bilayer membranes. Kinetic measurements using the fluorescent labeling method with lipid vesicles revealed that these molecules can efficiently mediate the transport of amino acids across lipid membranes at a very low channel-to-lipid ratio (EC₅₀ = 0.002 mol %). In several cases, chiral selectivity for amino acid enantiomers was achieved, which is one of the key functions of natural amino acid channels.

Full text of DOI:10.1021/ja312704e

Communication
pubs.acs.org/JACS
Chiral Selective Transmembrane Transport of Amino Acids through
Artificial Channels
Lei Chen,† Wen Si,† Liang Zhang, Gangfeng Tang, Zhan-Ting Li, and Jun-Li Hou*
Department of Chemistry, Fudan University, 220 Handan Road, Shanghai 200433, China
S
* Supporting Information
1).14 The nonpolar benzyl group of phenylalanine (Phe) on the
peptide chains was expected to increase the membrane-
ABSTRACT: Peptide-appended pillar[n]arene (n = 5, 6)
derivatives have been synthesized. 1H NMR and IR studies
revealed that the molecules adopt a tubular conformation
in solution and lipid bilayer membranes. Kinetic measure-
ments using the fluorescent labeling method with lipid
vesicles revealed that these molecules can efficiently
mediate the transport of amino acids across lipid
membranes at a very low channel-to-lipid ratio (EC50
=
0.002 mol %). In several cases, chiral selectivity for amino
acid enantiomers was achieved, which is one of the key
functions of natural amino acid channels.
he transmembrane transport of amino acids, which plays a
Tcrucial role in life systems, has been realized by specific
transporters,1 such as the ATP-binding cassette (ABC)
superfamily,2 the phospholemman (PLM) family,3 and
hypotonically activated amino acid channels (HAACs),4 with
high selectivity and efficiency. The development of artificial
transporters for amino acids is of importance fundamentally
and for practical applications. In recent years, many synthetic
systems that can transport various ions,5 water,6 and glucose7
have been described. However, the transport of amino acids by
synthetic molecules has been investigated only with simple U-
tube experiments across liquid phases.8 To the best of our
knowledge, no examples of amino acid transmembrane
transport by synthetic systems have been reported. Because
amino acids can pass membranes in the simple diffusion
mechanism,9 the design of any artificial tranporters should
require an efficiency that is much higher than that of this simple
diffusion process. This has been a challenge because for any
successful transporter, its concentration should be much lower
than that of the transported species. Herein we report that
pillar[n]arene (n = 5, 6) derivatives bearing short peptide
chains can serve as efficient artificial channels for trans-
membrane transport of amino acids that display high chiral
selectivity.
We have been interested in the construction of single-
molecule artificial transmembrane channels.10 Previously, we
found that hydrazide-appended pillar[5]arenes could form
single-molecule tubular structures that are induced by intra-
molecular N−H···OC hydrogen bonding of the hydrazide
unit.6c It was envisioned that peptide-attached pillar[n]-
arenes11−13 could give rise to similar tubular structures. Thus,
a library of pillar[n]arene (n = 5, 6) amino acid derivatives 1a−
c and 2a−d, which bear 10 and 12 peptide chains, respectively,
with different lengths and sequences were prepared (Figure
Figure 1. Structures of compounds 1a−c, 2a−d, 3a, and 3b.
insertion ability of the molecules. Both D- and L-Phe were
used to build the peptide chains to produce channels with
different chiral cavities.
The possibility of forming intramolecular hydrogen bonds in
bulk and lipid bilayer membranes was first investigated by using
as model compounds 3a and 1a (Figure 1), respectively, which
1
have the same main scaffold. In CD2Cl2, 3a produced a H
NMR spectrum of high resolution with the signals of the three
amide protons (from inside to outside) located at 7.67, 6.99,
and 7.62 ppm, respectively, as assigned by two-dimensional
(2D) correlation spectroscopy (COSY) and rotating-frame
Overhauser effect spectroscopy (ROESY) experiments [Figures
S51−S56 in the Supporting Information (SI)]. These signals
were shifted downfield significantly (0.7−1.3 ppm) relative to
the related signals of control 4 (Figure 2a,b). Diluting the
solution from 5 to 0.2 mM did not cause an observable shift in
the N−H signals. These results suggested that the peptide
chains of 3a engage in intramolecular hydrogen bonding, which
should induce the whole peptide chains to form a tubular
structure (Figure 2c and Figure S57).6c For further
investigation of the structure in the lipid phase, 1a and 5
were inserted into the membranes of large unilamellar vesicles
(LUVs) made from egg yolk L-α-phosphatidylcholine (EYPC).
Relative to that of CO(NH) of 5 at 1645 cm−1, the IR
stretch band of 1a shifted to 1660 cm−1 (Figure S58),
suggesting that 1a also assembles into a tubular structure in
lipid membranes.15 1H NMR and IR experiments for 1b and 3b
Received: December 31, 2012
Published: January 30, 2013
© 2013 American Chemical Society
2152
dx.doi.org/10.1021/ja312704e | J. Am. Chem. Soc. 2013, 135, 2152−2155

Journal of the American Chemical Society
Communication
slow transport of Gly was detected. Cryogenic transmission
electron microscopy (cryo-TEM) showed that in the presence
of 1b (x = 0.05%), the vesicles did not break over the course of
3 h (Figure S59), indicating that the vesicles were stable after
incorporation of 1b into their membranes. Thus, the increased
transport efficiency in the presence of 1b should be attributed
to the channel created by its tubular structure formed in the
membranes.
The transport kinetics was measured quantitatively using the
fluorescent labeling method.9a Thus, LUVs containing Gly (13
mM) were first prepared in HEPES buffer (pH 7.0) and put
into a dialysis tube immersed in the same buffer. Next, a DMSO
solution of 1b, 2b, 2c, or 2d (x = 0.05%) was injected into the
vesicle solution. Aliquots of the solution outside the dialysis
tube were taken out and then treated with fluorescent probe
fluorescamine.18 In this way, the Gly transported from the
inside of the vesicles was labeled with the probe by formation of
a lactam. By measuring the fluorescent intensity, the Gly
concentrations ([Gly]) were determined. Plots of [Gly] against
time are shown in Figure 4. For the systems containing 1b and
Figure 2. (a, b) Partial 1H NMR spectra of (a) 4 (25 mM) and (b) 3a
(5.0 mM) in CD2Cl2 at 298 K. (c) Molecular modeling (Gaussian 09,
semiempirical, PM6) structure of 3a (H atoms on the C atoms have
been omitted for clarity).
revealed similar results, indicatinig that these two compounds
form similar tubular structures. Since the peptide chains of 2a
and 2b possess the same arrangement as in 1a and 1b, it is
rational to assume that their peptide chains also form similar
tubular conformations in lipid membranes.
The potential of the tubular molecules for transport of Gly
was then explored first by using 13C NMR experiments.16 The
LUV solution was prepared in HEPES-buffered D2O (pD 6.5),
and then 1-13C-labeled Gly (Gly-1-13C, 150 mM) and the
chemical shift reagent EuCl3 (300 mM) were added. In the
presence of 1b [molar ratio relative to lipid (x) = 0.05%,
corresponding to 170 molecules of 1b per vesicle17], trans-
membrane transport of Gly was observed over the course of 80
min, as reflected by the appearance of a small peak at 171.7
Figure 4. Gly concentration outside the vesicles vs time after addition
⧫
■
▲
●
of pure DMSO ( ) or a solution of 1b ( ), 2b ( ), 2c ( ), or 2d
▼
(
) (x = 0.05%) in DMSO. Inset: schematic representation of the
transport process.
2b, [Gly] increased dramatically from 0 to 20 and 23 μM,
respectively, in 70 min, whereas for the systems containing the
shorter molecules 2c and 2d, the concentration increased to
only 6 and 7 μM, respectively. In the absence of tubular
compounds, the concentration increase was 5 μM. From fits of
the data using a nonlinear model,9a the Gly transport rate
constants (k) for 1b, 2b, 2c, and 2d were calculated to be (8.7
0.1) × 10−4, (9.0 0.2) × 10−4, (3.0 0.1) × 10−4, and (4.2
0.1) × 10−4 s−1, respectively. The transport rate constant in
the absence of channels, (k0) was determined to be (2.7 0.1)
× 10−4 s−1, indicating that Gly was transported across the lipid
membrane without channels, but slowly.9a In view of the very
low x value, the efficiency of the transport by 1b and 2b was
actually much higher than that of simple diffusion. The fact that
2b displayed a higher transport rate than 2c and 2d might be
rationalized by considering that the length of its hydrophobic
part (3.2 nm as estimated using a CPK model) better matched
the width of the hydrophobic part of the lipid membrane (3.5
nm).5b
13
ppm corresponding to the O C signal of Gly loaded in the
vesicles (Figure 3). However, in the absence of 1b, only very
Figure 3. Partial 13C NMR spectra indicating the transport of Gly-
1-13C cross the lipid membrane in the (a) absence and (b) presence of
1b (x = 0.05%).
Transport rates at different channel concentrations were also
measured for 1b. It was found that the transport activity (A ≡
k/k0) was strongly dependent on x. As x increased from 0 to
2153
dx.doi.org/10.1021/ja312704e | J. Am. Chem. Soc. 2013, 135, 2152−2155

Journal of the American Chemical Society
Communication
0.05%, A increased significantly. Further increasing x caused
only a slight increase in A (Figure S60). The effective
concentration needed for 50% activity (EC50) and the Hill
coefficient (n) were calculated to be 0.002% and 0.7,19
respectively (Figure S61). Such a low EC50 compared with
those of other artificial transporters of ions,20 water,6c and
glucose7 indicates that the new tubular channel is very effective
in transporting Gly, while the small n value shows that it works
in a single-molecule manner.19,21
The abilities of 1a, 1b, 2a, and 2b to transport other natural
amino acids, including Ala, Ser, Thr, Val, Leu, and Phe, were
also investigated using the same method (Figures S62−S68)
and were found to decrease in the order Gly > Ala > Ser > Thr
> Val > Leu > Phe, which is consistent with the increasing size
of the amino acids (Figure 5). For the same small amino acids
Figure 6. (a) Transport activities (A) of 2b (x = 0.05%) for L- and D-
amino acids. (b) Changes in L- or D-Phe concentration ([Phe]) outside
the vesicles with time in the absence or presence of channel 2b (x =
0.05%).
large enough to make the unfavorable transport negligible
compared with the “background” diffusion. Transport of the
two isomers of Phe by 1c, the enantiomer of 1a, was also
investigated and found to favor D-Phe.
It has been reported that natural transporters (e.g., solute
carrier families) carry amino acids but block Cl−,23 whereas
some other transporters (e.g., PLM and HAAC systems1a)
allow the transport of both species. The Cl− transport behavior
of the new artificial channels was then also studied using the
reported method.20 Thus, a solution of vesicles entrapping the
Cl−-sensitive dye lucigenin was added to a buffer containing
NaCl to produce a NaCl gradient. The flux of NaCl through
the channels into the vesicles was assessed by monitoring the
fluorescence intensity of lucegenin. Upon addition of 1a, 1b, 2a,
or 2b (x = 0.05%) to the above vesicle solution, the
fluorescence intensity increased significantly, reaching 91, 87,
99, and 52%, respectively, after 12 min (Figure 7a). This result
demonstrates that the channels are also capable of transporting
Cl−.20 When the vesicles contained Gly (10 mM), which was
shown by the above fluorescence experiments to pass the
Figure 5. Transport activities (A) of the channels 1a, 1b, 2a, and 2b (x
= 0.05%) for different natural amino acids (A = 1.0 indicates no
transport activity).
(Gly, Ala, and Ser), 1b and 2b showed very similar transport
activities. For the other larger amino acids, the transport activity
of 1b was decreased gradually in the order Thr > Val > Leu >
Phe, reflecting the increasing amino acid size. The transport
activity of 2b for these larger amino acids also decreased with
increasing amino acid size but was generally higher than that of
1b, which might be attributed to the larger cavity of 2b.22
Channels 1a and 2a transported only the smallest amino acids,
Gly and Ala (Figure 5). This result shows that the selectivity of
this kind of tubular channel for different amino acids can be
tuned simply by changing the sequence of the attached peptide
chains.
Because the tubular molecules bear chiral peptide chains, we
also investigated their capacity to discriminate enantiomers of
chiral Ala, Ser, Thr, Val, Leu, and Phe. It was revealed that
channel 2b could mediate the transport of all enantiomers of
the first five amino acids (Figure 6a). However, transport of the
D isomer was faster than that of the L isomer in every case, and
this difference increased with increasing amino acid size. For
Phe, only transport of the L isomer could be mediated (Figure
6a,b). Similar selective transport of D-Leu by 1a, 1b, and 2a and
D-Ser by 2a could also be realized (Table S2 in the SI). For
other amino acids, these channels transported both isomers
with an efficiency that depended on the structure of the amino
acid (Table S2). The observed high selectivity indicates that the
difference in the binding between the chiral amino acids and
the chiral environment formed by the peptide chains might be
Figure 7. (a, b) Changes in fluorescence intensity (λex = 372 nm, λem
=
503 nm) with time after addition of 1a, 1b, 2a, or 2b (x = 0.05%) to
vesicles (a) not containing or (b) containing Gly (10 mM). (c, d)
Schematic representations of (c) Cl− transport suppression for 2b and
(d) independent Cl− transport for 1a, 1b, and 2a in the presence of
Gly.
2154
dx.doi.org/10.1021/ja312704e | J. Am. Chem. Soc. 2013, 135, 2152−2155

Journal of the American Chemical Society
Communication
(5) (a) Akerfeldt, K. S.; Lear, J. D.; Wasserman, Z. R.; Chung, L. A.;
DeGrado, W. F. Acc. Chem. Res. 1993, 26, 191. (b) Gokel, G. W.;
Mukhopadhyay, A. Chem. Soc. Rev. 2001, 30, 274. (c) Fyles, T. M.
Chem. Soc. Rev. 2007, 36, 335. (d) Davis, A. P.; Sheppard, D. N.;
Smith, B. D. Chem. Soc. Rev. 2007, 36, 348. (e) Li, X.; Wu, Y.-D.; Yang,
D. Acc. Chem. Res. 2008, 41, 1428. (f) Matile, S.; Vargas Jentzsch, A.;
Montenegro, J.; Fin, A. Chem. Soc. Rev. 2011, 40, 2453. (g) Gale, P. A.
Acc. Chem. Res. 2011, 44, 216.
(6) (a) Percec, V.; Dulcey, A. E.; Balagurusamy, V. S. K.; Miura, Y.;
Smirdrkal, J.; Peterca, M.; Numellin, S.; Edlund, U.; Hudson, S. D.;
Heiney, P. A.; Duan, H.; Magonov, S. N.; Vinogradov, S. A. Nature
2004, 430, 764. (b) Duc, Y. L.; Michau, M.; Gilles, A.; Gence, V.;
Legrand, Y.-M.; Lee, A.; Tingry, S.; Barboiu, M. Angew. Chem., Int. Ed.
2011, 50, 11366. (c) Hu, X.-B.; Chen, Z.; Tang, G.; Hou, J.-L.; Li, Z.-
T. J. Am. Chem. Soc. 2012, 134, 8384.
(7) (a) Cho, H.; Widanapathirana, L.; Zhao, Y. J. Am. Chem. Soc.
2011, 133, 141. (b) Cho, H.; Zhao, Y. Langmuir 2011, 27, 4936.
(c) Zhang, S.; Zhao, Y. Chem.Eur. J. 2011, 17, 12444.
(8) For examples, see: (a) Demirel, N.; Bulut, Y.; Hosgoren, H.
Tetrahedron: Asymmetry 2004, 15, 2045. (b) Bozkurt, S.; Yilmaz, M.;
Sirit, A. Chirality 2012, 24, 129.
(9) (a) Chakrabarti, A. C.; Deamer, D. W. Biochim. Biophys. Acta
1992, 1111, 171. (b) Chakrabarti, A. C. Amino Acids 1994, 6, 213.
(10) Si, W.; Chen, L.; Hu, X.-B.; Tang, G.; Chen, Z.; Hou, J.-L.; Li,
Z.-T. Angew. Chem., Int. Ed. 2011, 50, 12564.
channels much quicker than the other amino acids (Figure 5),
the fluorescence enhancement of the 1a, 1b, and 2a systems
changed slightly (Figure 7b). However, for the 2b system, this
fluorescence enhancement was weakened remarkably (Figure
7b), indicating that the transport of Cl− by 2b was suppressed
considerably (Figure 7c). Fluorescent labeling experiments
showed that the transport of Gly by the four channels was not
affected by the presence of Cl− outside the vesicles.
The fact that Gly and Cl− did not affect each other with
respect to transport by 1a, 1b, and 2a (Figure 7d) might be
rationalized by the following considerations. The amount of
channel was much lower than that of lipid, meaning that there
was only a small chance for Gly or Cl− to approach an inserted
channel simultaneously. As a result, the apparent occupancy of
the inserted channels by either of them was low, so transport of
one did not suppress transport of the other. The transport of
Cl− by 2b was notably lower than that by 1a, 1b, and 2a, while
the transport of Gly by all of them was comparable. We might
propose that binding of Cl− to the inserted 2b is weaker, thus
increasing the probability of its transport being blocked by Gly.
In conclusion, we have developed a class of artificial
transmembrane amino acid channels from peptide-appended
pillar[n]arenes (n = 5, 6). The whole molecules are induced to
form a tubular architecture by the intramolecular hydrogen
bonding of the peptide chains. This unique shape enables
efficient transport of amino acids across membranes in a single-
molecule manner. In several cases, chiral selectivity was
realized, which is one of the key functions of natural amino
acid channels.1a Because the new single-molecule channels
possess relatively fixed diameters, we envision that they may
also mediate the transmembrane transport of longer peptides
consisting of the studied amino acids, which is currently being
investigated.
(11) (a) Ogoshi, T.; Kanai, S.; Fujinami, S.; Yamagishi, T.-A.;
Nakamoto, Y. J. Am. Chem. Soc. 2008, 130, 5022. (b) Li, C.; Xu, Q.; Li,
J.; Yao, F.; Jia, X. Org. Biomol. Chem. 2010, 8, 1568. (c) Cragg, P. J.;
Sharma, K. Chem. Soc. Rev. 2012, 41, 597. (d) Xue, M.; Yang, Y.; Chi,
X.; Zhang, Z.; Huang, F. Acc. Chem. Res. 2012, 45, 1294.
(12) Hu, X.-B.; Chen, L.; Si, W.; Yu, Y.; Hou, J.-L. Chem. Commun.
2011, 47, 4694.
(13) Yu, G.; Xue, M.; Zhang, Z.; Li, J.; Han, C.; Huang, F. J. Am.
Chem. Soc. 2012, 134, 13248.
(14) Theoretically, 1a, 1b, 2a, and 2b could exist as diastereomers.
1
However, their H NMR spectra displayed one set of signals of high
resolution. This result indicates that only one of the diastereomers was
formed for each of the compounds, although we do not know its
absolute configuration.
(15) Ghadiri, M. R.; Granja, J. R.; Buehler, L. K. Nature 1994, 369,
301.
(16) Davis, J. T.; Gale, P. A.; Okunola, O. A.; Prados, P.; Iglesias-
́
Sanchez, J. C.; Torroba, T.; Quesada, R. Nat. Chem. 2009, 1, 138.
(17) Mimms, L. T.; Zampighi, G.; Nozaki, Y.; Tanford, C.; Reynolds,
ASSOCIATED CONTENT
■
S
* Supporting Information
Synthetic procedures and characterization data for 1−5; IR and
1
2D H NMR spectra and cryo-TEM images; and detailed
procedures and measurement data for amino acid and Cl−
transport. This material is available free of charge via the
Internet at http://pubs.acs.org.
J. A. Biochemistry 1981, 20, 833.
(18) Udenfriend, S.; Stein, S.; Bohlen, P.; Dairman, W.; Leimgruber,
̈
W.; Weigele, M. Science 1972, 178, 871.
AUTHOR INFORMATION
Corresponding Author
houjl@fudan.edu.cn
■
(19) Matile, S.; Sakai, N.; Hennig, A. Transport Experiments in
Membranes. In Supramolecular Chemistry: From Molecules to Nanoma-
terials; Gale, P. A., Steed, J. W., Eds.; Wiley: Chichester, U.K., 2012;
Vol. 2, p 473 ff.
(20) Busschaert, N.; Wenzel, M.; Light, M. E.; Iglesias-Hernandez, P.;
Perez-Tomas, R.; Gale, P. A. J. Am. Chem. Soc. 2011, 133, 14136.
(21) Bhosale, S.; Matile, S. Chirality 2006, 18, 849.
(22) Han, C.; Ma, F.; Zhang, Z.; Xia, B.; Yu, Y.; Huang, F. Org. Lett.
2010, 12, 4360.
Author Contributions
†L.C. and W.S. contributed equally.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
(23) Carpenter, V. K.; Drake, L. L.; Aguirre, S. E.; Price, D. P.;
Rodriguez, S. D.; Hansen, I. A. J. Insect Physiol. 2012, 58, 513.
■
This work was supported by NSFC (20902012 and 91027008),
FANEDD (200930), and the Program for Changjiang Scholars
and Innovative Research Team in University (IRT1117).
REFERENCES
■
(1) (a) Saier, M. H., Jr. Microbiology 2000, 146, 1775. (b) Broer, S.
Physiol. Rev. 2008, 88, 249.
̈
(2) Saurin, W.; Hofnung, M.; Dassa, E. J. Mol. Evol. 1999, 48, 22.
(3) Chen, L. S. K.; Lo, C. F.; Numann, R.; Cuddy, M. Genomics 1997,
41, 435.
(4) Vieira, L. L.; Lafuente, E.; Gamarro, F.; Cabantchik, Z. I. Biochem.
J. 1996, 319, 691.
2155
dx.doi.org/10.1021/ja312704e | J. Am. Chem. Soc. 2013, 135, 2152−2155