Controlling the oxygenation level of hemoglobin by using a synthetic receptor for 2,3-bisphosphoglycerate

Article abstract of DOI:10.1002/anie.200351165

A designed synthetic receptor (1), which has guanidinium groups and a Cu^II center, is highly effective and selective for binding 2,3-bisphosphoglycerate (2,3-BPG) in physiological media (see picture). The affinity is high enough that the receptor can deprive hemoglobin of available 2,3-BPG, and thus control the active level of 2,3-BPG and the related oxygenation level of hemoglobin. Cu = Green, N = blue, C = gray, P = yellow, O = red.

Full text of DOI:10.1002/anie.200351165

Angewandte
Chemie
Receptor for Biological Biphosphate
Controlling the Oxygenation Level of
Hemoglobin by Using a Synthetic Receptor for
2,3-Bisphosphoglycerate**
Zhenlin Zhong and Eric V. Anslyn*
2,3-Bisphosphoglycerate (2,3-BPG) is an allosteric effector
that modulates the oxygenation level of hemoglobin. It is
present in human and most other mammalian erythrocytes,
normally in an approximately equimolar concentration to that
of hemoglobins. It is a glycolytic intermediate resulting from
the interconversion between 1,3-phosphoglycerate and 3-
phosphoglycerate catalyzed by phosphoglycerate mutase.[1]
As an allosteric effector, 2,3-BPG regulates the oxygen
binding and releasing ability of hemoglobin[2] by binding to
a central cavity created by the spatial arrangement of the four
heme proteins. Specifically, the docking of 2,3-BPG to this
cavity decreases the oxygen affinity of hemoglobin, because
2,3-BPG preferentially binds to deoxyhemoglobin.[3,4]
Various inherited diseases lead to abnormal concentra-
tions of 2,3-BPG and altered levels of oxygen transport.[5]
Increased levels of 2,3-BPG are present in patients suffering
from hypoxemia and anemia, because this facilitates the
oxygen release of oxyhemoglobin into tissue.[6] In contrast,
high 2,3-BPG levels in individuals with sickle cell anemia are
harmful, because 2,3-BPG favors sickling by facilitating the
polymerization of sickle hemoglobin.[7–9] Controlling the
physiological concentration of 2,3-BPG could yield a treat-
ment for these diseases. Our goal was to create a receptor[10]
that would have a sufficiently high affinity and selectivity for
2,3-BPG to effectively modulate its concentration in physio-
logical media, and thereby control hemoglobin oxygena-
tion.
Herein, we describe a rationally designed synthetic
receptor (1) that has a highly preorganized cavity composed
of a CuII center and four guanidinium groups that serve as
binding sites. The receptor displays high affinity and good
selectivity for 2,3-BPG. We further show that this receptor
will strip the 2,3-BPG from horse hemolyzate, and therefore
modulate the oxygenation level of the horse hemoglobin. The
[*] Prof. Dr. E. V. Anslyn, Dr. Z. Zhong
Department of Chemistry and Biochemistry
The University of Texas at Austin
Austin, TX 78712 (USA)
Fax: (+1)512-471-7791
E-mail: anslyn@ccwf.cc.utexas.edu
[**] This work was in part funded by the National Institutes of Health
(GM57306), an ARO-MURI grant, and the Welch Foundation.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2003, 42, 3005 – 3008
DOI: 10.1002/anie.200351165
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3005

Communications
strategy of sequestering a biological effector with a small
of 2,3-BPG binds to the CuII center while the phosphates of
2,3-BPG are flanked from above and below by the guanidi-
nium groups.
The synthesis of the receptor is shown in Scheme 1 (see
Supporting Information). Triaminomethyltriethylbenzene
(2)[15] was allowed to react with two equivalents of N-tBoc-
synthetic compound is an intriguing approach to controlling
physiological function, and represents an alternative to the
standard approach of creating small molecules that bind to
biological macromolecules, thereby modulating their activity.
2,3-BPG is a highly anionic molecule that is strongly
hydrated in water. Hence, several molecular-recognition
contacts are required in a receptor to successfully compete
with the solvation of this target. Nature competes with
solvation by forming several salt bridges between 2,3-BPG
and the cationic side chains of lysine and histidine in
deoxyhemoglobin.[4] Based upon this precedent, receptor 1
was designed. This receptor contains guanidinium groups
rather than ammonium and imidazolium groups as with the
natural receptor, because molecular recognition entities of
guanidiniums are well known to bind phosphates in both
biological and synthetic receptors.[11,12] Furthermore, receptor
1 has a single CuII binding site, designed to complex the
carboxylate or phosphate group of 2,3-BPG. The molecular
recognition entities converge creating a clam-shell-shaped
cavity. The clam shell is organized upon metal binding. Lastly,
the hexa substitution of the two benzene rings in 1 imparts a
thermodynamic preference for adjacent groups to be in an up
and down pattern,[13] thereby imparting convergence of the
guanidinium groups prior to binding 2,3-BPG.
Scheme 1. Synthesis of receptor 1.
2-methylthio-2-imidazoline[16] in MeOH/AcOH, thus giving
the bis(tBoc-guanidine)-monoamino-containing structure 3
with a yield of 33%. Reductive amination of pyridine-2,6-
dicarboxaldehyde with 3 by using sodium borohydride gave
the tBoc-protected ligand 4 in 76% yield, which was
deprotected in TFA/CH2Cl2 giving the free ligand 5 quanti-
tatively. Finally, the complexation of 5 with one equivalent of
CuCl2 gave the receptor 1.[17]
The molecular recognition behavior of 1 with 2,3-BPG
and its analogues was studied with an indicator displacement
assay[18] by using pyrocatechol violet (PV) in a 1:1 water/
methanol solution buffered with 10 mm HEPES at pH 7.4.
Figure 2a and 2b show the UV/Vis spectroscopic change of
the indicator upon complexation with 1 and displacement
from 1 by 2,3-BPG respectively. The binding of PV to 1 causes
an absorbance decrease at the lmax of 445 nm and an increase
at a new lmax of 590 nm with an isosbestic point at 490 nm
(Figure 2a), thus yielding a visual change from yellow to blue.
The titration data fits well to a typical 1:1 binding algorithm[19]
with a binding constant of 1.4 ” 106 mꢀ1. The addition of 2,3-
BPG to a solution of the complex between 1 and PV causes a
reversal of the spectral response, thus giving an increase at
445 nm and a decrease at 590 nm, and a color change back to
yellow. This implies that the indicator was cleanly displaced
from the binding cavity of receptor 1 by 2,3-BPG (Figure 2b).
This colorimetric response yields a visual response and a
quantitative sensing ensemble for 2,3-BPG.
Molecular modeling[14] shows that 1 has a cavity comple-
mentary in size, shape, and charge to 2,3-BPG (Figure 1). The
guest slips in between the clam shells, and fits tightly with a
large number of binding interactions. The modeling reveals
eight hydrogen bonds and five pairs of cation–anion inter-
actions between 2,3-BPG and 1. Specifically, the carboxylate
An independent UV/Vis titration of 1
directly with 2,3-BPG in water buffered
with 10 mm HEPES at pH 7.4 also shows
a 1:1 binding, which changes the CuII d–d
absorption. While free 1 has a maximum
Figure 1. A proposed structure of the complex of 1 and 2,3-BPG.
absorbance at 678 nm with
a e =
3006
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org Angew. Chem. Int. Ed. 2003, 42, 3005 – 3008

Angewandte
Chemie
displacement of the indicator PV from 1 does not show a
perfectly clean isosbestic point in pure water at pH 7.4,
possibly because of a small amount of aggregation. To avoid
this problem, a slight adjustment of the pH was necessary. At
pH 6.8,[20] a clean isosbestic point was found, and the binding
constant between 1 and 2,3-BPG in water was determined to
be 4 ” 107 mꢀ1. A drop in the binding constant of only 20-fold
occurs upon going to pure water from 50% methanol. Yet, at
pH 6.8 the phosphates of 2,3-BPG are protonated to a greater
extent than at pH 7.4. Although this binding constant is quite
high, it is actually slightly lower than it would be at pH 7.4.
The influence of receptor 1 on the oxygenation levels of
hemoglobin was studied spectrophotometrically with a ton-
ometer by using the method of Benesch.[21] Figure 3 shows the
Figure 2. UV/Vis absorbance spectra of a) 0.054 mm solution of PV in
the presence of 0–1.1 equivalents of 1 and b) 0.06 mm 1 and
0.054 mm PV solution in the presence of 0–0.08 mm of 2,3-BPG. Con-
ditions: 1:1 MeOH/H2O, 10 mm HEPES buffer, pH 7.4, 258C.
Figure 3. Oxygenation curves of hemoglobin in diluted hemolyzate of
horse red cells (0.04 mm hemoglobin, 10 mm phosphate buffer,
210mꢀ1 cmꢀ1, its 2,3-BPG complex has a maximum absorb-
ance at 660 nm with a e = 170mꢀ1 cmꢀ1. These results indicate
that the copper center plays an important role in the binding.
Because the binding is very strong, an accurate binding
constant was not obtained from this direct titration.
!
pH 7.2, 258C). (þ) Original hemolyzate, ( ) hemolyzate with 0.16 mm
~
^
1, ( ) hemolyzate with 0.16 mm additional 2,3-BPG, ( ) hemolyzate
with 0.16 mm additional 2,3-BPG and 0.16 mm 1.
Alternatively, the binding constant of 1 with 2,3-BPG in
1:1 water/methanol (v/v) was calculated to be 8.0 ” 108 mꢀ1 by
using the competitive spectrophotometric method[19] on the
basis of the spectrum changes shown in Figure 2b. Similarly,
the binding constants for analogues of 2,3-BPG in the same
solvent mixture were measured (Table 1). There is good
oxygenation curves of hemoglobin in horse red-cell hemoly-
zate diluted in 10 mm phosphate buffer solution at pH 7.2.
The curve for native hemolyzate (þ) has the classic shape
found in introductory biochemistry textbooks.[22] We first
studied the effect of just 1 on the horse hemolyzate. The
addition of 1 to the diluted hemolyzate causes the curve to
!
move to lower oxygen pressure (Figure 3 ), thus indicating
an increase of the oxygen affinity of hemoglobin. Therefore,
as anticipated, the addition of 1 leads to the stripping of the
natural 2,3-BPG from hemoglobin. Next, the addition of 2,3-
BPG to horse hemolyzate was studied to show that the
hemoglobin was responding to 2,3-BPG as it should. Addition
Table 1: Binding constants with receptor 1 in 1:1 water/methanol at
pH 7.4, 258C.
Analyte (as sodium salt)
Binding constant [mꢀ1
]
2,3-BPG
8”108 (4”107[a]
1.3”107
1.1”107
4.7”106
6”104
)
Phospho(enol)pyruvate
2-Phosphoglycerate
3-Phosphoglycerate
b-Glycerophosphate
Acetate
^
of 2,3-BPG was found to decrease the oxygen affinity ( ), as
expected. Lastly, the addition of an equivalent of 1 counter-
acted the added 2,3-BPG, thus revealing that the additional
~
2,3-BPG was captured by the receptor ( ).
7”103
All these results are in accordance with the binding
affinity we determined for 1. Because the binding constant of
receptor 1 with 2,3-BPG (near 107 mꢀ1) is significantly higher
than that of hemoglobin with 2,3-BPG (on the order of 104 to
105 mꢀ1),[3,23,24] the receptor depletes the 2,3-BPG from the
hemoglobin complex and thereby increases the oxygenation
level to that of striped hemoglobin.
In conclusion, we have created a designed synthetic
receptor that is highly effective and selective for binding 2,3-
BPG, even in physiological media. The affinity is high enough
that the receptor can deprive hemoglobin of available 2,3-
[a] in water at pH 6.8, 258C.
selectivity for binding with 1. For example, the binding
constants for the analogues of 2,3-BPG that have one less
phosphate are 80 to 180 times lower than that of 2,3-BPG,
while the binding constants for simple phosphates and
carboxylates are four to five orders of magnitude lower.
Apparently, the selectivity is related to the numbers of ion
pairs. A binding constant relevant to the use of 1 in blood was
desirable for the ultimate goal of this study. However,
Angew. Chem. Int. Ed. 2003, 42, 3005 – 3008
www.angewandte.org
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3007

Communications
[17] The stoichiometry of the complex of ligand 5 with CuCl2 was
determined to be 1:1 by the mole-ratio method by monitoring
the UV/Vis absorbance at 670 nm.
[18] S. L. Wiskur, H. Ait-Haddou, J. J. Lavigne, E. V. Anslyn, Acc.
Chem. Res. 2001, 34, 963 – 972.
[19] K. A. Connors, Binding Constants, The Measurement of Molec-
ular Complex Stability, Wiley, New York, 1987.
[20] At the adjusted pH, the phenols are essentially completely
protonated. E. Chiacchierini, R. Gocchieri, L. Sommer, Collect.
Czech. Chem. Commun. 1973, 38, 1478 – 1496.
BPG, and thus control the active level of 2,3-BPG and the
related oxygenation level of hemoglobin. The use of a small
organic receptor to bind and regulate the levels of a biological
effector, and thereby control a physiological function, repre-
sents a rarely explored means towards the development of a
pharmaceutical agent.[25] Traditionally, small organic mole-
cules are used to bind large biomolecules, and thereby
regulate biological function. We believe that the sequestering
of small effectors, agonists, and antagonists, by synthetic
receptors could play an alternative or complementary role.
[21] R. Benesch, G. Macduff, R. E. Benesch, Anal. Biochem. 1965,
11, 81 – 87.
[22] C. K. Mathews, K. E. Van Holde, Biochemistry, 2nd ed.,
Benjamin-Cummings, 1996, p. 233.
[23] M. K. Hobish, D. A. Powers, Proteins Struct. Funct. Genet. 1986,
1, 164 – 175.
[24] N. Hamasaki, Z. B. Rose, J. Biol. Chem. 1974, 249, 7896 – 7901.
[25] For another example, see: C. C. Huval, S. R. Holmes-Farley,
G. M. Whitesides (Geltex Pharmaceticals, Inc.), WO 0038664,
2000.
Experimental Section
Preparation and analysis of hemolyzate: Fresh defibrinated horse
blood (0.6 mL) was centrifuged at 4000 rpm for 5 minutes to isolate
the red cells. The supernatant was removed, and the cells were washed
four times with 0.9% sodium chloride (1.5 mL each time) in the
centrifuge tube. The isolated and washed red cells were hemolyzed by
adding distilled water (ca. 1.2 mL). Toluene (ca. 0.3 mL) was added
and mixed with the hemolyzate by shaking vigorously. After
centrifugation of the sample at 12000 rpm for 10 minutes, the lower
clear red layer was transferred by using a pipette to a new centrifuge
tube. The toluene-treatment procedure was repeated twice until the
interface between the toluene layer and the hemolyzate layer was
clear. The hemolyzate was stored in a refrigerator and used within
three days. Figure 3 was generated by using the method of Benesch.[21]
Received: February 12, 2003
Revised: April 29, 2003 [Z51165]
Keywords: bioinorganic chemistry · drug design ·
.
heme proteins · host–guest systems · receptors
[1] S. I. Winn, H. C. Watson, R. N. Harkins, L. A. Fothergill, Philos.
Trans. R. Soc. London Ser. B 1981, 293, 121 – 130.
[2] R. Benesch, R. E. Benesch, Nature 1969, 221, 618 – 621.
[3] A. Arnone, Nature 1972, 237, 146 – 149.
[4] R. Benesch, R. E. Benesch, C. I. Yu, Proc. Natl. Acad. Sci. USA
1968, 59, 526 – 532.
[5] M. Delivoria-Papadopoulos, F. A. Oski, A. J. Gottlieb, Science
1969, 165, 601 – 602.
[6] R. Juel, CRC Crit. Rev. Clin. Lab. Sci. 1979, 10, 113 – 146.
[7] W. N. Poillon, B. C. Kim, O. Castro, Blood 1998, 91, 1777 – 1783.
[8] W. N. Poillon, B. C. Kim, R. J. Labotka, C. U. Hicks, J. A. Kark,
Blood 1995, 85, 3289 – 3296.
[9] M.-C. Garel, N. Arous, M. C. Calvin, C. T. Craescu, J. Rosa, R.
Rosa, Proc. Natl. Acad. Sci. USA 1994, 91, 3593 – 3597.
[10] For a previous receptor from our group for 2,3-BPG, see M. D.
Best, E. V. Anslyn, Chem. Eur. J. 2003, 9, 51 – 57.
[11] S. Tobey, E. V. Anslyn in Kluwer-Plenum Symposium Series on
Anion-Recognition, in press.
[12] C. L. Hannon, E. V. Anslyn in Bioorganic Chemistry Frontiers
III (Ed.: H. Dugas), Springer, Heidelberg, 1993, pp. 193 – 256.
[13] D. J. Iverson, G. Hunter, J. F. Blount, J. R. Damewood, K.
Mislow, J. Am. Chem. Soc. 1981, 103, 6073 – 6083.
[14] Molecular modeling was performed with MacSpartan Pro
(Wavefunction, Inc.). The structure was energy-minimized by
Merck Molecular Force Field in gas phase.
[15] L. A. Cabell, M. D. Best, J. J. Lavigne, S. E. Schneider, D. M.
Perreault, M.-K. Monahan, E. V. Anslyn, J. Chem. Soc. Perkin
Trans. 2 2001, 315 – 323.
[16] S. R. Mundla, L. J. Wilson, S. R. Klopfenstein, W. L. Seibel, N. N.
Nikolaides, Tetrahedron Lett. 2000, 41, 6563 – 6566.
3008
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2003, 42, 3005 – 3008