Donor substrate binding to trans-sialidase of Trypanosoma cruzi as studied by STD NMR

Article abstract of DOI:10.1016/j.carres.2007.05.037

Using STD NMR experiments, we have studied the binding epitopes of p-nitrophenyl glycosides of sialic acid and analogs thereof when bound to Trypanosoma cruzi trans-sialidase (TSia). Time-dependent NMR spectra yielded data on the rate of substrate hydroly

Full text of DOI:10.1016/j.carres.2007.05.037

Carbohydrate Research 342 (2007) 1904–1909
Donor substrate binding to trans-sialidase of Trypanosoma cruzi
as studied by STD NMR
Astrid Blume,a, Bjo¨rn Neubacher,b, Joachim Thiemb and Thomas Petersa,*
aUniversity of Luebeck, Institute of Chemistry, Ratzeburger Allee 160, 23538 Lu¨beck, Germany
bUniversity of Hamburg, Faculty of Science, Department of Chemistry, Martin-Luther-King-Platz 6, 20146 Hamburg, Germany
Received 20 February 2007; received in revised form 22 May 2007; accepted 24 May 2007
Available online 9 June 2007
Abstract—Using STD NMR experiments, we have studied the binding epitopes of p-nitrophenyl glycosides of sialic acid and ana-
logs thereof when bound to Trypanosoma cruzi trans-sialidase (TSia). Time-dependent NMR spectra yielded data on the rate of
substrate hydrolysis in comparison to sialic acid transfer. Our experiments clearly demonstrate that shortening of the glycerol side
chain significantly favors the transfer reaction over hydrolysis. Our results extend the basis on which specific trans-sialidase inhib-
itors may be designed.
Ó 2007 Elsevier Ltd. All rights reserved.
Keywords: Trans-sialidase; STD NMR; Binding epitope; Sialic acid; Hydrolysis
1. Introduction
Although the protein transfers sialic acids, the TSia of
T. cruzi is considered to be a modified sialidase, rather
than a sialyltransferase, as it is distantly related to viral
and bacterial sialidases, and does not utilize CMP-N-
acetylneuraminic acid as the sialyl donor.4 Interestingly,
there is no other sialidase that is more efficient in trans-
ferring sialic acid residues rather than catalyzing hydro-
lysis. TSia shows high regiospecificity as it only transfers
a-(2!3) linked sialic acid residues with retention of con-
figuration from a terminal galactose of host sialyl glyco-
conjugates to terminal b-galactose residues on the
surface of the parasite.5
Because trypanosomes are incapable of sialic acid bio-
synthesis, TSia is a key enzyme in the pathogenesis of T.
cruzi. Therefore, TSia represents a promising target for
the development of therapeutics to treat Chagas disease,
especially as TSia is absent in mammals. Consequently,
the characterization of ligand binding properties of TSia
at atomic resolution is key for the design of such inhib-
itors, and also yields further insights into mechanistic as-
pects of TSia. With this aim, we studied the interactions
of different chemically-modified sialyl donor substrates
with TSia in solution, using saturation transfer differ-
ence (STD) NMR spectroscopy.6
Trypanosoma cruzi is a parasitic protozoa that is respon-
sible for over 20 million Chagas disease infections in
Central and South America. As part of the infection
process, the parasite expresses a developmental stage-
specific GPI-anchored trans-sialidase (TSia) that allows
it to survive in the bloodstream of its host.1 As trypan-
osomes are unable to synthesize sialic acids de novo,
they use their TSia to transfer sialic acids from exogenic
cell surface glycoconjugates of the host to mucin-like
acceptors present on the parasite’s plasma membrane.
This results in the acquisition of a negatively charged
glycopeptide coat that shields the parasite’s antigenic
structures.2 These sialylated surface glycoconjugates
have also been implicated in cell adhesion and invasion
mechanisms of T. cruzi.3
*
Corresponding author. Tel.: +49 451 500 4230; fax: +49 451 500
4241; e-mail: thomas.peters@chemie.mu-luebeck.de
These authors have contributed equally to this work.
0008-6215/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.carres.2007.05.037

A. Blume et al. / Carbohydrate Research 342 (2007) 1904–1909
1905
2. Material and methods
formed to allow the sample to come to equilibrium.
Spectra were performed with a sweep width of 5 kHz
and 32,768 data points. Reference spectra were acquired
using the same conditions but with only 768 transients.
Ligand protons were assigned using standard 1H–1H
COSY, 1H–1H TOCSY, 1H–13C HSQC and 1H–13C
HMBC experiments under the conditions described
above.
Determination of the hydrolysis rate and examination
of the enzyme kinetics was performed by NMR on a
Bruker DPX 250 NMR spectrometer. Samples consisted
of 15 lM protein and 5 mM donor substrate. For the
enzyme kinetics, 5 mM methyl b-lactoside was also
added. Each time point of the analysis results from a sin-
gle 1 D ‘pulse and acquire’ NMR experiment with 32
scans for acquisition and eight dummy scans each.
Decreasing signal intensities of isolated donor and
acceptor resonance signals were utilized to obtain the
time dependence of the reaction. The signals used were
the pNP signals and the acceptor signal at 4.10 ppm.
As a control the experiments were repeated in the
absence of protein. No hydrolysis of the pNP derivatives
was detected under these conditions.
2.1. Materials
All donor substrate ligands were synthesized as previ-
ously described.7,8 The methyl b-lactoside acceptor sub-
strate was purchased from Sigma (Germany).
TSia from T. cruzi was expressed and purified as de-
scribed earlier.8 For the STD NMR experiments, we em-
ployed deuterated buffers. Therefore, purified TSia was
dialysed against 12 mM phosphate buffer containing
100 mM sodium chloride in D2O, pH* 7.5 (uncorrected
2
reading for the presence of H+) for 3 h. The dialysed
protein was then applied to a PD-10 column (Amer-
sham) and eluted using 5 mL 12 mM phosphate buffer
containing 100 mM sodium chloride in D2O, pH* 7.5.
Fractions of five drops were collected and protein con-
centration was determined using the Bradford assay.
2.2. NMR experiments
Saturation transfer difference (STD) NMR spectra were
obtained as described previously9 at 5 °C using a 12.1 T
Bruker DRX 500 NMR spectrometer equipped with a
triple resonance probehead, incorporating gradients
along the z-axis. In brief, 600 lL samples containing
15 lM protein were mixed with the ligand dissolved in
D2O to yield a final ligand concentration of 5 mM. To
dissect STD signals that were due to direct irradiation,
experiments were repeated under the same conditions
but in the absence of protein. For the acquisition of
STD NMR spectra, a 1D sequence incorporating a
T1q filter was used.10 On-resonance irradiation was per-
formed at ꢀ1 ppm, and off resonance at 20 ppm. Irradi-
ation was performed using 50 Gaussian pulses with a 1%
truncation, and with a duration of 49 ms each. The
pulses were separated by a delay of 1 ms to give a total
saturation time of 2 s. The duration of the T1q filter was
15 ms. STD NMR spectra were acquired with a total of
1536 transients. Thirty two dummy scans were per-
3. Results
The hydrolysis of p-nitrophenyl sialic acid derivatives 1–
4 to TSia was monitored with one dimensional 1H NMR
spectra. To map the binding of these compounds to the
enzyme, we performed STD NMR experiments.
1
3.1. H NMR data reflect the hydrolysis and transfer
activity of T. cruzi TSia
As TSia is known to have a hydrolysis activity besides its
transfer activity, hydrolysis and transfer rates of p-nitro-
phenol N-acetyl neuraminic acid (pNP-Neu5Ac 1;
Scheme 1) were first investigated by time-dependent
1H NMR experiments. Changes of 1H NMR signal
Scheme 1. Structures of studied donor substrate ligands: pNP-Neu5Ac (1), pNP-Oct5Ac (2), pNP-Hept5Ac (3) and pNP-Neu5,9Ac2 (4).

1906
A. Blume et al. / Carbohydrate Research 342 (2007) 1904–1909
Figure 1. Time-dependent 1H NMR experiments of pNP-Neu5Ac (1) in the presence of TSia. Hydrolysis rates were detected by following the
decrease of the pNP signals from pNP-Neu5Ac 1 over time.
intensities of phenyl, acetyl and neuraminic acid signals
show that the enzyme is active under the NMR condi-
tions, even in the absence of acceptor substrate
(Fig. 1). When only the donor substrate is present,
tor is nearly the same for pNP-Hept5Ac 3 and the ‘natu-
ral’ donor pNP-Neu5Ac 1.
The overall enzyme activity in the presence of accep-
tor substrate is the sum of the rates of hydrolysis and
of sialyl transfer. The ratio between these two rates cer-
tainly is of biological interest. In the case of pNP-
Neu5,9Ac2 4, the rate of hydrolysis in the presence of
acceptor is slightly lower than the rate of sialyl transfer
whereas for pNP-Oct5Ac 2 the hydrolysis rate is slightly
larger than the transfer rate. For pNP-Hept5Ac 3, the
hydrolysis rate is significantly higher than the transfer
rate (Fig. 2c). This suggests that shortening of the exo-
cyclic side chain results in a reduction of the rate of sia-
lyl transfer as compared to hydrolysis.
1
time-dependent H NMR experiments show that after
20 min, ꢁ50% hydrolysis has occurred (Fig. 2). Next,
the rate of sialyl transfer was measured in the presence
of equimolar concentrations of pNP-Neu5Ac 1, and
methyl b-D-galactoside as an acceptor substrate. Inter-
estingly, in the presence of acceptor the overall turnover
is slightly decreased (Fig. 2a) and about 50% of the
enzyme activity is still devoted to hydrolysis.
Similar experiments were conducted with the synthetic
donor substrate analogues:7,8 pNP-N-acetyl octulosonic
acid (pNP-Oct5Ac 2, Scheme 1), pNP-N-acetyl heptulo-
sonic acid (pNP-Hept5Ac 3, Scheme 1) and pNP-5,9-di-
acetyl neuraminic acid (pNP-Neu5,9Ac2 4, Scheme 1).
The rates of enzymatic hydrolysis of these ligands in
the absence of acceptor are drastically reduced compared
to the rates of pNP-Neu5Ac 1. For pNP-Oct5Ac 2, 50%
hydrolysis is detected after 9 h, for pNP-Hept5Ac 3 after
3.5 h, and for pNP-Neu5, 9Ac2 4 after 5.5 h (Fig. 2b).
Furthermore, time-dependent 1H NMR experiments
show that the overall enzymatic activity of TSia depends
on the presence or absence of b-methyl lactoside accep-
tor, with the exception of pNP-Neu5,9Ac2 4 (Fig. 2b).
For the ‘natural’ donor pNP-Neu5Ac 1, a reduced over-
all enzymatic activity was detected (Fig. 2a) whereas for
pNP-Oct5Ac 2 and pNP-Hept5Ac 3, the overall enzy-
matic activity is significantly increased in the presence
of the acceptor. For pNP-Hept5Ac 3, the increase is even
more pronounced than for pNP-Oct5Ac 2. This may sug-
gest that the sensitivity of the TSia activity towards the
presence of acceptor is a function of the size of the side
chain. The overall TSia activity in the presence of accep-
3.2. STD NMR investigations on pNP-Neu5Ac 1 and
analogues 2–4
To obtain information about which parts of the ligands
are in contact with the protein, and which should there-
fore contribute to the binding energy, STD NMR exper-
iments were performed with ligands 1–4. As the
hydrolysis rate of pNP-Neu5Ac 1 is too fast to allow
observation of saturation transfer for the educt, only
the hydrolysis products, pNP and Neu5Ac, could be
studied separately. For p-nitrophenol, sizeable STD sig-
nals were detected (Fig. 3). As p-nitrophenol is a rather
hydrophobic component, ‘unspecific’ binding may con-
tribute to this response. For Neu5Ac no sizeable STD
signals were detected (Fig. 3), indicating that sialic acid
alone does not bind well to the enzyme. This is in agree-
ment with biochemical data that have shown that sialic
acid is not a substrate for TSia.
Because the rates of hydrolysis of pNP-Oct5Ac 2,
pNp-Hept5Ac 3, and pNP-Neu5,9Ac2 4 are significantly

A. Blume et al. / Carbohydrate Research 342 (2007) 1904–1909
1907
a
b
1
1
0.8
0.6
0.4
0.2
0
0.8
0.6
0.4
0.2
0
0
10
20
30
40
50
60
70
0
10
20
30
40
50
60
70
Time (min)
Time (min)
c
1
0.9
0.8
0.7
0.6
0.5
0
10
20
30
40
50
60
70
Time (min)
Figure 2. (a) TSia activity in the presence pNP-Neu5Ac 1 donor substrate and the absence or presence of methyl b-D-galactoside acceptor. Diamonds
denote the overall conversion in the absence of acceptor, asterisks the overall conversion in the presence of acceptor, triangles the transfer rate in the
presence of acceptor and squares denote the hydrolysis rate in the presence of acceptor. In the presence of acceptor, the overall conversion rate is
reduced. Transfer and hydrolysis rate are in the same range, indicating that about 50% of the conversion is still hydrolysis. (b) TSia activity in the
presence of different donor substrate analogues and the absence and presence of methyl b-D-galactoside acceptor. Filled squares denote the
conversion rate of pNP-Oct5Ac (2) in the absence of acceptor, open squares the conversion rate of pNP-Oct5Ac in the presence of acceptor, filled
triangles the conversion rate of pNP-Hept5Ac (3) in the absence of acceptor, open triangles the conversion rate of pNP-Hept5Ac in the presence of
acceptor, filled circles the conversion rate of pNP-Neu5,9Ac2 (4) in the absence of acceptor and open circles denote the conversion rate of pNP-
Neu5,9Ac2 4 in the presence of acceptor. In contrast to the ‘natural’ donor, pNP-Neu5Ac shows the analogues pNP-Oct5Ac 2 and pNP-Hept5Ac 3
shows a significant increased TSia activity in the presence of acceptor and pNP-Neu5,9Ac2 4 shows a slightly increased activity in the presence of
acceptor. (c) Transfer and hydrolysis rate in the presence of different donor substrate analouges in the presence of methyl-b-D-galactoside acceptor.
Filled squares denote the hydrolysis rate of pNP-Oct5Ac (2), open squares the transfer rate of pNP-Oct5Ac, filled triangles the hydrolysis rate of
pNP-Hept5Ac (3), open triangles the transfer rate of pNP-Hept5Ac, filled circles the hydrolysis rate of pNP-Neu5,9Ac2 (4) and open circles denote
the transfer rate of pNP-Neu5,9Ac2. For pNP-Oct5Ac 2 and pNP-Hept5Ac 3, the hydrolysis rate is significantly higher than the transfer rate,
whereas for pNP-Neu5,9Ac2 4, the hydrolysis rate is lower than the transfer rate.
reduced compared to compound 1, STD NMR experi-
ments of the non-hydrolyzed educts are possible. There-
fore, we have observed STD signals for ligands 2–4
(Fig. 4). Large STD effects were obtained for the pNP
moieties whereas the response protons attached to the
modified sialic acid residues were rather poor. More-
over, control experiments in the absence of TSia show
that the STD signals in the region between ꢁ1 and
4.5 ppm are due to direct irradiation. No direct irradia-
tion was observed for the aromatic signals.
4. Discussion
Recently, the analysis of the crystal structure of a T. cru-
zi TSia mutant11 has resulted in a model for the catalytic
mechanism,12 and in conjunction with NMR studies of
the acceptor specificity of TSia,13 some of the mechanis-
tic principles of T. cruzi TSia are now understood. A
crystallographic study where TSia had been co-crystal-
lized with 3-deoxy-2,3-didehydro N-acetyl neuraminic
acid (Neu5Ac2en) and lactose shows that Neu5Ac2en

1908
A. Blume et al. / Carbohydrate Research 342 (2007) 1904–1909
is located in a binding pocket that is in close proximity
to an adjacent binding pocket for lactose.11 Lactose is
oriented in such a fashion that its hydrophobic faces
are sandwiched between the side chains of the aromatic
amino acids Tyr 119 and Trp 312. This is in good agree-
ment with results of STD NMR studies for the binding
of sialyl lactose to an enzymatically inactive mutant of
TSia where several close contacts between protons of
the carbohydrate ligand and the protein were ob-
served.14 In particular, it was observed that the neuram-
inic acid residue and the region around the a-(2!3)-
glycosidic linkage is involved in molecular recognition.
In an extension to this work, it was found later that
binding of sialosides to this inactivated form of TSia
supports recognition of acceptor substrates in an alloste-
ric fashion.15
One part of our present study addresses the binding of
donor substrates 1–4 to enzymatically active TSia using
STD NMR experiments. In general, such experiments
identify ligand portions that strongly interact with the
enzyme’s binding pocket and will discriminate them
from those ligand fragments that are more distant.16
Surprisingly, our studies show that for the donor sub-
strates 1–4, the p-nitrophenyl residue dominates the
binding to TSia whereas the sugar moieties only receive
a rather small fraction of saturation from the protein.
Although we cannot readily explain these observations
on the basis of published results, our data do not neces-
sarily contradict previous findings. We suggest that the
molecular recognition of p-nitrophenyl glycosides of sia-
lic acid derivatives follows the same principles as the rec-
ognition of other ligands studied previously. If the p-
nitrophenyl residue is placed between the aromatic side
chains of Tyr 119 and Trp 312 instead of, for example, a
lactose moiety, one may envision that the neuraminic
acid residue is in a slightly different position as com-
pared to in the case of sialyl lactose.14 Also, in our study
we have used enzymatically active TSia that performs a
constant turnover of substrate molecules. One may spec-
ulate that under these conditions conformational
changes take place that on the time average yield a dif-
ferent binding epitope of donor substrates. Therefore, it
would be very interesting to compare the binding of the
donor substrates 1–4 to an inactive mutant of TSia.
A comparison of our data with results from binding
studies of donor substrates to active sites of, for exam-
ple, glycosyltransferases shows that similar recognition
principles must be at work. In those cases, it is the nucleo-
tide that guides the corresponding donor substrates into
the donor binding pocket.9,17 These obvious similarities
between glycosyltransferases and TSia may suggest a
common mechanism/origin or convergent evolution of
these enzymes, although TSia is distantly related to
sialidases.
Figure 3. Reference (top) and STD NMR (bottom) spectrum of pNP-
Neu5Ac (1) in the presence of TSia. Only hydrolysed pNP leads to
significant STD signals, whereas from pNP-Neu5Ac and Neu5Ac
alone, no signals are detectable. The signal at 1.7 ppm is an artifact due
to direct irradiation of the NAc group.
Figure 4. STD spectra of pNP-Oct5Ac 2 (top), pNP-Hept5Ac 3
(middle) and pNP-Neu5,9Ac2 4 (bottom). All theses ligands bind to
TSia with the pNP moiety dominating the binding. For the sugar
moiety, no sizeable STD signals were detected (the signals in the range
between ꢁ1 and 4.5 ppm are mostly due to direct irradiation; for a
discussion see text). In addition signals from hydrolyzed pNP are
visible.
Modifications of the glycerol side chain by shortening
or additional acetylation reduce both hydrolysis and

A. Blume et al. / Carbohydrate Research 342 (2007) 1904–1909
1909
transfer of sialyl residues. Nevertheless, substantial
References
transfer can still be observed for the modified substrates.
Although these structural features do not seem to be
crucial for the enzyme activity, they clearly have a mod-
ulating influence. Furthermore, our results show that
shortening of the glycerol chain favors hydrolysis over
the transfer reaction. A biological point of these findings
may point towards a regulatory function of natural
modifications of the glycerol side chain.
1. Schenkman, S.; Eichinger, D.; Pereira, M. E.; Nus-
senzweig, V. Annu. Rev. Microbiol. 1994, 48, 499–523.
2. Tomlinson, S.; Pontes de Carvalho, L. C.; Vandekerckh-
ove, F.; Nussenzweig, V. J. Immunol. 1994, 153, 3141–
3147.
3. Schenkman, S.; Jiang, M. S.; Hart, G. W.; Nussenzweig,
V. Cell 1991, 65, 1117–1125.
4. Paulson, J. C.; Colley, K. J. J. Biol. Chem. 1989, 264,
17615–17618.
The fact that complete transfer of sialyl residues is
not achieved, even if donor and acceptor substrate are
present at equimolar concentrations, implies that the
transfer is essentially controlled by the acceptor concen-
tration. Every sialic acid that binds to TSia is cleaved
from its glycoconjugate and if a glycoconjugate of the
parasite is available it is transferred to this structure; if
not, the sialic acid is transferred to water. Therefore,
the more the parasitic glycoconjugates are available,
the more effective the transfer of sialic acids by TSia
is. Compared to other sugar transferring enzymes, TSia
is relatively inefficient with regard to the transfer reac-
tion. It can be speculated that the ineffective transfer
of T. cruzi TSia is due to the missing evolutionary pres-
sure for an effective transfer reaction. As T. cruzi TSia
uses sialic acids attached to cells of the host organism
and does not have to produce these structures on its
own, it does not make any difference to the parasite
whether these sialic acid residues are transferred to an
acceptor or are just hydrolyzed. In conjunction with pre-
vious findings,5,11–15 our present data extend the basis
for a rational design of specific inhibitors of TSia, the
key enzyme in the T. cruzi life cycle.
5. Todeschini, A. R.; Mendonca-Previato, L.; Previato, J. O.;
Varki, A.; van Halbeek, H. Glycobiology 2000, 10, 213–
221.
6. Mayer, M.; Meyer, B. Angew. Chem., Int. Ed. 1999, 38,
1784–1788.
7. Neubacher, B.; Scheid, S.; Kelm, S.; Frasch, A. C.; Meyer,
B.; Thiem, J. ChemBiochem 2006, 7, 896–899.
8. Neubacher, B.; Schmidt, D.; Ziegelmuller, P.; Thiem, J.
Org. Biomol. Chem. 2005, 3, 1551–1556.
9. Blume, A.; Angulo, J.; Biet, T.; Peters, H.; Benie, A. J.;
Palcic, M.; Peters, T. J. Biol. Chem. 2006, 281, 32728–
32740.
10. Scherf, T.; Anglister, J. Biophys. J. 1993, 64, 754–761.
11. Buschiazzo, A.; Amaya, M. F.; Cremona, M. L.; Frasch,
A. C.; Alzari, P. M. Mol. Cell 2002, 10, 757–768.
12. Watts, A. G.; Damager, I.; Amaya, M. L.; Buschiazzo, A.;
Alzari, P.; Frasch, A. C.; Withers, S. G. J. Am. Chem. Soc.
2003, 125, 7532–7533.
13. Haselhorst, T.; Wilson, J. C.; Liakatos, A.; Kiefel, M. J.;
Dyason, J. C.; von Itzstein, M. Glycobiology 2004, 14,
895–907.
14. Todeschini, A. R.; Girard, M. F.; Wieruszeski, J. M.;
Nunes, M. P.; DosReis, G. A.; Mendonca-Previato, L.;
Previato, J. O. J. Biol. Chem. 2002, 277, 45962–45968.
15. Todeschini, A. R.; Dias, W. B.; Girard, M. F.; Wierusz-
eski, J. M.; Mendonca-Previato, L.; Previato, J. O. J. Biol.
Chem. 2004, 279, 5323–5328.
16. Mayer, M.; Meyer, B. J. Am. Chem. Soc. 2001, 123, 6108–
6117.
17. Blume, A.; Benie, A. J.; Stolz, F.; Schmidt, R. R.; Reutter,
W.; Hinderlich, S.; Peters, T. J. Biol. Chem. 2004, 279,
55715–55721.
Acknowledgement
T.P. and J.T. thank the Deutsche Forschungsgemeins-
chaft (DFG) for financial support within the project
program grant SFB 470.