Dynamic light scattering evidence for a ligand-induced motion between the two domains of glucoamylase G1 of Aspergillus niger with heterobivalent substrate analogues

Article abstract of DOI:10.1002/(SICI)1521-3773(19990401)38:7<974::AID-ANIE974>3.0.CO;2-K

Heterobifunctional ligands that bind at the same time to the catalytic domain and to the starch-binding domain of glucoamylase induce a conformational change of the protein, as shown by dynamic light scattering. The ligands consist of acarbose and β-cyclo

Full text of DOI:10.1002/(SICI)1521-3773(19990401)38:7<974::AID-ANIE974>3.0.CO;2-K

COMMUNICATIONS
1324 ± 1327. See also W. T. Lowther, D. A. McMillen, A. M. Orville,
B. W. Matthews, Proc. Natl. Acad. Sci. USA 1998, 95, 12153 ± 12157.
[13] For a recent review of protein myristoylation, see J. A. Boutin, Cell.
Signalling 1997, 9, 15 ± 35.
[14] T. Yoshida, Y. Kaneko, A. Tsukamoto, K. Han, M. Ichinose, S.
Kimura, Cancer Res. 1998, 58, 3751 ± 3756.
[15] For a scholarly review of substrate-directable chemical reactions, see
A. H. Hoveyda, D. A. Evans, G. C. Fu, Chem. Rev. 1993, 93, 1307 ±
1370.
[16] V. Van Rheenen, R. C. Kelly, D. Y. Cha, Tetrahedron Lett. 1976,
1973 ± 1976.
Dynamic Light Scattering Evidence for
a Ligand-Induced Motion between the
Two Domains of Glucoamylase G1 of
Aspergillus niger with Heterobivalent
Substrate Analogues**
Nathalie Payre, Sylvain Cottaz, Claire Boisset,
Redouane Borsali, Birte Svensson, Bernard Henrissat,
and Hugues Driguez*
[17] W. Langenbeck, O. Godde, L. Weschky, R. Schaller, Ber. Dtsch. Chem.
Ges. 1942, 75, 232 ± 236.
Glucoamylases (GAs) catalyze the hydrolytic release of b-
d-glucose from the nonreducing ends of starch and related
oligo- and polysaccharides. Most GAs possess a starch-
binding domain (SBD) separated from the catalytic domain
(CD) by a glycosylated peptide linker of variable length.[1]
Removal of the SBD reduces the activity of GA from
Aspergillus niger on insoluble starch but not on soluble
substrates.[2] We have previously shown that 6II-thiopanose
and its higher oligomers bind essentially to the SBD and
modulate GA activity on starch.[3] This raised the possibility of
an interaction between the CD and the SBD of GA, and these
observations have suggested that a cooperativity of the two
domains could be critical for optimal activity.[4] The only low
resolution structural information available so far on the entire
GA was obtained by scanning tunneling microscopy,[5] but the
possible mobility of the two domains induced by substrate
binding cannot be described by this technique. The three-
dimensional structure of the CD of the GA from Aspergillus
awamori X100 has been solved by X-ray crystallography,[6]
while that of the SBD of GA isolated from Aspergillus niger
has been recently determined by NMR spectroscopy.[7] Failure
to crystallize the entire GA has repeatedly been observed and
this is attributed to the inherent flexibility of the linker
peptide that connects the two constitutive domains. Co-
crystallization in the presence of a ligand targeted to both the
CD and SBD may stabilize one conformer. Recently, closure
of a flexible loop onto a substrate analogue has allowed the
crystallization of a cellulase.[8]
[18] 1-Diethylamino-1,3-butadiene can be prepared on a large scale in one
step from diethylamine and crotonaldehyde, see S. Hünig, H.
Kahanek, Chem. Ber. 1957, 90, 238 ± 245.
[19] Woodwardꢁs landmark synthesis of reserpine provides instructive,
early instances of this tactic, see R. B. Woodward, F. E. Bader, A. J.
Frey, R. W. Kierstead, Tetrahedron 1958, 2, 1 ± 57.
[20] Vinyl bromide 8 was prepared by a slight modification of the excellent
one-flask procedure of Corey et al., see E. J. Corey, J. Lee, B. E.
Roberts, Tetrahedron Lett. 1997, 38, 8915 ± 8918. See also E. J. Corey,
J. P. Dittami, J. Am. Chem. Soc. 1985, 107, 256 ± 257.
[21] For the use of Li(2-thienyl)CuCN, see B. H. Lipshutz, M. Koerner,
D. A. Parker, Tetrahedron Lett. 1987, 28, 945 ± 948.
[22] For reviews of the conjugate addition chemistry of organocuprates, see
a) B. H. Lipshutz, S. Sengupta, Org. React. 1992, 41, 135 ± 631; b) J. A.
Kozlowski in Comprehensive Organic Synthesis, Vol. 1 (Eds.: B. M.
Trost, I. Fleming, S. L. Schreiber), Pergamon, New York, 1991,
pp. 169 ± 198.
[23] Similar results were obtained when chlorotrimethylsilane was used in
place of BF3 ´ OEt2, followed by fluoride-induced cleavage of the
crude enol silyl ethers.
[24] a) C. H. Cummins, R. M. Coates, J. Org. Chem. 1983, 48, 2070 ± 2076;
b) R. M. Coates, C. H. Cummins, J. Org. Chem. 1986, 51, 1383 ± 1389.
[25] K. B. Sharpless, R. C. Michaelson, J. Am. Chem. Soc. 1973, 95, 6136 ±
6137.
[26] While our studies were underway, a closely related transformation was
described, see S. Amano, N. Ogawa, M. Ohtsuka, S. Ogawa, N. Chida,
Chem. Commun. 1998, 1263 ± 1264.
[27] This is a slight modification of a procedure employed in Corey and
Sniderꢁs pioneering synthesis of fumagillin (reference [6]).
[28] Structure 4 (E ˆ CO2Me) in optically active form is available from
(
)-quinic acid (D. F. McComsey, B. E. Maryanoff, J. Org. Chem.
1994, 59, 2652 ± 2654) or by Sharpless asymmetric dihydroxylation
followed by isopropylidene ketal formation (Z.-M. Wang, K. Kakiu-
chi, K. B. Sharpless, J. Org. Chem. 1994, 59, 6895 ± 6897).
Herein we describe the design and synthesis of high affinity
probes that bind the CD and SBD of GA at the same time,
and hence get new insights into the structure/activity relation-
ships of GA. The structural variations of GA between free
and bound states were monitored by quasi-elastic light
[*] Dr. H. Driguez, Dr. N. Payre, Dr. S. Cottaz, Dr. C. Boisset,
Dr. R. Borsali
Â
 Â
Centre de Recherches sur les Macromolecules Vegetales
(CERMAV-CNRS)
Â
Affiliated with the Universite Joseph Fourier, Grenoble
B.P. 53, F-38041 Grenoble cedex 9 (France)
Fax : (‡ 33)476-037-664
E-mail: hdriguez@cermav.cnrs.fr
Dr. B. Svensson
Department of Chemistry, Carlsberg Laboratory
Gamle Carlsberg Vej 10, DK-2500 Valby (Denmark)
Dr. B. Henrissat
Â
Architecture et Fonction des Macromolecules Biologiques
CNRS-IFR1
31 Chemin Joseph Aiguier, F-13402 Marseille cedex 20 (France)
[**] This work was supported in part by the European Union Biotechnol-
ogy Program (contract BIO4-CT98-0022).
974
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
1433-7851/99/3807-0974 $ 17.50+.50/0
Angew. Chem. Int. Ed. 1999, 38, No. 7

COMMUNICATIONS
scattering, a technique that has re-
cently been used with success for the
determination of the hydrodynamic
dimensions of a multidomain cellu-
lase.[9]
It is well known that acarbose is a
potent inhibitor of GA and that two
molecules of b-cyclodextrin may
mimic amylose binding on two sites
of the SBD but cannot fit into the
pocket-shaped active site.[10] These
observations indicate that an amylose
chain that interacts with the two
constitutive domains of GA could
be mimicked by an acarbose mole-
cule tethered to a cyclodextrin mole-
cule through flexible spacers of dif-
ferent lengths.
A few years ago we developed an
efficient strategy for the synthesis of
b-cyclodextrins that were substituted
at the 6-position with a thioglucosyl
moiety.[11] The X-ray structure of a
cyclodextrin glycosyltransferase com-
Scheme 1. Retrosynthetic pathway of the bifunctional ligands 1 ± 4.
plexed with one of these molecules was recently solved, and
indicated that the presence of a thioglucosidic linkage did not
hinder the binding to the protein.[12] These results encouraged
us to connect linkers to acarbose and b-cyclodextrin through
sulfur linkages.
Scheme 1 presents the retro-synthetic analysis on which the
synthesis was based. Disconnection at the indicated bonds led
to the key building blocks 5 ± 8 for the synthesis of 1 ± 3. The
longest probe 4 was best prepared by elongation of the linker
7 already coupled to b-cyclodextrin. To allow coupling with
6-deoxy-6-iodo-b-cyclodextrin 5 the bifunctional spacers 6
and 7 had to carry a thiol function and a protected hydroxyl
group that could be easily converted into a thiophilic group
for the establishment of the thioglycosidic linkage with S-
acetylacarbose (8).
Protection of the hydroxyl groups of commercially avail-
able monochlorotriethylene glycol 9 by tritylation (trityl ˆ
Trˆ triphenylmethyl), followed by halogen exchange and
nucleophilic displacement of iodine with potassium thioace-
tate, afforded the first spacer-arm 6 in 33% overall yield
(Scheme 2). Tetraethylene glycol 12 was treated with tosyl
chloride and the resulting monotosylate 13 was tritylated to
give 14. Substitution of the tosylate with thioacetate and
cleavage of the trityl group gave 16, which was coupled with
14 after S-deacetylation (!17). Mesylation (mesylˆMs ˆ
methanesulfonyl) of 17 gave 18, and substitution of the
mesylate group with thioacetate gave the bifunctional spacer
7 in 22% overall yield (from 12).
The S-deacetylation of compounds 6 or 7 gave the corre-
sponding sodium thiolates, which were coupled, without
characterization, with 6-deoxy-6-iodo-cyclodextrin 5 prepared
as already described[11] (Scheme 3). The branched cyclo-
dextrins 19 and 20 were obtained in 76 and 64% yield,
respectively. Detritylation, followed by mesylation and
displacement of the mesyl group with iodine gave the
Scheme 2. Preparation of the bifunctional spacers 6 ± 7. a) Ph3CCl, CH2Cl2,
4-DMAP, NEt3, RT, 20 h, 10: 70%, 14: 51%; b) NaI (1.1 equiv), DMF,
1008C, 3 h, 75%; c) KSAc (6 equiv), DMF, 1008C, 2 h, 6: 63%, 15: 87%;
d) TsCl (1 equiv), pyridine, 08C !RT, 12 h, 52%; e) HBF4 in H2O, CH3CN,
RT, 1 h, 95%; f) 1. 16, MeONa (1.3 equiv), MeOH, 08C, 2 h, argon; 2. 14
(0.6 equiv), DMF, 458C, RT, argon, 73%; g) MeSO3Cl, pyridine, 08C !RT,
84%; h) KSAc, DMF, RT, 2 h, argon, 87%. 4-DMAP ˆ dimethylamino-
pyridine.
Angew. Chem. Int. Ed. 1999, 38, No. 7
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
1433-7851/99/3807-0975 $ 17.50+.50/0
975

COMMUNICATIONS
The thermodynamic parameters of ligand binding to GA, as
well as to the CD and SBD of GA, have been studied previously
by isothermal titration calorimetry.[16] The DH0 values of
binding of the four heterobidentate ligands were, within
experimental error, equal to the sum of the DH0 values of the
binding of free acarbose and b-cyclodextrin to the CD and
SBD, respectively. These results showed that the catalytic and
starch binding sites of GA are in close proximity in solution
and this suggests considerable flexibility of the linker region.
Here the hydrodynamic dimensions of GA and the CD
were determined by quasi elastic light scattering experiments.
In accord with their three-dimensional structures the isolated
SBD and CD were approximated as spheres, while the dumb-
bell two-domain GA was approximated as an ellipsoid whose
minor semi-axis corresponded to the hydrodynamic radius of
the largest individual component, for example, the CD. The
experiments were performed at different angles and several
1
protein concentrations (0.3 ± 1.1 mgmL ). The autocorrela-
tion function was represented by a single exponential decay
for all concentrations and scattering angles. A very small
contribution from a slow process has been identified at high
concentration, which was attributed to the presence of minor
protein aggregates. Relaxation times were estimated from the
correlation functions, and the variation of the frequencies
versus the scattering angles allowed the determination of the
translational diffusion coefficients DT(c) at different concen-
trations. The values of the diffusion coefficients at infinite
dilution D0T ˆ DT(c)c!0, which represents the single particle
property, were converted into their hydrodynamic radii
(modified Stockes ± Einstein relation). Neither the GA or
CD systems showed any concentration dependence on the
diffusion coefficient (Figure 1).
Scheme 3. Preparation of branched cyclodextrins 21, 22, and 24. a) 6, 7, 22,
MeONa (10 equiv), 1 h, 08C, argon; b) 1. DMF, RT, 12 h, argon; 2. Ac2O,
pyridine, 4-DMAP, 608C, 12 h, 19: 76%, 20: 64%, 23: 70%; c) 1. HBF4 in
H2O, CH3CN, RT, 20 min; 2. MeSO3Cl, pyridine, 08C !RT, 12 h; 3. NaI,
DMF, 1008C, 4 h, 21: 51%, 22: 40%, 24: 43%.
corresponding iodo compounds 21 and 22
in overall yields of 51 and 40%, respec-
tively. This stepwise protocol gave a higher
yield than the direct OH !I substitu-
tion.[13] Coupling of the iodo derivative
22 with the thiolate arising from S-deace-
tylation of 7, as already described, gave
compound 23 in 70% yield. Transforma-
tion of 23 into 24 was obtained by sequen-
tial reactions as described for the synthesis
of 21 and 22. The longest synthon 24 was
obtained in an overall yield of 43%
starting from 22.
To synthesize donor 8, acarbose (25)
was first acetylated, then treated with
HBr/AcOH,[14] and the corresponding bro-
mide was transformed into 8 in 63%
overall yield (Scheme 4).
Coupling of the acarbose derivative 8
with acceptors 5, 21, 22, and 24 under the
conditions of von Itzstein et al.[15] led to
the acetylated target molecules 26 ± 29 in
yields of 63, 88, 83, and 74%, respectively.
The final products 1 ± 4 were obtained by
Scheme 4. Preparation of ligands 1 ± 4. a) 1. Ac2O, pyridine, RT, 24 h; 2. CH2Cl2, HBr/AcOH,
108C !08C, 1 h; 3. KSAc, DMF, RT, 18 h, 63%; b) Et2NH, DMF, RT, 4 h, 26: 63%, 27: 88%, 28:
83%, 29: 74%; c) MeONa, MeOH, RT, 1 h, then NH4OH (1m), RT, 12 h, 1: 78%, 2: 74%, 3: 91%,
4: 74%.
O-deacetylation and were purified by
reversed-phase chromatography on
C-18 column.
a
976
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
1433-7851/99/3807-0976 $ 17.50+.50/0
Angew. Chem. Int. Ed. 1999, 38, No. 7

COMMUNICATIONS
Keywords: carbohydrates ´ dynamic light scattering ´ en-
zyme inhibitors ´ protein structures ´ structure ± activity
relationships
[1] P. M. Coutinho, P. J. Reilly, Proteins Struct. Funct. Genet. 1997, 29,
334 ± 347.
[2] N. J. Belshaw, G. Williamson, FEBS Lett. 1990, 269, 350 ± 353.
[3] a) S. Cottaz, H. Driguez, B. Svensson, Carbohydr. Res. 1992, 228, 299 ±
305; b) C. Apparu, H. Driguez, G. Williamson, B. Svensson,
Carbohydr. Res. 1995, 277, 313 ± 320.
[4] B. Svensson, T. G. Pedersen, I. Svendsen, T. Sakai, M. Ottensen,
Carlsberg Res. Commun. 1982, 47, 55 ± 69.
[5] A. P. Gunning, V. J. Morris, G. Williamson, N. J. Belshaw, G. F. H.
Kramer, M. W. Kanning, Analyst 1994, 119, 1939 ± 1942, and refer-
ences therein.
[6] A. E. Aleshin, B. B. Stoffer, L. M. Firsov, B. Svensson, R. B. Honzat-
ko, Biochemistry 1996, 35, 8319 ± 8328, and references therein.
Figure 1. Variation of the translational diffusion coefficient DT(c) as a
~
&
È
[7] K. Sorimachi, A. J. Jacks, M.-F. Le Gal-Coeffet, G. Williamson, D. B.
*
^
function of concentration c of GA ( ), CD ( ), GA ‡ 1 ( ), GA ‡ 2 ( ),
&
Archer, M. P. Williamson, Mol. Biol. 1996, 259, 970 ± 987, and
references therein.
and GA ‡ 3 ( ).
[8] C. Reverbel-Leroy, G. Parsiegla, V. Moreau, M. Juy, C. Tardif, H.
Driguez, J. P. Belaich, R. Haser, Acta Crystallogr. Sect. D 1998, 54,
114 ± 118.
[9] C. Boisset, R. Borsali, M. Schülein, B. Henrissat, FEBS Lett. 1995, 376,
49 ± 52.
[10] B. W. Sigurskjold, B. Svensson, G. Williamson, H. Driguez, Eur. J.
Biochem. 1994, 225, 133 ± 141.
[11] S. Cottaz, H. Driguez, Synthesis 1989, 755 ± 757.
[12] A. K. Schmidt, S. Cottaz, H. Driguez, G. E. Schulz, Biochemistry 1998,
37, 5909 ± 5915.
The extrapolated values at infinite dilution gave hydro-
dynamic sizes of 78.5 Š (major semi-axis) and 30 Š (semi-
axis), for the GA and CD, respectively. These values are in
good agreement with those obtained previously with other
techniques (141 and 60 ± 65 Š for the major and minor axis of
GA, respectively).[5, 6] The diffusion coefficients for 1, 2, and 3
in the presence of GA show a decrease as a function of
concentration (Figure 1).
[13] B. Classon, Z. Liu, B. Samuelsson, J. Org. Chem. 1988, 53, 6126 ± 6130,
and references therein.
The three systems gave roughly the same concentration
dependence within experimental errors. The associated dif-
fusion coefficients decrease with the protein concentration
and the extrapolated value at infinite dilution gave a hydro-
dynamic size of 62 Š (major semi-axis). However a value of
72.5 Š was found with the longer probe 4. These values are
intermediate between systems with a CD and GA only; and is
in fact expected if the ligand somehow decreases the size of
the GA system. The negative value of the slope of DT versus c
shows that there is a strong attraction between the particles
and reflects the thermodynamics of the system; in general the
value of the slope is pH dependent. It appears from these
results that the presence of the bound bifunctional ligands
stabilizes a more compact conformation of GA. If this motion
of the two domains also occurs upon binding of the natural
substrate it may achieve the processivity of this exo-enzyme:
GA would bind to the polymer through its SBD and the
motion of the CD would allow the cleavage of several glucose
moieties at the nonreducing end of the same glucan chain
before the release of the enzyme. This new concept is under
investigation by using mutants of glucoamylase with linkers of
various lengths.
[14] B. Junge, H. Böshagen, J. Stoltefuss, L. Miller in Enzyme Inhibitors
(Ed.: U. Brodbeck) VCH, Basel, 1980, p. 123.
[15] S. Bennet, M. von Itzstein, M. J. Kiefel, Carbohydr. Res. 1994, 259,
293 ± 299.
[16] B. W. Sigurskjold, T. Christensen, N. Payre, S. Cottaz, H. Driguez, B.
Svensson, Biochemistry 1998, 37, 10446 ± 10452.
Metal-Free Haloperoxidases: Fact or Artifact?
Ole Kirk* and Lars Sparre Conrad
Haloperoxidases are enzymes that catalyze the oxidation of
a halide ion (X ) to the corresponding hypohalous acid
[Eq. (1)]. The hypohalous acid produced can further react
with different nucleophilic acceptors to form a diversity of
halogenated compounds.[1, 2] Different classes of haloperox-
Experimental Section
[*] Dr. O. Kirk
All compounds were homogeneous according to elemental analysis or HR-
Department of Bio-Organic Chemistry
Enzyme Research, Novo Nordisk A/S
Novo Alle, DK-2880 Bagsvaerd (Denmark)
Fax: (‡45)4442-3206
1
MS and H and 13C NMR spectra.
Dynamic light scattering experiments: Sample preparation, equipment,
and data analysis are essentially the same as described in ref. [9], except
that sodium acetate buffer (50mm, pH 4.2) and a temperature of 278C were
used. Equimolar ratios of ligands and enzyme were utilized.
E-mail: oki@novo.dk
Dr. L. Sparre Conrad
Department of Formulation Design
Enzyme Development and Application, Novo Nordisk A/S
Novo Alle, 2880 Bagsvaerd (Denmark)
Received: September 14, 1998 [Z12415IE]
German version: Angew. Chem. 1999, 111, 1027 ± 1030
Angew. Chem. Int. Ed. 1999, 38, No. 7
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
1433-7851/99/3807-0977 $ 17.50+.50/0
977