A potential fortuitous binding of inhibitors of an inverting family GH9 β-glycosidase derived from isofagomine

Article abstract of DOI:10.1039/c1ob05766a

Using structural insight, the binding mode of isofagomine-derived inhibitors with family GH9 glycosidases is achieved via the study of Alicyclobacillus acidocaldarius (AaCel9A) endoglucanase. In contrast to what was observed in the first report using thes

Full text of DOI:10.1039/c1ob05766a

View Article Online / Journal Homepage / Table of Contents for this issue
Organic &
Biomolecular
Chemistry
Dynamic Article Links
Cite this: Org. Biomol. Chem., 2011, 9, 5945
www.rsc.org/obc
COMMUNICATION
A potential fortuitous binding of inhibitors of an inverting family GH9
b-glycosidase derived from isofagomine†
Solange More´ra,*a Armelle Vigourouxa and Keith A. Stubbs*b
Received 16th May 2011, Accepted 7th June 2011
DOI: 10.1039/c1ob05766a
Using structural insight, the binding mode of isofagomine-
derived inhibitors with family GH9 glycosidases is achieved
via the study of Alicyclobacillus acidocaldarius (AaCel9A)
endoglucanase. In contrast to what was observed in the first
report using these compounds with inverting glycosidases
from family GH6, these inhibitors do not adopt a distorted
conformation in the active site.
putative transition-state mimics and the surrounding residues of
the enzyme.
Previously, we reported the structural characterization of a
novel endoglucanase, Cel9A, from Alicyclobacillus acidocaldarius
(AaCel9A).22 This enzyme, like all other members of glycoside
hydrolase family GH9, uses an inverting catalytic mechanism
where, in the case of AaCel9A, Glu515 acts as the catalytic acid and
Asp146, the catalytic base (Fig. 1). The enzyme has demonstrated
cellobiohydrolase activity and acts on carboxy-methylcellulose
(CMC), lichenan and pNP-cellooligosaccharides and displays a
The glycoside hydrolases are an important group of carbohydrate-
processing enzymes found in nature as they have roles in the
biosynthesis and degradation of glycoconjugates. One particular
set of enzymes that are receiving increased attention due to
biotechnological and environmental applications1 are the cellu-
lases which degrade the major plant cell wall polysaccharide,
cellulose. These enzymes display a variety of biological activity2
and are currently grouped into 11 of the 120 glycoside hydrolase
sequence-based families (GHs) within the Carbohydrate-Active
enZymes (CAZy) database.3–6
◦
temperature optimum of 70 C and a pH optimum of 5.5. Here,
we present a detailed analysis of the inhibition and structural
analysis of AaCel9A using a series of oligosaccharide inhibitors
1–3, derived from the parent azasugar, isofagomine 4.
One particular subset of cellulases are those from family
GH9. This family of glycoside hydrolases contains enzymes that
have endoglucanase, cellobiohydrolase, b-glucosidase and exo-b-
glucosaminidase activity. Previous structural studies of enzymes
in this family7–12 cover the four themes (A, B, C and D) describing
the modular arrangement of the GH9 enzymes.13 As a result of the
importance of these enzymes, much interest has been expressed in
the synthesis and testing of compounds as putative inhibitors. The
best known inhibitors of cellulases are the non-hydrolysable thio-
oligosaccharides9,14–18 and the azasugar-based oligosaccharides19–21
with the former already being used to study the binding mode
of substrates for family GH9 enzymes.9 However, to date no
structural insight has been obtained for GH9 enzymes using
iminosugar-based oligosaccharides which would be of interest
to understand not only the binding mode of these compounds
but would also allow insight into the protonation states of
Fig. 1 (A) AaCel9A uses a catalytic mechanism involving an inversion of
anomeric configuration. (B) Structures of the compounds 1–3 used in the
structural study and the known glycosidase inhibitor isofagomine 4.
The synthesis of the desired compounds proceeded using the
known Cbz-protected isofagomine 523 (Scheme 1) which can
be readily prepared from isofagomine 424 or prepared in situ
from 4. Using various protecting group manipulations the diol
6 could be obtained in good yield. The diol was selectively
acetylated to yield the alcohol 7 which was then treated with the
appropriate trichloroacetimidate 8–10 to give the corresponding
desired products 11–13, all of which had 1H and 13C NMR
characteristics consistent with the chemoenzymatic preparation
aLaboratoire d’Enzymologie et Biochimie Structurales (LEBS), CNRS,
Avenue de la Terrasse, 91198, Gif-sur-Yvette, France. E-mail: morera@
lebs.cnrs-gif.fr; Tel: +33 1 69823470
bChemistry M313, School of Biomedical, Biomolecular and Chemical
Sciences, The University of Western Australia, 35 Stirling Highway, Crawley,
WA, Australia, 6009. E-mail: kstubbs@cyllene.uwa.edu.au; Tel: +61 8 6488
2725
† Electronic supplementary information (ESI) available: Materials and
methods, refinement statistics and supplemental figures and tables. See
DOI: 10.1039/c1ob05766a
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 5945–5947 | 5945
©

View Article Online
conformation being more pronounced in the isofagomine residue.
These distortions appear to be driven through opportunistic
interactions with the enzyme. Indeed, the isofagomine in the
+1 subsite makes five interactions with the active site residues,
Glu515, His461 and Arg463, which is two additional interactions
compared to a glucose at this position.22
Both the glucose and isofagomine moieties in the -2 and -1
4
subsite respectively lie in undistorted C1 (chair) conformations
which is different to the distorted 2,5B (boat) conformation
observed for isofagomine for the inverting GH6 enzymes.19,26 The
observed binding mode has only been reported so far for retaining
GH5 enzymes.20,21 Similarly for 2 and 3, one inhibitor molecule
binds at the distal subsite with the isofagomine moiety in the
catalytic -1 subsite lying in an undistorted 4C1 conformation
(see Fig. S3, ESI†). For each structure, the isofagomine in the
catalytic -1 subsite makes the same interactions with AaCel9A.
For the structure of AaCel9A with 2, a cellobiose molecule, which
results from the cleavage of 2 by AaCel9A into isofagomine and
cellobiose, binds in the +1 and +2 subsites. For the structure of
AaCel9A with 3, compound 1, which results from the cleavage of
3, is bound in the +1/+2 subsites but is bound, quite distinctly, in
the opposite orientation to that observed for the parent compound
1 structure. This observation can be potentially rationalised by
the cleavage of 3 between the second and third glucose residues,
which liberates cellobiose and 1. Release of the cellobiose moiety
and binding of a second molecule of 3 is then able to occur in
the -4 to -1 subsites. This observation is consistent with what
has been reported for the catalytic activity and preferred binding
mode of substrates for this enzyme.27 Overall the structural results
observed here give credence to previous reports that the occupancy
of the -1/-2 binding sites are crucial to the catalytic activity of
the enzyme.22
Scheme 1 a) i. PhCH(OMe)2, CSA, CHCl3 ii. Ac2O, pyridine iii. AcOH,
H2O; b) AcCl, pyridine, CH2Cl2; c) TMSOTf, CH2Cl2; d) NaOH, CH3OH,
H2O.
of these compounds.23 Removal of the protecting groups gave the
desired materials 1–323 in good yields.
Compounds 1–4 were determined to be competitive inhibitors
of AaCel9A with Ki values, determined at the pH optimum for
catalysis (pH = 5.5), of 610 14 nM (1), 540 11 nM (2), 940
12 nM (3) and 87 4 mM (4) (see ESI†). These data reveal some
interesting changes in Ki values across the series. The monosaccha-
ride analogue, isofagomine 4, is a modest inhibitor of the enzyme
but the addition of a single b-glucosyl moiety residue to the 4-
position (yielding 1) improves binding dramatically. However, each
successive addition of b-glucosyl residues provides little binding
enhancement. These results though are in line with what has been
observed previously for this enzyme with it being demonstrated
that the -1 25 and -2 subsites22 are the strongest binding subsites
and only weak interactions are observed for glucosyl moieties
found in the -3 and -4 subsites. These observations suggest that
the enzyme has, primarily, cellobiohydrolase activity. Notably, the
Ki value obtained for 1 is consistent with that obtained for other
cellulases, from GH520 and GH6.19
The ring nitrogen of the isofagomine moiety in 1, 2 and 3
interacts with the side chain of Asp143 and Asp146 (Fig. 2),
consistent with electrostatic interactions of a protonated amine.
It overlaps with the O1 atom of the alpha-glucose (-1) observed
in the previously reported enzyme-product structures (PDB code
3H2W and 3H3K).22 Furthermore, the catalytic water molecule
which is very well defined in the unliganded structure of AaCel9A
(PDB code 3GZK22) is not found in the structures of AaCel9A
with 1–3 or a-glucose. Both the nitrogen atom and the O1 of
To gain a more detailed understanding of the molecular basis
for the binding of the compounds to AaCel9A, we determined
the three-dimensional structure of AaCel9A in complex with 1, 2
and 3 (Fig. 2 and Fig. S1 and S2, ESI†) by soaking experiments
˚
˚
˚
at 2.56 A, 1.99 A and 2.02 A resolution respectively (see Table,
ESI†). Following maximum-likelihood refinement, Fobs - Fcalc
“difference” electron density unambiguously reveals the position
of all carbohydrate associated hydrogen atoms for each structure.
Compound 1 binds in the -2/-1 subsites, and in the +1, +2
subsites with the isofagomine moiety in the catalytic -1 and +1
subsites. This, to our knowledge, is the first reported structure of
isofagomine bound to the +1 subsite of a cellulase and it is of note.
Both the isofagomine and glucose residue in the aglycone subsites
bind in a somewhat ‘twisted-chair’ conformation with the twisted
˚
the a-glucose are located only 1 A away from the position of the
catalytic water molecule in the unliganded structure and so overall
this occupation of the binding space by these atoms presumably
prevents binding of the catalytic water molecule. Interestingly, for
all the compounds tested, distortion of the isofagomine moiety
toward an oxocarbenium-like half chair is not observed although a
4H3 half chair conformation has been observed for a cellotetraose-
bound active site mutant GH9 cellulase from Clostridium ther-
mocellum (CbhA).28 The half chair conformation is thought to
facilitate oxocarbenium ion formation and subsequent substrate
cleavage. Both the position and 4C1 conformation of isofagomine
in AaCel9A resemble a product intermediate so far observed only
for retaining glycosidases.20,21,29
In conclusion, AaCel9A is another example of an invert-
ing enzyme obtained in complex with inhibitors-derived from
isofagomine. In contrast to the binding mode with inverting GH6
enzymes19,26 these inhibitors are not tight-binding inhibitors of
inverting GH9 enzymes by virtue of their bound conformation
Fig. 2 (A) Fobs - Fcalc electron density map contoured at 3s for 1 bound
in the active site cleft of AaCel9A, with the important active site residues
displayed. (B) Interactions between AaCel9a and the two molecules of 1.
Only the -2, -1 and +1 subsites are shown for simplicity. The catalytic
water molecule is not found in the structure of 1 with AaCel9A due to the
interaction of the ring nitrogen of isofagomine with Asp146.
5946 | Org. Biomol. Chem., 2011, 9, 5945–5947
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
in the active site, which does not reflect any potential transition-
state shape or “transition-state mimic” (as seen in GH6). They
potentially only rely on the putative transition state-mimicking
positive charge on the ring nitrogen. These subtle differences in
binding observed for these inhibitors will be critical for the future
development of selective inhibitors of these enzymes. Overall,
the structural details observed here for AaCel9A with 1–3 can
potentially be extrapolated to all family GH9 members.
13 E. A. Bayer, Y. Shoham and R. Lamed, in In The Prokaryotes: An
Involving Electronic Resource for the Microbial Community, ed. M.
Dworkin, S. Falkow, E. Rosenberg, K.-H. Schleifer & E. Stackebrandt,
Springer Verlag, New York, NY, 2000.
14 G. Parsiegla, M. Juy, C. Reverbel-Leroy, C. Tardif, J. P. Bela¨ıch, H.
Driguez and R. Haser, EMBO J., 1998, 17, 5551.
15 J. Zou, G. J. Kleywegt, J. Sta˚hlberg, H. Driguez, W. Nerinckx, M.
Claeyssens, A. Koivula, T. T. Teeri and T. A. Jones, Structure, 1999, 7,
1035.
16 A. Varrot, T. Frandsen, H. Driguez and G. J. Davies, Acta Crystallogr.,
Sect. D: Biol. Crystallogr., 2002, 58, 2201.
SM and AV are grateful to Beatriz Guimarez for help in data
collection on PROXIMA I at SOLEIL and KAS thanks the
Australian Research Council for funding.
17 A. Varrot, M. Schu¨lein, S. Fruchard, H. Driguez and G. J. Davies, Acta
Crystallogr., Sect. D: Biol. Crystallogr., 2001, 57, 1739.
18 G. Parsiegla, C Reverbel, C. Tardif, H. Driguez and R. Haser, J. Mol.
Biol., 2008, 375, 499.
19 A. Varrot, J. Macdonald, R. V. Stick, G. Pell, H. J. Gilbert and G. J.
Davies, Chem. Commun., 2003, 946.
20 A. Varrot, C. A. Tarling, J. M. Macdonald, R. V. Stick, D. L. Zechel,
S. G. Withers and G. J. Davies, J. Am. Chem. Soc., 2003, 125, 7496.
21 M. E. C. Caines, S. M. Hancock, C. A. Tarling, T. M. Wrodnigg, R.
V. Stick, A. E. Stutz, A. Vasella, S. G. Withers and N. C. J. Strynadka,
Angew. Chem., Int. Ed., 2007, 46, 4474.
22 K. Eckert, A. Vigouroux, L. Leggio and S. More´ra, J. Mol. Biol., 2009,
394, 61.
23 J. M. Macdonald, R. V. Stick, D. M. G. Tilbrook and S. G. Withers,
Aust. J. Chem., 2002, 55, 747.
24 E. D. Goddard-Borger and R. V. Stick, Aust. J. Chem., 2007, 60, 211.
25 G. J. Davies, K. S. Wilson and B. Henrissat, Biochem. J., 1997, 321, 557.
26 A. Varrot, S. Leydier, G. Pell, J. M. Macdonald, R. V. Stick, B.
Henrissat, H. J. Gilbert and G. J. Davies, J. Biol. Chem., 2005, 280,
20181.
Notes and references
1 M. E. Himmel and E. A. Bayer, Curr. Opin. Biotechnol., 2009, 20, 316.
2 H. J. Gilbert, Plant Physiol., 2010, 153, 444.
3 B. Henrissat and A. Bairoch, Biochem. J., 1993, 293, 781.
4 B. Henrissat and A. Bairoch, Biochem. J., 1996, 316, 695.
5 B. Henrissat and G. Davies, Curr. Opin. Struct. Biol., 1997, 7, 637.
6 B. L. Cantarel, P. M. Coutinho, C. Rancurel, T. Bernard, V. Lombard
and B. Henrissat, Nucleic Acids Res., 2009, 37, D233.
7 G. Parsiegla, A. Belaich, J. P. Belaich and R. Haser, Biochemistry, 2002,
41, 11134.
8 J. Sakon, D. Irwin, D. B. Wilson and P. A. Karplus, Nat. Struct. Biol.,
1997, 4, 810.
9 D. Mandelman, A. Belaich, J. P. Belaich, N. Aghajari, H. Driguez and
R. Haser, J. Bacteriol., 2003, 185, 4127.
10 M. Juy, A. G. Amit, P. M. Alzari, R. J. Poljak, M. Claeyssens, P. Beguin
and J. P. Aubert, Nature, 1992, 357, 89.
27 K. Eckert, F. Zielinski, L. L. Leggio and E. Schneider, Appl. Microbiol.
Biotechnol., 2002, 60, 428.
11 F. D. Schubot, I. A. Kataeva, J. Chang, A. K. Shah, L. G. Ljungdahl,
J. P. Rose and B. C. Wang, Biochemistry, 2004, 43, 1163.
12 S. Khademi, L. A. Guarino, H. Watanabe, G. Tokuda and E. F. Meyer,
Acta Crystallogr., Sect. D: Biol. Crystallogr., 2002, 58, 653.
28 F. D. Schubot, I. A. Kataeva, J. Chang, A. K. Shah, L. G. Ljungdahl,
J. P. Rose and B. Wang, Biochemistry, 2004, 43, 1163.
29 J. Sakon, D. Irwin, D. B. Wilson and P. A. Karplus, Nat. Struct. Biol.,
1997, 4, 810.
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 5945–5947 | 5947
©