Small-molecule glucosylation by sucrose phosphorylase: Structure-activity relationships for acceptor substrates revisited

Article abstract of DOI:10.1016/j.carres.2010.03.035

Sucrose phosphorylase catalyzes the O-glucosylation of awide range of acceptor substrates. Acceptors presenting a suitable 1,2-diol moiety are glucosylated exclusively at the secondary hydroxyl. Production of the naturally occurring compatible solute, 2-O-α-D-glucopyranosyl-sn-glycerol, from sucrose and glycerol is a notable industrial realization of the regio-and stereoselective biotransformation promoted by sucrose phosphorylase. The acceptor substrate specificity of sucrose phosphorylase was analyzed on the basis of recent high-resolution crystal structures of the enzyme. Interactions at the acceptor binding site, observed in the crystal (D-fructosyl) and suggested by results of docking experiments (glycerol), are used to rationalize experimentally determined efficiencies and regioselectivities of enzymatic glucosyl transfer.

Full text of DOI:10.1016/j.carres.2010.03.035

Carbohydrate Research 345 (2010) 1492–1496
Contents lists available at ScienceDirect
Carbohydrate Research
journal homepage: www.elsevier.com/locate/carres
Note
Small-molecule glucosylation by sucrose phosphorylase: structure–activity
relationships for acceptor substrates revisited
*
Christiane Luley-Goedl, Bernd Nidetzky
Institute of Biotechnology and Biochemical Engineering, Graz University of Technology, Petersgasse 12/1, 8010 Graz, Austria
a r t i c l e i n f o
a b s t r a c t
Article history:
Sucrose phosphorylase catalyzes the O-glucosylation of a wide range of acceptor substrates. Acceptors pre-
senting a suitable 1,2-diol moiety are glucosylated exclusively at the secondary hydroxyl. Production of the
naturally occurring compatible solute, 2-O-a-D-glucopyranosyl-sn-glycerol, from sucrose and glycerol is a
notable industrial realization of the regio- and stereoselective biotransformation promoted by sucrose
phosphorylase. The acceptor substrate specificity of sucrose phosphorylase was analyzed on the basis of
recent high-resolution crystal structures of the enzyme. Interactions at the acceptor binding site, observed
Received 15 January 2010
Received in revised form 25 March 2010
Accepted 28 March 2010
Available online 2 April 2010
Keywords:
Sucrose phosphorylase
Transglucosylation
Recognition of polyhydroxylated acceptors
Substrate-assisted catalysis
Reactive diol
in the crystal (D-fructosyl) and suggested by results of docking experiments (glycerol), are used to rational-
ize experimentally determined efficiencies and regioselectivities of enzymatic glucosyl transfer.
Ó 2010 Elsevier Ltd. All rights reserved.
Stacking interaction
Starting from the early 1940s, when the enzyme was first discov-
ered, a number of studies have demonstrated the usefulness of
The specificity of sucrose phosphorylase for reaction with glu-
cosyl acceptors is unusually relaxed. Accommodation of phosphate
sucrose phosphorylase in the synthesis of
sides.1–3 The natural reaction of the enzyme is conversion of sucrose
and phosphate into -glucose 1-phosphate and -fructose. The
catalytic mechanism of sucrose phosphorylase is that of a double
displacement-like reaction involving a b-glucosyl enzyme interme-
diate. In the absence of phosphate, alternative glucosyl acceptors
(e.g., monosaccharides, sugar alcohols) can take part in the reaction
by intercepting the glucosylated enzyme. This enzymatic transglu-
a
-configured O-gluco-
and D-fructose as acceptor substrate for the forward and reverse
directions of the natural reaction, respectively, is already remark-
able. However, aside from reaction with water (hydrolysis), the en-
zyme additionally promotes glucosyl transfer to hydroxy groups in
a diversity of compounds, including polyhydroxylated molecules
as well as aromatic alcohols.8 Under conditions in which a suitable
1,2-diol acceptor group is available for enzymatic reaction (e.g.,
glycerol, R-glycerate), the secondary hydroxyl is glucosylated with
essentially complete regioselectivity. However, polyhydroxylated
acceptors differ strongly in their reactivity toward glucosylated su-
crose phosphorylase. A molecular basis underpinning the observed
relationships between acceptor structure and reactivity is cur-
rently not available, and the high regioselectivity of sucrose phos-
phorylase is not well understood. Based on recent high-resolution
crystal structures of sucrose phosphorylase from Bifidobacterium
adolescentis,9 we compare here (bio)chemical evidence for the
enzymatic glucosyl transfer to structural properties of the acceptor
binding site of sucrose phosphorylase. Results are used to propose
a-
D
D
cosylation therefore leads to the formation of new a-D-glucosides.
Scheme 1 shows the kinetic mechanism of sucrose phosphorylase
for glucosyl transfer from sucrose to the reactive alcohol group of a
suitable acceptor molecule. Application of sucrose phosphorylase
as a biocatalyst for small-molecule glucosylation has recently at-
tracted renewed attention as it became clear that two naturally
occurring compatible solutes, 2-O-
a-D-glucopyranosyl-sn-glycerol
(GG)and 2-O- -glucopyranosyl-R-glycerate(GGA), canbesynthe-
a
-D
sized very efficiently by the enzyme (from Leuconostoc mesentero-
ides).4,5 Production of GG has been developed into a commer-
cialized process on industrial scale, and a solution containing 50%
GG has been introduced to the market as an active ingredientfor cos-
metics where it is available under the trade name Glycoin Extremi-
um.6 Both GG and GGA could find further application as protein
stabilizers.7
a
model of how the enzyme recognizes different acceptor
substrates.
Figure 1A is a close-up view of the substrate binding site in su-
crose phosphorylase, revealed by a crystal structure of the B. adole-
scentis enzyme bound with sucrose (PDB-entry 2gdu). The catalytic
machinery of this enzyme consists of Asp192 and Glu232. The Asp
functions as a nucleophile during formation of the covalent en-
zyme intermediate10 while the Glu serves a role as Brønsted acid
* Corresponding author. Tel.: +43 316 873 8400; fax: +43 316 873 8434.
E-mail address: bernd.nidetzky@tugraz.at (B. Nidetzky).
0008-6215/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.carres.2010.03.035

C. Luley-Goedl, B. Nidetzky / Carbohydrate Research 345 (2010) 1492–1496
1493
HO
HO
HO
O
Enzyme
Enzyme
+
R-OH
Enzyme
HO
D-Fructose
O
R
HO
HO
HO
Enzyme
kacceptor
kwater
O
HO
HO
HO
+
O
HO
HO
HO
HO
HO
O
OH
H2O
HO
+
O
O
OH
HO
OH
HO
Scheme 1. Kinetic mechanism of sucrose phosphorylase (enzyme) for glucosyl transfer from sucrose to various acceptors. R-OH, monosaccharide, sugar alcohol, acceptor
with phenolic, benzylic or other hydroxy groups.
and base during glucosylation and deglucosylation of the en-
zyme,11 respectively. Substrate turnover in the crystal was pre-
vented by substitution of Glu232 by Gln, which is not competent
as a catalytic Brønsted acid.
Additionally, there are stacking interactions of the fructofuranosyl
ring with Phe156 that could provide further stabilization of the
leaving group/acceptor molecule in a reactive orientation. We
now asked the question: to what extent are noncovalent en-
The mode of binding of the fructosyl moiety in Figure 1A sug-
gests three types of enzyme–ligand interactions, which are pre-
sumably decisive for discrimination between different leaving
zyme–substrate interactions utilized by D-fructosyl (or D-fructose)
exploitable by other glucosyl acceptor substrates of the enzyme?
Figure 1B shows results of an energy-minimized molecular dock-
ing experiment from which interactions of a glycerol acceptor with
the b-glucosyl enzyme intermediate of B. adolescentis sucrose phos-
phorylase are predicted. The crystal structure used for ligand dock-
ing is that of the wild-type enzyme (PDB-entry 2gdv) having a b-
glucosyl residue linked to Asp192. Note the skew boat conformation
of the covalently bound sugar. Glycerol is positioned such that the
C2–OH is ready for undergoing reaction (distance 3.06 Å) with the
anomeric carbon of the b-glucosyl residue. Glu232 has a bidentate
hydrogen bonding interaction with C1–OH and C2–OH of glycerol.
With a distance of 2.89 Å from one of its oxygens to the oxygen of
the glycerol C2–OH, the (ionized) side chain of Glu232 appears to
be placed suitably for providing general base catalysis to the nucle-
ophilic attack of the acceptor substrate. The high preference of su-
crose phosphorylase for glucosylation of the secondary hydroxyl as
compared to the sterically less hindered, hence chemically more ac-
tive primary hydroxyl of glycerol is thus plausibly explained by the
model in Figure 1B. The data also confirm the notion from above that
the role of the C1–OH is not just in acceptor substrate recognition. By
groups/acceptor substrates. The C1–OH of D-fructosyl appears to
assist in bringing the side chain of the native Glu232 into the
appropriate place for acid/base catalysis to occur. The C1–OH is
therefore expected to provide a major contribution to precise rela-
tive positioning of the reactive group of the substrate and the cat-
alytic Glu232 of the enzyme. The distances of the glycosidic oxygen
and the fructosyl C1–OH to the N/O atoms of the carboxamide of
Gln in the crystal structure are 3.15 and 2.59 Å, respectively.
Although the C1–OH does not directly participate in the catalytic
event, we propose that its role in the enzymatic reaction can be de-
scribed as ‘substrate-assisted facilitation’.
Figure 1A shows that there are further noncovalent enzyme–
substrate interactions that position the D-fructosyl moiety at the
binding site for the leaving group/acceptor in sucrose phosphory-
lase. The C6–OH is accommodated (‘sandwiched’) between the side
chains of Gln345 and Arg399, and the hydrogen bond between the
substrate hydroxyl and carboxamide side chain of the Gln is ex-
pected to stabilize the bound leaving group/acceptor substrate.
A
B
Arg399
Arg399
3.74
4.71
Asp290
Asp290
2.85
Phe156
Phe156
3.23
4.08
3.32
Gln345
3.06
Gln345
2.89
3.15
Asp192
Asp192
3.22
2.59
Glu232
Tyr196
Tyr196
Gln(Glu)232
Figure 1. (A) Close-up view of the substrate binding site in B. adolescentis sucrose phosphorylase (PDB-entry 2gdu, molecule A). Residues interacting with the fructosyl
moiety of sucrose (orange) are displayed. (B) Interactions of glycerol acceptor with the b-glucosyl enzyme intermediate of B. adolescentis sucrose phosphorylase (PDB-entry
2gdv, molecule A) derived from an energy-minimized flexible ligand-protein (cyan) docking experiment. Distances are in Å. (For interpretation of the references to color in
this figure legend, the reader is referred to the web version of this article.)

1494
C. Luley-Goedl, B. Nidetzky / Carbohydrate Research 345 (2010) 1492–1496
Table 1
pulling the side chain of Glu232 into the requisite position for ‘acti-
vation’ of the C2–OH, the primary hydroxyl could be directly auxil-
iary to the catalytic event. The model of bound glycerol shows
furthermore that in comparison with the observed binding mode
for b-fructosyl in Figure 1A, stacking interactions with Phe156 are
not present for obvious reasons. However, the hydrogen bond be-
tween the leaving group/acceptor substrate and Gln345 is retained
although it is clearly weaker for glycerol than the b-fructosyl moiety.
By way of comparison, Arg399 could play a role in accommodating
the carboxylate group in a glycerate acceptor substrate.4,5
Basedon Figure1 we canrationalizerecognitionofdifferentpoly-
hydroxylated acceptor substrates by sucrose phosphorylase and de-
velop structure–activity relationships for these molecules. The
compounds shown in Figure 2 were used in the analysis because
not only has their reactivity toward enzymatic glucosylation been
determined (Table 1) but also the site of chemical transformation
has been characterized. We use biochemical data for sucrose phos-
phorylase from L. mesenteroides. Note therefore that the residues
contributed to the substrate binding site in the B. adolescentis en-
zyme are fully conserved in the amino acid sequence of the L. mesen-
teroides sucrose phosphorylase. We examine the possible role in
enzyme–substrate recognition of the overall conformation of the
acceptor (open-chain vs pyranose or furanose ring) as well as the ori-
entation of individual hydroxy groups in it. Energy-minimized con-
formationsoftheacceptor substrates in solutionwere usedand from
the known binding mode of b-fructosyl (Fig. 1A), their probable ori-
entation in the substrate binding pocket of sucrose phosphorylase
could be inferred, as depicted in Figure 3.
Comparison of polyhydroxylated acceptors based on their reactivity to become
glucosylated by L. mesenteroides sucrose phosphorylase
Acceptor
Analytical synthetic
yielda (%)
Kinetic partition
coefficient (MÀ1
)
23.4
6.8
36.4
n.d.
n.d.
n.d.
n.d
27.8b
8.3b
D-Fructose
L-Sorbose
12.4
0
L-Arabinose
D
-Arabinose
-Glucose
2.4
D
24.3
25.0
L
-Arabinitol
-Arabinitol
Xylitol
-Glucitol
Glycerol
D
5.1
2.8
n.d.
0.2b
D
1.4
2.6
Analytical synthetic yields are taken from Ref. 14.
n.d., not determined.
a
The analytical synthetic yield is affected not only by the (intrinsic) reactivity of
the acceptor in the enzymatic transformation but also by prevalence of the reactive
conformation of the acceptor molecule in solution, especially in the case of
D-fructose.
b
Data are taken from Ref. 15.
D-Glucose is an interesting acceptor substrate of sucrose phos-
phorylase (from L. mesenteroides) because it is glucosylated at mul-
tiple sites, leading to the formation of two major transfer products,
maltose (4-O-a-D-glucopyranosyl glucose) and kojibiose (2-O-a-D-
glucopyranosyl glucose), in about equimolar amounts (Luley-Goedl
et al., unpublished results). Other authors have also reported nige-
The observation that
acceptor compared to -sorbose (Table 1) is explained by the pre-
dicted loss of interaction between C6–OH and Gln345 in bound
-sorbose. Despite its pyranose conformation, -arabinose is a better
acceptor than -sorbose. Ability of the C2–OH of -arabinose to partly
reinstall the hydrogen bond with Gln345 that was not available to
D-fructose behaves as a superior glucosyl
rose (3-O-
a
-
D
-glucopyranosyl glucose) production in significant
amounts.13 No conformation of
D-glucose appears to fulfill the
L
requirement of having a suitable diol moiety available for bonding
with Glu232. However, a plausible explanation is that recognition
L
L
L
L
of D-glucose does not involve neighboring hydroxyls and exploits
both the 4C1 and the energetically less favored 1C4 ring conforma-
tion in distinct orientations, as shown in Figure 3. According to the
L
-
sorbose may be responsible. An alternative possibility is that due to
its tilted pyranose ring conformation relative to the furanose ring
proposed modes of
C4–OH is rationalized. With
tion, the -anomeric sugar hydroxyl points toward the side chain
D
-glucose binding, glucosylation of C2–OH and
D
-glucose bound in the 4C1 conforma-
conformations of
ploits better than
D
-fructose and
L-sorbose (Fig. 3), L-arabinose ex-
L-sorbose the stacking interactions with Phe156.
a
D
-Arabinose is a very poor acceptor substrate for sucrose phosphor-
of Phe156, arguably generating steric conflict.
ylase, and only the b-anomeric conformation is reactive.12 In an ori-
A number of studies report on glucosylation of sugar alcohols by
sucrose phosphorylase.12,14,15 There is clear evidence supporting the
notion that glucosyl transfer is regioselective and orientation of the
individual hydroxyls affects the acceptor reactivity. Analytical syn-
entation that places the reactive diol moiety into a suitable position
(data not shown), there is no hydroxy group of
for bonding with Gln345.
D-arabinose available
CH2OH
CH2OH
OH
HO H2C
HO H2C
OH
O
OH
OH
O
O
O
HO
O
OH
HO
OH
CH2OH
HO
OH HO
OH
OH
OH
OH
OH
HO
OH
β-
D-Fructose
L-Arabinose
α-D-Glucose
α-
L-Sorbose
β-
D-Arabinose
CH2OH
CH2OH
OH
CH2OH
OH
CH2OH
OH
CH2OH
OH
HO
OH
OH
HO
HO
HO
HO
CH2OH
OH
OH
OH
CH2OH
CH2OH
CH2OH
CH2OH
L
-Arabinitol
Xylitol
Glycerol
D
-Arabinitol
D-Glucitol
Figure 2. Chemical structures of different polyhydroxylated acceptors for glucosyl transfer by sucrose phosphorylase. The site(s) of glucosylation, identified by X-ray
diffraction and NMR analysis,8 are indicated by an arrow.

C. Luley-Goedl, B. Nidetzky / Carbohydrate Research 345 (2010) 1492–1496
1495
C3
C2
C2
C4
D-Fructose
L-Sorbose
L-Arabinose
C2
C2 or C4
C2
4C1 Glucose
1C4 Glucose
Xylitol
L-Arabinitol
Figure 3. Proposed mode of recognition of polyhydroxylated acceptors by sucrose phosphorylase compared to the experimentally determined (PDB-entry 2gdu, molecule A)
orientation of the -fructosyl moiety in the acceptor/leaving group binding site of the enzyme (from B. adolescentis).
D
thetic yields are useful, however not always directly revealing (Table
1). Kinetic partition coefficients, obtained in assays in which break-
down of the glucosylated enzyme via reaction with acceptor and
reaction with water (hydrolysis) is compared, are direct measures
ofreactivity. Accordingto Table1, theglucosyltransferefficiencyde-
and between C –Cd in Glu232. Glycerol was set flexible with two
c
torsions (C1–C2, C2–C3). Three clusters were obtained using a ge-
netic algorithm with 10 runs and a root-mean-square-deviation
tolerance of 2.0 Å for clustering. The interactions of the enzyme
with bound glycerol in the obtained models were compared with
those of the bound fructosyl moiety in an experimental structure
of sucrose phosphorylase in complex with sucrose (PDB-entry
2gdu; molecule A). The model giving the best correspondence to
the experimental data was chosen for structural interpretation.
creases in the order
itol. Using data from synthesis,
fivefold over xylitol. The product of enzymatic glucosylation was
determined as 2-O-
-glucopyranosyl xylitol.14 However, the
D-fructose >
L-arabinitol >
D-arabinitol > D-gluc-
L
-arabinitol is preferred about
a-D
absolute configuration of the compound, in which two diastereo-
mers (3R and 3S) are possible, was not determined. Figure 3 depicts
how the energy-minimized solution structure of xylitol might be
accommodated in the acceptor binding site of sucrose phosphory-
lase. In the shown orientation, xylitol directs its C4–OH toward
Phe156andwouldthereforebeunabletoformahydrogenbondwith
Gln345. In the case of L-arabinitol for which unfortunately the site of
glucosylation is not known, we would predict a hydrogen bond be-
tween C4–OH and Gln345, implying reaction at C2–OH. Further-
1.2. Transglucosylation—determination of kinetic partition
coefficients
Purified sucrose phosphorylase from L. mesenteroides was pre-
pared by reportedmethods.10 The reactionmixture for kinetic studies
of transglucosylation contained 0.1 M a-D-glucose 1-phosphate, 0–
0.2 M acceptor, and 3 U mlÀ1 (determined as described in Ref. 15) of
sucrose phosphorylase dissolved in 20 mM MES buffer, pH 7.0. It
was incubated at 30 °C using an agitation rate of 550 rpm (Thermom-
ixer comfort; Eppendorf). Reactions were stopped after 10 min by
heating (99 °C, 5 min), and sample work-up included centrifugation
at 10,000 rpm for 10 min to remove precipitated protein. The amount
of released phosphate was determined colorimetrically at 850 nm.17
Glucose was measured using a coupled enzymatic assay in which
hexokinase and glucose 6-phosphate dehydrogenase were em-
ployed.18 Dataanalysis wasperformed accordingtoScheme 1. Kinetic
partition coefficients (kacceptor/kwater) were determined from plots of
more, glucosyl transfer to C4–OH is predicted for
whereby the C2–OH would interact with Gln345. The relatively poor
reactivity of -glucitol could be explained by structural similarity to
D-arabinitol
D
xylitol, that is, the absence of bonding with Gln345 due to ‘wrong’
orientation of C4–OH.
In conclusion, this communication brings structural informa-
tion on sucrose phosphorylase and biochemical data on the accep-
tor specificity of the enzyme into a coherent whole. We hope that it
will be considered useful in the selection and perhaps develop-
ment of novel acceptor substrates to become glucosylated by the
phosphorylase. The proposed model of enzyme–substrate recogni-
tion is expected to support molecular design approaches in which
phosphorylase mutants having altered acceptor specificities will be
generated.
the ratio of phosphate and glucose release rates (vphosphate/vglucose
)
versus the acceptor concentration using following equation: vphos-
phate/vglucose = 1 + (kacceptor/kwater) Â [acceptor]. Note: vphosphate is the
total reaction rate whereas vglucose measures the hydrolysis rate.
Acknowledgment
1. Experimental
Financial support from the Austrian Science Fund FWF (Project
L586-B03) is gratefully acknowledged.
1.1. Energy-minimized molecular docking
Spartan’02 (http://wavefun.com), AutoDockToolsv4,16 and Py-
MOL (http://pymol.sourceforge.net) were used for the determina-
tion of energy-minimized solution structures, enzyme–ligand
docking, and visualization, respectively. The X-ray crystal structure
of B. adolescentis sucrose phosphorylase having a b-glucosyl resi-
due covalently bound to Asp192 (PDB-entry: 2gdv, molecule A)
was used as macromolecule in a molecular docking experiment
that employed glycerol as the ligand. One twistable bond was al-
References
1. Mieyal, J. J.; Abeles, R. H. Disaccharides Phosphorylases. In The Enzymes, 3rd ed.;
Boyer, P. D. (Ed.); Academic Press: New York, 1972; Vol. 7, pp 515–532.
2. Vandamme, E. J.; van Loo, J.; Machtelinckx, L.; de Laporte, A. Adv. Appl.
Microbiol. 1987, 32, 163–201.
3. Silverstein, R.; Voet, J.; Reed, D.; Abeles, R. H. J. Biol. Chem. 1967, 242, 1338–
1346.
4. Goedl, C.; Sawangwan, T.; Mueller, M.; Schwarz, A.; Nidetzky, B. Angew. Chem.,
Int. Ed. 2008, 47, 10086–10089.
5. Sawangwan, T.; Goedl, C.; Nidetzky, B. Org. Biomol. Chem. 2009, 7, 4267–4270.
lowed in the side chain between Cb–C in Asp290 and Phe156
c

1496
C. Luley-Goedl, B. Nidetzky / Carbohydrate Research 345 (2010) 1492–1496
6. Jan Dekker International. The Spirit of Youth. In Official Show Preview of the In-
Cosmetics; Munich, Germany, 2009.
13. Kitao, S.; Yoshida, S.; Horiuchi, S.; Sekine, H.; Kusakabe, I. Biosci., Biotechnol.,
Biochem. 1994, 58, 790–791.
7. Sawangwan, T.; Goedl, C.; Nidetzky, B. Biotechnol. J. 2010, 5, 187–191.
8. Goedl, C.; Sawangwan, T.; Wildberger, P.; Nidetzky, B. Biocatal. Biotransform.
2009, 28, 10–21.
9. Mirza, O.; Skov, L. K.; Sprogoe, D.; van den Broek, L. A.; Beldman, G.; Kastrup, J.
S.; Gajhede, M. J. Biol. Chem. 2006, 281, 35576–35584.
14. Kitao, S.; Sekine, H. Biosci., Biotechnol., Biochem. 1992, 56, 2011–2014.
15. Goedl, C.; Schwarz, A.; Minani, A.; Nidetzky, B. J. Biotechnol. 2007, 129, 77–86.
16. Goodsell, D. S.; Morris, G. M.; Olson, A. J. J. Mol. Recognit. 1996, 9, 1–5.
17. Saheki, S.; Takeda, A.; Shimazu, T. Anal. Biochem. 1985, 148, 277–281.
18. Kunst, A.; Draeger, B.; Ziegenhorn, J. D-Glucose: UV-Methods with Hexokinase
10. Schwarz, A.; Nidetzky, B. FEBS Lett. 2006, 580, 3905–3910.
11. Schwarz, A.; Brecker, L.; Nidetzky, B. Biochem. J. 2007, 403, 441–449.
12. van den Broek, L. A.; van Boxtel, E. L.; Kievit, R. P.; Verhoef, R.; Beldman, G.;
Voragen, A. G. Appl. Microbiol. Biotechnol. 2004, 65, 219–227.
and Glucose-6-phosphate Dehydrogenase. In Methods of Enzymatic Analysis,
3rd ed.; Bergmeyer, H.U. (Ed.); Verlag Chemie: Weinheim, 1984; Vol. 6; pp
163–172.