Enzymatic liberation of lycotetraose from the Solanum glycoalkaloid α-tomatine

Article abstract of DOI:10.1016/j.carres.2004.07.004

The branched tetrasaccharide, O-β-D-glucopyranosyl-(1→2)-O- [β-D-xylopyranosyl-(1→3)]-O-β-D-glucopyranosyl-(1→4) -D-galactose (lycotetraose) is a key constituent of many biologically interesting natural products. Described herein is a convenient enzymatic preparation of lycotetraose from the readily available Solanum glycoalkaloid α-tomatine. The preparation makes use of the recombinant endo-glycosidase, tomatinase, from the plant pathogen Fusarium oxysporum f. sp. lycopersici.

Full text of DOI:10.1016/j.carres.2004.07.004

Carbohydrate
RESEARCH
Carbohydrate Research 339 (2004) 2325–2328
Note
Enzymatic liberation of lycotetraose from the Solanum glycoalkaloid
a-tomatine
a
Katherine Woods,a Chris J. Hamiltona, and Robert A. Field
*
aCentre for Carbohydrate Chemistry, School of Chemical Sciences and Pharmacy, University of East Anglia, Norwich NR4 7TJ, UK
Received 7 May 2004; received in revised form 15 June 2004; accepted 11 July 2004
Available online 13 August 2004
Abstract—The branched tetrasaccharide, O-b-D-glucopyranosyl-(1 ! 2)-O-[b-D-xylopyranosyl-(1 ! 3)]-O-b-D-glucopyranosyl-
(1 ! 4)-D-galactose (lycotetraose) is a key constituent of many biologically interesting natural products. Described herein is a con-
venient enzymatic preparation of lycotetraose from the readily available Solanum glycoalkaloid a-tomatine. The preparation makes
use of the recombinant endo-glycosidase, tomatinase, from the plant pathogen Fusarium oxysporum f. sp. lycopersici.
ꢀ 2004 Elsevier Ltd. All rights reserved.
Keywords: Lycotetraose; Tomatinase; endo-Glycosidase; Saponin; Enzymatic cleavage
1. Introduction
desgalactotigonin16 have previously been reported, but
the laborious nature of these preparations makes them
impractical for general use. The lycotetraose-based sapo-
nin a-tomatine plays an important role in the tomato
plantÕs defence against fungal pathogens, but several
pathogenic fungi produce specific glycosidases (tomatin-
ases) that inactivate it by cleavage of one or more sugars
from the lycotetraose side-chain.18 The tomatinase from
F. oxysporum f. sp. lycopersici is a Family 10 endo-glycos-
idase that selectively cleaves the b-linkage between the
galactose and tomatidine moiety of a-tomatine, to liber-
ate the fully intact tetrasaccharide from the sterol moiety
(tomatidine).19 The lengthy chemical synthesis of lyco-
tetraose15 prompted us to investigate its enzymatic prepa-
ration from commercially available a-tomatine, using the
tomatinase from F. oxysporum f. sp. lycopersici (Scheme
1). In a similar manner Nohara and co-workers have re-
cently utilized glycyrrhizin hydrolase to access the natural
oligosaccharides fabiotriose and mimosatriose during their
chemoenzymatic syntheses of novel artificial saponins.20
Saponins are a structurally diverse class of steroidal
glycosides found in a wide variety of plants. The general
saponin structure consists of either a steroidal or triterpe-
noid aglycone linked to one or more mono- or oligo-
saccharide glycone motifs. The branched tetrasaccha-
ride, O-b-D-glucopyranosyl-(1 ! 2)-O-[b-D-xylopyrano-
syl-(1 ! 3)]-O-b-D-glucopyranosyl-(1 ! 4)-D-galactose,
known as lycotetraose, is a recurrent glycone motif found
in a variety of saponins structures from a range of plant
species. Many of these lycotetraose-based saponins display
interesting biological properties including antifungal1–4
and anticancer activity3,5–9 as well as inhibition of cAMP
phosphodiesterase,10,11 Na,K-ATPase,12 tumor promoter-
induced lipid metabolism,13 and human spermatozoan
motility.14 The ability to prepare artificial saponins con-
taining the lycotetraose motif is therefore desirable. Lyco-
tetraosewasoriginally prepared ina0.9% isolatedyieldby
the partial hydrolysis of a-tomatine under acidic condi-
tions.17 Chemical syntheses of lycotetraose15 (11-steps in
10% overall yield) and the lycotetraose-derived saponin
2. Results and discussion
2.1. Over-expression of tomatinase
*
Corresponding author at present address: School of Chemistry,
QueenÕs University Belfast, David Keir Building, Stranmillis Rd,
Belfast BT9 5AG, UK. Tel.: +44-028-90974158; e-mail: c.hamilton@
qub.ac.uk
Recombinant tomatinase (14.5mg) was prepared from a
500mL culture of E. coli strain BL21(DE3) transformed
0008-6215/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.carres.2004.07.004

2326
K. Woods et al. / Carbohydrate Research 339 (2004) 2325–2328
tomatidine
Glc(I)
H
N
OH
O
H
Xyl
O
O
OH
O
Gal
O
HO
HO
OH
H
O
O
O
O
HO
HO
H
HO
O
OH
O
O
H
H
O
HO
HO
HO
HO
O
OH
O
O
OH
H
HO
Tomatinase,
pH 5.3, 12 hr
HO
HO
OH
OH
HO
HO
OH
Glc(II)
lycotetraose
(58%)
HO
OH
α−tomatine
Scheme 1.
with the expression vector (pFoTom1) encoding the
deca-histidine-tagged protein.19 As previously described,
sufficient tomatinase was expressed as a result of leaky
protein expression; induction of protein expression by
addition of IPTG has previously been shown to deposit
most of the tomatinase as insoluble inclusion bodies.19
The soluble recombinant protein was easily isolated
from the crude cell lysates by nickel affinity chromato-
graphy. Imidazole was removed from the eluted protein
fractions by extensive dialysis against de-ionized water;
dialysis against the standard assay buffer, 20mM NaO-
Ac pH5, caused the protein to precipitate. The enzyme
was stored as 25% glycerol stocks in un-buffered de-ion-
ized water at À18ꢁC. In our hands, storage of the en-
zyme in the absence of buffer salts had no adverse
effect on enzyme activity for at least three months.
identified by comparison with a-tomatine and the free
sterol. When the TLC-separated reaction mixture was
visualized by charring with an acidic cerium-(IV)-sul-
fate/molybdic acid stain, a-tomatine (Rf 0.18) and the
liberated sterol (Rf 0.84) were clearly visible, whereas ly-
cotetraose was not readily detected. When the TLC
plates were stained and charred using 5% (v/v) of sulf-
uric acid in ethanol, the tetrasaccharide (Rf 0.24) showed
up much more intensely than a-tomatine and the sterol
unit did not show up at all. Therefore, biotransforma-
tion reactions were routinely monitored by running par-
allel TLC plates, which were then visualized under these
two sets of conditions noted.
2.4. Enzymatic synthesis of lycotetraose
Apreparative scale enzymatic synthesis of lycotetraose
from 100mg of a-tomatine was then carried out. By
the time the reaction had run to completion (as judged
by TLC) most of the sterol had precipitated from the
reaction mixture. Isolation of lycotetraose from the
crude reaction mixture relied on manipulation of the dif-
ferent physical properties of the reaction components.
As the sterol was only partially soluble in the reaction
most of it was easily removed by filtration. Unlike lyco-
tetraose, both a-tomatine and tomatidine are soluble in
ether. Therefore, an ether extraction was used to remove
these components from the aqueous mixture. The aque-
ous layer was then passed through a mixed-bed ion
exchange resin to remove buffer salts and residual pro-
tein was finally removed by ultrafiltration to give essen-
tially pure lycotetraose in 58% yield.
2.2. Optimized reaction conditions
It was critical to be able to dissolve a-tomatine in the
reaction buffer at concentrations that would be appro-
priate for preparative scale biotransformations. In aque-
ous solution, a-tomatine is soluble to a maximum
concentration of 10mM up to pH4.5 and is much less
soluble at higher pH, whereas the optimum pH range
for tomatinase activity is pH5.5–7.0.21 The pH condi-
tions were manipulated to improve substrate dissolution
whilst retaining enzyme activity. The best results were
achieved when a solution of a-tomatine (10mM) in
20mM sodium acetate buffer at pH4.5 was diluted five-
fold into 20mM NaOAc at pH5.5 to give a final 2mM
reaction mixture of the enzyme and substrate (corre-
sponding to 2mg/mL of a-tomatine or 1.3mg/mL of
the tetrasaccharide thereof) at pH5.3. This initial
substrate concentration is almost twice the reported
Km value of 1.1mM for a-tomatine with wild type tom-
atinase.21
When purified in this manner, lycotetraose was iso-
lated as an anomeric mixture (a/b = 37:63). With the
1
aid of COSY and H–13C HSQC data, it was possible
to assign all of the anomeric centers and C-4 of galactose
(Table 1). Unlike the other sugar residues, there was no
doubling of the xylose resonances due to the mixed ano-
meric center at the reducing end hence it was possible to
2.3. TLCanalysis
1
assign all of the xylose H and 13C NMR signals. With
During the reaction, enzymatic cleavage of the saponin
was monitored by TLC and reaction products were
the obvious exception of the hemi-acetal center, all the
other assigned carbon NMR signals for lycotetraose

K. Woods et al. / Carbohydrate Research 339 (2004) 2325–2328
Table 1. 1H and 13C NMR assignment of the anomeric mixture of lycotetraose
2327
Position
Gal
1H (ppm)a
J (Hz)a
13C (ppm)a
a-Tomatine 13C (ppm)22
1
4
a (5.09, d)
b (4.43, d)
a (4.02, d)
b (3.96, d)
a (3.9)
b (7.7)
a (3.2)
b (3.0)
a 92.7
b 96.9
a 81.5
b 80.4
102.3
79.8
Glc (I)
Glc(II)
Xyl
1
1
a (4.519, d)
b (4.523, d)
a (7.9)
b (7.7)
a 103.4
b 103.2
104.7
104.9
a (4.87, d)
b (4.86, d)
a (7.9)
b (8.1)
a 102.4
b 102.4
1
(4.57, d)
(3.20, dd)
(3.32)b
(3.49)b
(7.7)
(9.1, 7.7)
103.2
73.8
76.3
74.0
104.8
74.9
78.5
70.6
2
3
Complex
Complex
(11.6, 5.4)
(11.6, 11.1)
4
5a
5b
(3.84, dd)
(3.14, dd)
65.7
67.2
a a/b Denomination refers to the anomeric configuration of the reducing end sugar.
b Multiplicities are unclear due to overlapping with other signals.
are in keeping with those previously assigned to the
same tetrasaccharide side-chain incorporated within
the parent glycoside, a-tomatine.22
4.2. Overexpression and purification of tomatinase
This was carried following a published procedure19 with
some modifications (as described earlier). In this manner
14.5mg of tomatinase was prepared from a 0.5L cell
culture.
3. Conclusions
Astraightforward, one-step preparation of lycotetraose
from the commercially available a-tomatine has been
established using recombinant His10-tagged a-toma-
tinase and simple purification methods. In future, this
should facilitate the preparation and studies of novel
lycotetraose-based saponins by the chemical coupling
of lycotetraose with other aglycone motifs using stand-
ard procedures.20
4.3. TLCanalysis
Biotransformations were monitored by TLC, eluted in a
solvent system consisting of acetic acid, ethyl acetate,
MeOH, and water (10:30:20:1).19,21 The plates were then
dipped in either an aqueous solution of cerium-(IV)-sul-
fate (1%, w/v), molybdic acid (1.5%, w/v), and sulfuric
acid (10%, v/v) or sulfuric acid (5%, v/v) in ethanol.
Compounds were then visualized by charring.
4. Experimental
4.1. General methods
4.4. O-b-D-Glucopyranosyl-(1 ! 2)-O-[b-D-xylopyranos-
yl-(1 ! 3)]-O-b-D-glucopyranosyl-(1 ! 4)-D-galactose
(lycotetraose)
E. coli strain BL21(DE3) containing the recombinant
tomatinase plasmid pFoTom1 was a generous gift from
Dr. M. Ruiz-Rubio, Universidad de Cordoba.19 a-Tom-
atine and tomatidine were purchased from Sigma. TLC
was performed on 250lm pre-coated glass-backed silica
plates (Whatman K6F). Disposable PD-10 desalting col-
umns and FPLC Ni-NTAsuperflow resin were pur-
chased from Amersham Biosciences and Qiagen,
respectively. Ultrafiltrations were carried out using
2mL Vivaspin concentrators with 10,000 MWCO
a-Tomatine (96mg, 0.093mmol) was dissolved in 20mL
of 20mM sodium acetate buffer at pH4.5 by sonication
at room temperature for 10min and was then diluted by
the addition of 80mL of 20mM sodium acetate buffer at
pH5.5. A25% glycerol stock of tomatinase (4.5mg) was
first eluted through a PD-10 desalting column (with buf-
fer at pH5.5) before addition to the reaction mixture.
This was left to stir overnight at 30ꢁC and a white pre-
cipitate had formed. The reaction mixture was filtered
through a plug of cotton wool and the filtrate concen-
trated to 20mL on a rotary evaporator at 40ꢁC.
This was then washed with ether (4 · 20mL) and the
aqueous layer was then passed through a column of
mixed-bed TMD-8 ion exchange resin (20mL) eluting
CTAmembranes (VivaScience AG, Germany).
NMR spectra were recorded on a Varian Unity Inova
1H
spectrometer at 600MHz and 13C NMR spectra were re-
corded on a Varian Unity Plus at 100MHz. H and 13
C
1
NMR spectra were referenced to MeOH (3.4 and
49.3ppm) as an internal standard.

2328
K. Woods et al. / Carbohydrate Research 339 (2004) 2325–2328
3. Sata, N.; Matsunaga, S.; Fusetani, N.; Nishikawa, H.;
Takamura, S.; Saito, T. Biosci. Biotechnol. Biochem. 1998,
62, 1904–1911.
4. Koketsu, M.; Kim, M.; Yamamoto, T. J. Agric. Food
Chem. 1996, 44, 301–303.
5. Candra, E.; Matsunaga, K.; Fujiwara, H.; Mimaki, Y.;
Kuroda, M.; Sashida, Y.; Ohizumi, Y. J. Pharm. Phar-
macol. 2002, 54, 257–262.
6. Mimaki, Y.; Watanabe, K.; Ando, Y.; Sakuma, C.;
Sashida, Y.; Furuya, S.; Sakagami, H. J. Nat. Prod.
2001, 64, 17–22.
7. Qiu, S. X.; Li, X. C.; Xiong, Y. S.; Dong, Y. M.; Chai, H.
Y.; Farnsworth, N. R.; Pezzuto, J. M.; Fong, H. H. S.
Planta Med. 2000, 66, 587–590.
8. Fattorusso, E.; Lanzotti, V.; Taglialatela-Scafati, O.; Di
Rosa, M.; Ianaro, A. J. Agric. Food Chem. 2000, 48,
3455–3462.
9. Mimaki, Y.; Kuroda, M.; Kameyama, A.; Yokosuka, A.;
Sashida, Y. Phytochemistry 1998, 48, 1361–1369.
10. Inoue, T.; Mimaki, Y.; Sashida, Y.; Nikaido, T.; Ohmoto,
T. Phytochemistry 1995, 39, 1103–1110.
11. Inoue, T.; Mimaki, Y.; Sashida, Y.; Nishino, A.;
Satomi, Y.; Nishino, H. Phytochemistry 1995, 40,
521–525.
with de-ionized water. The protein was then removed
from the eluent by centrifugation (4000g, 4ꢁC) in a
2mL Vivaspin concentrator (note: the Vivaspin car-
tridge was washed through with three 1mL aliquots of
de-ionized water to remove residual NaN3 and glycerol
in the CTAmembrane prior to use). The filtrate was
freeze-dried to give an anomeric mixture (a/b = 37:63)
of lycotetraose as a white powder (34mg, 58%);
28
22
½aŠ þ 3:0 (c 0.29, water), lit.17 ½aŠ þ 2 (c 0.5, water),
D
D
26
D
lit.15 ½aŠ þ 2 (c 1.2, water); dC (100MHz, D2O) 103.36
(aC-1, Glc(I)), 103.20 (bC-1, Glc(I)), 103.18 (C-1,
Xyl), 102.42 (bC-1, Glc(II)), 102.38 (aC-1, Glc(II)),
96.86 (bC-1, Gal), 92.66 (aC-1-Gal), 84.83, 84.76,
81.49 (aC-4, Gal), 80.44 (bC-4, Gal), 79.15, 79.13,
77.11, 76.31 (C-3, Xyl), 76.11, 75.85, 74.74, 74.20,
74.16, 74.00 (C-4, Xyl), 73.80 (C-2, Xyl), 73.09, 72.52,
70.63, 70.30, 70.12, 69.71, 68.41, 65.70 (C-5, Xyl),
62.93, 61.71 (CH2), 61.53 (CH2), 61.46 (CH2), 61.30
(CH2), 60.99 (CH2); [found: (ESI), 654.2453 (ammo-
nium adduct), C23H44NO20 requires 654.2451]; m/z 659
(100%, M + Na+).
12. Mirsalikhova, N. M.; Kravets, S. D.; Sokolova, S. F.;
Abubakirov, N. K. Chem. Nat. Compd. 1993, 29, 490–497.
13. Mimaki, Y.; Kanmoto, T.; Kuroda, M.; Sashida, Y.;
Nishino, A.; Satomi, Y.; Nishino, H. Chem. Pharm. Bull.
1995, 43, 1190–1196.
Acknowledgements
14. Wang, Y. F.; Li, X. C.; Yang, H. Y.; Wang, J. J.; Yang, C.
R. Planta Med. 1996, 62, 130–132.
15. Takeo, K.; Nakaji, T.; Shinmitsu, K. Carbohydr. Res.
1984, 133, 275–287.
16. Randolph, J. T.; Danishefsky, S. J. J. Am. Chem. Soc.
1995, 117, 5693–5700.
17. Kuhn, R.; Low, I.; Trischmann, H. Chem. Ber. 1957, 90,
203–218.
18. Morrissey, J. P.; Osbourn, A. E. Microbiol. Mol. Biol. Rev.
1999, 63, 708–724.
19. Roldan-Arjona, T.; Perez-Espinosa, A.; Ruiz-Rubio, M.
Mol. Plant Microbe In. 1999, 12, 852–861.
20. Ikeda, T.; Kinjo, J.; Kajimoto, T.; Nohara, T. Heterocy-
cles 2000, 52, 775–798.
These studies were supported by the MRC, the Weston
Foundation, and the Leverhulme Trust. We would like
to thank Prof. Manuel Ruiz-Rubio for the generous gift
of the E. coli transformant containing recombinant tom-
atinase, Dr. Alan Haines for his helpful discussions and
assistance with the TLC protocols and Dr. Colin Mac-
Donald for his assistance with the NMR analyses. We
also thank the EPSRC Mass Spectrometry Service Cen-
tre, Swansea for invaluable support.
References
21. Lairini, K.; Perez-Espinosa, A.; Pineda, M.; Ruiz-Rubio,
M. Appl. Env. Microbiol. 1996, 62, 1604–1609.
22. Willker, W.; Leibfritz, D. Magn. Reson. Chem. 1992, 30,
645–650.
1. Roddick, J. G. Phytochemistry 1974, 13, 9–25.
2. Bedir, E.; Khan, I. A.; Walker, L. A. Pharmazie 2002, 57,
491–493.