Unusually broad substrate tolerance of a heat-stable archaeal sugar nucleotidyltransferase for the synthesis of sugar nucleotides

Article abstract of DOI:10.1021/ja046070d

Herein, we report the first cloning, recombinant expression, and synthetic utility of a sugar nucleotidyltransferase from any archaeal source and demonstrate by an electrospray ionization mass spectrometry (ESI-MS)-based assay its unusual tolerance of heat, pH, and sugar substrates. The metalion-dependent enzyme from Pyrococcus furiosus DSM 3638 showed a relatively high degree of acceptance of glucose-1-phosphate (Glc1P), mannose-1-phosphate (Man1P), galactose-1-phosphate (Gal1P), fucose-1-phosphate, glucosamine-1-phosphate, galactosamine-1-phosphate, and N-acetylglucosamine-1-phosphate with uridine and deoxythymidine triphosphate (UTP and dTTP, respectively). The apparent Michaelis constants for Glc1P, Man1P, and Gal1P are 13.0 ± 0.7, 15 ± 1, and 22 ± 2 μM, respectively, with corresponding turnover numbers of 2.08, 1.65, and 1.32 s^-1, respectively. An initial velocity study indicated an ordered bi-bi catalytic mechanism for this enzyme. The temperature stability and inherently broad substrate tolerance of this archaeal enzyme promise an effective reagent for the rapid chemoenzymatic synthesis of a range of natural and unnatural sugar nucleotides for in vitro glycosylation studies and highlight the potential of archaea as a source of new enzymes for synthesis.

Full text of DOI:10.1021/ja046070d

Published on Web 11/19/2004
Unusually Broad Substrate Tolerance of a Heat-Stable
Archaeal Sugar Nucleotidyltransferase for the Synthesis of
Sugar Nucleotides
Rahman M. Mizanur, Corbin J. Zea, and Nicola L. Pohl*
Contribution from the Department of Chemistry and Plant Science Institute, Gilman Hall,
Iowa State UniVersity, Ames, Iowa 50011-3111
Received July 1, 2004; E-mail: npohl@iastate.edu
Abstract: Herein, we report the first cloning, recombinant expression, and synthetic utility of a sugar
nucleotidyltransferase from any archaeal source and demonstrate by an electrospray ionization mass
spectrometry (ESI-MS)-based assay its unusual tolerance of heat, pH, and sugar substrates. The metal-
ion-dependent enzyme from Pyrococcus furiosus DSM 3638 showed a relatively high degree of acceptance
of glucose-1-phosphate (Glc1P), mannose-1-phosphate (Man1P), galactose-1-phosphate (Gal1P), fucose-
1-phosphate, glucosamine-1-phosphate, galactosamine-1-phosphate, and N-acetylglucosamine-1-phosphate
with uridine and deoxythymidine triphosphate (UTP and dTTP, respectively). The apparent Michaelis
constants for Glc1P, Man1P, and Gal1P are 13.0 ( 0.7, 15 ( 1, and 22 ( 2 µM, respectively, with
corresponding turnover numbers of 2.08, 1.65, and 1.32 s-1, respectively. An initial velocity study indicated
an ordered bi-bi catalytic mechanism for this enzyme. The temperature stability and inherently broad
substrate tolerance of this archaeal enzyme promise an effective reagent for the rapid chemoenzymatic
synthesis of a range of natural and unnatural sugar nucleotides for in vitro glycosylation studies and highlight
the potential of archaea as a source of new enzymes for synthesis.
Introduction
The enzymes responsible for creating the incredible diversity
of carbohydrate structures that mediate many biological pro-
cesses are the glycosyltransferases.1 Most glycosyltransferases
that incorporate a wide range of sugars into both primary and
secondary metabolites rely on uridine- and thymidinediphospho
nucleotide sugars that are produced by sugar nucleotidyltrans-
ferases, also known as sugar pyrophosphorylases.2 Several
glycosyltransferases involved in secondary metabolism recently
have been shown to be very promiscuous with respect to their
nucleotidediphospho (NDP) sugar donor.2 A limiting factor in
exploiting this promiscuity to produce new compounds is the
lack of the required unusual NDP sugars.3 Since these activated
sugar donors are difficult to synthesize chemically, current
efforts to generate these structures center on using an array of
enzymes or extremely high enzyme concentrations to circumvent
poor acceptance of non-natural substrates.2a This prior work has
focused entirely on enzymes from eukaryotic and bacterial
sources. Our recent studies of a glycosyltransferase involved
in primary metabolism from the third branch of life, archaea,
have hinted that these untapped genomes could be a source of
enzymes with even greater carbohydrate substrate tolerances
Figure 1. Reaction scheme for the sugar nucleotidyltransferase UDPG-
PPase that couples glucose-1-phosphate and uridine triphosphate (UTP) to
form activated UDP-glucose as a Leloir pathway glycosyl donor.
than seen to date with enzymes cloned from larger genomes.4
In addition, archaea contain many extremophiles and therefore
can be a source of thermoactive and heat-stable enzymes for
biotechnology applications if the necessary substrate tolerances
can be found.5 As part of our program to understand the origins
of fidelity in carbohydrate biosynthesis, we report the first study
of a sugar nucleotidyltransferase from any archaeal source, the
hyperthermophile Pyrococcus furiosus DSM 3638, and show
that its inherent substrate tolerance allows its use for the
synthesis of commercially unavailable non-natural NDP sugars,
such as uridinediphospho (UDP) mannose, with over 300-fold
fewer units of enzyme than previously reported.2a
Most sugar nucleotidyltransferases catalyze the transfer of a
sugar-1-phosphate to nucleotidetriphosphate (NTP) with the
concomitant release of pyrophosphate (Figure 1). Chemoenzy-
matic synthesis of a number of natural and unusual UDP and
deoxythymidine (dTDP) sugars has been carried out using wild-
(1) (a) Saxon, E.; Bertozzi, C. R. Annu. ReV. Cell DeV. Biol. 2001, 17, 1-23.
(b) Sears, P.; Wong, C.-H. Cell. Mol. Life Sci. 1998, 54, 223-252. (c)
Bertozzi, C. R.; Kiessling, L. L. Science 2001, 291, 2357-2364. (d) Koeller,
K. M.; Wong, C.-H. Nat. Biotechnol. 2000, 18, 835-841.
(2) (a) Jiang, J.; Biggins, J. B.; Thorson, J. S. J. Am. Chem. Soc. 2000, 122,
6803-6804. (b) Thorson, J. S.; Hosted, T. J.; Jiang, J.; Biggins, J. B.; Ahlert,
J.; Rupen, M. Curr. Org. Chem. 2001, 5, 139-167.
(4) Zea, C. J.; MacDonell, S. W.; Pohl, N. L. J. Am. Chem. Soc. 2003, 125,
13666-13667.
(5) Sterner, R.; Liebl, W. Crit. ReV. Biochem. Mol. Biol. 2001, 36, 39-106.
(3) Barton, W. A.; Biggins, J. B.; Jiang, J.; Thorson, J. S.; Nikolov, D. B.
Proc. Natl. Acad. Sci. U.S.A. 2002, 9, 13397-13402.
9
10.1021/ja046070d CCC: $27.50 © 2004 American Chemical Society
J. AM. CHEM. SOC. 2004, 126, 15993-15998
15993

A R T I C L E S
Mizanur et al.
type and engineered variants of R-D-glucopyranosyl phosphate
thymidylyltransferase (also known as Ep) from the bacteria
Salmonella enterica, LT 2.2a,3 The wild-type enzyme and a
number of its mutants have been used for the synthesis of NDP
sugar libraries with high enzyme concentrations (3.5 units; a
unit is defined as the amount of enzyme needed to produce one
micromole of nucleotidediphospho sugar per minute).2a Al-
though no kinetic studies have been carried out with this
Salmonella enzyme with any non-natural substrates, our results
with a related enzyme from Escherichia coli6 suggest that
substrate binding affinity differences are not the major problem,
and high enzyme concentrations are necessary to override the
limited turnover of non-natural substrates. Expansion and
diversification of both natural and non-natural NDP sugar
libraries still await the discovery of enzymes with wide glycosyl
phosphate acceptability, ideally without the introduction of
mutations or active site engineering and with a tolerance of
robust working conditions. The study of sugar nucleotidyltrans-
ferase has also attracted interest for their potential as antibiotic
targets.7 For example, UDP-glucose, synthesized by UDP-R-
D-glucose pyrophosphorylases (UDPG-PPases), has a number
of vital cellular functions, including the synthesis of glycogen,8
the synthesis of the carbohydrate moiety of glycolipids,9
glycoproteins,10 and proteoglycans,11 the entry of galactose into
glycolysis,12 the synthesis of UDP-glucuronic acid,13 and the
pathogenesis of a number of bacteria.14
archaea. The amino acid sequence of a putative archaea UDP-
R-D-glucose pyrophosphorylase (UTP: R-D-glucose-1-phosphate
uridylyltransferase, EC 2.7.7.9) from P. furiosus DSM 3638 was
compared to related bacterial and eukaryotic sequences using
BLAST (www.ncbi.nih.gov) to identify conserved residues. P.
furiosus was chosen as a representative archaea because a large
structural proteomics effort centered on this extremophile
promises protein data to match biochemical studies.18 Pairwise
alignment of the enzyme showed a high degree of residue
identity between archaea and bacteria ranging from 40 to 51%.
Sixty-two residues are strictly conserved among UDPG-PPases
from a variety of evolutionary diverse organisms. Among the
conserved residues, glycine is the most abundant, followed by
glutamate, lysine, and proline. However, among the amino acids
of P. furiosus UDPG-PPase, glutamate is the most prominent,
followed by leucine, isoleucine, lysine, and glycine. The
percentage of Glu and Lys increased in the hyperthermophilic
proteomes, and the percent of Gln and His decreased, so that
(Glu + Lys)/(Gln + His) is greater than 4.5.19 This ratio for
the P. furiosus UDPG-PPase is 4.8, which would predict
thermostability. A single motif can be clearly identified at
alignment position 5-30 (Figure S4 of the Supporting Informa-
tion), which is in the N-terminal position of the enzyme. All
other conserved residues are distributed throughout the sequence.
Amino acid sequence alignments among the UDPG-PPases from
Saccharomyces cereVisiae, E. coli, and P. furiosus were also
done using BLAST. As shown in Figure 2, 38% overall
sequence identity was found between the E. coli and P. furiosus
enzymes (Figure 2). In contrast, the archaeal enzyme showed
no sequence similarity with the yeast enzyme. Therefore, the
archaea enzyme could be expected to have significantly different
properties than the related bacterial and eukaryotic enzymes and
thereby could be a good target for exploring its potential for
chemoenzymatic synthetic strategies.
To assess the potential of archaeal proteins in the synthesis
of sugar nucleotides, this open reading frame predicted to encode
the P. furiosus UDPG-PPase (852 base pairs) was amplified
by PCR and cloned into a pET21a vector for expression of the
protein in E. coli cells with a hexahistidine tag at the C-terminus.
To purify the enzyme, the cells were lysed by sonication, heat-
treated to denature native E. coli proteins, and centrifuged. The
resulting cell-free extract was passed through a Ni-affinity
column and eluted with imidazole buffer. SDS-PAGE analysis
of the purified enzyme showed a protein with an apparent
molecular mass of 36 kDa (see the Supporting Information),
which was in good agreement with the calculated molecular
mass of the gene product (31.2 kDa). The enzyme was used
directly for determination of its activity and substrate specificity
or stored at -20 °C as a glycerol stock solution for future use
with little loss of activity after 90 days.
Results and Discussion
Preparation of an Archaeal Sugar Nucleotidyltransferase.
To date, UDPG-PPase has been isolated and characterized from
both eukaryotic15 and prokaryotic16 sources. On the basis of
sequence similarities, UDPG-PPases have been classified into
two groups despite their similar catalytic properties: the
prokaryotic enzymes with about 300 amino acids residues and
the eukaryotic enzymes with about 500 amino acids residues.17
Surprisingly, no study has been conducted so far on sugar
nucleotidyltransferases from the third class of living beings,
(6) Zea, C. J.; Pohl, N. L. Anal. Biochem. 2004, 328, 196-202.
(7) (a) Brown, K.; Pompeo, F.; Dixon, S.; Mengin-Lecreulx, D.; Cambillau,
C.; Bourne, Y. EMBO J. 1999, 18, 4096-4107. (b) Blankenfeldt, W.;
Asuncion, M.; Lam, J. S.; Naismith, J. H. EMBO J. 2000, 19, 6652-6663.
(c) Sulzenbacher, G.; Gal, L.; Peneff, C.; Fassy, F.; Bourne, Y. J. Biol.
Chem. 2001, 276, 11844-11851.
(8) (a) Alonso, M. D.; Lomako, W. M.; Whelan, W. J. FASEB J. 1995, 9,
1126-1137. (b) Hellerstein, M. K.; Letscher, A.; Schwarz, J. M.; Ce´sar,
D.; Shackleton, C. H. L.; Turner, S.; Neese, R.; Wu, K.; Bock, S.; Kaempfer,
S. A. J. Physiol. 1997, 272, E155-E162.
(9) Sandhoff, K.; van Echten, G.; Schro¨der, M.; Schnabel, D.; Suzuki, K.
Biochem. Soc. Trans. 1992, 20, 695-699.
(10) (a)Verbert, A. Glycoproteins; Elsevier: Amsterdam, 1995; pp145-152.
(b) Roth, J. Glycoproteins; Elsevier: Amsterdam, 1995; pp 287-312.
(11) Silbert, J. E.; Sugumaran, G. Biochim. Biophys. Acta 1995, 1241, 371-
384.
(12) Frey, P. A. FASEB J. 1996, 10, 461-470.
(13) (a) Aw, T. Y.; Jones, D. P. Anal. Biochem. 1982, 127, 32-36. (b) Aw, T.
Y.; Jones, D. P. Biochem. J. 1984, 219, 707-712.
Enzymatic Analysis. The difficulty of assays, such as those
based on HPLC,20 needed to measure sugar nucleotidyltrans-
ferase activity with non-natural substrates has been a significant
deterrent to kinetic studies. Our recent development of an
electrospray ionization mass spectrometry assay has significantly
shortened the analysis time, and since it is not substrate-specific,
(14) (a) Chang, H. Y.; Deng, W. L.; Lee, J. H.; Fu, T. F.; Peng, H. L. Microb.
Pathog. 1996, 20, 255-261. (b) Crater, D. J.; Van de Rijn, I. J. Biol. Chem.
1995, 270, 18452-18458.
(15) (a) Hansen, R. G.; Albrecht, G. J.; Bass, S. T.; Seifert, L. L. Methods
Enzymol. 1966, 8, 248-253. (b) Gustafson, G. L.; Gander, J. E. J. Biol.
Chem. 1972, 247, 1387-1397. (c) Pannbacker, R. G. Biochemistry 1967,
6, 1287-1293. (d) Villar-Palasi, C.; Larner, J. Arch. Biochem. Biophys.
1960, 86, 61-66.
(16) (a) Kamogawa, A.; Kurahashi, K. J. Biochem. (Tokyo) 1965, 57, 758-
765. (b) Nikae, T.; Nikaido, H. J. Biol. Chem. 1971, 246, 4386-4396. (c)
Nikae, T.; Nikaido, H. J. Biol. Chem. 1971, 246, 4397-4403. (d) Hossain,
S. A.; Tanizawa, K.; Kazuta, Y.; Fukui, T. J. Biochem. (Tokyo) 1994, 115,
965-972.
(17) Weissborn, A. C.; Liu, Q.; Ruimley, M. K.; Kennedy, E. P. J. Bacteriol.
1994, 176, 2611-2618.
(18) Adams, M. W. W.; Dailey, H. A.; DeLucas, L. J.; Luo, M.; Prestegaard, J.
H.; Rose, J. P.; Wang, B.-C. Acc. Chem. Res. 2003, 36, 191-198.
(19) Farias, S. T.; Bonato, M. C. M. Genet. Mol. Res. 2003, 2, 383-393.
(20) (a) Lazarowski, E. R.; Harden, T. K. J. Pharmacol. 1999, 127, 1272-
1278. (b) Jiang, J.; Biggins, J. B.; Thorson, J. S. Angew. Chem., Int. Ed.
2001, 40, 1502-1505.
9
15994 J. AM. CHEM. SOC. VOL. 126, NO. 49, 2004

Substrate Tolerance of a Heat-Stable Archaeal Sugar
A R T I C L E S
Figure 2. Amino acid sequence alignment of P. furiosus UDPG-PPase with an E. coli homolog. Pfu: P. furiosus (accession no. NC003413.1). Eco: E. coli
(accession no. AP002556.1). The identical amino acids are indicated in bold. The yeast enzyme used in this study showed no sequence similarity with the
P. furiosus enzyme.
Table 1. Effects of Divalent Cations on Enzymatic Activity of the
Sugar Nucleotidyltransferase from P. furiosus
it can be used for any sugar nucleotidyltransferase.6 With this
assay, the activity of both crude protein extracts and purified
divalent
cationa
relative
enzyme was determined at 90 °C using glucose-1-phosphate
(1, Glc1P) and uridinetriphosphate (UTP) as initial trial
substrates. Formation of UDP-glucose at these elevated tem-
peratures in both the crude extract and the purified protein
confirmed the presence of an active UDPG-PPase, thereby
verifying the tentatively assigned function of this gene. Because
uridinemonophosphate transfer to Glc1P releases pyrophosphate,
which can serve as an inhibitor, a commercially available
thermostable inorganic pyrophosphatase (IPP) was added to the
reaction mixture. Although heat treatment should have denatured
all native proteins, an additional control experiment was run
with extracts of E. coli carrying a pET21a vector without an
inserted gene. No additional protein band corresponding to 36
kDa was detected by SDS-PAGE analysis of the crude extract
after passing through the Ni-affinity column, and no UDPG-
PPase activity was detected at 90 °C. No product formation
was observed in the absence of UDPG-PPase, sugar phosphates,
Mg2+, or NTP.
activity (%)
Mg2+
Mn2+
Cu2+
Co2+
Ca2+
Zn2+
None
100
72
44
14
9
9
0
a The purified enzyme (0.007 U) was incubated with standard assay
mixtures at pH 7.5 with the addition of various metal ions at 3 mM (chloride
form) and incubated for 5 min at 90 °C. The UDP-glucose that formed was
analyzed by ESI-MS, and the relative activity is expressed as the percentage
of the activity measured in the presence of MgCl2. “None” indicates that
the reaction was carried out in the absence of any divalent cations.
tidyltransferase, Ep,22 suggest that Mg2+ plays a structural role
in organizing the substrate binding region of the Ep around itself
to fix the nucleotidetriphosphate at an optimal position for
catalysis.
Substrate Specificity Studies. We are particularly interested
in developing enzymatic routes to natural and non-natural UDP
and dTDP sugars, ideally without the use of multiple enzymes
and without the need for extremely high enzyme concentrations.
In addition, the discovery of enzymes that demonstrate intrinsic
catalytic competence with a range of substrates provides clues
to how biosynthetic machinery has evolved to maintain fidelity
in carbohydrate biosynthesis. Although the P. furiosus enzyme
catalyzed the formation of UDP-glucose, as predicted from gene
sequence analysis, experiments with the glycogen synthase4 from
this organism demonstrated the danger of testing only one
substrate pair to assign the chemical function of a gene.
Therefore, a range of substrates (Figure 3) were assayed with
the enzyme. Mixtures containing the reaction components were
initiated by addition of 0.01 units of the enzyme and incubated
for 1 h at 90 °C. The percent conversion of sugar-1-phosphate
with UTP and dTTP was then measured by mass spectrometry.
In addition to UTP, dTTP was found to serve as a nucleoside
monophosphate donor (Table 2), similar to data reported from
enzymes from E. coli and Salmonella typhimurium.22,23 No
The temperature, pH ranges, and divalent cation requirements
of the sugar nucleotidyltransferase were ascertained next. This
first purified archaeal UDPG-PPase showed maximum activity
at 99 °C (Figure S1 of the Supporting Information) with little
loss of activity at 110 °C in phosphate buffer with glycerol.
The enzyme exhibited activity between pH 5.5 and 9.5, with a
maximum around pH 7.5-8.0 in phosphate buffer (Figure S2
of the Supporting Information). The shape and maxima of the
pH velocity curves, however, were found to vary somewhat with
the nature of the buffer ion present. The enzyme showed an
absolute requirement of divalent cations for activity, as estab-
lished for other UDP-glucose pyrophosphorylases.15,16 Magne-
sium, manganese, cobalt, and calcium were found to be effective
in decreasing order (Table 1). In the direction of UDP-glucose
synthesis, a magnesium concentration of 3 mM elicited the
maximum reaction rate (Figure S3 of the Supporting Informa-
tion); enzyme activity did not decrease extensively, however,
even at higher concentrations of magnesium. Although molec-
ular and structural data are not available for any UDPG-PPase,
it has been suggested that the major function of magnesium in
enzyme catalysis is in the formation of MgUTP2- complexes.21
However, recent structural data of the Salmonella sugar nucleo-
(22) Barton, W. A.; Lesniak, J.; Biggins, J. B.; Jeffrey, P. D.; Jiang, J.;
Rajashanker, K. R.; Thorson, J. S.; Nikolov, D. B. Nat. Struct. Biol. 2001,
8, 545-551.
(23) Weissborn, A. C.; Liu, Q.; Ruimley, M. K.; Kennedy, E. P. J. Bacteriol.
1994, 176, 2611-2618.
(21) Gustafson, G. L.; Gander, J. E. J. Biol. Chem. 1972, 247, 1387-1397.
9
J. AM. CHEM. SOC. VOL. 126, NO. 49, 2004 15995

A R T I C L E S
Mizanur et al.
Table 2. Comparison of Kinetic Parameters of UDPG-PPase from S. cerevisiae, E. coli, and P. furiosus
S. cerevisiaea
E. colia
Vmax
M/min)
P. furiosusb
Vmax
KM
M)
Vmax
k
cat/KM
KM
M)
k
cat/KM
KM
M)
kcat/KM
1
1
1
substrate
(
µ
(
µM/min)
(
µ
M-1 s-
)
(
µ
(µ
(
µ
M-1 s-
)
(
µ
(
µM/min)
(
µ
M-1 s-
)
UTP-Glc1P
dTTP-Glc1P
ATP-Glc1P
UTP-Gal1P
UTP-Man1P
7 ( 1
3.8 ( 0.1
1.7 ( 0.1
10.9
12 ( 2
1.2 ( 0.1
0.9 ( 0.1
1.45
1.05
13 ( 1
2.8 ( 0.1
2.6 ( 0.1
0.16
0.09
not accepted
7 ( 1
13 ( 2
19 ( 2
4.8
not accepted
not acceptedb
not acceptedb
not acceptedb
22 ( 2
15 ( 1
not acceptedb
not acceptedb
2.00 ( 0.07
2.10 ( 0.04
0.06
0.11
a Previously reported (ref 6). b From this study. Kinetic parameters were determined in 25 mM phosphate buffer (pH 7.5) at 90 °C using 2.5 × 10-4
U
of UDPG-PPase. Reported error bars represent the standard deviation, which was obtained from three independent experiments. UDPG-PPases from E. coli
and S. cereVisiae were run in triplicate at varying concentrations of Man/Gal1P and at fixed concentrations of UTP using the same enzyme unit concentration
as was used for the P. furiosus enzyme.
Table 3. Conversion of Sugar-1-Phosphates by UDPG-PPase
from P. furiosus
UTP
conversion
(%)b
dTTP
conversion
(%)b
sugar-1-phosphatea
glucose-1-phosphate (1)
mannose-1-phosphate (2)
99
92
96
82
66
28
84
91
36
15
10
43
5
galactose-1-phosphate (3)
fucose-1-phosphate (4)
glucosamine-1-phosphate (5)
galactosamine-1-phosphate (6)
N-acetylglucosamine-1-phosphate (7)
56
a The purified enzyme (0.01 U) was incubated in a 50 µL reaction volume
with 3 mM Mg2+, 5 mM UTP or dTTP, and 5 mM sugar-1-phosphate,
including all other reaction components. The reaction was incubated at 90
°C for 60 min. NDP-sugars that formed were analyzed by ESI-MS. b Percent
conversion is defined as 100 times the ratio of the amount of sugar-1-
phosphate remaining in the reaction mixture over the amount of initial sugar-
1-phosphate. No substrate hydrolysis was seen under these conditions.
Figure 3. Sugar-1-phosphates used in this study.
Control experiments were run to ensure that the archaeal
protein was solely responsible for this unusually broad substrate
turnover. It has been reported that small amounts of UDP-
mannose/galactose/xylose were formed in the presence of
UDPG-PPase purified from calf liver.27 However, coelution of
other contaminating sugar nucleotidyltransferases in the purified
liver enzyme preparation has not been ruled out. Control
experiments using heat treatment of the crude extract, as well
as a plasmid without insert, eliminate the possibility of
contaminating E. coli proteins as a source of alternate sugar
nucleotidyltransferase activities in the case of the archaeal
protein. In addition, we have tested the recombinant E. coli
UDPG-PPase6 activity at 90 °C, and no activity was observed
at this elevated temperature. The presence of only P. furiosus
UDPG-PPase can account for the conversion of the various
sugars to activated sugar donors.
Enzymatic Mechanisms. Relative substrate specificity stud-
ies are useful for ascertaining the potential of an enzyme for
chemoenzymatic synthesis; unfortunately, such data provide no
clues as to the reasons why one substrate is preferred over
another. Therefore, the mechanism of this thermostable archaeal
enzyme was probed next. Previous functional investigations of
a related thymidylyltransferase have provided alternate hypo-
theses on the catalytic mechanism.28 On one hand, the bimo-
lecular group transfer reaction has been described by a ping-
pong bi-bi mechanism, based on the analysis of double
reciprocal plots of initial velocity as a function of different
activity was recorded against all other nucleotide triphosphates
in those cases. Although several enzymes from P. furiosus
exhibit an unusual usage of nucleotides,4,24 our archaeal sugar
nucleotidyltransferase also showed this discrimination against
other NTPs. However, unlike other reported UDPG-PPases and
dTDPG-PPases, a substantial amount of activity against man-
nose-1-phosphate (2, Man1P) and galactose-1-phosphate (3,
Gal1P) with UTP and dTTP was found when the reaction
mixture was incubated for 30 min with only 0.007 units of the
enzyme. This finding is in stark contrast to results obtained with
the related Salmonella enzyme Ep in which 500 times more units
of enzyme were used to get only 18% conversion of Man1P
into the non-natural UDP-mannose donor.2a In addition, fucose-
1-phosphate (Fuc1P), glucosamine-1-phosphate (GlcN1P), ga-
lactosamine-1-phosphate (GalN1P), and N-acetylglucosamine-
1-phosphate (GlcNAc1P) also demonstrated significant product
formation of their respective UDP and dTDP sugars when the
reaction mixture was incubated with 0.01 units of the enzyme
(Table 3). For comparison, very low rates of activity (<1%
compared to those of Glc1P) were evident against Gal1P and
Man1P with UTP of the enzyme purified from human eryth-
rocyte25 and <5% activity was reported against Gal1P and
Man1P with the enzyme purified from Thermus caldophilus.26
(24) (a) Kengen, S. W. M.; Tuininga, J. E.; de Bok, F. A. M.; Stams, A. J. M.;
de Vos, W. M. J. Biol. Chem. 1995, 270, 30453-30457. (b) Mai, X.;
Adams, M. W. W. J. Bacteriol. 1996, 178, 5897-5903. (c) Tuininga, J.
E.; Verhees, C. H.; Van der Ooost, J.; Kengen, S. W. M.; Stams, A. J. M.;
De Vos, W. M. J. Biol. Chem. 1999, 274, 21023-21028.
(25) Tsuboi, K. K.; Fukunaka, K.; Petricciana, J. C. J. Biol. Chem. 1969, 244,
1008-1015.
(27) Albrecht, G. J.; Bass, S. T.; Seifert, L. L.; Hansen, R. G. J. Biol. Chem.
1966, 241, 2968-2975.
(28) (a) Lindquist, L.; Kaiser, R.; Reeves, P. R.; Lindberg, A. A. Eur. J. Biochem.
1993, 211, 763-770. (b) Zuccotti, S.; Zanardi, D.; Rosano, C.; Sturla, L.;
Tonetti, M.; Bolognesi, M. J. Mol. Biol. 2001, 313, 831-843.
(26) Kim, J. S.; Koh, S.; Shin, H.; Lee, D.; Lee, S. Y. Biotechnol. Appl. Biochem.
1999, 29, 11-17.
9
15996 J. AM. CHEM. SOC. VOL. 126, NO. 49, 2004

Substrate Tolerance of a Heat-Stable Archaeal Sugar
A R T I C L E S
Figure 5. (A) Mechanism of the reaction catalyzed by UDPG-PPase. (B)
Distinct chemical groups that form a ternary complex with the protein.
substrate to bind to the enzyme and is also the last to be released.
Therefore, the archaeal enzyme follows the same mechanism
previously demonstrated by bacterial and eukaryotic sugar
nucleotidyltransferases.7,22,25 This type of reaction has been
suggested as an SN2 type, with dTTP binding to the protein,
followed by Glc1P, which acts as the R-phosphate of dTTP.20
The â- and γ-phosphates of dTTP are displaced and leave as
pyrophosphate. In SN2-type reactions, the leaving group and
the attacking nucleophile (the oxygen atom of the R-phosphate)
reside at opposite faces of the central atom. On the basis of the
above evidence and argument, the catalytic mechanism of
UDPG-PPase is diagrammed in Figure 5.
Figure 4. Steady-state kinetic parameters of UDPG-PPase from P. furiosus.
(A) The double reciprocal plots are from assays with varying glucose-1-
phosphate (1, Glc1P) concentrations as a function of UTP (µM): 20 (O),
50 (b), 100 ()), 200 (0), and 300 (4). (B) The double reciprocal plots are
from assays with varying UTP concentrations as a function of Glc1P
concentration (µM): 5 (O), 10 (b), 20 ()), 40 (0) and 80 (4). The enzymatic
reaction mixtures also contained phosphate buffer (25 mM, pH 7.5),
inorganic pyrophosphorylase (0.2 U), MgCl2 (3 mM), and the sugar
nucleotidyltransferase (0.00025 U), with a final volume of 50 µL. Reactions
were carried out at 90 °C for 5 min. Error bars represent the standard
deviation of three independent averaged data points.
Given this sequential bi-bi mechanism, we next wanted to
determine if the different relative conversions of sugar substrates
were a function of differences in substrate binding to the enzyme
active site or a difference in turnover rates of bound substrates.
Therefore, the kinetic properties of the enzyme with a number
of glycosyl phosphates and UTP and dTTP were determined
(Table 2). The concentrations of Glc1P (2-80 µM) or Man1P/
Gal1P (5-200 µM) were varied with a fixed concentration of
UTP/dTTP (400 µM). A Michaelis-Menten plot of the velocity
versus sugar-1-phosphate ion concentration was determined
using ESI-MS for the P. furiosus UDPG-PPase in the presence
of UTP for all three sugar-1-phosphates and dTTP for Glc1P
(Supporting Information). Nonlinear regression analysis of the
data yields KM and Vmax values of 13.0 ( 1 µM and 2.8 ( 0.1
µM min-1, respectively, for the enzyme with Glc1P and UTP.
These values are similar to values for the related enzymes from
yeast and E. coli (Table 2).6 Surprisingly, the KM and kcat/KM
values for glucose-, galactose-, and mannose-1-phosphate were
all comparable. This finding indicates that the archaeal sugar
nucleotidyltransferase truly is unusual in its ability to bind and
turnover a range of substrates.
concentrations of the two substrates.28a Achievement of a ping-
pong-type catalytic mechanism would require the availability
of reactive (nucleophilic) residues within the structural sur-
roundings of the dTTP R-phosphate group to yield the first
covalent enzyme-dTMP intermediate. Alternatively, the achieve-
ment of a sequential bi-bi mechanism requires proper nonco-
valent binding of dTTP and Glc1P within the active site
(yielding an enzyme substrate ternary complex) for the onset
of catalysis.28b In both cases, activation of the substrate reactive
phosphate groups (i.e., R-phosphate on dTTP and phosphate-1
on Glc1P) for the group transfer reaction through local interac-
tions with active site residues is required. To determine the
kinetic mechanism of the archaeal enzyme, double reciprocal
plots of the initial velocity were obtained by varying the
concentration of one substrate at fixed concentrations of the
other. The initial kinetic patterns of velocity with the Glc1P as
the variable substrate are shown in Figure 4A, and those with
UTP as the variable substrate are shown in Figure 4B. The clear
intersecting patterns left of the 1/V axis is indicative of a
sequential bi-bi mechanism in which the nucleotide is the first
Conclusion
This initial study of any sugar nucleotidyltransferase from
an archaeal source has uncovered a valuable heat-stable enzyme
for the chemoenzymatic synthesis of a range of activated sugar
nucleotides. The catalytic and mechanistic properties of the P.
furiosus enzyme are almost the same as those of the prokaryotic
and eukaryotic enzymes based on our findings. However, unlike
9
J. AM. CHEM. SOC. VOL. 126, NO. 49, 2004 15997

A R T I C L E S
Mizanur et al.
bacterial, plant, and animal enzymes, P. furiosus UDPG-PPase
has an unusual tolerance of high temperatures. More surpris-
ingly, this archaeal sugar nucleotidyltransferase can accept an
unusually broad range of sugar phosphates with reasonably
efficient turnover rates, unlike previously reported enzymes that
carry out this reaction. The archaeal UDPG-PPase demonstrated
a unique ability to activate a range of sugar substrates and,
therefore, should serve as an ideal biocatalytic tool in order to
simplify the synthesis of useful nucleotide sugars. In the context
of significant promiscuity displayed by a number of glycosyl-
transferases and the availability of appropriate aglycones, the
UDPG-PPase-catalyzed production of activated sugar donor
libraries is a particularly promising approach toward the
generation of a diverse library of glycosylated structures. With
a Salmonella sugar nucleotidyltranferase, a number of engi-
neered variants had to be obtained in order to have a wider
acceptance of substrates.23 At the same time, approximately
350-25 000-fold higher enzyme units, in comparison to those
of the P. furiosus UDPG-PPase, were used to get a reasonable
turnover and thereby an increased product yield.1,30 A slight
increase of inorganic pyrophosphorylase and UDPG-PPase
should allow >99% conversion of a wide range of natural and
non-natural substrates by the P. furiosus enzyme. Future work
with the archaeal enzyme will assess its tolerance for substrates
that include, specifically, reactive groups for chemoselective
ligation reactions or other tags to allow for the downstream
diversification or labeling of natural products and glycosylated
proteins using sugar nucleotide reactions.29,30 The catalytic
efficiency and thermostability of this enzyme also make it
potentially attractive for industrial applications in the biochemi-
cal synthesis of various sugar nucleotides.
Ambiguous enzyme activities may well reside in the genomes
of organisms as a result of either the partial retention of an earlier
optimized catalytic function or the redundancy of molecular
structures in the many modern metabolic pathways.31 Whether
the broader substrate acceptance of the archaeal sugar nucleo-
tidyltransferase reported here is a general characteristic of this
class of life, just characteristic of enzymes from smaller
genomes, or a special case requires further investigation.
However, the first substrate specificity studies of an archaeal
enzyme that activates carbohydrate substrates show the value
of more extensive probing of archaeal genomes to expand the
range of chemoenzymatic synthesis strategies.
Acknowledgment. We thank an NSF-CAREER award, the
Plant Sciences Institute, a Shimadzu University Research Grant,
and the Herman Frasch Foundation (American Chemical
Society) for their support of this research. N.L.P. is a Cottrell
Scholar of Research Corporation.
Supporting Information Available: Full experimental details
and additional data. This material is available free of charge
via the Internet at http://pubs.acs.org.
(29) (a) Hang, H. C.; Chong, Y.; Pratt, M. R.; Bertozzi, C. R. J. Am. Chem.
Soc. 2004, 126, 6-7. (b) Helm, J. S.; Hu, Y.; Chen, L.; Gross, B.; Walker,
S. J. Am. Chem. Soc. 2003, 125, 11168-11169. (c) Kolb, H. C.; Finn, M.
G.; Sharpless, K. B. Angew. Chem., Int. Ed. 2001, 40, 2004-2021.
(30) Fu, X.; Albermann, C.; Jiang, J.; Liao, J.; Zhang, C.; Thorson, J. S. Nat.
Biotechnol. 2003, 21, 1467-1469.
JA046070D
(31) (a) Miller, B. G.; Raines, R. T. Biochemistry 2004, 43, 6387-6392. (b)
Kraut, D. A.; Carroll, K. S.; Hershlag, D. Annu. ReV. Biochem. 2003, 72,
517-571.
9
15998 J. AM. CHEM. SOC. VOL. 126, NO. 49, 2004