Fidelity and Promiscuity of a Mycobacterial Glycosyltransferase

Article abstract of DOI:10.1021/jacs.6b04481

Members of the genus Mycobacterium cause devastating human diseases, including tuberculosis. Mycobacterium tuberculosis can resist some antibiotics because of its durable and impermeable cell envelope. This barrier is assembled from saccharide building bl

Full text of DOI:10.1021/jacs.6b04481

Subscriber access provided by UNIVERSITY OF KENTUCKY
Article
Fidelity and Promiscuity of a Mycobacterial Glycosyltransferase
Kenzo Yamatsugu, Rebecca A Splain, and Laura L Kiessling
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.6b04481 • Publication Date (Web): 14 Jun 2016
Downloaded from http://pubs.acs.org on June 15, 2016
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of the American Chemical Society is published by the American Chemical
Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 13
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Fidelity and Promiscuity of a Mycobacterial
ferase
Glycosyltrans-
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Kenzo Yamatsugu†§, Rebecca A. Splain†∫, and Laura L. Kiessling†,‡,*
†Department of Chemistry, University of Wisconsin–Madison, 1101 University Ave., Madison, WI 53706
‡Department of Biochemistry, University of Wisconsin–Madison, 433 Babcock Dr., Madison, WI 53706
§Current address: Pharmaceutical Sciences, University of Tokyo, 7-3-1, Hongo, Bunkyo-ku, Tokyo 113-0033, JAPAN
∫Current address: API Chemistry, GlaxoSmithKline, King of Prussia, PA
*Address correspondence to kiessling@chem.wisc.edu
Abstract: Members of the mycobacterium genus cause devastating human diseases, including tuberculosis (TB). Myco-
bacterium tuberculosis can resist some antibiotics because of its durable and impermeable cell envelope. This barrier is
assembled from saccharide building blocks not found in mammals, including galactofuranose (Galf). Within the cell enve-
lope, Galf residues are linked together to afford an essential polysaccharide, termed the galactan. The formation of this
polymer is catalyzed by the glycosyltransferase GlfT2, a processive carbohydrate polymerase, which generates a sequence-
specific polysaccharide with alternating regioisomeric β(1–5) and β(1–6) Galf linkages. GlfT2 exhibits high fidelity in link-
age formation, as it will terminate polymerization rather than deviate from its linkage pattern. These findings suggest
that GlfT2 would prefer an acceptor with canonical β(1–5) and β(1–6) Galf sequence. To test this hypothesis, we devised a
synthetic route to assemble oligosaccharides with natural and non-natural sequences. GlfT2 could elongate each of these
acceptors, even those with non-natural linkage patterns. These data indicate that the glycosyltransferase is surprisingly
promiscuous in its substrate preferences. However, GlfT2 did favor some substrates: it preferentially acted on those in
which the lipid-bearing Galf residue was connected to the sequence by a β(1–6) glycosidic linkage. The finding that the
relative positioning of the lipid and the non-reducing end of the acceptor influences substrate selectivity is consistent
with a role for the lipid in acceptor binding. The data also suggest that the fidelity of GlfT2 for generating alternating β(1–
5) and β(1–6) pattern of Galf residues arises not from preferential substrate binding but during processive elongation.
These observations suggest that inhibiting the action of GlfT2 will afford changes in cell wall structure.
Introduction Many glycans essential for microbial
pathogenesis, virulence, or viability contain non-
mammalian building blocks. Blocking the biosynthesis of
these unique glycans could lead to new antibacterial ther-
apies. The galactan of mycobacteria is composed of Galf
residues, which are not found in mammalian glycans
(Figure 1). The galactan is essential for mycobacterial via-
bility, yet no enzymes that assemble the galactan are tar-
geted by current antibiotics. Understanding the catalytic
properties of galactan biosynthetic enzymes could lead to
new strategies to target mycobacteria.
losis, GlfT2, which is encoded by the glfT2 (or Rv3808c)
gene, is essential.5 GlfT2 promotes the formation of linear
polysaccharides consisting of about 30-40 Galf residues
that are connected by alternating β(1-5) and β(1-6) linkag-
es.6 The product polymer, the galactan, extends from the
peptidoglycan and displays pendant arabinofuranose se-
quences that are linked ultimately to mycolic acids.
The natural glycosyl acceptor for GlfT2 is the tetra-
sacharide sequence D-Galf–β(1–5)–D–Galf–β(1–4)–L–
Rhaα(1–3)-D-GlcNAc linked by a pyrophosphate to a pol-
yprenol lipid (compound 1).7-9 This compound is generat-
ed through the action of the GlfT1. The ability of GlfT2 to
generate polymers depends on the nature of the lipid sub-
stituent,10,11 but the glycosytransferase can readily process
A key carbohydrate polymerase involved in galactan
biosynthesis is GlfT2 (Figure 2). While most glycosyltrans-
ferases catalyze the formation of one type of glycosyl
bond, some can generate multiple linkage types to yield
oligosaccharides or polysaccharides of controlled se-
quence.1-4 Little is known, however, regarding the mecha-
nism or substrate specificity of these multifunctional en-
zymes. GlfT2 is one such glycosyltransferase. In
Mycobacterium tuberculosis, GlfT2 (EC 2.4.1.288), GlfT2 is
one such glycosyltransferase. In Mycobacterium tubercu-
simple diGalf motifs with lipophilic anomeric substitu-
ents.6,7,11-13 As a result, synthetic substrates have been used
to probe the mechanism and products of GlfT2. GlfT2-
catalyzed elongation results in polysaccharides similar in
length to those isolated from mycobacteria.10,14 During
elongation, GlfT2 catalyzes the formation of multiple Gal
linkages before product release; it is a processive poly-
f
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 13
merase.10,14 Chain terminating nucleotide UDP-Galf deriv-
sequence in elongation, we focused on synthesizing tri-
1
2
3
4
5
6
7
8
atives have been used to reveal that GlfT2 faithfully gen-
erates alternating β(1–5) and β(1–6) Galf linkages.15,16 The-
se studies revealed GlfT2’s fidelity for assembly products
with alternating regioisomeric linkages. The structure of
GlfT2, determined by X-ray crystallography, indicates it
can form a tetramer, with each monomer unit possessing
a single active site.17 Site-directed mutagenesis studies
support a model in which one active site facilitates for-
mation of both glycosidic linkages.14 Sequence fidelity of
the enzyme could be manifested by selectivity in sub-
strate binding, in the process of elongation, or in both. To
explore these possibilities, we sought to synthesize and
test substrates whose sequences vary.
saccharides 4 and 5 with naturally alternating linkages,
trisaccharides 6 and 7 with non-natural consecutive link-
ages, and tetrasaccharide 8 with natural alternating link-
ages (Figure 3). The comparison of natural substrates (i.e.,
4 and 5) to their non-natural (i.e., 6 and 7) counterparts
would test whether GlfT2 can tolerate a “mistake” in the
initial acceptor. In this way, we could assess the effect of
different structural elements of the acceptor (internal and
terminal glycosidic linkage) on the efficiency of polymeri-
zation by GlfT2.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
We developed a divergent approach to produce oli-
gogalactofuranoside acceptors with different sequences.
The product oligosaccharides were assessed as substrates
for GlfT2 by analyzing the efficiency of their elongation
and the lengths of the product polysaccharides. Unex-
pectedly, substrates with perfectly alternating sequence
were not preferred over those with “errors”. It was the
nature of the first glycosidic linkage following the lipid
that had the largest influence on GlfT2-promoted glyco-
sylation. The data indicate that the preferential formation
of alternating β(1–6)Galf then β(1–5)Galf residues arises
not from substrate binding but during catalytic processive
polymerization.
Natural
OR2
OR2
HO
HO
O
β(1–5)
O
HO
O
HO
O
OH
OH
HO
HO
O
OH
HO
β(1–6)
OH
3
O
HO
HO
2
OH
OR2
OR2
HO
HO
β(1–5)
O
O
HO
O
HO
O
OH
OH
HO
O
O
OH
HO
β(1–6)
OH
β(1–6)
O
β(1–5)
Figure 1. The mycobacterial cell envelope contains
an essential polymer composed of Galf residues. A. Sche-
matic of the mycobacterial cell envelope showing that the
galactan (red) links the mycolic acids to the peptidogly-
can. B. Mycobacteria generate the lipid-linked carbohy-
drate polymer, which then is covalently linked to the pep-
tidoglycan.
O
HO
HO
O
O
OH
HO
HO
HO
OH
4
5
OH
OH
HO
OR2
HO
O
Non-natural
HO
OR2
HO
OH
β(1–5)
O
O
O
HO
HO
β(1–6)
β(1–5)
O
O
OH
Our oligosaccharide targets were predicated on find-
ings that recombinant GlfT2 can extend acceptor mimics,
such as 2, to afford polymers with lengths similar to those
isolated from mycobacteria (Figure 3).11 To test the role of
HO
O
HO
OH
O
OH
β(1–6)
OH
HO
HO
O
OH
HO
OH
O
HO
HO
6
7
OH
OR2
HO
O
HO
Natural Tetrasaccharide
OH
β(1–6)
O
HO
O
β(1–5)
O
HO
O
HO
OH
OH
OH
8
O
HO
O
β(1–6)
HO
HO
OH
O
9
R2
=
Ph
Figure 3. Oligosaccharides synthesized to probe GlfT2 speci-
ficity. Trisaccharides 4 and 5 and tetrasaccharide 8 have nat-
ural alternating glycosidic linkages, while trisaccharides 6
ACS Paragon Plus Environment
and 7 have non-natural consecutive glycosidic linkages.

Page 3 of 13
Journal of the American Chemical Society
Another feature of our strategy is that it uses inter-
mediates that can be converted into acceptors with either
1. allyl alcohol
BF3⋅OEt2, 78%
O
O
OAc
NEt3
MeOH
BzO
BzO
BzO
O
O
O
O
O
1
2
3
4
5
6
7
8
BzO
AcO
BzO
HO
HO
OBz
OBz
OBz
2. AcCl, MeOH
87%
a free O5 or O6 hydroxyl group. Specifically, the acetyl
group from a compound such as 13 could be removed to
yield an acceptor with C6 hydroxyl group (e.g., compound
12). Compounds with free C5 hydroxyl groups (e.g., 10)
were envisioned to arise from benzoyl group migration
from the C5 to the C6 hydroxyl group. Thus, oligosaccha-
ride substrates could be assembled with either alternating
or consecutive regioisomeric glycosidic linkages using
only a small set of building blocks.
85%
15
16
17
BzO
NEt3
MeOH
SEt
SEt
BzO
BzO
1. TMSSEt, ZnI2, quant
2. AcCl, MeOH 79%
BzO
HO
HO
OBz
OBz
74%
18
19
BzO
N
N
S
HS
OAc
AcO
S
AcO
O
S
O
AcO
BF3⋅OEt2
9
AcO
AcO
OAc
OAc
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
98%
20
21
AcO
Route to oligosaccharide acceptors. The afore-
mentioned retrosynthetic strategy was applied to produce
the target monosaccharide building blocks (Scheme 1).
Both allyl glycoside nucleophile 16 with its C6 hydroxyl
group free and 17 with its C5 hydroxyl group free were
generated from acetoxy glycoside 15.11,20 Compound 15 was
converted into the protected allyl glycoside. Acid-
mediated acetate removal from the C6 hydroxyl group
provided 16, which is poised to form a 1,6-linkage. To ob-
tain glycosidic acceptor 17 with a free C5 hydroxyl group,
we sought to promote the migration of the benzoyl group
from O5 position to O6 position. The desired transfor-
mation occurred readily when 16 was treated with tri-
ethylamine in methanol to afford 17 in 85% yield.
Similarly, compound 15 was efficiently converted to
the S-ethyl thioglycosides 18 and 19. As anticipated, the S-
ethyl anomeric group of 15 survives the acidic conditions
for acetate removal. Compounds 18 and 19 can be used
first as glycosyl acceptors, while the thioethyl group re-
mains inert. The product of these reactions can function
subsequently as an electrophile in glycosylation. By ex-
ploiting acetate group removal from O6 positions under
mildly acidic conditions and the triethylamine-catalyzed
benzoyl group migration, we could access each of the
requisite monosaccharide building blocks.
We next examined the utility of these intermediates
for the production of trisaccharide and tetrasaccharide
substrates for GlfT2. The synthetic route to trisaccharide
4 is representative (Scheme 2). Glycosylation of 21 with
acceptor 19 provided disaccharide 22. Only the thiazolinyl
glycoside was activated with silver triflate, while the thi-
oethyl functionality of acceptor 22 was undisturbed.18,19
When the disaccharide 22 was exposed to catalytic silver
triflate and N-iodosuccinimide (NIS) it underwent cou-
pling to 16 to generate a trisaccharide with the natural
alternating β(1–6)Galf-β(1–5)Galf linkage pattern. We em-
ployed a cross metathesis reaction to append the 11-
phenoxy-1-undecene moiety. Ester group removal provid-
ed the target trisaccharide acceptor 4. In an analogous
manner, trisaccharides 5-7 were synthesized (See: Sup-
porting Information).
Scheme 1. Synthesis of monosaccharide building
blocks
HO
OR
HO
O
OR
HO
O
O
HO
HO
OUDP
O
HO
O
HO
O
HO
OH
O
OH
m
HO
OH
HO
OH
OH
OH
GlfT2
1
O
HO
OH
O
O
OH
HO
HO
HO
H
10
O
O
Our route for trisaccharide assembly exploits com-
mon building blocks, yet it is divergent. It could be ex-
ploited construct onger oligomers, as illustrated by the
production of tetrasaccharide 8 (Scheme 3). Compounds
16 and 23 were exposed to silver triflate to afford a disac-
charide bearing a C6 acetate. Acetate removal and ben-
zoyl group migration afforded glycosyl acceptor 24, which
is poised for an O5 glycosylation reaction. Glycosylation
with disaccharide 25 generated a tetrasaccharide that was
transformed into target tetrasaccharide 8. We previously
used a chemoenzymatic route to generate this compound
but it afforded only small quantities (1-5 mg) of the tetra-
saccharide.10 The synthetic route described herein is ro-
bust and scalable. Its use of orthogonally activatable thio-
glycosides and the application of the mild benzoyl group
migration strategy to synthesize both O5 and O6 glyco-
sylation acceptors enabled production of the tetrasaccha-
ride glycolipid 8 in quantities sufficient to characterize
GlfT2 elongation (vide infra).
O
O
P
AcHN
O
O
P
O
O
R =
O
O
HO
OH
Figure 2. GlfT2 mediates galactan formation by catalyzing
the addition of 30–40 Galf residues to oligosaccharide-lipid
precursor 1. GlfT2 is a bifunctional glycosyltransferase that
can generate alternating Galf-β(1–5)Galf and Galf-β(1–6)Galf
linkages.
Retrosynthesis. The retrosynthesis of tetrasaccha-
ride 8 illustrates our general strategy to assemble oli-
gogalactofuranosides that vary in sequence (Figure 4). To
maximize convergence, the tetrasaccharide was divided
into disaccharide units 9 and 10. These building blocks
could arise from monosaccharide precursors 11 and 12.
Thioethylglycoside 12 was selected as a key intermediate
for its ability to function either as an acceptor or as a do-
nor. Specifically, the thioethyl anomeric group is suffi-
ciently stable to allow a glycosylation reaction with more
reactive donor 11.18,19 The disaccharide 9 that would result
from the first glycosylation reaction could be induced to
undergo a subsequent reaction to build oligosaccharides.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 13
Oligosaccharide elongation catalyzed by GlfT2.
Scheme 2. Synthesis of trisaccharide 4
1
2
3
4
5
6
7
8
We evaluated trisaccharides 4-7 and tetrasaccharide 8 in
GlfT2-catalyzed polymerization reactions. To test their
utility as substrates, recombinant GlfT2 from M. tubercu-
losis was mixed with the synthetic oligosaccharides and
UDP-Galf.11 After 20 hours at room temperature, the crude
enzymatic reaction mixtures were analyzed by MALDI-
TOF mass spectrometry. The results were unexpected: All
of the substrates were elongated including non-natural
trisaccharides 6 and 7 (Figure 5). The polysaccharide
products that result have as many Galf residues as found
in the natural galactan (38-44 Galf residues).6 Given that
GlfT2 processes all sequences, the data highlight the en-
zyme’s promiscuity.
SEt
BzO
SEt
O
BzO
N
HO
O
S
O
AcO
AcO
OBz
19
O
O
S
OBz
OBz
BzO
AcO
AcO
AcO
AcO
OAc
OAc
AgOTf, 95%
21
22
O
BzO
O
1.
O
HO
9OPh
O
BzO
HO
HO
OBz
16
OH
β(1–6)
9
O
AgOTf, NIS, 93%
HO
O
β(1–5)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
O
HO
O
OH
4
(CH9)9OPh
2.
3.
HO
HO
OH
Grubbs I, 57%
OH
NaOMe, MeOH, 69%
closest to the lipid (reducing end). Both acceptors 4 and 7
terminate with a β(1-5) linkage, but 4, with a reducing end
β(1–6) linkage exhibited 2-fold smaller apparent Km and 4-
fold higher kcat than did acceptor 7 with a β(1–5) linkage at
the reducing end. Similarly, acceptor 6 with the internal
β(1–6) linkage exhibited 5-fold more favorable apparent
Km and 7-fold higher kcat than did acceptor 5 with the in-
ternal β(1–5) linkage. These results indicate that the rela-
tive positioning of the acceptor lipid substituent impacts
substrate binding to GlfT2 and subsequent catalytic turn-
over. A comparison of the kcat/Km values for all of the oli-
gosaccharide substrates indicates that the acceptors with
a β(1–6) near the reducing end (4, 6, 8) are superior sub-
strates to those with a corresponding β(1–5) linkage (5, 7).
Thus, the positioning or orientation of the saccharide
relative to the anomeric lipid substituent is more im-
portant than whether the sequence alternates.
Kinetic analyses of GlfT2 with oligosaccharide accep-
tors. The intensity of peaks in the mass spectra suggests
that some substrates (e.g., 4) are elongated more effi-
ciently than others (e.g., 5). Surprisingly, however, the
favorable substrates were not those that had “natural”
alternating sequences. To quantify the activity of GlfT2
with different substrates, we used a continuous coupled
assay to monitor the rates of GlfT2-catalyzed glycosyla-
tion of disaccharides 2 and 3, trisaccharides 4-7, and tetra-
saccharide 8. Because GlfT2 produces UDP upon the addi-
tion of each Galf residue to an acceptor, turnover can be
assessed by linking UDP production to NADH
oxidation.11,21-23 In agreement with previous results,11 GlfT2
exhibited a kinetic lag phase with disaccharide acceptors
2 and 3 (Figure 6). Though trisaccharide 7 exhibited a
minor lag phase, none was observed with trisaccharides 4-
6 nor with tetrasaccharide 8. As the products of GlfT1 are
tetrasaccharides and pentasaccharides,9 endogenous in-
termediates should be rapidly processed by GlfT2.
DISCUSSION
Data from our group and others indicate that GlfT2
has high fidelity for generating alternating β(1–5) and β(1-
6) linkages.15,16 The alternating linkages could be required
for subsequent steps in cell envelope biosynthesis or the
stability of the cell envelope itself. We therefore examined
whether GlfT2 can process acceptors with “mistakes” in
the acceptor sequence. To address this issue, we designed
oligogalactofuranose glycolipids 4-8 with different sac-
charide sequences (Figure 2). Synthetic trisaccharide ac-
We monitored the initial rate of UDP production at
different acceptor substrate concentrations to determine
apparent Km and kcat values for GlT2 (Table 1). Acceptors
bearing anomeric lipids with longer chains bound more
tightly to the enzyme and afforded higher catalyst turno-
ver number than those with an octyl glycoside at the re-
ducing end.13 These results are consistent with the influ-
ence of the anomeric lipid substitutent on acceptor
elongation by GlfT2.
Scheme 3. Synthesis of tetrasaccharide 8.
The specific saccharide sequence of each compound
influenced its ability to serve as an acceptor. Surprisingly,
the differences were not based on whether the acceptor
had a natural or non-natural sequence. For example, a
large difference in the kinetic parameters for trisaccha-
rides 4 and 5 was observed, but both have the natural al-
ternating linkages. Specifically, trisaccharide 4 with a β(1–
6) glycosidic linkage at the reducing end (i.e., near the
lipid) had more favorable apparent Km values (13±1 µM)
and higher catalyst turnover (kcat of 3.7±0.1 s-1) than the
trisaccharide 5 wiith a β(1–5) glycosidic linkage at the re-
ducing end (apparent Km of 190±86 µM and kcat of
0.72±0.15 s-1).
O
BzO
1.
O
O
BzO
BzO
HO
O
OBz
BzO
SEt
BzO
16
OBz
O
AgOTf, NIS, 86%
O
BzO
AcO
BzO
O
OBz
HO
2. AcCl, MeOH, 82%
3. NEt3, MeOH, 85%
23
24
OBz
BzO
1.
SEt
BzO
O
BzO
OBz
25
O
HO
β(1–6)
8
9 OPh
O
O
AcO
O
HO
AcO
OH
OAc
O
HO
β(1–5)
O
AcO
2.
AgOTf, NIS, 93%
O
HO
O
OH
(CH9)9OPh
Comparison between 4 and 7 or between 5 and 6 fur-
ther emphasizes the influence of the glycosidic linkage
HO
OH
Grubbs I, 75%
OH
β(1–6)
3. NaOMe, MeOH,98%
O
HO
O
HO
OH
HO
ACS Paragon Plus Environment

Page 5 of 13
Journal of the American Chemical Society
tion are the two central strategic features that made it
ceptor substrates for GlfT2 have been studied previously,
1
2
3
4
5
6
7
but the reported products of GlfT2-catalyzed reactions
did not approach the length of the native galactan.
Moreover, previous investigations focused on substrates
with natural alternating glycosidic linkages.13 We there-
fore developed a divergent approach to synthesize GlfT2
substrates that vary in sequence.
possible to rapidly access acceptors with either natural or
non-natural sequences to evaluate with GlfT2. Because
the synthetic route outlined relies on a small number of
reactions that can be repeated iteratively, we anticipate
that it could be used to access longer oligosaccharides
with any desired combination of glycosidic linkage pat-
tern.
13
Access to the oligogalactofuranose glycolipids was at-
tained by using orthogonally activatable glycosyl donors
and a facile method to quickly afford acceptors with dif-
ferent sequences of β(1–6) and β(1–5) Galf linkages. One
key to the strategy was its reliance on S-thiazolinyl and S-
ethyl glycosides as glycosyl donors with either high or
moderate reactivity.18,19 The stability of thioglycosides is
an advantage compared to more acid sensitive, reactive
intermediates, such as trichloroacetimidate donors.20,24
The second important feature of our strategy was the mi-
gration of the benzoyl group to the least hindered O6
position.
8
9
It is striking that each substrate tested, trisaccharides
4-7 and tetrasaccharide 8, was elongated by GlfT2. The
length of the product carbohydrate polymers was similar
to that of the endogenous mycobacterial galactan (30-40
Galf residues).6 Acceptor 4 afforded slightly shorter galac-
tan than the other three trisaccharide substrates. The
observed truncated oligosaccharide could result from
efficient consumption of the donor, which would limit the
available quantity of UDP-Galf as the polymerization pro-
ceeds.6 We previously demonstrated that the presence
and features of the lipid substituent of the initiating ac-
ceptor substrates influenced the product carbohydrate
polymer length, and we proposed a tethering mechanism
for length control by a template-independent polymer-
ase.11 The findings shown herein that all the oligosaccha-
ride acceptor substrates were elongated to the length of
the galactan regardless of the different carbohydrate mo-
tifs lends additional support to the model. Our results
also indicate that GlfT2 exhibits fidelity in the sequence
generated during elongation, yet the carbohydrate poly-
merase is far more promiscuous in processing substrates
of different sequences.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Acyl group migration is often viewed as a problem in
glycan synthesis, but our route exploits it. A previous syn-
thetic approach to galactofuranose derivatives employed a
selective protecting group strategy to afford an acceptor
nucleophile with free C5 hydroxyl group. The key inter-
mediate was synthesized by non-selective removal of pri-
mary acetyl group from C6 position, which resulted in a
concomitant benzoyl group migration in low to moderate
yield.20,24 We found a base-catalyzed benzoyl group mi-
gration provided the means to distinguish between adja-
cent hydroxyl groups in different steric environments.
The orthogonal glycosyl donors and the acyl group migra-
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 13
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5. MALDI-TOF mass spectra of GlfT2-catalyzed polymerization reactions with trisaccharide acceptors 4-7 and tetra-
saccharide acceptor 8. Synthetic oligosaccharides were incubated at rt with GlfT2 in the presence of UDP-Galf. After 20 h, the
resulting mixtures were analyzed by MALDI-TOF mass spectrometry. Peaks that correspond to m/z values of [M+Na]+ are
labeled with the value of n, which is the number of Galf residues in the product.
Previous studies demonstrated that GlfT2 can elon-
gate disaccharide acceptors with an arabinose residue at
the nonreducing end13 or even a glycolipid bearing a sin-
gle Galf residue.20 We now show that trisaccharides with
non-natural glycosidic linkages are elongated by GlfT2.
Once GlfT2 begins elongation, its sequence fidelity is
high.15 While this finding was unexpected, in mycobacte-
rial cells, the only available Galf glycolipid initiators are
those generated from the upstream galactofuranosyltrans-
ferase GlfT1.7,8,25 This enzyme exhibits high specificity for
either the native acceptor or structurally similar analogs.
Thus, the high specificity of GlfT1 should ensure that
GlfT2 can access the “correct” initial acceptor. The gener-
ation of polysaccharides with alternating linkages by
GlfT2 appears to occur during processive elongation of
the substrate. Specifically, as GlfT2 acts on the substrate,
the growing polymer remains bound to the enzyme.10 In
this way, the orientation of the nucleophilic nonreducing
end of the growing polymer could be poised to generate
the alternating linkages in the enzyme active site.14,17
We found previously that GlfT2 can act on both β(1–
6)- and β(1–5)-linked disaccharide acceptors 2 and 3 to
form products with similar degrees of polymerization.20
ACS Paragon Plus Environment

Page 7 of 13
Journal of the American Chemical Society
posed of a tetrasaccharide or pentasaccharide generated
There is, however, a significant difference in their initial
1
2
3
reaction rate; β(1–6)-linked disaccharide 2 has a reaction
rate 4-fold faster
by GlfT1. Thus, natural acceptor would fill the enzyme
subsites, so that GlfT2 could efficiently elongate the initial
acceptor.
4
5
6
7
8
9
We postulate that rapid processive elongation of en-
dogenous acceptors by GlfT2 is inextricably correlated
with high fidelity in formation of the alternating β(1–5)
and β(1–6) regioisomeric linkages. Our finding that GlfT2
can process a variety of Galf-containing acceptors of vary-
ing sequence indicate that substrate binding is not re-
sponsible for sequence selectivity. It is in the processive
formation of multiple glycosidic linkages that leads to
carbohydrate polymer of defined sequence.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
CONCLUSION
A robust and divergent synthetic approach has been
developed to assemble lipid-modified oligogalacto-
furanose glycans. Two key features, orthogonally activat-
able thioglycosides and a benzoyl group migration strate-
gy enabled the production of both alternating (natural)
and non-alternating (non-natural) regioisomeric oligo-
saccharide acceptors for the mycobacterial galacto-
furanosyltransferase GlfT2. Polymerization and kinetic
analyses of the synthetic acceptors revealed that the sac-
charide sequence of the initial acceptor influences glyco-
syltransferase elongation. All of the oligosaccharide ac-
ceptors were efficiently processed. These results support
the tethering model previously proposed for the length
Figure 6. Kinetics of GlfT2-catalyzed polymerization with
oligosaccharide acceptors 2–8. GlfT2 turnover was monitored
by coupling UDP production to NADH oxidation in a con-
tinuous assay.
(5.7±3.9 µM/min) than that observed for β(1–5)-linked
disaccharide 3 (1.3±0.4 µM/min, Figure 6). The data indi-
cate that the ability of an acceptor with a β(1–6) linkage
near the reducing end is a superior substrate for GlfT2
compared to a β(1–5)-linked acceptor. These results using
either perfectly alternating or mismatched acceptors
with β(1–6) near the reducing end indicates that larger
contribution to both acceptor binding and catalyst turn-
over than the glycosidic linkage at the non-reducing end.
These observations are consistent with the saturation
transfer difference (STD) NMR studies that indicate a
reducing end Galf residue in trisaccharide acceptors
makes more intimate contacts with GlfT2 than a non-
reducing end Galf residue.26 They suggest that the rela-
tive positioning of the lipid and the glycan is critical for
elongation.
Table 1. Apparent Km and kcat for oligosaccharides 2 and
4-8 determined by the continuous assay (Figure 6).
kcat/Km
(103 M-1.s-1)
Km
Acceptor
(μM)
66±2
13±2
kcat (s-1)
1.0±0.03
3.7±0.1
2 Galfβ(1-6)Galfa
4 Galf(1-5)Galf(1-6)Galf
5 Galf(1-6)Galf(1-5)Galf
15±1
280±20
190±86 0.72±0.15 3.8±1.9
5.3±1.3 160±80
0.87±0.01 30±1
2.9±0.1 220±20
6 Galf(1-6)Galf(1-6)Galf 34±16
7 Galf(1-5)Galf(1-5)Galf
8 Tetrasaccharide
29±1
13±1
We previously found GlfT2 exhibited a kinetic lag
phase with a disaccharide acceptor, while no such lag
occurred with a chemoenzymatically synthesized tetra-
saccharide.10 The data suggested that the lag phase arose
from the lower affinity of the disaccharide acceptor for
GlfT2. These findings led us to propose that GlfT2 has
subsites, and oligosaccharide acceptors that fill the sub-
sites would not exhibit a kinetic lag phase. This substrate
subsite model also has been proposed for other carbohy-
drate polymerases.27-33 It was, however, unknown how
many carbohydrate residues are necessary to fully occupy
the subsites of GlfT2. Our results indicate that neither
trisaccharides 4-6 nor tetrasaccharide 8 give rise to a ki-
netic lag phase. Kinetic analyses of trisaccharide 4 and
tetrasaccharide 8 revealed that both have similar apparent
Km values, and additional Galf residues did not lead to
large rate gains during the course of GlfT2-catalyzed gly-
cosylation. This observation is interesting because the
natural acceptor substrate for GlfT2 is a glycolipid com-
aData for compound 2 were obtained from reference 11.
control by a template-independent carbohydrate poly-
merase.11 The polymerization results revealed the se-
quence promiscuity of GlfT2 for its initial acceptor. More-
over, the results reveal that the nature of the glycosidic
linkage closest to the reducing end of the acceptor makes
a larger contribution to the efficiency of the catalysis by
GlfT2. These observations provide further support for the
importance of the lipid for catalysis. They also suggest
that inhibitors that promote GlfT2 release of substrate
could afford either shortened galactan or galactan in
which deviations from the alternating β(1–5) and β(1–6)
linkages occur. The possibility that such defects in the cell
envelope would be detrimental provides impetus to seek
GlfT2 inhibitors.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 13
(10) Levengood, M. R.; Splain, R. A.; Kiessling, L. L. J. Am.
1
2
3
4
5
6
7
8
Chem. Soc. 2011, 133, 12758-66.
(11) May, J. F.; Splain, R. A.; Brotschi, C.; Kiessling, L. L. Proc.
Natl. Acad. Sci. U.S.A. 2009, 106, 11851-6.
ASSOCIATED CONTENT
Supporting Information
(12) Kremer, L.; Dover, L. G.; Morehouse, C.; Hitchin, P.;
Everett, M.; Morris, H. R.; Dell, A.; Brennan, P. J.; McNeil,
M. R.; Flaherty, C.; Duncan, K.; Besra, G. S. J. Biol. Chem.
2001, 276, 26430-40.
(13) Rose, N. L.; Completo, G. C.; Lin, S.-J.; McNeil, M.; Palcic,
M. M.; Lowary, T. L. J. Am. Chem. Soc. 2006, 128, 6721-9.
(14) May, J. F.; Levengood, M. R.; Splain, R. A.; Brown, C. D.;
Kiessling, L. L. Biochemistry 2012, 51, 1148-59.
Routes employed for the syntheses of trisaccharide 5-7, ex-
perimental protocols for the preparation of all new com-
pounds, and spectral data for all new compounds. Experi-
mental procedures for protein production and enzyme assays
are also included. The Supporting Information is available
free of charge on the ACS Publications website.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(15) Brown, C. D.; Rusek, M. S.; Kiessling, L. L. J. Am. Chem.
Soc. 2012, 134, 6552-5.
(16) Poulin, M. B.; Zhou, R. K.; Lowary, T. L. Org. Biomol.
Chem. 2012, 10, 4074-87.
(17) Wheatley, R. W.; Zheng, R. B.; Richards, M. R.; Lowary, T.
L.; Ng, K. K. S. J. Biol. Chem. 2012, 287, 28132-43.
(18) Demchenko, A. V.; Pornsuriyasak, P.; De Meo, C.;
Malysheva, N. N. Angew. Chem. Int. Ed. 2004, 43, 3069-72.
(19) Pornsuriyasak, P.; Demchenko, A. V. Chem. Eur. J. 2006, 12,
6630-46.
(20) Splain, R. A.; Kiessling, L. L. Biorg. Med. Chem. 2010, 18,
3753-9.
(21) Gosselin, S.; Alhussaini, M.; Streiff, M. B.; Takabayashi, K.;
Palcic, M. M. Anal. Biochem. 1994, 220, 92-7.
(22) Chen, L.; Men, H.; Ha, S.; Ye, X.-Y.; Brunner, L.; Hu, Y.;
Walker, S. Biochemistry 2002, 41, 6824-33.
(23) Rose, N. L.; Zheng, R. B.; Pearcey, J.; Zhou, R.; Completo,
G. C.; Lowary, T. L. Carbohydr. Res. 2008, 343, 2130-9.
(24) Completo, G. C.; Lowary, T. L. J. Org. Chem. 2008, 73, 4513-
25.
(25) Alderwick, L. J.; Dover, L. G.; Veerapen, N.; Gurcha, S. S.;
Kremer, L.; Roper, D. L.; Pathak, A. K.; Reynolds, R. C.;
Besra, G. S. Protein Expr. Purif. 2008, 58, 332-41.
(26) Szczepina, M. G.; Zheng, R. B.; Completo, G. C.; Lowary, T.
L.; Pinto, B. M. Biorg. Med. Chem. 2010, 18, 5123-8.
(27) Schwartz, B.; Markwalder, J. A.; Seitz, S. P.; Wang, Y.;
Stein, R. L. Biochemistry 2002, 41, 12552-61.
(28) Chen, L.; Walker, D.; Sun, B.; Hu, Y.; Walker, S.; Kahne, D.
Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 5658-63.
(29) Jing, W.; DeAngelis, P. L. J. Biol. Chem. 2004, 279, 42345-9.
(30) Forsee, W. T.; Cartee, R. T.; Yother, J. J. Biol. Chem. 2006,
281, 6283-9.
(31) Vionnet, J.; Vann, W. F. Glycobiology 2007, 17, 735-43.
(32) Barrett, D.; Wang, T.-S. A.; Yuan, Y.; Zhang, Y.; Kahne, D.;
Walker, S. J. Biol. Chem. 2007, 282, 31964-71.
(33) Wang, T.-S. A.; Lupoli, T. J.; Sumida, Y.; Tsukamoto, H.;
Wu, Y.; Rebets, Y.; Kahne, D. E.; Walker, S. J. Am. Chem.
Soc. 2011, 133, 8528-30.
Author Information
The authors declare no competing financial interest.
ACKNOWLEDGMENT
NIH is acknowledged for financial support (AI063596). K.Y.
thanks JSPS Postdoctoral Fellowships for Research Abroad,
and R.A.S. thanks American Chemical Society Division of
Medicinal Chemistry Graduate Fellowship respectively for
financial supports. All compound characterizations were
performed at the University of Wisconsin-Madison Chemis-
try Instrument Center. The NMR facility is supported by the
NSF (CHE-9208463), the Mass Spectrometry Facility by the
NIH (NCRR 1S10RR024601-01) and NSF (CHE-9974839). We
thank A. Justen, V. Winton, G. Kincaid, and D. Wesener for
helpful comments on the manuscript.
REFERENCES
(1) Whitfield, C.; Trent, M. S. Annu. Rev. Biochem. 2014, 83,
99-128.
(2) Yother, J. Annu. Rev Microbiol 2011, 65, 563-81.
(3) Lombard, V.; Ramulu, H. G.; Drula, E.; Coutinho, P. M.;
Henrissat, B. Nucleic Acids Res. 2014, 42, D490-D5.
(4) Cantarel, B. L.; Coutinho, P. M.; Rancurel, C.; Bernard, T.;
Lombard, V.; Henrissat, B. Nucleic Acids Res. 2009, 37,
D233-D8.
(5) Sassetti, C. M.; Boyd, D. H.; Rubin, E. J. Mol. Microbiol.
2003, 48, 77-84.
(6) Besra, G. S.; Khoo, K.-H.; McNeil, M. R.; Dell, A.; Morris, H.
R.; Brennan, P. J. Biochemistry 1995, 34, 4257-66.
(7) Beláňová, M.; Dianišková, P.; Brennan, P. J.; Completo, G.
C.; Rose, N. L.; Lowary, T. L.; Mikušová, K. J. Bacteriol.
2008, 190, 1141-5.
(8) Mikušová, K.; Beláňová, M.; Korduláková, J.; Honda, K.;
McNeil, M. R.; Mahapatra, S.; Crick, D. C.; Brennan, P. J. J.
Bacteriol. 2006, 188, 6592-8.
(9) Martinez Farias, M. A.; Kincaid, V. A.; Annamalai, V. R.;
Kiessling, L. L. J. Am. Chem. Soc. 2014, 136, 8492-5.
ACS Paragon Plus Environment

Page 9 of 13
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
TOC Graphic
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
9
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 13
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1
207x242mm (300 x 300 DPI)
ACS Paragon Plus Environment

Page 11 of 13
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5
219x237mm (300 x 300 DPI)
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 12 of 13
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 6
156x108mm (300 x 300 DPI)
ACS Paragon Plus Environment

Page 13 of 13
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
TOC
47x27mm (300 x 300 DPI)
ACS Paragon Plus Environment