The α-Thioglycoligase Derived from a GH89 α-N-Acetylglucosaminidase Synthesises α-N-Acetylglucosamine-Based Glycosides of Biomedical Interest

Article abstract of DOI:10.1002/adsc.201601091

We report here on the preparation of a novel α-thioglycoligase that can be used for the fast and efficient synthesis of α-N-acetylglucosamine-based glycosides. Using the α-N-acetyl-glucosaminidase from Clostridium perfringens of family GH89 (according to the Carbohydrate Active Enzymes classification) as starting point, we prepared mutants in the acid/base residue glutamic acid 483 that were found to have different synthetic efficiencies (maximal yields >80% after 24 h) in the presence of an activated donor and suitable acceptors. The synthetic potential of the Glu483 alanine mutant as an α-thioglycoligase – only the third biocatalyst with this stereospecificity reported to date, and the first selective for the N-acetylglucosamine moiety – was demonstrated by producing for the first time N-acetyl-α-d-glucosaminyl azide and N-acetylglucosamine α-thioglycosides in high yields. To showcase the application of such compounds, we show that they offer the wild-type CpGH89 protection from thermal unfolding, demonstrating their potential as pharmacological chaperones for the treatment of mucopolysaccharidosis IIIB (Sanfilippo syndrome). (Figure presented.).

Full text of DOI:10.1002/adsc.201601091

FULL PAPERS
DOI: 10.1002/adsc.201601091
The a-Thioglycoligase Derived from a GH89 a-N-Acetylglucosa-
minidase Synthesises a-N-Acetylglucosamine-Based Glycosides of
Biomedical Interest
Ndivhuwo Olga Tshililo,+b Andrea Strazzulli,+a Beatrice Cobucci-Ponzano,a
Luisa Maurelli,a Roberta Iacono,a Emiliano Bedini,c Maria Michela Corsaro,c
Erick Strauss,b,* and Marco Moraccia,d,*
a
Institute of Biosciences and Bioresources – National Research Council of Italy, Via P. Castellino 111, 80131, Naples, Italy
Fax : (+39)-081-613-2646; phone: (+39)-081-613-2271; e-mail: marco.moracci@ibbr.cnr.it
Department of Biochemistry, Stellenbosch University, Private Bag X1, 7602 Matieland, South Africa
b
Fax : (+27)-21-808-5863; phone: (+27)-21-808-5866; e-mail: estrauss@sun.ac.za
Department of Chemical Sciences, University of Naples “Federico II”, Complesso Universitario di Monte S. Angelo, Via
c
Cupa Nuova Cinthia 21, 80126 Napoli, Italy
Department of Biology, University of Naples “Federico II”, Complesso Universitario di Monte S. Angelo, Via Cupa
d
Nuova Cinthia 21, 80126 Napoli, Italy
+
These authors contributed equally to the work
Received: October 4, 2016; Revised: November 27, 2016; Published online: && &&, 0000
Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/adsc.201601091.
Abstract: We report here on the preparation of for the N-acetylglucosamine moiety – was demon-
a novel a-thioglycoligase that can be used for the strated by producing for the first time N-acetyl-a-d-
fast and efficient synthesis of a-N-acetylglucosamine- glucosaminyl azide and N-acetylglucosACTHNUGRTENUNGamine a-thio-
based glycosides. Using the a-N-acetyl-glucosamini- glycosides in high yields. To showcase the application
dase from Clostridium perfringens of family GH89 of such compounds, we show that they offer the
(according to the Carbohydrate Active Enzymes wild-type CpGH89 protection from thermal unfold-
classification) as starting point, we prepared mutants ing, demonstrating their potential as pharmacological
in the acid/base residue glutamic acid 483 that were chaperones for the treatment of mucopolysacchari-
found to have different synthetic efficiencies (maxi- dosis IIIB (Sanfilippo syndrome).
mal yields >80% after 24 h) in the presence of an
activated donor and suitable acceptors. The synthetic Keywords: carbohydrate active enzymes; chemo-en-
potential of the Glu483 alanine mutant as an a-thio- zymatic synthesis; glycoside hydrolases; pharmaco-
glycoligase – only the third biocatalyst with this ste- logical chaperones; reaction mechanism of glycoside
reospecificity reported to date, and the first selective hydrolases
Introduction
a positive role in pharmacological chaperone therapy
(PCT) used in lysosomal storage disorders (LSDs)
The inhibition of glycoside hydrolases (GHs) is alone or in combination with the enzyme replacement
a proven strategy for the study of these interesting en- therapy (ERT), helping to maintain the target
zymes, the discovery of GH-binding molecules, and enzyme in its folded, active form.[3]
treatment or study of a variety of metabolic, genetic,
The use of GH inhibitors as pharmacological chap-
and infective diseases.[1] For example, the drugs migli- erones has become the main treatment option for
tol and voglibose are used to treat type 2 diabetes LSDs where ERT is not possible. A case in point is
based on their competitive inhibition of the a-glucosi- mucopolysaccharidosis (MPS), a very serious autoso-
dase enzyme responsible for the breakdown of starch mal recessive lysosomal storage disorder caused by
and smaller oligosaccharides to glucose in the small the impaired function of enzymes involved in the deg-
intestine.[2] However, GH inhibitors can also play radation of heparan sulphate, leading to accumulation
Adv. Synth. Catal. 0000, 000, 0 – 0
1
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

FULL PAPERS
asc.wiley-vch.de
of undegraded glycosaminoglycans in lysosomes. In
particular, defective mutations of the human gene en-
coding for an a-N-acetyl-glucosaminidase (NAGLU)
causes MPS IIIB, or Sanfilippo syndrome, a tragic
rare disease leading to severe neurological disorders.[4]
There are no current treatments for MPS IIIB, as
gene therapy is in its infancy, and ERT cannot be
adopted because the pharmacological enzyme cannot
cross the blood–brain barrier to replace the defective
enzyme in the nervous system. However, small-mole-
cule chaperones can interact with mutant enzymes to
favour their correct conformation and enhance their
stability.[5] In this manner pharmacological chaperone
therapy (PCT) becomes a viable alternative to ERT
in cases where suitable molecules with high affinity
for the target enzyme can be identified.
Whether for use as inhibitors or in PCT, the ability
of small molecules to selectively bind to their respec-
tive targets in the presence of other enzymes with
similar substrates would be highly beneficial from
a therapeutic perspective. One strategy whereby this
can be achieved is by using compounds that have the
same anomeric configuration as the natural substrates
of the targeted enzymes. However, in the case of a-
GHs, this is not easily accomplished. Remarkably, the
chemical synthesis of a-linked glycosides of especially
N-acetyl-glucosamine (GlcNAc) is generally less
tractable than that of their b-counterparts. This has
severely hampered the discovery and the production
in good yields of inhibitors that are selective for a-N-
acetyl-glucosaminidases, and that can serve as leads
for the development of molecules with the potential
for use in a clinical setting. Consequently, those com-
pounds that have been shown to inhibit a-N-acetyl-
glucosaminidase enzymes are in fact repurposed in-
hibitors of b-N-acetyl-glucosaminidases, and do not
show selectivity at the anomeric position.[6] Since such
a lack of specificity is an undesirable pharmacological
feature, the search for a-specific N-acetyl-glucosam-
Scheme 1. Mechanisms of a-glycoside hydrolase enzymes
and of the biocatalysts that are derived from them. A:
Mechanism of a retaining a-GH showing the active site resi-
dues acting as nucleophile and acid/base respectively. R can
be a sugar or an alcohol. B: Mechanism of an a-glycosyn-
thase (a-GS) derived from the a-GH shown in panel A
after exchange of its active site nucleophile for a non-nucle-
ophilic residue. Such enzymes can use two types of sub-
strates: if the leaving group is in the a-configuration (–OR1
in green; with R1 usually 2-, 4-nitrophenyl, or 2,4-dinitro-
phenyl), a small external nucleophile (Nu in blue) such as
azide or formate is added to allow the reaction with an ac-
ceptor (R2 can be a sugar or an alcohol) to take place, while
substrates with small leaving groups (green X=fluoride or
azide) in the b-configuration can react directly. In both cases
the product is an a-glycoside. C: Mechanism of an a-thiogly-
coligase (a-TGL) derived from the a-GH shown in panel A
by exchange of its active site acid/base residue for a non-
protic/non-nucleophilic residue. In this case the leaving
group (X in green) can be 2-, 4-nitrophenyl, 2,4-dinitrophen-
yl, fluoride or azide, and a thiolate (in orange) is used as ac-
ceptor, which acts as nucleophile to form an a-thioglycoside
product.
ACHTUNGTRENNUNGinidase inhibitors continues. In this regard the discov-
ery of more efficient methods for the reliable synthe-
sis of a-linked glycosides would be extremely useful,
and would enable the production of new compounds
for testing in such a context.
To address some of the constraints associated with
the chemical synthesis of oligosaccharides, such as
protection/deprotection strategies and selective acti-
vation of the anomeric centre, the use of engineered
retaining GH enzymes (referring to the anomeric con-
figuration of the product relative to the substrate)
was pioneered.[7] This is achieved by mutating either
the active site nucleophile or acid/base residues that
are required for normal GH catalysis (Scheme 1A) to cosynthases (GSs), mutation of the active site nucleo-
non-nucleophilic residues, thereby removing the en- phile is countered by the use of an activated donor
zymesꢁ ability to hydrolyse glycosidic bonds. To con- substrate with an appropriate leaving group at the
vert the mutant enzymes into biocatalysts, two ap- anomeric position (Scheme 1B). If the leaving group
proaches are followed: in the first, which leads to gly- is in the same anomeric configuration as the substrate
Adv. Synth. Catal. 0000, 000, 0 – 0
2
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

FULL PAPERS
asc.wiley-vch.de
of the wild-type enzyme an external nucleophile such glycoligases capable of producing glycoside analogues
as sodium azide or sodium formate is added to chemi- when used with suitable acceptors. The potential of
cally rescue the loss of the active site residue.[8] With such compounds as pharmacological chaperones was
some small leaving groups (e.g., fluoride and azide) in subsequently explored by determining the effects of
a configuration opposite to that of the substrate, these novel products on parental CpGH89. Taken to-
chemical rescue is not required (for reviews see gether, these observations demonstrated that the first
ref.[9]). In either case, the product formed accumulates a-TGL is capable of efficiently producing a-N-acetyl-
due to the impaired hydrolytic activity of the mutated glucosaminide-containing thioglycosides and N-acetyl-
GH enzyme. In the second approach, which was de- a-d-glucosaminyl azide (a-GlcNAc-N3), a compound
veloped specifically to produce non-hydrolysable thio- that is not commercially available, yet holds promise
glycosides, the exchange of the acid/base residue of for the use in bio-orthogonal click chemistry reac-
GHs is again countered by the use of an activated tions. Such compounds can find broad application in
donor with the same anomeric configuration as the biomedicine for the synthesis of inhibitors and lead
substrate but, in this instance, also by an acceptor compounds of a-GlcNAc-active enzymes.
containing a thiol serving as nucleophile (Scheme 1C).
The thioglycoside product accumulates in the reac-
tion, as it is resistant toward enzymatic hydrolysis.[10] Results and Discussion
Such enzymes are therefore dubbed thioglycoligases
(TGL), and their thioglycoside products usually serve Preparation of a-N-Acetyl-glucosaminosynthases
as GH inhibitors.
Despite the potency of these approaches – which CpGH89 was selected for investigation as potential
recently led to an engineered b-hexosaminidase with chassis to produce mutant enzymes able to synthesise
improved transglycosidase activity[11] – so far a-GSs N-acetyl-a-d-glucosaminide
(a-GlcNAc)
derived
are limited to a-fucosynthases [from families GH29 products. The gene encoding for CpGH89 was pro-
and GH95 according to the classification of carbohy- duced as a truncated version of the wild type, which
drate active enzymes (CAZy, www.cazy.org[12]], a-glu- comprised residues 26–916 as reported previously.[6]
cosynthases
(GH31),
and
a-galactosynthases The enzyme purified to homogeneity by a single-step
(GH36).[8a,9a,13] In striking contrast, thioglycoligases purification by immobilised metal affinity chromatog-
are even less common. To the best of our knowledge, raphy showed the steady state kinetic constants listed
only two a-TGLs have been produced so far, a thio- in Table 1. Previous studies demonstrated that the
glucoligase and a thioxyloligase, both belonging to enzyme followed a double-displacement retaining
family GH31.[14]
mechanism and that the residues glutamic acid 601
Considering the need for selective inhibitors and (Glu601) and 483 (Glu483) were the nucleophile and
molecular chaperones of the medicinally relevant a- the acid/base of the reaction, respectively, with no evi-
N-acetyl-glucosaminidase enzymes, we decided to dence of the participation in catalysis of the N-acetyl
embark on the modification of the catalytic residues group of the substrate (Figure 1A and B).[6] To test if
of an a-N-acetyl-glucosaminidase to produce a novel CpGH89 could be converted into an a-GS by follow-
a-GS and a-TGL able to synthesise a-N-acetyl-gluco- ing the approach that was demonstrated to be suc-
saminide-containing glycosides. In CAZy, a-N-acetyl- cessful for GHs from several families[7] the nucleo-
glucosaminidases (EC 3.2.1.50) are reported only in phile Glu601 was exchanged for non-nucleophilic resi-
family GH89,[12] in which the enzyme from the bacte- dues producing the alanine, glycine, and serine mu-
rium Clostridium perfringens (CpGH89) has been tants (E601A/G/S). These mutant enzymes had no de-
identified as a useful model system of the human tectable activity on 2-nitrophenyl-N-acetyl-a-d-
NAGLU. Studies on the 3D-structure of CpGH89 glucosaminide (2NP-a-GlcNAc) and 2,4-dinitrophen-
demonstrated that the enzymes belonging to this yl-N-acetyl-a-d-glucosaminide (2,4DNP-a-GlcNAc)
family followed the classical retaining reaction mecha- as donor substrates. This is in agreement with previ-
nism as opposed to b-hexosaminidases from families ous studies on CpGH89[6] and with the 103-fold reduc-
GH18, 20, 25, 56, 84, and 85, which hydrolyse sub- tion of specific activity observed in other retaining
strates containing an N-acetyl (acetamido) or N-gly- glycoside hydrolases mutated in the nucleophile of
colyl group at the 2-position acting as nucleophile and the reaction.[17] However, remarkably, attempts to
forming an oxazolinium ion intermediate.[15] In addi- chemically rescue the enzymatic activity of the mu-
tion, mutations giving rise to MPS IIIB were also tants acting on 2- and 2,4DNP-a-GlcNAc at 25 and
identified from a homology model of NAGLU.[6,16] We 378C by using high concentrations (up to 1M) of ex-
show here that CpGH89 mutants in the nucleophile ternal nucleophiles (sodium azide, sodium formate,
of the reaction were recalcitrant to act as a-glycosyn- sodium chloride and sodium fluoride) were all unsuc-
thases, but that modification of the acid/base with cessful and remained so even with the use of excess
non-nucleophilic amino acids led to efficient a-thio- of enzyme, different reaction temperatures, buffer/pH
Adv. Synth. Catal. 0000, 000, 0 – 0
3
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

FULL PAPERS
asc.wiley-vch.de
Table 1. Steady-state kinetic parameters of wild type CpGH89 and its Glu483 mutants.
CpGH89 mutant (substrate)
2NP-a-GlcNAc
KM [mM]
kcat [sꢀ1]
378C
kcat/KM [sꢀ1 mMꢀ1]
258C
378C
258C
258C
378C
wild type
E483Q
E483A
E483S
E483G
4NP-a-GlcNAc
416ꢁ56
11ꢁ5
190ꢁ10
10ꢁ1
50ꢁ3
62ꢁ4
68ꢁ5
2.31ꢁ0.11
23ꢁ0.3
5.6
3.1
1.7
1.9
0.3
121
0.033ꢁ0.001
0.076ꢁ0.001
0.064ꢁ0.014
0.022ꢁ0.001
0.029ꢁ0.001
0.051ꢁ0.001
0.093ꢁ0.002
0.046ꢁ0.001
2.90
1.02
1.49
0.67
46ꢁ5
34ꢁ5
79ꢁ14
wild type
E483Q
E483A
E483S
E483G
2,4DNP-a-GlcNAc
1049ꢁ79
237ꢁ19
459ꢁ43
443ꢁ25
820ꢁ220
1800ꢁ300
200ꢁ20
800ꢁ40
700ꢁ70
600ꢁ70
0.17ꢁ0.01
0.35ꢁ0.01
0.16
0.19
0.0062ꢁ0.0001
0.0130ꢁ0.0004
0.0158ꢁ0.0004
0.0027ꢁ0.0002
0.0065ꢁ0.0001
0.0146ꢁ0.0002
0.0240ꢁ0.0078
0.0482ꢁ0.0163
0.026
0.028
0.036
0.003
0.028
0.017
0.035
0.080
wild type
E483Q
E483A
3.2ꢁ0.3
nd[a]
nd
6.2ꢁ0.6
nd
nd
5.5ꢁ0.1
nd
nd
9.9ꢁ0.2
nd
nd
1718
nd
nd
1596
nd
nd
E483S
nd
nd
nd
nd
nd
nd
E483G
nd
nd
nd
nd
nd
nd
2,4DNP-a-GlcNAc+250 mM NaN3
E483Q
E483A
E483S
E483G
2.4ꢁ0.3
0.7ꢁ0.1
1.1ꢁ0.2
1.7ꢁ0.4
1.3ꢁ0.1
1.8ꢁ0.3
1.8ꢁ0.2
3.2ꢁ0.4
1.5ꢁ0.1
1.1ꢁ0.1
2.1ꢁ0.1
1.0ꢁ0.1
1.6ꢁ0.2
2.3ꢁ1.2
2.4ꢁ0.7
1.9ꢁ0.9
616
1289
1307
1326
591
1642
2191
582
[a]
nd=not detectable.
systems, substrate acceptors, and donor/acceptor type), on 4-nitrophenyl-N-acetyl-a-D-glucosaminide
ratios (data not shown). In addition, no synthesis was (4NP-a-GlcNAc, up to 70-fold less) and on 2,4DNP-
observed by using N-acetyl-b-d-glucosaminyl azide a-GlcNAc (up to 30-fold less) (Table 1), confirming
(b-GlcNAc-N3) as donor, in an attempt to apply the what was previously reported for the CpGH89–
same strategy that was successful for other a-GSs E483A mutant on 4NP-a-GlcNAc as substrate.[6] We
(Scheme 1B).[8a,9a] We confirmed that this lack of ac- confirmed that this lack of activity is not due to struc-
tivity was not due to changes in the stability (Fig- ture-based defects introduced by the mutation
ure 2A) or structure (Supporting Information, Fig- through circular dichroism (CD) analysis of the pro-
ure S1A) of the E601A/G/S mutants. More likely, the teinꢁs secondary structure elements (Supporting Infor-
shape of the CpGH89 active site prevents entry and/ mation, Figure S1B) which produced mutants with
or placement of the acceptor substrates in catalytical- very low enzymatic activity on 2NP-a-GlcNAc (up to
ly relevant positions (Figure 1B). This renders the 780-fold less than the wild type), 4-nitrophenyl-N-
enzyme resistant to conversion to an a-GS, as has acetyl-a-d-glucosaminide and by comparative analysis
been reported for similar cases.[18] We therefore of their temperature melting curves (Figure 2B). No
turned our attention to the conversion of CpGH89 significant differences could be observed, indicating
into a novel a-TGL.
that the loss of activity can be directly attributed to
the mutation.
Preparation of a-N-Acetyl-thioglycosaminoligases
Chemical Rescue of the Acid/Base Mutants with
a-TGLs are mutants in which the general acid/base NaN3
catalytic residue of the enzyme has been replaced by
a non-nucleophilic amino acid.[10] The mutant, in the To test if the inactivation caused by mutation of the
presence of a sugar donor with an excellent leaving general acid/base residue could be chemically res-
group and suitable acceptors with a thiol goup, cataly- cued, we used sodium azide (NaN3) as external nucle-
ses the formation of S-glycosidic linkages in high ophile. In an initial test, the rate of the release of 2,4-
yields. The replacement of the general acid/base of dinitrophenolate from the more reactive donor
CpGH89 with non-acidic residues (E483A/S/Q/G) 2,4DNP-a-GlcNAc at 378C was measured in the pres-
produced mutants with very low enzymatic activity on ence of increasing concentrations of sodium azide as
2NP-a-GlcNAc (up to 780-fold less than the wild catalysed by the Glu483 mutants (Figure 3A). The re-
Adv. Synth. Catal. 0000, 000, 0 – 0
4
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

FULL PAPERS
asc.wiley-vch.de
Figure 1. CpGH89 catalytic mechanism and active site
structure. A: Mechanism of the initial hydrolytic step cata-
lyced by CpGH89, showing the roles of the active site nucle-
ophile (Glu601) and acid/base (Glu483) residues. B: Sub-
strate binding pocket of CpGH89 shown as a section
through the solvent-accessible surface with GlcNAc (stick
structure with carbon atoms in green) bound. The anomeric
carbon that is the point of initial attack by Glu601 (nucleo-
phile) is indicated by an arrow. Both Glu601 and Glu483 are
shown as stick structures with carbon atoms in grey.
Figure 2. CpGH89 structure-based stability. A: Melting
curves of CpGH89 and its Glu601 mutants measured by fol-
lowing the change in CD absorption at 220 nm by heating
from 258C to 858C in intervals of 18C each minute. B: As
for A, but for the Glu483 mutants.
GlcNAc-N3, the closest commercially available ana-
sults clearly show that addition of azide restored the logue of the expected product a-N-acetyl-d-glucosa-
mutantsꢁ activity, with the CpGH89–E483A and mide azide (a-GlcNAc-N3) (data not shown). These
E483S mutants showing the best performance under same conditions were subsequently used to determine
these conditions. This finding was confirmed by deter- the synthetic efficiency of all the mutants (calculated
mining the steady-state kinetic constants of the mu- as the % products formed from the donor assuming
tants acting on 2,4DNP-a-GlcNAc in the presence of the amount of GlcNAc available in 2,4DNP-a-
250 mM sodium azide (Table 1). TLC analysis of GlcNAc as 100%) by analysis with high performance
these reactions incubated for 16 h showed that the anionic exchange chromatography using pulsed am-
donor was converted to a glycoside with a retention perometric detection (HPAEC-PAD). The efficiency
factor (Rf) very similar to that of the control b- was found to be remarkably high for all the mutants,
Adv. Synth. Catal. 0000, 000, 0 – 0
5
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

FULL PAPERS
asc.wiley-vch.de
These high yields prompted us to reinvestigate the
activity of the mutants at either 25 or 378C with all
three donors and either low (5 mM) or high
(100 mM) concentrations of NaN3. TLC analysis
showed spots corresponding to a-GlcNAc-N3 for all
the mutants but only when 100 mM NaN3 was used
with either 2NP-a-GlcNAc or 2,4DNP-a-GlcNAc as
donors; we found that no reaction took place with
4NP-a-GlcNAc (Figure 4). Interestingly, the most
productive reactions were found to be those using
either 2NP-a-GlcNAc as donor at 378C or 2,4DNP-a-
GlcNAc at 258C. The results also show that when
2NP-a-GlcNAc was used as donor at 378C or
2,4DNP-a-GlcNAc at either temperature, hydrolysis
(i.e., formation of GlcNAc) was the predominant re-
action being catalysed. These findings show that
donor reactivity (as mediated by the nature of the
leaving group or the reaction temperature) is the pri-
mary consideration for activity. However, when reac-
tivity is too high, or in the absence of sufficient ac-
ceptor (NaN3), donor hydrolysis becomes a significant
problem. Optimised product formation would there-
fore entail finding a balance between donor reactivity
in the presence of a specific acceptor, and hydrolysis.
Taking these results together, we concluded that
the Glu483 mutants were promising candidates for
application as a-TGLs and set out to demonstrate this
using a range of thiol acceptors.
Evaluating a-TGL Activity using Selected Thiol
Acceptors
Encouraged by the chemical rescue of the activity of
the CpGH89 acid/base mutants using sodium azide,
we decided to perform a comparative evaluation of
their a-TGL activity. This was achieved by measuring
their synthetic efficiencies when incubated at 378C
using 2NP-a-GlcNAc as donor (the preferred choice,
since it is commercially available and showed high
levels of product formation in the reactions with
NaN3). As acceptors, the thiolated esters methyl or
ethyl thioglycolate were used, since the pKa values of
their thiol groups (~8.08 at 258C for methyl thiogly-
colate – from the National Center for Biotechnology
Information. PubChem Compound Database; CID=
16907, https://pubchem.ncbi.nlm.nih.gov/compound/
16907, accessed on August 5, 2016) would ensure that
at least half of the molecules in the reaction mixture
would be in the more nucleophilic thiolate form at
pH 8.0, i.e., the conditions under which the tests were
Figure 3. Thioglycoligase activity of CpGH89-derived acid/
base mutants. A: Chemical rescue of the activity of the
CpGH89–Glu483 mutants in the presence of increasing
amounts of sodium azide. The activity was measured at
378C using 2,4DNP-a-GlcNAc as donor under standard con-
ditions. B: Synthetic efficiency of the CpGH89–Glu483 mu-
tants using 2NP-a-GlcNAc as donor and either methyl thio-
glycolate (Me) or ethyl thioglycolate (Et) as acceptors. The
reaction mixtures were analysed by HPAEC-PAD after 24
and 48 h to quantify the amount of product formed. The
synthetic efficiency of the mutants was determined assuming
the amount of GlcNAc available in 2NP-a-GlcNAc as
100%. Bars indicate the average value obtained from dupli-
cate experiments; the error bars indicate the standard devia-
tion.
being 70.3ꢁ0.1%, 78.6ꢁ0.1%, 75.0ꢁ1.5% and 81.4ꢁ performed. Mutants were incubated using >20-fold
0.8% for E483A/S/Q/G, respectively, while the blank molar excess of the donor, and the amount of thiogly-
mixture, containing all the reagents but the enzyme, coside product formed was measured by HPAEC-
revealed only GlcNAc resulting from the spontaneous PAD analysis of the reaction mixtures after 24 h and
hydrolysis of the donor (Supporting Information, Fig- 48 h. The yield of converted product in each case was
ure S2).
calculated based on the amount of GlcNAc available
Adv. Synth. Catal. 0000, 000, 0 – 0
6
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

FULL PAPERS
asc.wiley-vch.de
Figure 4. Products of the chemical rescue of CpGH89–E483 mutants from different donors as analysed by thin layer chro-
matography (TLC). b-GlcNAc-N3, GlcNAc, and the respective donor are loaded in the first three lanes of each TLC plate
as references. Three sets of reaction mixtures were analysed in each case, with either no NaN3, or 5 mM or 100 mM NaN3
present, and with each set containing wild-type enzyme (WT), E483Q (Q), E483A (A), E483S (S) mutant, and a blank reac-
tion without enzyme (B). Reaction mixtures were incubated at either 258C or 378C, as indicated. Spots that were UV-visible
are indicated by dotted circles; other spots were visualised by staining with 10% sulphuric acid in methanol, followed by in-
cubation at 1508C for 15 min. A: Using 2NP-a-GlcNAc as donor at 258C. B: Using 2NP-a-GlcNAc as donor at 378C. C:
Using 4NP-a-GlcNAc as donor at 258C. D: Using 4NP-a-GlcNAc as donor at 378C. E: Using 2,4DNP-a-GlcNAc as donor
at 258C. F: Using 2,4DNP-a-GlcNAc as donor at 378C.
in 2NP-a-GlcNAc set as 100%. Blanks showed only efficient a-TGL, with a yield of 81.8ꢁ10.2% after
unreacted donor (Supporting Information, Figure S3). only 24 h (Figure 3B). With the larger ethyl thioglyco-
The results show that in the case of methyl thioglyco- late acceptor the efficiency of the mutants was re-
late, the CpGH89–E483A mutant is by far the most duced, and in this case the E483S mutant gave the
Adv. Synth. Catal. 0000, 000, 0 – 0
7
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

FULL PAPERS
asc.wiley-vch.de
best results (64.6ꢁ1.70%, also after only 24 h). The by means of four sequential steps that removed the
E483Q mutant was found to show the least efficient protein, buffer and nitrophenol side-product, respec-
activity, whereas the activity of the E483G mutant tively, before the final product was purified by flash
corresponded to those of the alanine and serine mu- chromatography on silica. The purified products were
1
tants with ethyl and methyl thioglycolate, respectively. analysed by H and 13C NMR and HR-MS analysis to
In the light of these results, we chose the CpGH89– confirm their structures, including the expected a-con-
E483A and E483S mutants to investigate their activity figuration of the newly-formed glycosidic bond. The
as a-TGLs using a variety of thiol acceptors at con- preparative scale yields were also determined based
centrations of between 2.5 mM and 100 mM. Reac- on the amount of donor used. These values, summar-
tions were performed at 378C using 5 mM 2NP-a- ised in Table 2, show that the small scale synthetic ef-
GlcNAc as donor in 100 mM sodium phosphate ficiencies were excellent predictors of the mutants
buffer. Based on previous reports that the pKa of the preparative scale abilities, with the CpGH89–E483A
thiol group is an important consideration in its ability mutant also giving preparative yields in excess of
to act as acceptor in a thioglycoligase reaction,[19] we 80%. The lower synthetic yields observed in some
chose the thiols to span a range of pKa values, and cases are due to less favourable transglycosylation/hy-
also performed the reactions at various pHs to evalu- drolysis ratios (compare the columns showing donor
ate if the predominance of the thiolate anion is a pre- hydrolysis and product formation for each enzyme
requisite for the TGL reaction to proceed. Reactions and substrate in Table 2) and, in some cases to diffi-
were evaluated by TLC and scored for product for- culties encountered in the purification of the products.
mation and hydrolysis as a side-reaction (Table 2). However, it is worth mentioning that the overall effi-
The results show that a-TGL activity is usually ob- ciency of the chemo-enzymatic approach shown here
served with thiols that have pKa values of 8 and is much better than the classical chemical synthesis
lower, i.e., thiols that would be at least 50% ionised strategy that needs many more steps. Moreover, the
at pH 8.0. We found that promoting the ionisation of yields obtained for the compounds prepared in this
thiols with higher pKa values was not practically pos- study are comparable to the overall yield of 35% re-
sible, since the hydrolysis of the donor becomes a sig- ported for one of these compounds (entry 1, Table 2)
nificant side reaction at higher pH values. In fact, al- prepared in six steps by traditional synthetic meth-
though the 2NP-a-GlcNAc donor (as an acetal) is ex- ods.[21] These findings confirm that the CpGH89–
pected to show enhanced stability at increased pH, E483A mutant is an excellent a-TGL biocatalyst. In
the observed side reaction likely is the result of the addition, this mutant also provides a means for the
higher concentration of HOꢀ by acting as nucleophile preparation of a-GlcNAc-N3, a compound that is cur-
on the enzyme-GlcNAc intermediate. Indeed, we ob- rently not commercially available and, according to
served that for the thiosugar methyl 4-deoxy-4-thio- the chemical database SciFinder, has never been de-
glucoside and for cyclohexanethiol, both of which scribed before. Nevertheless, it yet has multiple po-
have pKa values of 9.5 and above, GlcNAc was the tential uses – including as azide substrate in the ex-
major product formed. These findings indicated that ceedingly versatile “click” reactions such as cop-
the CpGH89–E483A and E483S mutants have excel- per(I)-catalysed- and strain-promoted azide–alkyne
lent a-TGL activity, but only if the thiol acceptor is cycloadditions (CuAAC and SPAAC).[22] Taken to-
carefully chosen.
gether, the CpGH89–E483A mutant opens routes to
the facile preparation of a variety of a-linked S-glyco-
sides and triazoles of high potential value that were
previously not synthetically tractable.
CpGH89–E483A/S Mutants as Preparative Scale a-
TGL Biocatalysts
The excellent synthetic efficiency of the CpGH89– Effect of the a-TGL Products on CpGH89
E483A mutant motivated us to determine if its a-
TGL activity would also translate to reactions on Since GH enzymes cannot hydrolyse the S-glycosidic
a preparative (multimilligram) scale, and allow the bond, we next sought to demonstrate the potential of
synthesis and purification of a-GlcNAc-N3 (using the biocatalytically-prepared GlcNAc a-thioglycosides
NaN3 as external nucleophile) or a-GlcNAc thiogly- as inhibitors of N-acyl-glucosaminidases. This was
cosides (using thiol acceptors with pKa values of 8 or done by evaluating the ability of the compounds to
lower). Towards this end, we performed biocatalysis bind to and inhibit the GH activity of wild-type
reactions using CpGH89–E483A and/or E483S mu- CpGH89. We included in these tests commercially ob-
tants using ~0.08 mmol 2NP-a-GlcNAc as donor and tained b-GlcNAc-N3 to demonstrate the importance
a 20-fold molar excess of acceptor at 378C. Reaction of the configuration of the glycosidic bond, as well as
progress was monitored by TLC, and once being GlcNAc-a(1!4)-(methyl
4-deoxy-4-thioglucoside)
judged as complete, the reaction product was purified [GlcNAc-S-(Me4-thioGlc)], the expected product of
Adv. Synth. Catal. 0000, 000, 0 – 0
8
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

FULL PAPERS
asc.wiley-vch.de
Table 2. The a-thioglycoligase activity of CpGH89–E483A and E483S with various thiol acceptors or sodium
azide, and 2NP-a-GlcNAc as donor.
[a]
Values taken from the National Center for Biotechnology Information (NCBI) PubChem Compound Database
(https://pubchem.ncbi.nlm.nih.gov/), or from ref.[20] The pKa value for methyl 4-deoxy-4-thioglucoside was cal-
culated using ChemAxonꢁs pKa predictor.
[b]
Relative extent of donor hydrolysis based on the intensity of the spots representing GlcNAc and 2NP-a-
GlcNAc as visualised from the TLC analyses of the reaction mixtures.
Relative extent of product formation based on the intensity of the anticipated product spot (identified by com-
[c]
parison with a standard, if available) as visualised from the TLC analyses of the reaction mixtures.
Yield of purified product prepared from preparative scale biocatalytic transformations, based on the amount of
[d]
2NP-a-GlcNAc donor used as starting material.
Adv. Synth. Catal. 0000, 000, 0 – 0
9
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

FULL PAPERS
asc.wiley-vch.de
the TGL reaction with methyl 4-deoxy-4-thiogluco-
side as acceptor. While we were not able to prepare
this compound through biocatalysis, we prepared it by
chemical synthesis to act as standard for the TGL
evaluation tests, and it was therefore available for
testing.
To quantify the inhibitory effect we measured the
specific activity at standard conditions of wild-type
CpGH89 at 600 mM of each compound. This concen-
tration was chosen based on this amount of O-(2-acet-
amido-2-deoxy-d-glucopyranosylidene)amino-N-phe-
nylcarbamate (PUGNAc), a known competitive in-
hibitor of CpGH89.[6] leading to 20% residual activity.
Among the a-TGL products tested, the products of
the reactions with methyl and ethyl thioglycolate
(GlcNAc-S-CH2CO2Me and GlcNAc-S-CH2CO2Et,
respectively) showed the highest inhibition with 79%
and 62% of residual activity, respectively (Figure 5A).
In order to evaluate if the products obtained by
chemical rescue and by the TGL approach could pro-
tect wild-type CpGH89 from unfolding by binding to
the enzyme, we performed differential scanning fluo-
rimetry (DSF) experiments in the 25–948C tempera-
ture range and in the presence of 10 mM of each
product, the highest concentration used for allosteric
pharmaceutical chaperones (PCs).[23] As positive con-
trols we used PUGNAc (0.2 mM) and two other
known N-acetylglucosaminidase inhibitors: 2-acetami-
do-1,2-dideoxynojirimycin (2AcDNJ, 0.1 mM) and 6-
acetamido-6-deoxycastanospermine
(6AcCAS,
2.8 mM), i.e., at concentrations 30-fold higher than
their respective Ki or Kd values.[6] The thermal stabili-
ty analysis (Figure 5B) revealed that the three inhibi-
tors 6AcCAS, PUGNAc and 2AcDNJ increased the
melting temperature (Tm) of wild-type CpGH89 by
6.218C, 3.768C, and 2.118C, respectively (Figure 5C).
Among the products of the a-TGL, the best stabilis-
ing effect was produced by a-GlcNAc-N3 (DTm =
4.028C), while b-GlcNAc-N3 did not show any stabili-
sation of the wild-type CpGH89. This finding demon-
strates the importance of the configuration of the gly-
cosidic bond for binding, and as a potential factor in
Figure 5. Effect of the a-thioglycoligase products on
CpGH89 activity and stability. A: Effect by known inhibi-
tors and a-TGL products on the activity of wild type
CpGH89. Bars indicate the average value obtained from du-
plicate experiments; the error bars indicate the standard de-
viation. B: Denaturation curves followed by differential
scanning fluorimetry (DSF) of wild type CpGH89 in the ab-
sence of ligands (control) and in the presence of the
6AcCAS inhibitor and a-GlcNAc-N3 are shown as case ex-
amples. C: The increasing melting temperatures (Tm) of wild
type CpGH89 in the presence of the indicated ligands as de-
termined by DSF. Bars indicate the average value obtained
from triplicate experiments; the error bars indicate the stan-
dard deviation.
Adv. Synth. Catal. 0000, 000, 0 – 0
10
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

FULL PAPERS
asc.wiley-vch.de
was prepared by chemical synthesis using a previously pub-
lished protocol.[26] Methyl 4-deoxy-4-thioglucoside and
GlcNAc-S-(Me4-thioGlc) were prepared by chemical syn-
thesis as outlined in the Supporting Information. General
chemicals, reagents, and media were purchased from Sigma–
Aldrich or Merck Chemicals (Darmstadt, Germany) and
were of the highest purity. Solvents used for reactions were
Chromasolv HPLC grade solvents (Sigma–Aldrich), and the
ethyl acetate (EtOAc), methanol (MeOH) and water used
for purification were purchased from Merck Chemicals. Pro-
tein purifications were conducted on an ꢂkta Prime system
(GE Healthcare). Spectrophotometric assays were per-
formed on a Cary 100 Scan spectrophotometer (Varian, Aus-
tralia), coupled with a thermally controlled Peltier system.
High performance anionic exchange chromatography with
pulsed amperometric detection (HPAEC-PAD) was per-
formed on a Thermofisher Scientific system equipped with
a CarboPac PA200 Analytical (2ꢃ205 mm) column. Solid
phase extraction (SPE) columns (Phenomenex Strataꢄ)
were obtained from Separations (Cape Town, South Africa).
establishing selectivity. GlcNAc, the product of the
hydrolysis reaction, also significantly improved the
thermal stability of the enzyme (DTm = +3.388C)
compared to galactose (DTm = +0.708C), indicating
that the stabilising effect of a carbohydrate-derivative
molecule could be directly related to the donor ꢀ1
subsite.[16] It is worth noting that a-GlcNAc-N3, which
is the second-best PC after 6AcCAS, showed no in-
hibition of CpGH89 at 0.6 mM concentration (Fig-
ure 5A), a characteristic highly recommended in
second generation PCs as they can promote the cor-
rect folding of the defective target enzyme with no
detrimental effect on its activity in the lysosome.[3,24]
In addition, as already stated above, a-GlcNAc-N3
can be used as azide substrate in “click” reactions,
thus, showing interesting potential as lead for novel
a-linked PC molecules to test in pharmacological
chaperone therapies for MPS IIIB.
Cloning and Mutagenesis of CpGH89 Wild Type and
Mutants
Conclusions
Stable a-linked analogues of GlcNAc glycosides hold
great potential for the development of new selective
therapies for a range of diseases, for example, com-
pounds that can act as inhibitors of the enzymes in-
volved in the biosynthesis of essential cofactors in
pathogenic organisms,.[25] or by stabilising the struc-
tures of defective mutant variants of key metabolic
enzymes whose deficiency lead to genetic diseases.[4,5]
However, such compounds are not easily obtained
through chemical means due to the requirement for
multistep synthesis with several protection/deprotec-
tion protocols. Here we have demonstrated that a-
GlcNAc-N3 and several a-GlcNAc thioglycosides can
be obtained by preparative scale biocatalysis using an
engineered mutant of the a-N-acetyl-glucosaminidase
from Clostridium perfringens (CpGH89) in which its
catalytically essential acid/base residue (Glu483) has
been replaced by alanine. The development of this a-
N-acetyl-thioglycosaminoligase lowers the barrier to
the access of a-linked GlcNAc glycoside analogues
significantly, and should facilitate the investigation of
such compounds as therapeutic agents. In this study
we highlight the potential scope for such compounds
by showing their ability to inhibit or stabilise the
wild-type form of CpGH89, a proxy for the NAGLU
enzyme whose deficiency causes mucopolysaccharido-
sis IIIB, or Sanfilippo syndrome.
The synthetic CpGH89 wild type and E601A mutant genes
were custom-made by GeneArt Gene Synthesis (Thermo-
fisher Scientific) in pET28a vector, with N-terminal 6ꢃHis
tag. The mutants E601G/S and E483G were prepared from
the pET28a-CpGH89wt by using the GeneTailor Site-Di-
rected Mutagenesis system (Invitrogen) with the following
synthetic oligonucleotides (PRIMM, Milan, Italy).
E601G_fwd:
5’-TATTGGTATTACACCGGGCGCCATTAATAC
CAATCCGC-3’
E601G_rev: 5’-GCGGATTGGTATTAATGGCGCCCGGTGTAATAC
CAATA-3’
E601S_fwd:
5’-TGGGTATTGGTATTACACCGTCAGCCATTAA
TAC-3’
E601S_rev: 5’-CGGTGTAATACCAATACCCACCATATGTTC-3’
E483G_fwd: 5’-ATGGTGTGGATCCGTTTCATGGTGGTGGTAA
TACCGGTGATCT-3’
E483G_rev: 5’-AGATCACCGGTATTACCACCACCATGAAACG
GATCCACACCAT-3’
E483A_fwd: 5’-GGATCCGTTTCATGCAGGTGGTAATACCGG-3’
E483A_rev: 5’-CCGGTATTACCACCTGCATGAAACGGATCC-3’
E483S_fwd: 5’-GGATCCGTTTCATTCAGGTGGTAATACCGG-3’
E483S_rev: 5’-CCGGTATTACCACCTGAATGAAACGGATCC-3’
E483Q_fwd: 5’-GGATCCGTTTCATCAAGGTGGTAATACCGG-3’
E483Q _rev: 5’-CCGGTATTACCACCTTGATGAAACGGATCC-
3’
PCR-generated constructs were verified by sequencing.
Expression and Purification of CpGH89 Wild Type
and Its mutants
CpGH89 wild type and mutants were expressed in E. coli
strain JM109 (DE3) or BL21*(DE3). The cells transformed
with the appropriate construct were grown at 378C in 2 L of
Luria-Bertani (LB) broth supplemented with kanamycin
(50 mg.mLꢀ1). Gene expression was induced by the addition
of 1 mM IPTG when the culture reached an OD600 of 1.0.
Growth was allowed to proceed for 16 h, and cells were har-
vested by centrifugation at 5000 g. The resulting cell pellet
Experimental Section
General Materials and Methods
The donor substrates 2- and 4-NP-a-GlcNAc were obtained
from Carbosynth (Compton, U.K.) while 2,4DNP-a-GlcNAc
Adv. Synth. Catal. 0000, 000, 0 – 0
11
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

FULL PAPERS
asc.wiley-vch.de
was resuspended in 50 mM sodium phosphate buffer,
pH 8.0, 300 mM NaCl and 1% TRITON-X100 with a ratio
of 5 mLgꢀ1 cells, followed by incubation at 378C for 1 h
with 20 mg of lysozyme (Fluka) and 25 Ugꢀ1 cell of DNAseI
(SIGMA). Cells were lysed by 5 sonication cycles and cell
debris was removed by centrifugation at 12000 g for 30 min.
The cell free extract (CFE) was loaded on a HisTrap FF
column (GE-Healthcare) equilibrated with 50 mM sodium
phosphate buffer, pH 8.0, 300 mM NaCl (Buffer A). After
an initial wash-step (20 column volumes) with Buffer A, the
protein was eluted at 250 mM imidazole in Buffer A. The
active fractions were pooled, dialysed against 20 mM
sodium phosphate buffer, pH 7.3, 150 mM NaCl (PBS
buffer) and stored at 48C. The protein concentration was
determined with the Bradford assay (Bio-Rad). After this
procedure the enzymes were more than 95% pure by SDS-
PAGE, with an approximate final yield of 20 mgLꢀ1 of cul-
ture.
enzyme(s) were incubated at the same conditions and used
to subtract the spontaneous hydrolysis of the donor(s).
Chemical Rescue with Sodium Azide
The chemically rescued hydrolase activity of the four
CpGH89 acid/base mutants (E483A/S/Q/G) was measured
in 100 mM sodium phosphate buffer at pH 7.3 using 5 mM
2,4DNP-a-GlcNAc and increasing the concentrations of
sodium azide (between 50 mM and 1.5M) at 378C. The spe-
cific activity was measured as described for the hydrolytic
activity assay above. Blank mixtures, containing all the re-
agents but the enzyme(s) were incubated at the same condi-
tions and used to subtract the spontaneous hydrolysis of the
donor(s). Aliquots of the reaction mixtures (20 mL) incubat-
ed for 16 h were analysed by thin layer chromatography
(TLC) as described previously.[27] Briefly, reaction mixtures
were loaded on a silica gel 60 F254 plate by using EtOAc/
MeOH/H2O (70:20:10 v/v) as eluent and were detected by
exposure to 10% sulphuric acid in methanol, followed by
charring at 1708C. The efficiency of each mutant was evalu-
ated by HPAEC-PAD analysis at 358C using a flow rate
0.4 mLminꢀ1 and 16 mM sodium hydroxide (NaOH) as
eluent. For quantification, a calibration curve was created
with three different amounts (0.25, 0.75 and 1.5 nmol) of the
a-GlcNAc-N3, and cellobiose (0.75 nmol) was included as in-
ternal standard. The synthetic efficiency of the mutants was
determined assuming the amount of GlcNAc available in
2,4DNP-a-GlcNAc as 100%. All the samples were analysed
in duplicate.
Comparative Structural Analysis
CD spectroscopy and thermal analysis were conducted using
a Photophysics Chirascan-plus spectrometer at 258C. The
samples contained 0.5 mgmLꢀ1 protein in 100 mM sodium
phosphate buffer (pH 7.3). The optimum wavelength for
analysis of thermal stability was determined by a CD ab-
sorption scan from 200–300 nm at 258C. The thermal stabili-
ty experiments were conducted by heating the samples from
258C to 1008C and measuring the changes in CD absorption
at 220 nm.
Evaluating the a-Thioglycoligase Activity of
CpGH89–E483A/S with Selected Thiols
CpGH89 Hydrolytic Activity Assay
Hydrolytic activity assays for CpGH89 wild type and mu-
tants at 258C were conducted using a microplate-based con-
tinuous spectrophotometric assay measuring the release of
the nitrophenolate from the respective donor that was used.
For assays conducted at 258C, reactions were performed in
triplicate in a final reaction volume of 250 mL. The assays
contained 2 mM enzyme, 100 mM sodium phosphate buffer
(pH 7.3), and substrate concentrations in a range varying be-
tween 0–2.5 mM. The liberation of the nitrophenolate ion
was monitored at 400 nm (e=1.8 mMꢀ1 cmꢀ1 and
13.3 mMꢀ1 cmꢀ1 for 2- and 4-nitrophenolate, respectively)
for 10 min. Assays conducted at 378C were performed in
a final reaction volume of 1 mL. The assays typically con-
tained 220 nM of wild type and 1.3–4.11 mM of mutant en-
zymes, in 100 mM sodium phosphate buffer (pH 7.3), and
substrate concentrations in a range varying between 0.5 mM
and 5 mM. The liberation of the nitrophenolate ion was
The a-thioglycoligase activities of the CpGH89–E483A/S
mutants were evaluated using 2NP-a-GlcNAc as the donor.
All thioglycoligase reactions were performed using 100 mM
sodium phosphate buffer, with pH ranging from 5.8–8.0
(values indicated in Table 2), 1.33 mM enzyme and 5 mM
2NP-a-GlcNAc in a final volume of 200 mL. The final thiol
concentration ranged from 2.5 mM to 100 mM. The reac-
tions were incubated at 378C, and the mixtures were ana-
lysed by TLC with ethyl acetate/methanol/water (EtOAc/
MeOH/H2O) (70:20:10 v/v) using UV light to detect the ni-
trophenol group, and staining with 10% sulphuric acid in
methanol, followed by incubation at 1508C for 15 min, to
visualise all other components. For each reaction 4ꢃ5 mL of
the solution were spotted on the TLC.
Synthetic Efficiency of CpGH89–E483 Mutants with
Methyl and Ethyl Thioglycolate as Acceptor
monitored
at
405 nm
(e=3.63 mMꢀ1 cmꢀ1,
16.39 mMꢀ1 cmꢀ1and 11.92 mMꢀ1 cmꢀ1 for 2-, 4- and 2,4-dini-
trophenol, respectively) for 10 min. The rates of product for-
mation were calculated using linear regression analysis of
the initial linear portions of the progress curves; these were
plotted against the substrate concentration to obtain the ac-
tivity profiles. The kinetic parameters kcat and KM were ob-
tained by performing a least squares fitting of the data to
the Michealis–Menten equation. One unit of enzyme activity
was defined as the amount of enzyme catalysing the hydrol-
ysis of 1 mmol of substrate in 1 min under the conditions de-
scribed. Blank mixtures, containing all the reagents but the
To determine the synthetic efficiency of the four CpGH89
acid/base mutants (CpGH89–E483A/S/Q/G), reactions were
performed at 378C using 100 mM sodium phosphate buffer
(pH 8.0), 1.33 mM enzyme, 5 mM 2NP-a-GlcNAc as donor
and 106 mM methyl or ethyl thioglycolate as acceptor (sup-
plied with 17.9 mM DTT) in a final volume of 200 mL. At
time intervals (0–48 h) aliquots (10 mL) were withdrawn, di-
luted 1:100 with water and analysed by HPAEC-PAD at
358C using a flow rate of 0.4 mL.minꢀ1 and the following
elution program: [segment 1] 0–12’ isocratic 16 mM NaOH,
[segment 2] 12–42’ up to 96 mM NaOH, [segment 3] 42–52’
Adv. Synth. Catal. 0000, 000, 0 – 0
12
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

FULL PAPERS
asc.wiley-vch.de
step to 160 mM NaOH. The total amount of the expected
thioglycoside product was determined by using a calibration
curve with three different amounts (0.53, 1.06 and 2.12 nmol
of a-GlcNAc-S-CH2CO2Me and 0.25, 0.75 and 1.5 nmol of
a-GlcNAc-S-CH2CO2Et, respectively) of the product, and
cellobiose (0.73 nmol) as internal standard. The authentic
compounds used to obtain the calibration curves were ob-
tained by preparative scale biocatalysis as outlined below.
The synthetic efficiency of the mutants was determined as-
suming the amount of GlcNAc available in 2NP-a-GlcNAc
as 100%. All the samples were analysed in duplicate.
formed at 0.28Cminꢀ1 and the SYPRO Orange fluorescence
was normalised to maximum fluorescence value within each
scan to obtain relative fluorescence. All the measurements
were performed in triplicate. The Tm value represents the
flection point of the transition curve, as described by the
Boltzmann equation.[28]
Acknowledgements
This work was supported by the project “Esobiologia e ambi-
enti estremi: dalla Chimica delle Molecole alla Biologia degli
Estremofili – ECMB” n. 2014-026-R.0 of the Italian Space
Agency. N.T.M., E.S., and M.M. were supported by an Exec-
utive programme of scientific and technological co-operation
between the Italian Republic and the Republic of South
Africa for the years 2011–2013, entitled: “Novel a-glycosyn-
thases for the preparation of Mycothiol analogues as poten-
tial antituberculosis agents”, funded by the Directorate for
Cultural Promotion and Cooperation, Ministry of Foreign
Affairs, Italy and the National Research Foundation of South
Africa.
Preparative Scale Biocatalysis
Preparative scale a-TGL reactions were carried out in
15 mL of 100 mM sodium phosphate buffer (pH 8.0). 2NP-
a-GlcNAc (0.026 g, 0.076 mmol) was added as the donor,
followed by addition of thiol (20 equiv.) or 100 mM sodium
azide as the acceptor. DTT (0.160 mM, 0.370 mg,
0.0024 mmol,) was added as reducing agent to prevent disul-
phide bond formation in the case of the thiol acceptors. The
reaction was initiated by addition of the thioglycoligase cat-
alyst (1.33 mM of either CpGH89–E483A or CpGH89–
E483S) to the reaction mixture followed by incubation at
378C. Reaction progress was monitored and analysed by
TLC (EtOAc/MeOH/H2O; 70:20:10 v/v). Upon completion
(usually after overnight incubation), the reaction was fil-
tered through a syringe filter (13 mm GHP Acrodiscꢄ filter
with 0.2 mm GHP membrane) to remove the protein. The fil-
tered mixture was loaded onto an SPE column (Phenomen-
ex Strata C18-U, 55 mm, 70 A, 10 g/60 mL) and the crude
product was eluted with methanol. The fractions with the
crude product were pooled together and concentrated by
rotary evaporation. In the case of reactions with acceptors
that do not proceed close to completion (as in the case of
thiophenol), the crude product was treated with wild-type
CpGH89 for 72 h under standard conditions to hydrolyse
any residual donor that may be present. The enzyme was
subsequently removed as described above. The crude prod-
uct was purified by flash column chromatography on silica
(EtOAc/MeOH/H2O; 70:20:10 v/v). Any residual nitrophe-
nol was removed by passing the purified product through
Phenomenex Strataꢄ Phenyl (55mm 70 ꢅ) SPE column, elut-
ing with water. The final purified products were obtained
following by freeze drying for three days to remove any
References
[1] a) H. C. Dorfmueller, D. M. F. van Aalten, FEBS Lett.
2010, 584, 694–700; b) G. J. Davies, S. J. Williams, Bio-
chem. Soc. Trans. 2016, 44, 79–87.
[2] S. R. Joshi, E. Standl, N. Tong, P. Shah, S. Kalra, R.
Rathod, Expert Opin. Pharmacother. 2015, 16, 1959–
1981.
[3] G. Parenti, M. Moracci, S. Fecarotta, G. Andria, Future
Med. Chem. 2014, 6, 1031–1045.
[4] G. Yogalingam, J. J. Hopwood, Hum. Mutat. 2001, 18,
264–281.
[5] G. Parenti, EMBO Mol. Med. 2009, 1, 268–279.
[6] E. Ficko-Blean, K. A. Stubbs, O. Nemirovsky, D. J. Vo-
cadlo, A. B. Boraston, Proc. Natl. Acad. Sci. USA 2008,
105, 6560–6565.
[7] a) L. F. Mackenzie, Q. P. Wang, R. A. J. Warren, S. G.
Withers, J. Am. Chem. Soc. 1998, 120, 5583–5584; b) M.
Moracci, A. Trincone, G. Perugino, M. Ciaramella, M.
Rossi, Biochemistry 1998, 37, 17262–17270; c) C. Malet,
A. Planas, FEBS Lett. 1998, 440, 208–212.
[8] a) B. Cobucci-Ponzano, F. Conte, E. Bedini, M. M. Cor-
saro, M. Parrilli, G. Sulzenbacher, A. Lipski, F. Dal
Piaz, L. Lepore, M. Rossi, M. Moracci, Chem. Biol.
2009, 16, 1097–1108; b) I. Park, H. Lee, J. Cha, Biotech-
nol. Lett. 2014, 36, 789–796.
water, and were fully characterised by H NMR, 13C NMR
and MS-ESI or HR-MS (see the Supporting Information for
data).
1
CpGH98 Wild Type Inhibition assay and Thermal
Stability
The inhibitory effects of the different compounds (0.6 mM
each) were assayed by measuring the specific activity of
CpGH89 wild type on 2NP-a-GlcNAc in 100 mM phosphate
buffer (pH 7.3) at 378C as reported above. Thermal stability
of CpGH89 wild type was measured by DSF as previously
reported[23] with some modifications. Briefly, CpGH89 wild
type (2.5 mg) was incubated in citrate phosphate buffer and
SYPRO Orange dye 5X, in a total volume of 25 mL in the
presence of the different compounds under test. Samples
were heated from 25 to 948C in a Real-Time Light Cycler
(Bio-Rad, Milan, Italy). Thermal stability scans were per-
[9] a) B. Cobucci-Ponzano, C. Zorzetti, A. Strazzulli, S. Ca-
rillo, E. Bedini, M. M. Corsaro, D. A. Comfort, R. M.
Kelly, M. Rossi, M. Moracci, Glycobiology 2011, 21,
448–456; b) Z. Armstrong, S. G. Withers, Biopolymers
2013, 99, 666–674.
[10] M. Jahn, J. Marles, R. A. J. Warren, S. G. Withers,
Angew. Chem. 2003, 115, 366; Angew. Chem. Int. Ed.
2003, 42, 352.
[11] K. Slamova, J. Krejzova, P. Marhol, L. Kalachova, N.
Kulik, H. Pelantova, J. Cvacka, V. Kren, Adv. Synth.
Catal. 2015, 357, 1941–1950.
Adv. Synth. Catal. 0000, 000, 0 – 0
13
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

FULL PAPERS
asc.wiley-vch.de
[12] V. Lombard, H. Golaconda Ramulu, E. Drula, P. M.
Coutinho, B. Henrissat, Nucleic Acids Res. 2014, 42,
D490–495.
[20] a) U. Srinivasan, P. A. Mieyal, J. J. Mieyal, Biochemistry
1997, 36, 3199–3206; b) S. Zhang, J. Baker, P. Pulay, J.
Phys. Chem. A 2010, 114, 432–442.
[21] D. W. Gammon, D. J. Steenkamp, V. Mavumengwana,
M. J. Marakalala, T. T. Mudzunga, R. Hunter, M. Mu-
nyololo, Bioorg. Med. Chem. 2010, 18, 2501–2514.
[22] a) J. E. Hein, V. V. Fokin, Chem. Soc. Rev. 2010, 39,
1302–1315; b) N. E. Mbua, J. Guo, M. A. Wolfert, R.
Steet, G. J. Boons, ChemBioChem 2011, 12, 1912–1921.
[23] C. Porto, M. C. Ferrara, M. Meli, E. Acampora, V.
Avolio, M. Rosa, B. Cobucci-Ponzano, G. Colombo, M.
Moracci, G. Andria, G. Parenti, Mol. Ther. 2012, 20,
2201–2211.
[24] R. E. Boyd, G. Lee, P. Rybczynski, E. R. Benjamin, R.
Khanna, B. A. Wustman, K. J. Valenzano, J. Med.
Chem. 2013, 56, 2705–2725.
[25] V. K. Jothivasan, C. J. Hamilton, Nat. Prod. Rep. 2008,
25, 1091–1117.
[13] a) M. Okuyama, H. Mori, K. Watanabe, A. Kimura, S.
Chiba, Biosci. Biotechnol. Biochem. 2002, 66, 928–933;
b) J. Wada, Y. Honda, M. Nagae, R. Kato, S. Wakatsu-
ki, T. Katayama, H. Taniguchi, H. Kumagai, M. Kitao-
ka, K. Yamamoto, FEBS Lett. 2008, 582, 3739–3743.
[14] Y. W. Kim, A. L. Lovering, H. M. Chen, T. Kantner,
L. P. McIntosh, N. C. J. Strynadka, S. G. Withers, J. Am.
Chem. Soc. 2006, 128, 2202–2203.
[15] D. J. Vocadlo, S. G. Withers, Biochemistry 2005, 44,
12809–12818.
[16] M. Fujita, A. Tsuchida, A. Hirata, N. Kobayashi, K.
Goto, K. Osumi, Y. Hirose, J. Nakayama, T. Yamanoi,
H. Ashida, M. Mizuno, J. Biol. Chem. 2011, 286, 6479–
6489.
[17] D. L. Zechel, S. G. Withers, Curr. Opin. Chem. Biol.
2001, 5, 643–649.
[18] V. M. A. Ducros, C. A. Tarling, D. L. Zechel, A. M.
Brzozowski, T. P. Frandsen, I. von Ossowski, M. Schu-
lein, S. G. Withers, G. J. Davies, Chem. Biol. 2003, 10,
619–628.
[19] M. Almendros, D. Danalev, M. Francois-Heude, P.
Loyer, L. Legentil, C. Nugier-Chauvin, R. Daniellou, V.
Ferrieres, Org. Biomol. Chem. 2011, 9, 8371–8378.
[26] H. M. Chen, S. G. Withers, ChemBioChem 2007, 8,
719–722.
[27] B. Cobucci-Ponzano, V. Aurilia, G. Riccio, B. Henrissat,
P. M. Coutinho, A. Strazzulli, A. Padula, M. M. Cor-
saro, G. Pieretti, G. Pocsfalvi, I. Fiume, R. Cannio, M.
Rossi, M. Moracci, J. Biol. Chem. 2010, 285, 20691–
20703.
[28] F. H. Niesen, H. Berglund, M. Vedadi, Nat. Protoc.
2007, 2, 2212–2221.
Adv. Synth. Catal. 0000, 000, 0 – 0
14
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!

FULL PAPERS
The a-Thioglycoligase Derived from a GH89 a-N-
Acetylglucosaminidase Synthesises a-N-Acetylglucosamine-
Based Glycosides of Biomedical Interest
15
Adv. Synth. Catal. 2017, 359, 1 – 15
Ndivhuwo Olga Tshililo, Andrea Strazzulli,
Beatrice Cobucci-Ponzano, Luisa Maurelli, Roberta Iacono,
Emiliano Bedini, Maria Michela Corsaro, Erick Strauss,*
Marco Moracci*
Adv. Synth. Catal. 0000, 000, 0 – 0
15
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÞÞ
These are not the final page numbers!