Chemoenzymatic synthesis of arabinomannan (AM) glycoconjugates as potential vaccines for tuberculosis

Article abstract of DOI:10.1016/j.ejmech.2020.112578

Mycobacteria infection resulting in tuberculosis (TB) is one of the top ten leading causes of death worldwide in 2018, and lipoarabinomannan (LAM) has been confirmed to be the most important antigenic polysaccharide on the TB cell surface. In this study, a convenient synthetic method has been developed for synthesizing three branched oligosaccharides derived from LAM, in which a core building block was prepared by enzymatic hydrolysis in flow chemistry with excellent yield. After several steps of glycosylations, the obtained oligosaccharides were conjugated with recombinant human serum albumin (rHSA) and the ex-vivo ELISA tests were performed using serum obtained from several TB-infected patients, in order to evaluate the affinity of the glycoconjugate products for the human LAM-antibodies. The evaluation results are positive, especially compound 21 that exhibited excellent activity which could be considered as a lead compound for the future development of a new glycoconjugated vaccine against TB.

Full text of DOI:10.1016/j.ejmech.2020.112578

Journal Pre-proof
Chemoenzymatic Synthesis of Arabinomannane (AM) Glycoconjugates as Potential
Vaccines for Tuberculosis
Zhihao Li, Teodora Bavaro, Sara Tengattini, Roberta Bernardini, Maurizio Mattei,
Francesca Annunziata, Richard B. Cole, Changping Zheng, Matthieu Sollogoub,
Lucia Tamborini, Marco Terreni, Yongmin Zhang
PII:
S0223-5234(20)30550-X
DOI:
https://doi.org/10.1016/j.ejmech.2020.112578
EJMECH 112578
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 6 May 2020
Revised Date: 11 June 2020
Accepted Date: 12 June 2020
Please cite this article as: Z. Li, T. Bavaro, S. Tengattini, R. Bernardini, M. Mattei, F. Annunziata, R.B.
Cole, C. Zheng, M. Sollogoub, L. Tamborini, M. Terreni, Y. Zhang, Chemoenzymatic Synthesis of
Arabinomannane (AM) Glycoconjugates as Potential Vaccines for Tuberculosis, European Journal of
Medicinal Chemistry, https://doi.org/10.1016/j.ejmech.2020.112578.
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. Please note that,
during the production process, errors may be discovered which could affect the content, and all legal
disclaimers that apply to the journal pertain.
© 2020 Elsevier Masson SAS. All rights reserved.

Chemoenzymatic Synthesis of Arabinomannane (AM)
Glycoconjugates as Potential Vaccines for Tuberculosis
Zhihao Li,a Teodora Bavaro,b Sara Tengattini,b Roberta Bernardini,c Maurizio
Mattei,c,d Francesca Annunziata,e Richard B. Cole,a Changping Zheng,a Matthieu
Sollogoub,a Lucia Tamborini,e Marco Terrenib* and Yongmin Zhanga*
a
Sorbonne Université, CNRS, Institut Parisien de Chimie Moléculaire, UMR 8232, 4 Place
b
Jussieu, 75005 Paris, France. Drug Sciences Department, University of Pavia, Viale
Taramelli 12, 27100 Pavia, Italy. c Italy Centro Servizi Interdipartimentale – STA, University
of Rome “Tor Vergata”, Rome, Italy. d Dept. of Biology, University of Rome “Tor Vergata”,
e
Rome, Italy. Department of Pharmaceutical Sciences, University of Milan, via Mangiagalli
25, 20133 Milan, Italy.
Abstract
Mycobacteria infection resulting in tuberculosis (TB) is one of the top ten leading causes of
death worldwide in 2018, and lipoarabinomannan (LAM) has been confirmed to be the most
important antigenic oligosaccharide on the TB cell surface. In this study, a convenient synthetic
method has been developed for synthesizing three branched oligosaccharides derived from LAM,
in which a core building block was prepared by enzymatic hydrolysis in flow chemistry with
excellent yield. After a series of steps of glycosylations, the obtained oligosaccharides were
conjugated with recombinant human serum albumin (rHSA) and the ex-vivo ELISA tests were
performed using serum obtained from several TB-infected patients, in order to evaluate the
affinity of the glycoconjugate products for the human LAM-antibodies. The evaluation results are
positive, especially compound 21 that exhibited excellent activity which could be considered as a
lead compound for the future development of a new glycoconjugated vaccine against TB.
Keywords: lipoarabinomannan; glycoconjugate; chemoenzymatic synthesis; tuberculosis; vaccine
1. Introduction
Ranking above HIV in 2018, around 1.5 million people died from tuberculosis (TB), which is
among the top 10 causes of death. Moreover, 25% of the world's population has latent TB,

meaning civilians have been infected by TB bacteria without showing signs of illness. Due to the
serious situation, ending the TB epidemic becomes one of the health targets of the Sustainable
Development Goals by 2030[1]. Although tuberculosis is curable by administering medicine, the
efficacy of drugs is limited by multiple drug resistance (MDR) and extensively-drug resistance
(XDR)[2]. As an example, it is estimated that during 2018 there were 484000 new cases resistant
to the most effective first-line drug rifampicin, leading to vaccination as still being the ideal
method of fighting against TB. However, BCG (Bacillus Calmette-Guérin) is the only licensed
vaccine against TB and it is only effective for children[3]. Thus, there remains an urgency to
develop novel vaccines for TB prevention.
Carbohydrates are commonly found in nature as one of most complicated and diverse classes of
biomolecules. They play important roles in fundamental biological processes with the outcome
that many pharmaceuticals and vaccines use them as building blocks[4], and today glycoconjugate
vaccines are being regarded as one of the safest and most effective anti-infective vaccines being
manufactured in the last few decades[5, 6]. Commercial glycoconjugate vaccines are commonly
obtained from antigenic sugar isolated from pathogens[4, 7, 8]. For Mycobacterium tuberculosis
(M.TB) the most important sugar-antigen is the glycolipid lipoarabinomannan (LAM) that is also
common to other Mycobacterium species.
Now the complex structure of the M.TB membrane oligosaccharides has been well
established[9]. In the case of LAM, the arabinomannane (AM) is the antigenic structure composed
by an internal mannan backbone, made of α(1→6) and α(1→2)
long and branched arabinan domain. The non-reducing terminus of AM is composed by branched
motifs including -Ara-β(1→2) disaccharide units linked with -arabinose by α(1→5) and α(1→3)
bonds (Figure 1 A). This motif can also be end capped with -mannose or -Man-α(1→2) -Man
units in LAM[10]. Branched motifs of -arabinose by α(1→5) and α(1→3) bonds are also
involved in different parts of the arabinan domain of LAM (Figure 1 B). Moreover, these -Ara
D-mannose units and covered by a
D
D
D
D
D
D
D
motifs of LAM are common with arabinogalactane (AG), another important oligosaccharide in
M.TB membrane which is crucial for cell growth, viability, and virulence[11, 12].

Figure 1. Structure of the branched motifs in AM
Since different glycoforms of LAM exposed outside the M.TB membrane were proven to
stimulate the production of specific antibodies[13] and this antigen is one of the most important in
the BCG vaccine, the conjugation of the AM part with protein has attracted the interests of many
researchers and has been proposed as a strategy for the development of new and efficient
TB-vaccines[14, 15] targeting the LAM antigen, with the advantage that these products could be
active against different species of virulent mycobacteria in humans (such as Mycobacterium avium
Spp.). Moreover, for the complexity of LAM structure, scientists also studied different types of
synthetic AM-analogues and interestingly these analogues became antigenic after conjugating with
suitable carrier proteins or lipids[16, 17]. However, until now there is no evidence to show which
are the epitopes for this immunological activity. Thus, more fundamental study of AM based
vaccine is crucial for the development of efficient sub-unit glycoconjugated vaccines.
In carbohydrate chemistry, the synthesis of oligosaccharides can be strongly limited by the
complex chemical procedures involving protection/deprotection strategies by different protecting
groups. As one of the most widely used protecting groups, acetyl is widely used in sugar chemistry
and different chemical approaches have been reported to deprotect it. Compared to the chemical
procedure, biocatalytic deacetylation strategies can selectively deprotect acetyl groups in
peracetylated sugars in the desired position, using mild condition and aqueous medium[18, 19].
Moreover, the development of flow chemistry using immobilized enzyme on packed bed reactors
provides wider applications in material discovery, future medicine and compound synthesis[20,
21], which saves time and labor to a significant degree. In addition, for the development of
glycoconjugated vaccines, the selection of an adequate chemical linker in the anomeric sugar for

protein glycosylation should be considered and the synthetic strategy designed accordingly[7, 22,
23].
In the present work, acetyl was used as the only protecting group to avoid the redundant
procedures and enzymatic regioselective hydrolysis was conducted to furnish arabinose building
blocks with free hydroxy groups in one or two desired positions starting from peracetylated
arabinose thioglycosides. Furthermore, in order to improve the performances of the enzymatic
reactions, the use of flow-based bioreactors has been also considered. After the required steps of
glycosylations, the motif containing
D
-arabinose branched with α(1→5) and α(1→3) -Ara and
D
some analogues were prepared considering the combination of different arabinose and mannose
units, and then glycosylated with a mannose building block possessing a terminal thiocyanomethyl
group at the anomeric position[24]. In fact, after one-pot deprotection and activation of the
anomeric linker, the glycans obtained were combined with recombinant human serum albumin
(rHSA) protein. Finally, the three obtained glycoconjugates (in Figure 2) were submitted to
biological evaluation by ex-vivo tests performed on serum obtained from several infected patients
in order to evaluate their affinity for the human polyclonal antibodies against TB.
Figure 2. Glyconjugated products as AM analogues
2. Results and discussion
2.1 Synthesis of the oligosaccharide

HO
AcO
OAc
O
AcO
SCH2CN
14
AcO
OAc
O
AcO
OAc
O
AcO
AcO
AcO
OAc
O
O
AcO
OAc
O
AcO
O
OAc
O
OAc
OAc
OAc
O
O
O
O
OAc
O
O
OAc
O
OAc
O
O
OAc
O
O
OAc
O
O
18
19
O
AcO
AcO
AcO
20
AcO
AcO
AcO
AcO
AcO
OAc
AcO
O
SCH2CN
SCH2CN
SCH2CN
AcO
AcO
O
AcO
OAc
O
O
AcO
AcO
AcO
AcO
AcO
AcO
OAc
OAc
O
AcO
AcO
AcO
O
AcO
AcO
OAc
O
OAc
O
O
OAc
O
O
O
OAc
O
OAc
OAc
O
OAc
O
STol
17
15
STol
16
O
O
OAc
STol
OAc
O
O
AcO
AcO
AcO
O
AcO
O
O
O
AcO
AcO
OAc
AcO
AcO
AcO
AcO
OAc
O
AcO
AcO
AcO
AcO
OAc
O
HO
OAc
O
O
NH
10
O
NH
OAc
STol
OH
CCl3
CCl3
8
6
AcO
AcO
OAc
O
AcO
OAc
O
O
HO
OAc
O
O
NH
CCl3
OAc
13
STol
OAc
4
Figure 3. Retrosynthetic analysis for protected AM compounds 18, 19 and 20.
For the synthesis of three acetylated oligosaccharides, intermediates needed are shown in
Figure 3 based on the retrosynthetic analysis. First, synthesis of tetrasaccharide 18 would start
with trisaccharide 15 and monosaccharide 14, while the trisaccharide 15 is constructed from
monosaccharides 6 and 8. Then for compound 19, using the same monosaccharide 14, a
pentasaccharide 16 would be used which can be constructed by disaccharide 13 and
monosaccharide 6, where 13 is synthesized from the two monosaccharides 4 and 10. Finally,
trisaccharide 17 and monosaccharide 14 are required for preparing 20, in which building block 17
is constructed by compound 6 and 10.

HO
OAc
AcO
OAc
O
HO
OH
O
a
O
c
OH
R1
R
R2
OAc
b
OH
4: R1 = STol, R2 = OAc
5: R1 = SEt, R2 = OAc
6: R1 = STol, R2 = OH
7: R1 = SEt, R2 = OH
1
2: R = STol
3: R = SEt
AcO
OAc
O
NH
O
CCl3
OAc
8
Scheme 1. Reagents and conditions: Transformations (a) and (b) have been conducted using reported
protocols[25,27]; (c) CALB (Nvozyme 435), t-BuOH (4% water) (4 and 5) / t-BuOH (8% water) (6 and 7), 60 °C,
1000 rpm, 15 h (4 and 5) / 80 h (6 and 7), 65% (4), 43% (5), 65% (6), 48% (7), respectively.
As shown in Scheme 1, the synthesis of monosaccharides 2 and 3 started with commercially
available
D
-arabinose 1. Since both furanose and pyranose forms can be obtained during the one
-arabinose, hydroxy group at C-1 or C-5 position needs to be protected first.
step acetylation of
D
Thus, peracetylated arabinofuranose was synthesized by three steps which are anomeric methoxy
introduction, acetylization and anomeric acetylization[25]. Then, Lewis acid was used to activate
the anomeric position introducing the sulfur group to afford compounds 2 and 3 at low
temperature. After that, enzymatic reactions catalyzed by immobilized CALB (Nvozyme 435)
were performed to selectively remove the acetyl group only in C-5 (products 4 and 5) and in both
1
1
C-3 and C-5 (products 6 and 7). From H NMR and H-1H COSY spectra (see supporting
information), the acetyl on C-5 position was first deprotected before the removal of C-3 acetyl.
And compared to taking monosaccharide 3 as raw material, compound 2 can remove both acetyls
on C-3 and C-5 positions with higher yield implying that building block 2 is more suitable as an
intermediate for further synthesis. Afterwards, the conditions for the optimization of the selective
hydrolysis of substrate 2 in C-5 position to obtain products 4, and in C-5 and C-3 for synthesis of
intermediate 6, were tested and reported in the supporting information (table S1 and S2). From
results obtained by adjusting water and enzyme amounts, as well as controlling the reaction time,
it was deduced that larger amounts of water and enzyme along with longer reaction time promoted
the deprotection of two acetyls.
A flow chemistry approach was also considered for the enzymatic deprotection of thioglycosyl
arabinose 2 since immobilized CALB was reported to dramatically improve performance in the

regioselective hydrolysis of peracetylated arabinose[26]. Thus, immobilized CALB was packed
into a glass column and a solution of 2 in t-BuOH/water flowed through it. Reaction parameters
(i.e., concentration, amount of water, temperature, pressure and residence time) have been
optimized for the preparation of intermediates 4 and 6 according to the experiment reported in the
additional information (Table S3 and S4). All the optimization studies were performed with the
same bioreactor without observing any relevant loss of activity. Thus the column reactor can be
used to perform in continuous flow reaction for at least 100 h in the condition used (working in
t-BuOH with water between 2% and 10%; 60°C and atmospheric pressure). In the hydrolysis of
thiglycosyl arabinose 2, the reaction outcome improved compared with the results obtained by
batch (Table 1). Thus, 67% yield of product 4 was obtained in only one half hour, while substrate
2 was completely converted into product 6 (98% by HPLC analysis of the reaction) after 18 hours
of recirculation in flow condition. The system was then tested in the synthesis of intermediate 6 on
the preparative gram scale. Recirculating the substrate solution for 48 h (Table 1), less than 350
mg of biocatalyst was able to efficiently hydrolyse 880 mg of 2 allowing the production of 6 in
80% yield after purification.
Table 1. Performances obtained in the hydrolysis of substrate 2 catalyzed by immobilized CALB
Entry
Substrate 2
30 mg
Enzyme
150 mg
345 mg
150 mg
345 mg
345 mg
Product
Condition
Batch (a)
Flow (b)
Batch (c)
Flow (d)
Flow (e)
Time
15 h
0.5 h
80 h
18 h
48 h
Yields
65%
1
2
3
4
5
4
4
6
6
6
90 mg
67%
36 mg
65%
75 mg
98% (HPLC)
80%
880 mg
Batch: (a) t-BuOH (water 8%), volume: 1.2 mL, 1000 rpm, 60 °C; (c) t-BuOH (water 8%), volume: 1.2 mL, 1000
rpm, 60 °C; Flow: All the conditions use bioreactor volume: 1.54 mL. 60°C; substrate 2 concentration: 25 mg/mL:
(b) t-BuOH (water 4%), flow rate 52 µL/min, pressure: Atm. (d) t-BuOH (water 4%); Recirculation: flow rate 52
µL/min, pressure: Atm. Conversion calculated by HPLC analysis. (e) t-BuOH (water 2%); Recirculation: flow rate
104 µL/min ; pressure: 75 psi.
On the other hand, metal catalyst was applied to remove the anomeric acetyl in peracetylated
arabinofuranose followed by imidate introduction to prepare monosaccharide 8 according to the

known approaches[27].
OAc
HO
O
4
STol
c
OAc
AcO
AcO
AcO
OAc
O
AcO
AcO
OAc
O
HO
HO
HO
OH
a
O
AcO
OAc
O
O
O
OH
NH
10
C
9
STol
OAc
11
CCl3
b
d
OAc
AcO
AcO
AcO
OAc
AcO
AcO
AcO
O
O
e
HO
AcO
AcO
OAc
O
OAc
O
OAc
O
O
O
OH
NH
CCl3
O
12
OAc
13
OAc
SCH2CN
14
Scheme 2. Reagents and conditions: Transformations (a) and (b) have been conducted using reported
protocols[23,28]; (c) BF3·Et2O, 4Å MS, CH2Cl2, -60 °C, 30 min, 55%; (d) NIS, AgOTf, TTBP, wet CH2Cl2, 0 °C -
RT, 1 h, 90%; (e) CCl3CN, DBU, CH2Cl2, 0 °C, 30 min, 90%.
In Scheme 2,
D-mannose was used as starting material to synthesize monosaccharide
trichloroacetimidate 10 as intermediate by the reported methods[28]. After that, glycosylation was
performed using donor 10 and acceptor 4 catalyzed by BF3·Et2O at low temperature. In this step,
we found that reaction temperature has an effect on the preferred reaction pathway, and low
temperature can avoid the byproduct to some degree. When reaction was carried out at 0 °C,
almost no targeting product was obtained because the Lewis acid activated the STol group in
compound 4, and when the temperature decreased to -30 °C and -60 °C, the yield came to 40%
and 55%, respectively. However, no increase of the output was observed when temperature
continued declining implying that -60 °C is the proper temperature balancing the yield and
economic cost. Afterwards, the sulfur group in compound 11 was removed by NIS/AgOTf and
2,4,6-Tri-tert-butylpyrimidine (TTBP) in wet CH2Cl2 obtaining compound 12 with free hydroxy at
the anomeric center, and finally intermediate 12 was transformed into trichloroacetimidate 13 at
0 °C with CCl3CN and DBU for further use. On the other hand, the known compound 14 was
synthesized by several steps, including a final enzymatic hydrolysis catalyzed by Candida rugose
lipase for deprotection in C-6, according to the literature[23].

Scheme 3. Reagents and conditions: (a) BF3·Et2O, 4Å MS, CH2Cl2, 0 °C, 30 min, 40%; (b) BF3·Et2O, 4Å MS,
CH2Cl2, 0 °C, 1 h, 32%; (c) BF3·Et2O, 4Å MS, CH2Cl2, 0 °C, 1 h, 37%; (d) NIS, AgOTf, 4Å MS, CH2Cl2, -30 °C,
50 min, 60%; (e) NIS, AgOTf, 4Å MS, CH2Cl2, -30 °C, 50 min, 49%; (f) NIS, AgOTf, 4Å MS, CH2Cl2, -30 °C, 50
min, 55%
With the three donors prepared (8, 10 and 13), glycosylations were carried out with acceptor 6,
(Scheme 3). Since catalyst and temperature may affect the reaction, different conditions were
studied. For the reason that both BF3·Et2O and TMSOTf were taken as promoters for the
glycosylation, with BF3·Et2O providing more products than TMSOTf, only the former promoter
was studied in Table 2. In Entry 1, since BF3·Et2O activated the STol group in acceptor 6 at
-30 °C, hardly any product 15 was obtained. Then with temperature cooled to -50 °C and -60 °C in
Entry 2 and 3, the yield went up to 29% and 40%, respectively. However, a continued reduction
of temperature to -70 °C did not increase the yield in Entry 6, indicating that -60 °C is proper for
producing 15. In addition to this, the quantity of catalyst was also studied in the Table. In Entry 4,
less product was obtained by adding more catalyst at -60 °C. With less BF3·Et2O promoter (Entry
5), the yield also appeared to decrease, because it needs a longer time to finish the reaction which
is detrimental for a stable experimental condition.

Similarly, for constructing compound 17, low temperature is beneficial to obtain product 17 in
decent yield. Compared to Entry 7-9, Entry10 provided a higher yield (23%) using the same
amount of catalyst. However, deceasing the temperature in Entry 13 did not result in a better
output than in Entry 10. In addition, comparison of Entry 10, 11 and 12 demonstrated that
reducing the catalyst quantity could enhance the reaction yield, whereas increasing the quantity
may decrease the yield and the best result acquired is shown in Entry 12.
Subsequently, different conditions were also studied for synthesizing pentasaccharide 16. From
the results, in Entry 14, 15 and 18, -60 °C is the optimal temperature to construct 16 with yield
reaching 37%. In Entry 17, doubling the amount of promoter cut down the yield from 37% to
25% at the same temperature, while decreasing it to 0.2 equiv. did not affect the yield in Entry 16.
Table 2. Conditions studied to furnish compounds 15-17
Catalyst
Entry
Donor
Acceptor
Temperature
Product
Yield
(equiv.)
0.5
0.5
0.5
1
1
2
8
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
-30 °C
-50 °C
-60 °C
-60 °C
-60 °C
-70 °C
-30 °C
-50 °C
-60 °C
-70 °C
-70 °C
-70 °C
-90 °C
-50 °C
-60 °C
-60 °C
-60 °C
-70 °C
15
15
15
15
15
15
17
17
17
17
17
17
17
16
16
16
16
16
< 10%
29%
40%
33%
37%
35%
< 10%
< 10%
18%
23%
11%
32%
17%
35%
37%
37%
25%
34%
8
3
8
4
8
5
8
0.2
0.5
0.5
0.5
0.5
0.5
1
6
8
7
10
10
10
10
10
10
10
13
13
13
13
13
8
9
10
11
12
13
14
15
16
17
18
0.2
0.5
0.5
0.5
0.2
1
0.5
*In the table, all the conditions are donor being 3 equiv. and acceptor being 1 equiv.
Finally, promoted by NIS/AgOTf, glycoylations were performed between the three synthesized
donors (15, 16 and 17) and acceptor 14, respectively. In Scheme 3, monosaccharide 14 is a known

compound synthesized by enzymatic hydrolysis[23], where there is a SCH2CN group at the
anomeric centre in 14 that may be damaged by using the same reagent. Since in SCH2CN the
cyano group is an electron withdrawing group, it can reduce the electronic cloud density around
sulfur. Besides, furanose (donors 15, 16 and 17) is general more activated than pyranose (acceptor
14) in glycosylation. We propose the catalyst would activate the STol groups in the donor first. In
this step, the reactions were carried out at -10 °C, -30 °C and -50 °C, respectively, and -30 °C
provided the best yield. Finally, two tetrasaccharides (18 and 20) and one hexasaccharide (19)
were obtained for further protein conjugation via thiocyanomethyl group.
2.2 Preparation of the glycoconjugates
Scheme 4. Reagents and conditions: (a) 21-23, MeONa/MeOH, 35 °C, 48 h; (b) rHSA, sodium tetraborate buffer,
glycosidic reagent/rHSA (molar ratio 200 : 1), pH 9.5, 25 °C, 24 h.
As shown in Scheme 4, compounds 18-20 were submitted to one pot deacetylation and
activation to obtain the corresponding 2-deprotected iminomethoxyethyl (IME) thioglycosides
18a-20a. The reactions were performed according to a previously reported procedure[17,23],
working at 35°C in order to achieve a conversion around 50% of each intermediate into the
corresponding IME derivative (Table 3), as resulting from ESI-MS analysis (see Figure S1-S3 in
supplementary information).
The coupling reactions between activated glycans 18a-20a and the carrier protein, namely
rHSA, were carried out according to a previously reported protocol for IME-activated saccharides
that is specific for lysine residues[17]. Starting from literature conditions, different reaction
parameters (temperature, saccharide/protein ratio, reaction temperature and time, presence of

organic solvent) were tested in order to reach satisfactory conjugation yields. For this purpose,
reaction mixtures were analysed by direct injection in Hydrophilic Interaction Liquid
Chromatography (HILIC-UV). In HILIC chromatography, the stationary phase (for our column
composed of amide-derivatized silica) interacts with the sugar moieties of glycoproteins or
glycoconjugates thanks to its high hydrophilicity, and the resulting elution order reflects the
degree of glycosylation. Nowadays, HILIC analysis at intact protein level is considered a
consolidated approach in glycoprofiling studies[29].
Collecting all the different observations, the final glycosylation protocol was developed. The
HILIC-UV trace of neo-glycoprotein 21-23 obtained in the optimized conditions were reported in
Figure S4 supplementary data, where the absence of chromatographic peaks at the retention time
of rHSA indicated that all three glycosylations were completed after 24 h (Table 3). Finally,
MALDI-ToF analysis was performed for the characterization of the different neoglycoproteins.
Based upon the mass difference between un-conjugated HSA and each of the 3 conjugated
compounds, it was possible to estimate the average level of conjugation (average number of
oligosaccharides attached) for each sample (Table 3).
Table 3. Results of the coupling reaction by 18-20
Compound
IME-derivative (%) (a) Product (%) (b)
n° sugar
14.5
18
19
20
18a (49%)
19a (53%)
20a (54%)
21 (>98%)
22 (>98%)
23 (>98%)
12.4
14.3
2.3 Biological ex-vivo evaluation of the glycoconjugates
Ex-vivo antigenic evaluation has been performed with neo-glycoproteins 21-23, in order to
evaluate the affinity of putative epitopes for human antibodies. This approach is an elective
method for the screening of new antigens because it is directly related to the recognition of human
epitopes and allows an appropriate selection of candidates for successive in vivo testing. Thus,
indirect ELISA (Figure 4) was applied using serum obtained from infected patient (30 individuals)
to detect the potential of 21 (ManAra3-rHSA), 22 (Man3Ara3-rHSA) and 23 (Man3Ara-rHSA) as

antigens and 16 healthy members' serum as controls (details such as age and nationality are shown
in Table S5). The results from this analysis demonstrated a statistically significant increase of the
antibody reactivity in TB subjects compared with healthy control ones for neoglycoproteins 21
and 22 (p≤0.0004 and p≤0.01 respectively).
In the test, the obtained cut-off value of the rHSA-based ELISA was 0.380 (mean = 0.128, SD =
0.084), while the cut-off value of the 21, 22 and 23-based ELISA was 0.482 (mean = 0.168, SD =
0.105), 0.298 (mean = 0.106, SD = 0.064), and 0.307 (mean = 0.121, SD = 0.062), respectively. In
these conditions on a total of 30 TB serum samples, 17 were determined as positive for product 21
(mean = 0.767, SD = 0.619), 13 for neoglycoprotein 22 (mean = 0.471, SD = 0.539) and only 2 for
23 (mean = 0.175, SD = 0.125).
Figure 4. Antibodies (IgG) levels reactive against neoglycoproteins 21, 22 and 23 in 30 TB
patients and 16 healthy subjects (CTR).
In the figure, rHSA was used as protein control. Panel A: Histograms report OD mean values and standard
deviation of TB and control samples. Panel B: Distribution of IgG level of samples. Horizontal lines represents the
cut off values calculated for each antigens used in the study as the average of negative controls + (3 × SD of
negative controls). The sera above horizontal line were considered positive and sera below the line were negative

according to the cut off value. OD = Optical Density; CTR: Controls; TB = Tuberculosis. P ≤ 0.05 = *.
The analysis of antibody specificity to different neoglycoproteins of sera from M.TB patients
demonstrated that the
compound 21, is well recognized by antibodies from 17 out of 30 patients. The replacement of the
two -Ara residues with two -Man α(1→5) units (compound 23), completely abrogated the
reactivity of antibodies. The activity is partially maintained when two additional units of
-Man-α(1→5) are introduced on the branched -Ara motif to obtain compound 22.
Interestingly, the analysis of the 13 patients positively responding to the neoglycoprotein 22 are
D-Ara motif branched with
D-Ara α(1→5) and α(1→3), contained in
D
D
D
D
almost all revealing positive for glycoconjugate 21 (in Figure 5) (except for one patient that is
purely responding towards both sugars), while only about 70% of the 17 patients responding for
antigen 21 are positive for 22. These results may indicate that the epitope of 22 is correlated with
the branched
D-Ara motif contained in 21 and the introduction of two additional mannose
molecules at the end of the chain reduces the recognition of this epitope.
Figure 5. OD values obtained with the M.TB patients that respond positively against compound
22
As one of the most convenient and rapid technologies in immunoassay, ELISA plays an
essential role in TB ex-vivo biological evaluation because of its sensitivity and specificity for
human antibodies. However, most of the research used mice as the experimental model[16, 30, 31],
which may lead to different results than humans epitopes. In this work, we use the serum obtained
from different infected patients to evaluate the prepared compounds and our study demonstrates

that the branched -Ara motif included in different parts of the AM structure can have significance
D
as a human epitope and thus relevant in developing novel TB vaccine candidates.
3. Conclusion
A novel enzymatic approach was developed to selectively remove acetyls from a peracetylated
arabinose thioglycoside furnishing a key building block for synthesizing three branched AM
analogues. The flow chemistry application through immobilized CALB (Nvozyme 435) in the
column bioreactor not only shortened the reaction time, but also enhanced the output significantly
on the gram scale. After required steps of glycosylations, three oligosaccharides derived from AM
were furnished and conjugated with rHSA. In the ELISA test, the three glycoconjugates showed
different affinities to human antibodies present in the serum of M.TB infected patients, with
compound 21 obtaining the best results. The fact that the number of patients responding to 21 and
the mean response observed by ex-vivo assay to 21 were much greater than the responses obtained
by the two other analogues prepared with the same synthetic approach, demonstrates that the
branched arabinose motif contained in neo-glycoprotein 21 is a putative human epitope of LAM.
Thus, neo-glycoprotein based on branched -Ara motif are prone to induce production of specific
D
human antibodies against LAM and are worthy of further investigation.
In order to find out the best epitopes of AM and design more effective antituberculosis vaccine
candidates, 21 can be regarded as the lead compound to prepare more related compounds
including the use of suitable carrier proteins for in-vivo evaluation. It is noteworthy that this is the
first work using the chemo-enzymatic approach and it represents a method to rapidly synthesize
new branched analogues in order to better elucidate the antigenic structures of AM. In this context,
the use of biological evaluation by ex-vivo tests is also helpful to identify the optimal human
epitope(s) and design novel efficient AM-based TB glycovaccine candidates for further evaluation.
Furthermore, studies such as in vivo evaluations on suitable animal models will also be considered
in the future for selected compounds.
4. Experimental Section
4.1 General method

All chemicals were purchased as reagent grade and used without further purification. All
reactions were carried out under Ar atmosphere and anhydrous conditions with freshly distilled
solvents, unless otherwise noted. Reactions were monitored by thin layer chromatography (TLC)
on a pre-coated plate of silica gel 60 F254 (Merck) and detection by charring with sulfuric acid.
Solvents were evaporated under reduced pressure and below 40 °C (water bath). Column
chromatography was performed on silica gel 60 (230-400 mesh, Merck). All the new compounds
1
were fully characterized by 1H and 13C NMR, as well as HRMS. H NMR and 13C NMR spectra
were recorded at 400 MHz and 100 MHz with Bruker AVANCE DRX 400 spectrometer,
respectively. The chemical shifts were referenced to the solvent peak, 7.26 ppm (1H) and 77.16
ppm (13C) for CDCl3 at 25 °C, and coupling constants were given in Hz. Complete assignment of
all NMR signals was performed using a combination of 1H,1H-COSY and 1H,13C-HSQC
experiments. For characterization of the synthetic intermediates and oligosaccharides
high-resolution mass spectra (HRMS) were recorded with a Bruker Micro-TOF spectrometer in
electrospray ionization (ESI) mode, using Tuning-Mix as reference. The continuous flow reactions
were performed using a R2+/R4 series flow reactor, commercially available from Vapourtec
equipped with Omnifit glass columns. For the preparation of the different neo-glycoprotein, rHSA
(recombinant human serum albumin) is derived from rice grains (gene-modified rice endosperm)
and obtained by chromatographic purification (>99% by SDS-PAGE) was provided by Healtgen
biotechnology Corp (Wuhan, China).
4.2 Synthesis of the AM analogues
4.2.1 4-Methylphenyl 2,3-di-O-acetyl-1-thio-α- -arabinofuranoside (4)
D
4.2.1.1 Batch method: To a solution of t-BuOH (12 mL) with 4% water (0.84 mL) added,
compound 2 (600 mg, 1.6 mmol) and CALB enzyme (1.5 g) were added. Then the reaction was
stirred vigorously at 1000 rpm at 60 °C for 15 h. After filtrating the enzyme, the solution was
concentrated under vacuum before purification by column chromatography (cyclohexane: ethyl
acetate = 2:1), with white foam 4 obtained (360 mg, 65%).
4.2.1.2 Flow method: A solution (25 mg/mL) of substrate 2 was prepared in a mixture
water/tert-butanol (96:4). The bioreactor (prepared packing 345 mg CALB into a glass column;
volume: 1.54 mL) was washed with the flow stream (water/tert-butanol 96:4) at 0.5 mL/min for 5
min at RT and 5 min at 60 °C. The substrate solution (3.6 mL) was then recirculated at 60 °C
through the bioreactor (flow 52 µL/min). The reaction was monitored by HPLC and after for 18 h
the bioreactor was washed with water/tert-butanol (96:4) at 0.5 mL/min for 10 min at 60 °C. The
solvent was evaporated under reduced pressure and the crude was purified by flash
chromatography (cyclohexane: ethyl acetate = 1:1) to give pure 4 (53 mg, 67%). Rf = 0.6
(cyclohexane: ethyl acetate = 1:1). [α]D25 +174.1 (c 1, CHCl3). 1H NMR (400 MHz, CDCl3) δ 5.47
(d, J = 2.5 Hz, 1H, H-1), 5.30 (t, J = 2.5 Hz, 1H, H-2), 5.12 (m, 1H, H-3), 4.35 (m, 1H, H-4),
3.93-3.79 (m, 2H, H-5, H-5'), 2.34 (s, 3H, Tol), 2.14 (s, 3H, Ac), 2.09 (s, 3H, Ac), 2.02-1.99 (m,
1H, OH). 13C NMR (100 MHz, CDCl3) δ 170.46, 169.67, 138.16, 132.83, 129.85, 129.58, 91.10
(C-1), 82.59, 81.71, 77.08, 61.57, 21.14, 20.81, 20.76. HR ESI-TOF MS (m/z): calcd for

C16H20O6SNa [M+Na]+, 363.0873; found, 363.0880.
4.2.2 Ethyl 2,3-O-acetyl-1-thio-α- -arabinofuranoside (5)
D
To a solution of t-BuOH (12 mL) with 4% water (0.84 mL) added, compound 2 (510 mg, 1.6
mmol) and CALB enzyme (1.5 g) were added. Then the reaction was stirred vigorously at 1000
rpm at 60 °C for 15 h. After filtrating the enzyme, the solution was concentrated under vacuum
before purification by column chromatography (cyclohexane: ethyl acetate = 3:2), with white
25
foam 5 obtained (191 mg, 43%). Rf = 0.5 (cyclohexane: ethyl acetate = 1:1). [α]D +162.5 (c 1,
CHCl3). 1H NMR (400 MHz, CDCl3) δ 5.33 (m, 1H, H-1), 5.14 (t, J = 2.0 Hz, 1H, H-2), 5.08 (m,
1H, H-3), 4.25 (m, 1H, H-4), 3.92-3.79 (m, 2H, H-5, H-5'), 2.74-2.61 (m, 2H, SCH2), 2.11 (s, 3H,
Ac), 2.09 (s, 3H, Ac), 2.10 (m, 1H, OH), 1.3 (t, J = 7.4 Hz, 1H, CH3). 13C NMR (100 MHz, CDCl3)
δ 170.51, 169.76, 87.57 (C-1), 82.28, 82.17, 77.29, 61.66, 25.17, 20.81, 20.79, 14.72. calcd for
C11H18O6SNa [M+Na]+, 301.0716; found, 301.0724.
4.2.3 4-Methylphenyl 2-O-acetyl-1-thio-α- -arabinofuranoside (6)
D
4.2.3.1 Batch method: To a solution of t-BuOH (12 mL) with 8% water (0.84 mL) added,
compound 2 (600 mg, 1.6 mmol) and CALB enzyme (1.5 g) were added. Then the reaction was
stirred vigorously at 1000 rpm at 60 °C for 80 h. After filtrating the enzyme, the solution was
concentrated under vacuum before purification by column chromatography (cyclohexane: ethyl
acetate = 1:1), with white foam 6 obtained (304 mg, 65%).
4.2.3.2 Flow method: A solution of 2 (25 mg/mL) was prepared in a mixture water/tert-butanol
(98:2). The bioreactor (prepared as above described) was washed with the flow stream
(water/tert-butanol 98:2) at 0.5 mL/min for 5 min at RT and 5 min at 60 °C (P = 75 psi). The
substrate solution (36 mL) was then recirculated through the bioreactor at 104 µL/min (60 °C, P
=75 psi). The reaction was monitored by HPLC and after 48 h the bioreactor was washed with
water/tert-butanol (98:2) at 0.5 mL/min for 10 min (60 °C). The solvent was evaporated under
reduced pressure and the crude was purified by flash chromatography (cyclohexane: ethyl acetate
25
= 1:1) to yield 550 mg of pure 4 (80% yield). Rf = 0.28 (cyclohexane: ethyl acetate = 1:1). [α]D
1
+184.9 (c 0.5, CHCl3). H NMR (400 MHz, CDCl3) δ 5.51 (d, J = 3.2 Hz, 1H, H-1), 4.93 (t, J =
3.3 Hz, 1H, H-2), 4.30 (m, 1H, H-4), 4.15 (m, 1H, H-3), 3.91 (m, 1H, H-5), 3.80 (m, 1H, H-5'),
3.49 (s, 1H, OH), 2.33 (s, 3H, Tol), 2.12 (s, 3H, OAc). 13C NMR (100 MHz, CDCl3) δ 172.00,
138.32, 132.93, 129.96, 129.67, 89.66 (C-1), 86.84, 82.74, 77.48, 77.16, 76.84, 76.25, 61.29,
21.26, 20.95. HR ESI-TOF MS (m/z): calcd for C14H18O5SNH4 [M+NH4]+, 316.1213; found,
316.1214.
4.2.4 Ethyl 2-O-acetyl-1-thio-α-
D
-arabinofuranoside (7)
To a solution of t-BuOH (12 mL) with 8% water (0.84 mL) added, compound 2 (600 mg, 1.6

mmol) and CALB enzyme (1.5 g) were added. Then the reaction was stirred vigorously at 1000
rpm at 60 °C for 80 h. After filtrating the enzyme, the solution was concentrated under vacuum
before purification by column chromatography (cyclohexane: ethyl acetate = 1:1), with white
foam 7 obtained (212 mg, 48%). Rf = 0.28 (cyclohexane: ethyl acetate = 1:1). [α]D25 +164.8 (c 1,
CHCl3). 1H NMR (400 MHz, CDCl3) δ 5.31 (d, J = 2.8 Hz, 1H, H-1), 4.73 (t, J = 3.0 Hz, 1H, H-2),
4.13 (m, 1H, H-4), 4.03 (m, 1H, H-3), 3.85 (m, 1H, H-5), 3.72 (m, 1H, H-5'), 3.29 (d, J = 3.9 Hz,
1H, OH), 2.70-2.54 (m, 2H, SCH2), 2.06 (s, 3H, Ac), 1.25 (t, J = 7.4 Hz, 3H, CH3). 13C NMR (100
MHz, CDCl3) δ 171.80, 87.18 (C-1), 86.06, 82.59, 76.52, 61.37, 25.13, 20.84, 14.68.
C11H18O6SNa [M+Na]+, 259.0611; found, 259.0621.
4.2.5 4-Methylphenyl
(2,3,4,6-tetra-O-acetyl-α-
D
-mannopyranosyl)-(1→5)-2,3-di-O-acetyl-1-thio-α- -arabinofuranoside
D
(11)
The dried compound 4 (300 mg, 0.88 mmol), 10 (650 mg, 1.32 mmol) and activated 4Å MS (1 g)
were mixed together in anhydrous CH2Cl2 (10 mL) stirred for 1 h at RT under argon. Then
BF3 Et2O (63 μL, 0.44 mmol) was added and the reaction was left for 30 min at -60 °C before
quenching by triethylamine. After filtering by Celite and concentration, the resulting crude product
was purified by column chromatography (cyclohexane : ethyl acetate = 3:2) getting 11 (325 mg,
55%) as colorless syrup. Rf = 0.25 (cyclohexane: ethyl acetate = 3:2). [α]D25 +133.8 (c 1, CHCl3).
1H NMR (400 MHz, CDCl3) δ 7.40 (d, J = 8.1 Hz, 2H, Ph), 7.13 (d, J = 8.1 Hz, 2H, Ph), 5.47 (s,
1H, H-1Ara), 5.31-5.27 (m, 4H, H-2Ara, H-2Man, H-3 Man, H-4 Man), 5.14 (m, 1H, H-3Ara), 4.90 (d, J =
1.0 Hz, 1H, H-1Man), 4.39 (m, 1H, H-4Ara), 4.02 (m, 1H, H-5Ara), 3.82 (m, 1H, H-5'Ara), 4.30 (m,
1H, H-6Man ), 4.10 (m, 1H, H-6' Man), 4.07 (m, 1H, H-5 Man), 2.33 (s, 3H, Ac), 2.14 (s, 3H, Ac), 2.13
(s, 3H, Ac), 2.09 (s, 3H, Ac), 2.04 (s, 3H, Ac), 1.97 (s, 3H, Ac). 13C NMR (100 MHz, CDCl3) δ
170.62, 170.15, 170.04, 169.89, 169.67, 169.63, 138.06, 132.84, 129.81, 129.46, 98.03 (C-1Man),
91.15(C-1Ara), 81.75, 81.50, 77.50, 69.29, 69.06, 68.70, 66.38, 65.98, 62.28, 21.13, 20.85, 20.75,
20.73, 20.70, 20.65, 20.63. C30H38O15SNa [M+Na]+, 693.1824; found, 693.1831.
4.2.6 4-Methylphenyl
(2,3,5-tri-O-acetyl-α-
D
-arabinofuranosyl)-(1→3)-[2,3,5-tri-O-acetyl-α-
D
-arabinofuranosyl-(1→5)-
]-2-O-acetyl-1-thio-α-
D
-arabinofuranoside (15)
The dried compound 6 (100 mg, 0.34 mmol), 8 (420 mg, 1.00 mmol) and activated 4Å MS (600
mg) were mixed together in anhydrous CH2Cl2 (5 mL) stirred for 1 h at RT under argon. Then
BF3 Et2O (25 μL, 0.17 mmol) was added and the reaction was left for 30 min at -60 °C before
quenching by triethylamine. After filtering by Celite and concentration, the resulting crude product
was purified by column chromatography (cyclohexane: ethyl acetate = 1:1) getting 15 (110 mg,
40%) as colorless syrup. Rf = 0.3 (cyclohexane: ethyl acetate = 1:1). [α]D25 +177.6 (c 0.6, CHCl3).
1H NMR (400 MHz, CDCl3) δ 7.40 (d, J = 8.1 Hz, Ph), 7.11 (d, J = 8.1 Hz, Ph), 5.47 (m, 1H,
H-1Ara-A), 5.27 (s, 1H, H-1Ara-B), 5.21 (t, J = 1.9 Hz, 1H), 5.15 (m, 1H), 5.12 (m, H-1Ara-B'), 5.12 (m,
1H), 5.01 (m, 1H), 4.96 (m, 1H), 4.44-4.35 (m, 3H), 4.27-4.18 (m, 5H), 3.92 (dd, J1 = 4.3 Hz, J2 =

11.6 Hz, 1H), 3.76 (dd, J1 = 4.3 Hz, J2 = 11.6 Hz, 1H), 2.99 (s, 3H, Tol), 2.12 (s, 3H, Ac), 2.12 (s,
3H, Ac), 2.10 (s, 3H, Ac), 2.09 (s, 3H, Ac), 2.09 (s, 3H, Ac), 2.08 (s, 3H, Ac), 2.08 (s, 3H, Ac). 13C
NMR (100 MHz, CDCl3) δ 170.53, 170.47, 170.11, 169.94, 169.52, 169.41, 137.76, 132.39,
130.22, 129.74, 105.62 (C-1Ara-B), 105.29 (C-1Ara-B'), 91.02 (C-1Ara-A), 82.41, 81.32, 81.12, 80.90,
80.80, 80.68, 77.21, 76.86, 65.20, 63.21, 63.02, 21.11, 20.86, 20.75, 20.71, 20.63. HR ESI-TOF
MS (m/z): calcd for C36H46O19SNa [M+Na]+, 837.2246; found 837.2216.
4.2.7 4-Methylphenyl
(2,3,4,6-tetra-O-acetyl-α-
3)-[(2,3,4,6-tetra-O-acetyl-α-
(1→5)-]-2-O-acetyl-1-thio-α-
D
-mannopyranosyl)-(1→5)-(2,3-di-O-acetyl-α-
-mannopyranosyl)-(1→5)-(2,3-di-O-acetyl-α-
-arabinofuranoside (16)
D
-arabinofuranosyl)-(1→
D
D
-arabinofuranosyl)-
D
To a solution of 11 (830 mg, 1.24 mmol) in 16 mL wet CH2Cl2 (20% H2O) was added NIS (558
mg, 2.48 mmol), AgOTf (318 mg, 1.24 mmol) and TTBP (744 mg, 3 mmol). The solution was
stirred for 1 h until no 11 left. The crude was purified by column chromatography (cyclohexane:
ethyl acetate = 1:2) to give 12 (629 mg, 90%) as colorless syrup. All the 12 was immediately put
into 6 mL dry CH2Cl2 under argon and then CCl3CN (1.1 mL, 11.1 mmol) and DBU (166 μL, 1.11
mmol) were added to the solution at ice bath. The solution was stirred for 1 h at 0 °C and solution
was removed under vacuo and purified by column chromatography (cyclohexane: ethyl acetate =
1:1) with 1% triethylamine added to afford 13 (705 mg, 90%) as colorless syrup. 13 should be
used without delay for its poor stability. The dried compound 6 (100 mg, 0.34 mmol), 13 (705 mg,
1.00 mmol) and activated 4Å MS (1.5 g) were mixed together in anhydrous CH2Cl2 (10 mL)
stirred for 1 h at RT under argon. Then BF3 Et2O (25 μL, 0.17 mmol) was added and the reaction
was left for 30 min at -60 °C before quenching by triethylamine. After filtering by Celite and
concentration, the resulting crude product was purified by column chromatography (cyclohexane:
ethyl acetate = 1:2) getting 16 (172 mg, 37%) as colorless syrup. Rf = 0.4 (cyclohexane: ethyl
25
acetate = 1:2). [α]D +84.5 (c 0.5, CHCl3). 1H NMR (400 MHz, CDCl3) 7.42 (d, J = 8.1 Hz, 1H,
Ph), 7.12 (d, J = 8.1 Hz, 1H, Ph), 5.48 (s, 1H, H-1Ara-A), 5.33 (m, 7H), 5.29 (s, 1H, H-1Ara-B), 5.25
(t, J = 1.8 Hz, 1H), 5.19 (d, J = 1 Hz, 1H), 5.15-5.14 (m, 2H), 5.13 (s, 1H, H-1Ara-B'), 5.05 (d, J =
4.9 Hz, 1H), 4.92 (m, 2H, H-1Man, H-1Man'), 4.46 (m, 1H), 4.34-4.31 (m, 2H), 4.23 (dd, J1 = 1.4 Hz,
J2 = 5.6 Hz, 1H), 4.17-4.11 (m, 6H), 4.04-4.01 (m, 2H), 3.91- 3.88 (m, 1H), 3.81-3.75 (m, 3H),
2.34 (s, 3H, Tol), 2.17 (s, 6H, Ac), 2.17 (s, 3H, Ac), 2.14 (s, 3H, Ac), 2.13 (s, 3H, Ac), 2.13 (s, 3H,
Ac), 2.12 (s, 6H, Ac), 2.10 (s, 3H, Ac), 2.06 (s, 3H, Ac), 2.06 (s, 3H, Ac), 1.99 (s, 3H, Ac), 1.98 (s,
3H, Ac). 13C NMR (100 MHz, CDCl3) δ 170.64, 170.37, 170.28, 170.12, 169.99, 169.91, 169.89,
169.70, 169.64, 169.61, 137.68, 132.37, 130.37, 129.72, 105.46 (C-1Ara-B), 105.21 (C-1Ara-B'),
98.14 (C-1Man), 98.12 (C-1Man') 91.09 (C-1Ara-A), 82.55, 82.40, 82.27, 81.53, 81.36, 81.22, 80.82,
77.45, 76.89, 69.26, 69.11, 69.07, 68.75, 68.68, 66.34, 66.23, 65.96, 65.36, 62.27, 21.10, 20.86,
20.74, 20.71, 20.64, 20.62. HR ESI-TOF MS (m/z): calcd for C60H78O35SNa [M+Na]+, 1413.3937;
found 1413.3970.
4.2.8 4-Methylphenyl
(2,3,4,6-tetra-O-acetyl-α-
D
-mannopyranosyl)-(1→3)-[2,3,4,6-tetra-O-acetyl-α-
D
-mannopyranosyl
-(1→5)-]-2-O-acetyl-1-thio-α-
D
-arabinofuranoside (17)

The dried compound 6 (100 mg, 0.34 mmol), 10 (490 mg, 1.00 mmol) and activated 4Å MS (1 g)
were mixed together in anhydrous CH2Cl2 (7 mL) stirred for 1 h at RT under argon. Then
BF3 Et2O (25 μL, 0.17 mmol) was added and the reaction was left for 50 min at -70 °C before
quenching by triethylamine. After filtering by Celite and concentration, the resulting crude product
was purified by column chromatography (cyclohexane : ethyl acetate = 2:3) getting 17 (172 mg,
32%) as colorless syrup. Rf = 0.4 (cyclohexane: ethyl acetate = 2:3). [α]D25 +117.8 (c 0.5, CHCl3).
1H NMR (400 MHz, CDCl3) δ 5.45 (s, 1H, H-1Ara), 5.37-5.25 (m, 6H), 5.18 (s, 1H), 5.09 (d, J =
1.6 Hz, 1H, H-1Man), 4.94 (d, J =1.0 Hz, 1H, H-1Man'), 4.53-4.52 (m, 1H), 4.32-4.23 (m, 3H),
4.20-4.19 (m, 1H), 4.13-4.05 (m, 2H ), 4.04-3.99 (m, 2H), 3.87 (t, J = 3.9 Hz, 1H), 2.32 (s, 3H,
Tol), 2.17 (s, 3H, Ac), 2.13 (s, 3H, Ac), 2.10 (s, 3H, Ac), 2.08 (s, 3H, Ac), 2.08 (s, 3H, Ac), 2.05 (s,
3H, Ac), 2.04 (s, 3H, Ac), 2.00 (s, 3H, Ac), 1.96 (s, 3H, Ac). 13C NMR (100 MHz, CDCl3) δ
170.55, 170.40, 170.03, 169.86, 169.77, 169.74, 169.70, 169.62, 138.03, 132.91, 129.81, 129.63,
98.10 (C-1Man), 97.48 (C-1Man'), 91.06 (C-1Ara), 82.34, 82.19, 81.53, 69.40, 69.32, 69.13, 68.93,
68.74, 68.65, 66.24, 65.98, 62.59, 62.32, 21.13, 20.84, 20.68, 20.65, 20.60. HR ESI-TOF MS
(m/z): calcd for C42H54O23SNa [M+Na]+, 981.2669; found 981.2682.
4.2.9 Cyanomethyl
(2,3,5-tri-O-acetyl-α-
D
-arabinofuranosyl)-(1→3)-[2,3,5-tri-O-acetyl-α-
D
-arabinofuranosyl-(1→5)-
-mannopyranoside (18)
]-(2-O-acetyl-α- -arabinofuranosyl)-(1→6)-2,3,4-tri-O-acetyl-1-thio-α-
D
D
The dried 15 (40 mg, 0.049 mmol), 14 (27 mg, 0.073 mmol) and activated 4Å MS (100 mg) were
mixed together in anhydrous CH2Cl2 (2 mL) for 1 h at RT. The reaction was put at -30 °C for 5
min before NIS (13 mg, 0.059 mmol) and AgOTf (15 mg, 0.059 mmol) were added. Upon for
stirring at -30 °C for 50 min, the reaction was diluted with CH2Cl2, after filtering by Celite, the
organic phase was washed with saturated aq. Na2S2O3, water and dried over MgSO4, and then
concentrated under reduced pressure. The residue was finally purified by silica gel column
chromatography with cyclohexane: ethyl acetate (1:2) as the eluent to afford 18 (31 mg, 60%) as
25
1
foamy solid. Rf = 0.3 (cyclohexane: ethyl acetate = 1:2). [α]D +196.0 (c 0.5, CHCl3). H NMR
(400 MHz, CDCl3) δ 5.47 (s, 1H, H-1Man), 5.37 (m, 1H), 5.23 (m, 1H, H-1Ara-B), 5.22-5.20 (m, 2H),
5.14 (m, 1H, H-1Ara-B'), 5.14-5.10 (m, 2H), 5.05-5.03 (m, 1H), 4.98 (m, 1H, H-1Ara-A), 4.98-4.95
(m, 2H), 4.46-4.42 (m, 2H), 4.29-4.18 (m, 7H), 3.91-3.85 (m, 2H), 3.74 (dd, J1 = 2.3 Hz, J2= 11.6
Hz, 1H), 3.57 (d, J = 17.3 Hz, SCH2CH), 3.50 (dd, J1 = 2.3 Hz, J2 = 11.6 Hz, 1H), 3.36 (d, J =
17.3 Hz, SCH2CH), 2.17 (s, 3H, Ac), 2.12 (s, 3H, Ac), 2.10 (s, 18H, Ac), 2.06 (s, 3H, Ac), 1.99 (s,
3H, Ac). 13C NMR (100 MHz, CDCl3) δ 170.55, 170.45, 170.23, 170.13, 170.11, 169.83, 169.71,
169.65, 169.39, 115.98, 105.73 (C-1Ara-B), 105.52 (C-1Ara-B'), 104.74 (C-1Ara-A), 82.13, 81.45,
81.28, 80.81 (C-1Man), 80.78, 80.58, 80.14, 79.60, 70.49, 69.47, 66.53, 65.56, 65.13, 63.23, 62.73,
20.85, 20.73, 20.69, 20.63, 20.54, 14.77. HR ESI-TOF MS (m/z): calcd for C43H57NO27SNa
[M+Na]+, 1074.2731; found 1074.2735.
4.2.10 Cyanomethyl
(2,3,4,6-tetra-O-acetyl-α-
3)-[(2,3,4,6-tetra-O-acetyl-α-
(1→3)-]-(2-O-acetyl-α- -arabinofuranosyl)-(1→6)-2,3,4-tri-O-acetyl-1-thio-α-
D
-mannopyranosyl)-(1→5)-(2,3-di-O-acetyl-α-
D
-arabinofuranosyl)-(1→
-arabinofuranosyl)-
-mannopyranosid
D-mannopyranosyl)-(1→5)-(2,3-di-O-acetyl-α-
D
D
D

e (19)
The dried 16 (45 mg, 0.032 mmol), 14 (18 mg, 0.049 mmol) and activated 4Å MS (100 mg) were
mixed together in anhydrous CH2Cl2 (2 mL) for 1 h at RT. The reaction was put at -30 °C for 5
min before NIS (8.5 mg, 0.038 mmol) and AgOTf (10 mg, 0.038 mmol) were added. Upon for
stirring at -30 °C for 50 min, the reaction was diluted with CH2Cl2, after filtering by Celite, the
organic phase was washed with saturated aq. Na2S2O3, water and dried over MgSO4, and then
concentrated under reduced pressure. The residue was finally purified by silica gel column
chromatography with cyclohexane: ethyl acetate (1:3) as the eluent to afford 18 (26 mg, 49%) as
25
1
foamy solid. Rf = 0.25 (cyclohexane: ethyl acetate = 1:3). [α]D +70.4 (c 0.5, CHCl3). H NMR
(400 MHz, CDCl3) δ 5.45 (s, 1H, H-1Man-A), 5.35 (m, 1H), 5.28 (m, 6H), 5.21(m, 1H, H-1Ara-A),
5,21-5.17 (m, 2H), 5.15-5.10 (m, 3H), 5.00-4.97 (m, 2H), 5.10 (m, 1H, H-1Ara-B), 4.97 (m, 1H,
H-1Ara-B'), 4.89 (m, 2H, H-1Man-B, H-1Man-B'), 4.31-4.23 (m, 3H), 4.18 (m, 2H), 4.13-4.07 (m, 6H),
4.02-3.96 (m, 2H), 3.89-3.77 (m, 3H), 3.72-3.68 (m, 2H), 3.58 (d, J = 17 Hz, 1H), 3.50 (m, 1H),
3.36 (d, J = 17 Hz, 1H), 2.15 (s, 3H, Ac), 2.13 (s, 9H, Ac), 2.10 (s, 3H, Ac), 2.08 (s, 15H, Ac),
2.04 (s, 3H, Ac), 2.02 (s, 6H, Ac), 1.96 (s, 3H, Ac), 1.95 (s, 3H, Ac), 1.94 (s, 3H, Ac). 13C NMR
(100 MHz, CDCl3) δ 170.60, 170.39, 170.33, 170.06, 169.90, 169.88, 169.85, 169.66, 169.62,
169.58, 116.00, 105.58 (C-1Ara-B), 105.46 (C-1Ara-B'), 104.69 (C-1Ara-A), 98.12 (C-1Man-B, C-1Man-B'),
82.42, 82.04, 81.94, 81.75, 81.22, 81.07, 80.50 (C-1Man-A), 79.75, 77.53, 70.53, 69.47, 69.22,
69.10, 69.06, 68.74, 68.67, 66.51, 66.44, 66.08, 65.92, 65.66, 65.21, 62.24, 20.84, 20.71, 20.69,
20.60, 20.53, 14.77. HR ESI-TOF MS (m/z): calcd for C67H89NO43SNH4 [M+NH4]+, 1645.4867;
found 1645.4871.
4.2.11 Cyanomethyl
(2,3,4,6-tetra-O-acetyl-α-
D
-mannopyranosyl)-(1→3)-[2,3,4,6-tetra-O-acetyl-α-
D
-mannopyranosyl
-mannopyranosid
-(1→5)-]-(2-O-acetyl-α-
D
-arabinofuranosyl)-(1→6)-2,3,4-tri-O-acetyl-1-thio-α-
D
e (20)
The dried 17 (40 mg, 0.041 mmol), 14 (22 mg, 0.061 mmol) and activated 4Å MS (100 mg) were
mixed together in anhydrous CH2Cl2 (2 mL) for 1 h at RT. The reaction was put at -30 °C for 5
min before NIS (11 mg, 0.049 mmol) and AgOTf (12.5 mg, 0.049 mmol) were added. Upon for
stirring at -30 °C for 50 min, the reaction was diluted with CH2Cl2, after filtering by Celite, the
organic phase was washed with saturated aq. Na2S2O3, water and dried over MgSO4, and then
concentrated under reduced pressure. The residue was finally purified by silica gel column
chromatography with cyclohexane: ethyl acetate (1:2) as the eluent to afford 18 (27 mg, 55%) as
25
1
foamy solid. Rf = 0.3 (cyclohexane: ethyl acetate = 1:2). [α]D +129.7 (c 0.5, CHCl3). H NMR
(400 MHz, CDCl3) δ 5.46 (m, H-1Man-A), 5.39 (dd, J1 = 1.3 Hz, J2 = 3.2 Hz, 1H), 5.33-5.27 (m,
6H), 5.24-5.17 (m, 3H), 5.10 (m, H-1Man-B), 5.00 (m, 1H), 4.99-4.98 (m, 2H, H-1Man-B', H-1Ara),
4.35-4.21 (m, 5H), 4.14-4.06 (m, 2H), 4.03-3.82 (m, 5H), 3.60-3.54 (m, 2H), 3.49 (d, J = 17 Hz,
1H), 2.18 (s, 3H, Ac), 2.16 (s, 3H, Ac), 2.15 (s, 3H, Ac), 2.11 (s, 3H, Ac), 2.10 (s, 3H, Ac), 2.10 (s,
3H, Ac), 2.07 (s, 3H, Ac), 2.05 (s, 3H, Ac), 2.04 (s, 3H, Ac), 1.99 (m, 6H, Ac), 1.96 (s, 3H, Ac).
13C NMR (100 MHz, CDCl3) δ 170.56, 170.40, 170.23, 170.03, 169.87, 169.85, 169.80, 169.71,
169.68, 169.66, 169.58, 116.17, 105.52 (C-1Ara-A), 98.44 (C-1Man-B), 97.32 (C-1Man-B'), 81.83,
81.70, 81.56, 80.73 (C-1Man-A), 70.47, 69.40, 69.37, 69.15, 69.02, 68.93, 68.80, 66.54, 66.17,

66.03, 65.79, 65.66, 62.49, 62.41, 20.85, 20.81, 20.75, 20.70, 20.66, 20.63, 20.60, 20.58, 20.53,
14.85. HR ESI-TOF MS (m/z): calcd for C67H89NO43SNH4 [M+NH4]+, 1213.3599; found
1213.3605.
4.3 Preparation of neo-glycoproteins
Compounds 18–20 (0.005-0.01 mmol) were dissolved in 3 mL of dry methanol and 1 wt% of
sodium methoxide (0.18 mmol) was added. The reaction was reacted for 48 h under nitrogen at
35 °C, after which time the reaction mixture was concentrated in vacuo and the solid formed was
characterized by ESI-MS analysis (See Figures S1-S3 in the additional information). The MS
analysis were carried out on a LTQ-MS (Thermo Electron, San Jose, CA) with an electrospray
ionization (ESI) source controlled by X-calibur software 1.4 (Thermo Finnigan, San Jose, CA).
The samples were dissolved in methanol and introduced into the mass spectrometer with a syringe
pump at a flow rate of 10 μL/min. Full scan MS experiments were carried out under the following
instrumental conditions: positive ion mode, mass range 300–1500 m/z, source voltage 4.6 kV,
capillary voltage 36 V, sheath gas 6, auxiliary gas 1, capillary temperature 250 °C, tube lens
voltage 80 V.
rHSA derived from gene-modified rice endosperm, was selected as carrier protein in order to
have a pure (>99% by SDS-PAGE) homogeneous and non-antigenic protein. The reaction
solutions were analysed by Hydrophilic Interaction Liquid Chromatography (HILIC) using an UV
detector. Thus, an aliquot of 10 µL of each reaction mixture was brought to pH 6 by adding 1M
HCl (2.3 µL), diluted with water to reach a final concentration of 0.5 mg/mL and directly injected
in the chromatographic system. Chromatographic separations were performed on an Agilent
HPLC series 1100 system, equipped with mobile-phase online degasser, quaternary pump,
autosampler, column thermostated compartment, and diode array detector. For data acquisition
and analysis, the ChemStation software version Rev. B.04.01 was used in a Microsoft Windows
XP environment.
The chromatographic conditions were optimized starting from an already reported method[29].
The selected HILIC column was an AdvanceBio Glycan Mapping column (150 × 2.1 mm, 2.7 μm)
from Agilent Technologies (Palo Alto, CA, USA). The mobile phase was composed of ACN and
water both containing 0.1% TFA. The elution gradient was from 28% to 51% B in 35 min. The
injection volume was fixed at 5 μL, the column temperature at 60 °C and the flow rate was set at
0.25 mL/min. According to the results obtained in the optimization study, the final glycosylation
method entails a rHSA concentration of 2.0 mg/mL, a glycoside/protein molar ratio of 200/1 and
the use of sodium tetraborate buffer (100 mM, pH 9.5). Thus, rHSA was dissolved in the buffer
and the resulting solution was mixed with IME-glycoside. The reaction mixture was vortexed for
1 min and incubated for 24 h at 25 °C under continuous stirring. At the end of the reaction, the
samples were purified by ultrafiltration in order to remove reagents and salts (seven 10 min steps
of centrifugation at 13000 g and 4 °C) using centrifuge 5804-R (Eppendorf s.r.l., Milan, Italy) and
Millipore's Amicon® Ultra filters with a nominal molecular weight limit of 10 kDa and load
capacity of 500 μL. HSA conjugates were finally collected in water and lyophilized. MALDI-ToF
MS analysis was performed for the characterization of the different neo-glycoproteins (see Figure
S5).

The experiments were performed on an Autoflex III instrument (Bruker Daltonics, Inc., Bremen,
Germany), equipped with a SmartBeam Nd:YAG laser emitting at 355 nm. Each mass spectrum
represents an accumulation of 3000 laser shots obtained in the linear mode, with a repetition rate
of 100 Hz, using DHB as the matrix. Calibration was performed using ProMix 3 (LaserBio Labs,
Valbonne, FR).
Based upon the mass difference between the peak summit for the un-conjugated rHSA and each
of the 3 conjugated saccharides, using the molecular weight of the different IME-oligosaccharide, it
was estimated the average level of conjugation (average number of oligosaccharides attached): 21
(ManAra3-rHSA) 14.24; 22 (Man3Ara3-rHSA) 12.21; 23 (Man3Ara) 14.22.
4.4 Biological ex-vivo Immunological evaluation
4.4.1 Studied population: The population considered in the study included 30 subjects with
newly diagnosed and untreated active pulmonary TB, as well as 16 healthy individuals without
any history of TB exposure (hereafter indicated as TB unexposed controls). The diagnosis of
active TB was confirmed by M. tuberculosis culture isolation in all cases. A detailed analysis of
the main demographic features of each study group are shown in the Supplementary information
(Table S5).
4.4.2 ELISA assay: Antibodies directed against neo-glycoproteins were determined as previously
described by ELISA assay with minor modifications[32]. ELISA assay was performed in 96-well
microtiter plates (Nunc Maxisorp, Invitrogen, Italy) coated with rHSA, 21, 22 and 23 (1
microg/well) diluted in 100 µL of carbonate buffer (pH = 9.6) and incubated 1 h at RT and
overnight at 4 °C. The wells were then washed (2 times with 200 µL phosphate-buffered saline
(PBS) containing 0.05% Tween-20 (PBS-T). The plate was incubated (RT, 1 h) with 100 mL of
serum (1:100). Plates were incubated 1 h at RT with anti-human IgG Secondary Antibody HRP
conjugated (Invitrogen, Italy). All samples were measured in duplicate and for each serum,
difference in mean ODs of antigen- and buffer-coated wells was calculated.
4.4.3 Statistical analysis and cut off determination: Data was reported as the mean values ±
standard deviation (SD). Comparisons between two groups were performed by Student's t-test. p
values<0.05 were considered to be statistically significant. GraphPad Prism version 5.0 (GraphPad
Software, Inc., La Jolla, CA) was used for all statistical analyses and graphs. By utilizing the
formula (Average of negative controls + (3 X Standard deviation of negative controls)), the cut of
value was set for each protein/glycoprotein.
4.4.4 Ethics statement: The Ethics Committee of the University of Rome “Tor Vergata” (Rome,
Italy) approved the study protocol (protocol number 173/19).
Acknowledgements
We thank the China Scholarship Council (CSC) for Ph.D. fellowships to Zhihao Li and Changping
Zheng. Financial supports from the Centre National de la Recherche Scientifique (CNRS) and the
Sorbonne Université in France are gratefully acknowledged. We also thank Linyan Wang, the

account Manager of Healthgen Biotechnology Corp (Wuhan, China) for the technical assistance in
the use of rHSA.
References
[1] Global tuberculosis report 2019, 2019. https://www.who.int/tb/publications/global_report/en/.
[2] N.R. Gandhi, P. Nunn, K. Dheda, H.S. Schaaf, M. Zignol, D. Van Soolingen, P. Jensen, J. Bayona,
Multidrug-resistant and extensively drug-resistant tuberculosis: a threat to global control of
tuberculosis, The Lancet, 375 (2010) 1830-1843.
[3] H. McShane, Tuberculosis vaccines: beyond bacille Calmette–Guérin, Philosophical Transactions of
the Royal Society B: Biological Sciences, 366 (2011) 2782-2789.
[4] R.D. Astronomo, D.R. Burton, Carbohydrate vaccines: developing sweet solutions to sticky
situations?, Nature Reviews Drug Discovery, 9 (2010) 308-324.
[5] M. Tontini, F. Berti, M. Romano, D. Proietti, C. Zambonelli, M. Bottomley, E. De Gregorio, G. Del
Giudice, R. Rappuoli, P. Costantino, Comparison of CRM197, diphtheria toxoid and tetanus toxoid as
protein carriers for meningococcal glycoconjugate vaccines, Vaccine, 31 (2013) 4827-4833.
[6] R. Adamo, A. Nilo, B. Castagner, O. Boutureira, F. Berti, G.J. Bernardes, Synthetically defined
glycoprotein vaccines: current status and future directions, Chemical Science, 4 (2013) 2995-3008.
[7] P. Stallforth, B. Lepenies, A. Adibekian, P.H. Seeberger, Carbohydrates: a frontier in medicinal
chemistry, Journal of Medicinal Chemistry, 52 (2009) 5561-5577.
[8] O. Finco, R. Rappuoli, Designing vaccines for the twenty-first century society, Frontiers in
Immunology, 5 (2014) 12.
[9] T.L. Lowary, Twenty years of mycobacterial glycans: furanosides and beyond, Accounts of Chemical
Research, 49 (2016) 1379-1388.
[10] A.K. Mishra, N.N. Driessen, B.J. Appelmelk, G.S. Besra, Lipoarabinomannan and related
glycoconjugates: structure, biogenesis and role in Mycobacterium tuberculosis physiology and
host–pathogen interaction, FEMS Microbiology Reviews, 35 (2011) 1126-1157.
[11] M. Jankute, S. Grover, H.L. Birch, G.S. Besra, Genetics of mycobacterial arabinogalactan and
lipoarabinomannan assembly, Molecular Genetics of Mycobacteria, (2014) 535-557.
[12] A. Viljoen, M. Richard, P.C. Nguyen, P. Fourquet, L. Camoin, R.R. Paudal, G.R. Gnawali, C.D.
Spilling, J.-F. Cavalier, S. Canaan, Cyclipostins and cyclophostin analogs inhibit the antigen 85C from
Mycobacterium tuberculosis both in vitro and in vivo, Journal of Biological Chemistry, 293 (2018)
2755-2769.
[13] G. Källenius, A. Pawlowski, B. Hamasur, S.B. Svenson, Mycobacterial glycoconjugates as vaccine
candidates against tuberculosis, Trends in Microbiology, 16 (2008) 456-462.
[14] Z. Li, C. Zheng, M. Terreni, L. Tanzi, M. Sollogoub, Y. Zhang, Novel vaccine candidates against
tuberculosis, Current Medicinal Chemistry, (2020). doi: 10.2174/0929867326666181126112124.
[15] Z. Li, C. Zheng, M. Terreni, T. Bavaro, M. Sollogoub, Y. Zhang, A Concise Synthesis of
Oligosaccharides Derived From Lipoarabinomannan (LAM) with Glycosyl Donors Having a
Nonparticipating Group at C2, European Journal of Organic Chemistry, 14 (2020) 2033-2044.
[16] L. Wang, S. Feng, L. An, G. Gu, Z. Guo, Synthetic and immunological studies of mycobacterial
lipoarabinomannan oligosaccharides and their protein conjugates, The Journal of Organic Chemistry,
80 (2015) 10060-10075.

[17] T. Bavaro, M. Filice, C. Temporini, S. Tengattini, I. Serra, C. Morelli, G. Massolini, M. Terreni,
Chemoenzymatic synthesis of neoglycoproteins driven by the assessment of protein surface reactivity,
RSC Advances, 4 (2014) 56455-56465.
[18] A. Dunne, J.M. Palomo, Efficient and green approach for the complete deprotection of
O-acetylated biomolecules, RSC Advances, 6 (2016) 88974-88978.
[19] M. Filice, J.M. Guisan, M. Terreni, J.M. Palomo, Regioselective monodeprotection of peracetylated
carbohydrates, Nature Protocols, 7 (2012) 1783.
[20] J.C. Pastre, D.L. Browne, S.V. Ley, Flow chemistry syntheses of natural products, Chemical Society
Reviews, 42 (2013) 8849-8869.
[21] R. Yuryev, S. Strompen, A. Liese, Coupled chemo (enzymatic) reactions in continuous flow,
Beilstein Journal of Organic Chemistry, 7 (2011) 1449-1467.
[22] P. Costantino, R. Rappuoli, F. Berti, The design of semi-synthetic and synthetic glycoconjugate
vaccines, Expert Opinion on Drug Discovery, 6 (2011) 1045-1066.
[23] C. Temporini, T. Bavaro, S. Tengattini, I. Serra, G. Marrubini, E. Calleri, F. Fasanella, L. Piubelli, F.
Marinelli, L. Pollegioni, G. Speranza, G. Massolini and M. Terreni. Liquid chromatography–mass
spectrometry structural characterization of neo glycoproteins aiding the rational design and synthesis
of a novel glycovaccine for protection against tuberculosis, Journal of Chromatography A, 1367 (2014)
57-67.
[24] T. Bavaro, S. Tengattini, L. Piubelli, F. Mangione, R. Bernardini, V. Monzillo, S. Calarota, P. Marone,
M. Amicosante, L. Pollegioni, C. Temporini and M. Terreni. Glycosylation of recombinant antigenic
proteins from Mycobacterium tuberculosis: in silico prediction of protein epitopes and ex vivo
biological evaluation of new semi-synthetic glycoconjugates, Molecules, 22 (2017) 1081.
[25] S.J. Jun, M.S. Moon, S.H. Lee, C.S. Cheong, K.S. Kim, Selective monodeacetylation of methyl 2, 3,
5-tri-O-acetyl-D-arabinofuranoside using biocatalyst, Tetrahedron Letters, 46 (2005) 5063-5065.
[26] T. Bavaro, A. Pinto, F. Dall’Oglio, M.J. Hernáiz, C.F. Morelli, P. Zambelli, C. De Micheli, P. Conti, L.
Tamborini, M. Terreni, Flow-based biocatalysis: Application to peracetylated arabinofuranosyl-1,
5-arabinofuranose synthesis, Process Biochemistry, 72 (2018) 112-118.
[27] M.K. Gurjar, L.K. Reddy, S. Hotha, Synthesis of Oligosaccharides of Motifs D and E of
Arabinogalactan Present in Mycobacterium t uberculosis, The Journal of Organic Chemistry, 66 (2001)
4657-4660.
[28] M. Tosin, P.V. Murphy, Synthesis of structurally defined scaffolds for bivalent ligand display based
on glucuronic acid anilides. The degree of tertiary amide isomerism and folding depends on the
configuration of a glycosyl azide, The Journal of Organic Chemistry, 70 (2005) 4107-4117.
[29] S. Tengattini, E. Domínguez-Vega, C. Temporini, T. Bavaro, F. Rinaldi, L. Piubelli, L. Pollegioni, G.
Massolini, G.W. Somsen, Hydrophilic interaction liquid chromatography-mass spectrometry as a new
tool for the characterization of intact semi-synthetic glycoproteins, Analytica Chimica Acta, 981 (2017)
94-105.
[30] Y. Chang, X. Meng, Y. Li, J. Liang, T. Li, D. Meng, T. Zhu, P. Yu, Synthesis and immunogenicity of the
Mycobacterium tuberculosis arabinomannan–CRM197 conjugate, MedChemComm, 10 (2019)
543-553.
[31] L. Wang, S. Feng, S. Wang, H. Li, Z. Guo, G. Gu, Synthesis and Immunological Comparison of
Differently Linked Lipoarabinomannan Oligosaccharide–Monophosphoryl Lipid A Conjugates as
Antituberculosis Vaccines, The Journal of Organic Chemistry, 82 (2017) 12085-12096.
[32] R. Bernardini, R. Aufieri, A. Detcheva, S. Recchia, R. Cicconi, M. Amicosante, C. Montesano, P.

Rossi, H.K. Tchidjou, B. Petrunov, Neonatal protection and preterm birth reduction following maternal
group B streptococcus vaccination in a mouse model, The Journal of Maternal-Fetal & Neonatal
Medicine, 30 (2017) 2844-2850.

Highlights
Three lipoarabinomannan analogues were synthesized as TB vaccine candidates.
A core intermediate was furnished by chemoenzymatic approach.
ELISA was applied using TB infected patients’ serum to detect the activities.
A glycoconjugate performed excellently which should be considered as a lead compound.

Declaration of interests
☒ The authors declare that they have no known competing financial interests or personal relationships
that could have appeared to influence the work reported in this paper.
☐The authors declare the following financial interests/personal relationships which may be considered
as potential competing interests: