An α-D-galactan and a β-D-glucan from the mushroom Amanita muscaria: Structural characterization and antitumor activity against melanoma

Polysaccharides α-D-galactan (GAL-Am) and β-D-glucan (GLC-Am) were obtained from Amanita muscaria fruiting bodies. They were purified using different methodologies, such as Fehling precipitation (for both fractions), freeze-thawing process and ultrafiltration (for GLC-Am).

Available online 8 September 2021

0144-8617/© 2021 Elsevier Ltd This article is made available under the Elsevier license (http://www.elsevier.com/open-access/userlicense/1.0/)

Structural characterization and antitumor activity against melanoma

Matheus Zavadinacka, Daniel de Lima Bellanb, Jessica Loren da Rocha Bertagea,

Shayane da Silva Milhorinia, Edvaldo da Silva Trindadeb, Fernanda Fogagnoli Simasb,

Guilherme Lanzi Sassakia, Lucimara M.C Cordeiroa, Marcello Iacominia,*

aDepartment of Biochemistry and Molecular Biology, Federal University of Paran´a, Curitiba, PR CEP 81531-980, Brazil

bDepartment of Cell Biology, Federal University of Paran´a, Curitiba, PR CEP 81531-980, Brazil

A R T I C L E I N F O

Keywords:

A muscaria

Polysaccharides

α-Galactan

β-Glucan

Chemical structure

Antimelanoma properties

A B S T R A C T Polysaccharides α-D-galactan (GAL-Am) and β-D-glucan (GLC-Am) were obtained from Amanita muscaria fruiting

bodies They were purified using different methodologies, such as Fehling precipitation (for both fractions), freeze-thawing process and ultrafiltration (for GLC-Am) Results showed that the GAL-Am has (1 → 6)-linked

Galp main chain branched at O-2 by terminal Galp units and has not been previously reported Besides, GLC-Am has (1 → 3)-linked Glcp in the main chain, substituted at O-6 by (1 → 6)-linked β-Glcp units Both are water-

soluble, with 9.0 × 103 g/moL and 1.3 × 105 g/moL, respectively GAL-Am and GLC-Am presented a selec-tive proliferation reduction against B16-F10 melanoma cell line, not affecting non tumoral BALB/3T3 fibroblast cell line Furthermore, both fractions reduced clonogenic capacity of melanoma cell line over an extended period

of time These results were obtained without modulations in B16-F10 cell adhesion, reinforcing the biological activities towards cell proliferation impairment and eliciting these polysaccharides as promising compounds to further exploration of their antimelanoma properties

1 Introduction

The fungal fruiting body, popularly known as mushroom, is an

important source of nutrients and fibers, such as β-glucans (Abreu et al.,

2019), being considered a delicacy food in many cultures and countries

(White et al., 2019)

Commonly known as fly agaric mushroom, A muscaria (L.:Fr) has

toxic and hallucinogenic molecules in its composition such as muscimol,

muscarina, ibotenic acid and muscazone, besides heterocyclic alkaloids

in small amounts (Kondeva-Burdina et al., 2019; Ruthes et al., 2013a,b)

Despite crude mushroom show toxicity, it is consumed in regions of

Europe, North America and Japan after boiling in salt water and

steeping in vinegar (Coville, 1898; Kiho et al., 1992)

In addition, edible mushrooms present a diverse range of biological

properties, including immunomodulatory, hypolipidemic, antibacterial,

anti-inflammatory, hepatoprotective and antitumor effects (Morales

et al., 2020)

Amongst mushrooms biological activities, their antitumor properties

are especially relevant and explored Cancer stands out as the second

leading cause of death worldwide (Bray et al., 2018) One of the few cancer types with a rising incidence rate is melanoma, the most lethal form of skin cancer given its fast progression to the metastatic stage (Ward & Farma, 2017) Although recent advances in targeted and im-munotherapies, cancer and particularly melanoma treatment are mostly palliative and accompanied of several debilitating adverse effects (Ramos-Casals et al., 2020), thus engendering the search for new ther-apeutic approaches Polysaccharides obtained from mushrooms have already demonstrated its potential to modulate malignancy related traits

in cancer models, reducing cell proliferation, inducing selective cyto-toxicity, impairing migration and metastasis and decreasing tumor growth (Kothari et al., 2018; Ren et al., 2012) Their pharmacological relevance already reached cancer clinical trials in China and Japan, as illustrated by PSP - a heteropolysaccharide mainly constituted of D- glucose with a main chain α-(1 → 4) and β-(1 → 3)-linked units - and PSK – mainly constituted of glucan β-(1 → 4)-linked with (1 →

6)-β-gluco-pyranosidic side chains – both obtained from the mushroom Coriolus

versicolor (Habtemariam, 2020; Kobayashi et al., 1995)

It is known that polysaccharides biological properties are

* Corresponding author

E-mail address: iacomini@ufpr.br (M Iacomini)

Contents lists available at ScienceDirect Carbohydrate Polymers

journal homepage: www.elsevier.com/locate/carbpol

https://doi.org/10.1016/j.carbpol.2021.118647

Received 31 March 2021; Received in revised form 3 September 2021; Accepted 4 September 2021

intrinsically linked to their composition and chemical structure (Ruthes

et al., 2013a,b; Vetvicka & Yvin, 2004) and there is limited information

in the literature regarding polysaccharides from A muscaria Ruthes

et al (2013a,b) isolated a fucomannogalactan formed by a (1 → 6)-

linked α-D-galactopyranosyl main chain partially substituted for non-

reducing end units at O-2 mainly by α-L-Fucp and β-D-Manp and a (1

→ 3),(1 → 6) β-D-glucan Regarding biological effects of these

polysaccharides, the (1 → 3),(1 → 6) β-D-glucan showed potent inhibi-tion of inflammatory pain (Ruthes et al., 2013a,b) and significant anti-tumor activity against Sarcoma 180 in mice (Kiho et al., 1992) Therefore, in the present study we advance in the knowledge of

A muscaria polysaccharides, describing for the first time the chemical

structure of an unmethylated α-D-galactan, not yet reported for mush-rooms, as well as the structure of a β-D-glucan Moreover, their possible

Fig 1 Extraction and purification scheme of the β-D-glucan (GLC-Am) and the α-D-galactan (GAL-Am) from A muscaria (L.:Fr)

Fig 2 HSQC spectra of the α-galactan (GAL-Am) The sample was analyzed in D2O at 70 ◦C in a Bruker Avance III 600 MHz (chemical shifts are expressed in δ ppm)

Fig 3 2D NMR analysis of the GAL-Am fraction COSY (A),

TOCSY (B), HMBC and 1H (C), and HSQC-TOCSY - in blue and green - with superimposed HSQC - in red - (D) corre-lation maps of the α-D-galactan (GAL-Am) from A muscaria

(L.:Fr) The sample was analyzed in D2O at 70 ◦C in a Bruker Avance III 600 MHz (chemical shifts are expressed in δ ppm) n = carbon number from the sugar ring The symbols ' and " indicate the units 2,6→)-α-D-Galp-(1 → and α-D-Galp-

(1→, respectively, while the absence of them represent the 6→)-α-D-Galp-(1 → units

antimelanoma activities were further explored against the murine

mel-anoma B16-F10 cell line

2 Material and methods

2.1 Biological material

A muscaria fruiting bodies were collected at the Biological Sciences

Sector gardens (25◦26′48.4′′S 49◦14′16.5′′W) of the Federal University

of Paran´a in Curitiba, Paran´a - Brazil The respective mushrooms were

cleaned, freeze-dried and then grounded to a powder

2.2 General extraction and purification processes

General extraction and purification processes performed on the

A muscaria are represented in the flowchart (Fig 1) The extraction processes were carried out using 270.86 g dehydrated fruiting bodies The material was defatted with chloroform-methanol (2:1; v/v) at 65 ◦C for 3 h using a Soxhlet apparatus Then, the residue was submitted to aqueous extraction at room temperature (25 ◦C) for 6 h (5×, 4 L each) under stirring The aqueous extract was concentrated, and the poly-saccharides precipitated by ethanol addition (3:1, v/v), collected by centrifugation (20 min, 8000 rpm, 25 ◦C) and dialyzed (6–8 kDa membrane) against tap water for 24 h Insoluble materials that formed during the dialysis process were removed by centrifugation (20 min,

Fig 3 (continued)

Fig 4 2D NOESY analysis of the GAL-Am fraction The sample was analyzed in D2O at 70 ◦C in a Bruker Avance III 600 MHz (chemical shifts are expressed in

δ ppm)

10,000 rpm, 25 ◦C) The soluble polysaccharides were named S-Am and

submitted to Fehling treatment (Brito et al., 2018) Cu2+-complexed

polysaccharides recovered by centrifugation (20 min, 10,000 rpm,

25 ◦C) gave rise to the fraction (I-1) while the Cu2+-uncomplexed

polysaccharides were named as S-1 fraction Both were neutralized with

HOAc, dialyzed (2 kDa cut-off membrane) and deionized with mixed ion

exchange resins and then freeze dried The latter one (S-1) was subjected

to freeze-thawing (Gorin & Iacomini, 1984) cycles until no more cold-

water insoluble polysaccharides appear After each cycle, the

precipi-tated fraction was recovered by centrifugation (10,000 rpm, 20 min,

4 ◦C) and the soluble portion was named as S-2 This was again treated

with Fehling solution and Cu2+-uncomplexed polymers (fraction S-3)

were further submitted to ultrafiltration (Millipore®; polyethersulfone;

3 kDa membrane) on a filter holder (Sartorius – Model 16,249) with

compressed air at 10 psi carrier gas, generating the retained (GLC-Am)

and the eluted polysaccharide

Sample I-1, due to the presence of contaminating glucans, suffered a

second precipitation with Fehling reagent, originating a purified

gal-actan polysaccharide (GAL-Am)

2.3 Monosaccharide composition

Polysaccharides (5 mg) were hydrolyzed with 2M TFA at 100 ◦C for

8 h and analyzed by GC–MS as alditol acetate derivatives using a Varian

(model CP-3800) gas chromatograph coupled to an Ion-Trap 4000 mass

spectrometer using a VF5-MS column programmed from 100 to 280 ◦C

at 10 ◦C/min, with He as the carrier gas The monosaccharides obtained

were identified by their typical retention times and electron impact

profiles in comparison with standards (Sassaki et al., 2008)

2.4 Methylation analysis by GC–MS

GAL-Am and GLC-Am were subjected to methylation process adapted from Ciucanu and Kerek (1984) by dissolving the sample in DMSO followed by addition of NaOH powder and methyl iodide, kept under stirring for 30 min at 25 ◦C Samples were maintained overnight; the reaction was neutralized with acetic acid and lyophilized The

methylation process was repeated thrice for each sample The per-O-

methylated derivatives were hydrolyzed with 45% formic acid (1 mL) at

100 ◦C for 15 h (Carbonero et al., 2012) Acid excess was removed by lyophilization followed by reduction with NaBD4 and acetylation

generating a mixture of partially O-methylated alditol acetates

de-rivatives which were analyzed by GC–MS using a Varian (model 4000) gas chromatograph equipped with VF5-MS capillary columns The injector temperature was maintained at 210 ◦C and the oven tempera-ture increased from 50 ◦C (maintained 2 min) to 90 ◦C (20 ◦C/min, then maintained for 1 min), 180 ◦C (5 ◦C/min, then maintained for 2 min) and to 210 ◦C (3 ◦C/min, then maintained for 5 min) Helium was used

as the carrier gas at a flow rate of 1.0 mL/min Partially O-methylated alditol acetates were identified by the ion m/z by comparing their

pos-itive ions with standards The results are expressed as a relative per-centage of each component (Sassaki et al., 2005)

2.5 Controlled Smith degradation of the β-glucan (GLC-Am)

Fraction GLC-Am (50 mg) was solubilized in 25 mL of distilled water and oxidized with 25 mL of 0.1 M sodium periodate at room temperature

in the dark under stirring for 72 h (Delgobo et al., 1998) The material was then dialyzed (2 kDa, cut-off membrane) for 24 h against tap water Subsequently it was reduced with NaBH4 to pH ~8 and kept at room

Table 1

Chemical shiftsa of the correlation between the 1H/1H and 1H/13C of the adjacent carbons in monosaccharides units of the GAL-Am

6→)-α-Galp-(1→ Correlations 2,6→)-α-Galp-(1→ Correlations α-Galp-(1→ 1 H/ 13 C

1 H-1/ 1 H-2 4.98/3.88 b 1 H-1/ 1 H-2 d 5.02/3.97 b 1 H-1/ 1 H-2 5.15/3.88 b

1 H-2/ 13 C-3 3.88/68.4 c 13 C-2/ 1 H-3 75.1/3.98 c 1 H-2/ 13 C-3 3.88/69.5 b

1 H-3/ 1 H-4 3.85/4.03 b 1 H-3/ 13 C-4 3.98/66.2 c 1 H-3/ 1 H-4 3.94/4.03 d

1 H-4/ 1 H-5 4.03/4.17 b 1 H-4/ 1 H-5 4.25/4.17 b 1 H-4/ 1 H-5 4.03/4.18 d

1 H-5/ 13 C-6a 3.91/66.7 c 13 C-5/ 1 H-6a 68.8/3.91 c 1 H-5/ 1 H-6c 4.18/3.76 c

aAssignments are based on 1H and HSQC analysis and are expressed as ppm

b Signals in COSY spectrum

cSignals in HSQC-TOCSY/HSQC spectrum

dSignals in TOCSY spectrum

Table 2

1H and 13C NMR chemical shiftsa of fractions GAL-Amb, GLC-Amc and GLC-Smd from A muscaria (L.:Fr)

GAL-Am 6→)-α-D-Galp-(1→ 13 C 98.1 69.6 68.4 69.5 68.8 66.7 66.7

2,6→)-α-D-Galp-(1→ 13 C 98.0 75.1 66.9 66.2 68.8 66.7 66.7

3→)-β-D-Glcp-(1→ 13 C 103.2 72.8 86.3/86.7 68.7 76.4 60.9 60.9

3,6→)-β-D-Glcp-(1→ 13 C 103.2 72.8 86.1 68.7 74.8 68.5 68.5

GLC-Sm 3→)-β-D-Glcp-(1→ 13 C 103.0 72.9 86.2 68.4 76.4 60.9 60.9

aAssignments are based on 13C NMR, 1H and HSQC analysis and are expressed as ppm

b GAL-Am: α-D-galactan isolated from A muscaria

cGLC-Am: β-D-glucan isolated from A muscaria

dGLC-Sm: GLC-Am that was submitted to a controlled Smith degradation, according to material and methods, Section 2.5

temperature for 20 h, neutralized with acetic acid and dialyzed (2 kDa)

for 24 h Then, it was hydrolyzed with TFA pH 2.0 at 100 ◦C for 30 min,

dialyzed (2 kDa) (Ruthes et al., 2015), subsequently precipitated by

ethanol (3:1, v/v) and lyophilized The residual polysaccharide was

named GLC-Sm and analyzed by NMR spectroscopy

2.6 Determination of homogeneity and relative molecular weight

The homogeneity and relative Mw were determined by high

perfor-mance steric exclusion chromatography (HPSEC) using a refractive

index (RI) detector and 4 Waters Ultrahydrogel columns – 120, 250, 500

and 2000 were coupled in series The eluent was 0.1 M NaNO3,

con-taining 0.5 g/L NaN3 Each sample was dissolved in the solvent (1 mg/

mL) and filtered through a membrane (0.22 μm, Millipore) The relative

Mw was determined using a calibration curve of dextran standards (9.4

kDa, 17.2 kDa, 40.2 kDa, 72.2 kDa, 124 kDa, 266 kDa and 487 kDa, from

Sigma) The data obtained were analyzed by the Wyatt Technology

ASTRA program

2.7 Nuclear magnetic resonance (NMR) spectroscopy

NMR analyses were performed on a Bruker Avance III HD

spec-trometer (Bruker), at base frequencies of 400 MHz (1H), and 100 MHz

(13C), and on a Bruker Avance III 600 MHz spectrometer (Bruker

Ger-many) An aliquot of each sample (20 mg) was dissolved in 500 μL of

solvent (D2O or Me2SO-d 6) and left under stirring in a magnetic stirrer

during 24 h and transferred to a 5 mm NMR tube The analyses were carried out at temperature of 70 ◦C The chemical shifts are expressed in

δ (ppm) relative to signals from solvent Me2SO-d 6 at δ 39.7 (13C) and 2.5 (1H) or external reference of acetone (at δ 30.2 and 2.22 for 13C and 1H, respectively)

2.8 Total sugar and protein content determination

Total sugar content present in GAL-Am and GLC-Am fractions was determined spectrophotometrically following the method of Dubois (Dubois et al., 1956) Galactose and glucose were used to construct the calibration curve for determination of the total sugar content in GAL-Am and GLC-Am, respectively

The protein content present in the samples was determined using a 96-well plate by the Bradford method (Bio-Rad, Hercules, CA, USA) (Bradford, 1976) with bovine serum albumin (BSA) as the standard and performed according to the manufacturer's instructions

2.9 In vitro antitumor activity of polysaccharides 2.9.1 Cell culture and polysaccharide preparation

The B16-F10 murine melanoma (BCRJ, Code 0046) and the non- tumorigenic fibroblast BALB/3T3 clone A31 (ATCC, Code CCL-163) cell lines were used Cells were cultivated in Dulbecco's Modified Ea-gle's Medium (DMEM; Gibco, Cat 12800-017, Waltham, Massachusetts, USA), supplemented with 10% fetal bovine serum (FBS; Gibco, Cat

Table 3

Partially O-methylated alditol acetates of the α-D-galactan (GAL-Am) and β-D-glucan (GLC-Am) polysaccharides purified from A muscaria

Partially O-methylated alditol acetatesa Rtb Relative peak area % c Linkage types

GAL-Am d GLC-Am e

aGC–MS analysis on a Varian 4000 capillary column;

bRetention time (min)

cBased on derived O-methylalditol acetates

dGAL-Am: α-D-galactan isolated from A muscaria

eGLC-Am: β-D-glucan isolated from A muscaria

Fig 5 HSQC spectra of the β-glucan (GLC-Am) The sample was analyzed in Me2SO-d 6 at 70 ◦C, at base frequencies of 400 MHz (1H) (chemical shifts are expressed in

δ ppm)

12657029, Waltham, Massachusetts, USA), 0.25 μg/mL penicillin/

streptomycin (Thermo Fisher, Cat 15140122, Waltham, Massachusetts,

USA), 1.57 g/L sodium bicarbonate (Merck, Cat 36486, Kenilworth,

New Jersey, USA), and maintained in a humidified incubator at 37 ◦C

and 5% CO2 For subculture and seeding, cells were always used with

flask confluence not higher than 80%, detached with trypsin/EDTA

(Gibco, Cat 15400054, Waltham, Massachusetts, USA) and counted in

hemocytometer GAL-Am and GLC-Am were dissolved at a final

con-centration of 5 mg/mL in DMEM without FBS and sterilized in 0.22 μm

membranes (Millipore, Cat SLGV033RS, Kenilworth, New Jersey, USA)

following the preparation of serial dilutions in the desired

concentra-tions DMEM without FBS was used as control in the same volume as the

polysaccharides

2.9.2 Cytotoxicity and proliferation analysis

B16-F10 cells (0.5 × 103 cells/well) or BALB/3T3 cells (0.2 × 104

cells/well) were seeded in 96 well plates and exposed or not (control

group) to 10, 100 and 1000 μg/mL of GAL-Am or GLC-Am for 72 h

Cytotoxicity was assessed by the Neutral Red (NR) assay (Borenfreund &

Puerner, 1985), while cell proliferation was analyzed using Crystal

Vi-olet (CV) dye (Gillies et al., 1986)

2.9.3 Clonogenic capacity assay

B16-F10 clonogenic capacity was analyzed in two different

ap-proaches: in the first model, B16-F10 cells (1.2 × 104 cells/well) were

seeded in 6 well plates and then exposed to 10 or 100 μg/mL GAL-Am or

GLC-Am for 72 h, following cell detachment, subsequent seeding in low

density (5.5 × 102 cells/well, 6 well plate) and maintained in culture

without any treatments for more 96 h In the second approach, non-pre-

treated B16-F10 cells were seeded in the same low density conditions in

the presence of 10 or 100 μg/mL GAL-Am or GLC-Am for 96 h Colonies

were fixed, stained with CV and imaged, following area measurement and counting using ImageJ Fiji Software (Schindelin et al., 2012)

2.9.4 Cell adhesion assay

Cell adhesion ability was analyzed on plastic and on pre-coated Matrigel® (BD Biosciences, Cat 356234, San Jose, California, USA) plates For plate coating, 50 μL of 20 μg/mL Matrigel® diluted in cold PBS was pipetted per well on 96 well plates and kept overnight at 4 ◦C Wells were washed once with PBS before cell seeding For cell treatment, B16-F10 cells (1.2 × 104 cells/well) were seeded in 6 well plates and then exposed to 10 or 100 μg/mL GAL-Am or GLC-Am for 72 h, following cell detachment with 2 mM EDTA and counting with a hemocytometer

3 × 104 cells were seeded per well and kept for 2h30m in a cell incu-bator After the adhesion period, wells were washed with PBS (37 ◦C) for removal of non-adherent cells Remaining adherent cells were fixed and stained with CV, following dye elution and absorbance reading in a microplate reader at 570 nm

2.9.5 Statistical analysis

Significant differences between experimental groups were

deter-mined by unpaired t-test with Welch's correction, using GraphPad Prism

6 software Data presented as median ± interquartile range unless stated otherwise

3 Results and discussion

3.1 Chemical structure of the α -D-galactan (GAL-Am) from A muscaria

The purification of the galactan (GAL-Am fraction) was developed sequentially and are depicted in Fig 1 The crude soluble fraction S-Am (Suppl Fig 1A), obtained after ethanolic precipitation and dialysis, was

Fig 6 GAL-Am and GLC-Am cell cytotoxicity and proliferation (A) Neutral red and (B) crystal violet B16-F10 melanoma cell line (C) Neutral red and (D) crystal

violet – Balb/3T3 non-tumor cell line These results represent the set of at least three biologically independent experiments Control represented as a dotted line

treated with Fehling solution, generating the Cu2+-complexed fraction I-

1 (Suppl Fig 1B), which after a new treatment with Fehling solution,

gave rise to fraction GAL-Am, containing the purified galactan (Fig 2)

Results showed that the total sugar content was 95% and protein was

absent in GAL-Am

GAL-Am showed a homogeneous elution profile in HPSEC analysis,

with M w 9.08 × 103 g/mol (Suppl Fig 2A) and composed of 99% of

galactose (Suppl Fig 3A) NMR analyses (1H, 13C, Suppl Figs 4 and 5,

HSQC, COSY, TOCSY, HSQC-TOCSY, HMBC and NOESY spectroscopy,

Figs 2, 3 and 4) contributed to elucidate the α-D-galactan structure

through the signals and intermolecular correlations observed The

α-configuration was confirmed by the coupling constant J C-1,H-1 =173

Hz observed in coupled HSQC spectrum (Perlin & Casu, 1969) It is possible to observe in COSY (Fig 3A) and TOCSY (Fig 3B) experiments signals at δ 4.98/3.85, 5.02/3.97 and 5.15/3.88 indicating the corre-lation between H-1 and H-2 of the three different monosaccharide groups present in the HSQC spectra (Fig 2) HMBC with superimposed

1H (Fig 3C) and NOESY (Fig 4) were performed to confirm the linkage between the different monosaccharide units In HMBC experiment, it is possible to observe signals that correlate the C-2 of the 2,6→)-α-D-Galp-

(1→ units with the H-1 of the α-D-Galp-(1→ units and between the C-1

and the H-6 of the 6→)-α-D-Galp-(1→ units, which indicates the

pres-ence of the linkage types 1 → 2 and 1 → 6, respectively, which can also

be observed in the result of the Nuclear Overhauser enhancement

Fig 7 GAL-Am and GLC-Am reduce B16-F10 colony formation capacity (A) Representative images of pre-treated colony formation assay (a- Control, b- GAL-Am

100 μg/mL, c- GLC-AM 100 μg/mL) (B) Colony counting – pre-treatment (C) Colony area – pre-treatment (D) Colony counting – simultaneous treatment (D) Colony area – simultaneous treatment These results represent the set of at least three biologically independent experiments Data shown as median ± interquartile range for

B and D, and as mean ± SD for C and E

experiment, indicating a spatial molecular proximity relationship

be-tween the cited molecules In HSQC-TOCSY spectrum with

super-imposed HSQC (Fig 3D) the correlation between H-6 and C-5 (δ 3.91/

68.8, 3.71/68.8, 3.91/68.8, 3.71/68.8, 3.76/71.0 for 5/6a, 5/6b, 5/6a,

5/6b, 5/6c, respectively) is indicated for the three monosaccharides

units Through the results obtained, we conclude that the anomeric

signals at δ 98.1/4.98 and 98.0/5.02 in the HSQC spectra correspond to

6-O-linked and 2,6-O-linked α-D-Galp units, respectively, and at δ 95.7/

5.15 corresponds to C-1 of terminal α-D-Galp units linked to C-2 of Galp

units in the main chain Signals at δ 66.7/3.91 and 66.7/3.71 were

assigned to substituted C-6, while those at δ 61.2/3.76 corresponding to

unsubstituted C-6 from terminal units (Brito et al., 2018; Carbonero

et al., 2008) The signals at δ 75.1/3.97 and at δ 69.6/3.88 correspond to

substituted and non-substituted C-2 from Galp units, respectively

The other observed correlations are indicated in the Table 1

ac-cording to signals already reported by Brito et al (2018), Zhang et al

(2013), Carbonero et al (2008) and Rosado et al (2003) Thus, the

assignments of all carbons and hydrogen from the GAL-Am are

sum-marized in Table 2

The analysis of methylated derivatives of GAL-Am fraction (Table 3

and Suppl Fig 6) agrees with NMR analyses The mainly derivative

observed was the 1,5,6-O-Ac3-2,3,4-Me3-galactitol (71%) from (1 → 6)-

linked Galp units from main chain Approximately 14% of the main

chain units were 2-O-substituted by terminal Galp units (15%), which

could be confirmed by the presence of 1,2,5,6-O-Ac4-3,4-Me2-galactitol

and 1,5-O-A2-2,3,4,6-Me4-galactitol derivatives, respectively

Polysaccharides having (1 → 6)-linked Galp units partially

substituted at O-2 by different side chains, constituting

hetero-polysaccharides, have already been reported for A muscaria (Ruthes

et al., 2013a,b) and other mushrooms, such as Agaricus bisporus (Komura

et al., 2010), Flammulina velutipes (Zhang et al., 2012), Fomitella fraxinea

(Cho, Yun, et al., 2011) and Ganoderma atrum (Zhang et al., 2014) It is

worth to note that all of them had distinct chemical characteristics when

compared with the α-D-galactan reported in the present work As far as

we know, α-D-galactan homopolymer has not been described from the

mushroom A muscaria until now Besides, α-D-galactans have

previ-ously been observed in mushrooms of the genus Pleurotus, having (1 →

6)-linked and partially 3-O-methylated α-D-Galp in the main chain (Brito

et al., 2018; Carbonero et al., 2008; Rosado et al., 2003) Conversely,

none of them showed 2-O-substitution by α-D-Galp non-reducing end

units Finally, (1 → 6)-linked α-D-Galp units 2-O-substituted and without

the presence of methyl groups at C-3 have not yet been reported in the

literature

3.2 Chemical structure of the (1 → 3),(1 → 6)-β-D-glucan (GLC-Am) from A muscaria

For the GLC-Am purification, the Cu2+-uncomplexed fraction S-1 (Suppl Fig 7A) obtained after the treatment with Fehling solution of S-

Am fraction has undergone to freezing and thawing process (2×) and a new treatment with Fehling solution (Fig 1), generating the Cu2+- uncomplexed fraction S-3 (Suppl Fig 7B) In order to separate the mixture of polysaccharides present in the S-3 fraction, an ultrafiltration (3 kDa membrane) was performed, which gave rise the retained fraction GLC-Am, containing the purified glucan (Fig 5) Results showed that the total sugar content was 96% and the protein was absent in this fraction

GLC-Am showed a M w of 1.3 × 105 g/mol with a homogeneous elution profile in the HPSEC analysis (Suppl Fig 2B) It was composed

of 98.8% of glucose (Suppl Fig 3B), confirming the presence of a glucan In the methylation data (Table 3 and Suppl Fig 8) it is possible

to identify and characterize a (1 → 3),(1 → 6)-linked D-glucan repre-sented by the presence of 1,5-O-Ac2-2,3,4,6-Me4-glucitol (19%), 1,3,5- O-Ac3-2,4,6-Me3-glucitol (45%), 1,5,6-O-Ac3-2,3,4-Me3-glucitol (13%) and 1,3,5,6-O-Ac4-2,4-Me2-glucitol (23%) derivatives, suggesting a β-D- glucan with (1 → 3)-linked main chain, according to the analysis of residual polysaccharide (GLC-Sm fraction) obtained after controlled Smith degradation (Suppl Fig 9) The six signals observed at δ 103.0/ 4.53, 86.2/3.49, 76.4/3.27, 72.9/3.31, 68.4/3.25 and 60.9/3.71;3.48 are relatives to C1/H1, C3/H3, C5/H5, C2/H2, C4/H4, C6/H6a;b, respectively, also referenced by de Jesus et al (2018); Morales et al (2020) and characteristic of a linear (1 → 3)-β-D-glucan (Table 2) The HSQC spectra (Fig 5) of GLC-Am presented signals relative to the (1 → 3),(1 → 6)-linked D-glucan, with anomeric signals at δ 103.2/ 4.55 and 103.2/4.27 (C1/H1), substituted C3/H3 signals at δ 86.3/3.50, and substituted C6/H6 signals at δ 68.5/4.11 and 68.5/3.58 The H6/C6

signals from terminal β-D-Glcp units can be observed at δ 60.9/3.72 and

60.5/3.50 The NMR chemical shifts are present in Table 2 and were based on 1H, 13C (Suppl Figs 10 and 11), HSQC (Fig 5) and comparison with literature data (Bhanja et al., 2014; Kono et al., 2017; Morales

et al., 2020; Ruthes et al., 2015; Zhu et al., 2015)

(1 → 3),(1 → 6) β-D-glucans had already been isolated from the

mushroom A muscaria and presented different M w and branching degree when compared to GLC-Am fraction Kiho et al (1992) obtained a glucan with estimated Mw of 9.5 × 103 g/mol and (1 → 3)-linked main

chain substituted by two (1 → 6)-linked D-Glcp units at every seven

monosaccharide units in the main chain Besides, Ruthes et al (2013b)

isolated a (1 → 3),(1 → 6) β-D-glucan with Mw 1.6 × 104 g/mol

substituted at O-6 mostly by terminal β-D-Glcp units In addition, GLC-

Am appears to have a Mw similar to the (1 → 3),(1 → 6) β-D-glucan

Fig 8 B16-F10 cell adhesion (A) Cell adhesion on plastic (B) Cell adhesion on Matrigel® These results represent the set of four biologically independent

ex-periments Control represented as a dotted line

isolated from Lactarius rufus (Ruthes et al., 2013a), but differs in the

degree of branching and in the composition of the side chains

3.3 GAL-Am and GLC-Am selectively reduce melanoma cells viability

and proliferation

GAL-Am and GLC-Am significantly reduced murine melanoma B16-

F10 neutral red uptake in all tested concentrations, as showed in

Fig 6A (26.2%, 27.9% and 26.8% for GAL-Am; 22.5%, 23.5% and

32.7% for GLC-Am; concentrations of 10, 100 and 1000 μg/mL,

respectively) Similarly, both polysaccharides reduced B16-F10

prolif-eration at all tested concentrations (Fig 6B; 21%, 26.3% and 34.4% for

GAL-Am; 27.4%, 34.2% and 37.5% for GLC-Am; concentrations of 10,

100 and 1000 μg/mL, respectively) Strikingly, although GAL-Am and

GLC-Am induced significant reductions in B16-F10 proliferation (and

consequently decrease the neutral red color measure), they did not

impaired murine fibroblast 3T3 cell line in the same manner, with only

GAL-Am at 1000 μg/mL decreasing cell viability The smaller

concen-trations (10 and 100 μg/mL) that did not induce any effects on 3T3 cell

line were chosen to be used in the next assays

3.4 GAL-Am and GLC-Am cause B16-F10 colony formation impairment

in extended cultivation periods that persists with treatment withdraw

GAL-Am and GLC-Am treatment were able to affect B16-F10 colony

formation capacity, both when cells were pre-treated and seeded to form

colonies for 96 h without the polysaccharides (Fig 7A–B; 24.8% and

26.2% reduction in comparison to control group for GAL-Am and GLC-

Am, respectively, at 100 μg/mL for both compounds) and when cells

were seeded to form colonies in the polysaccharide presence (Fig 7D;

32.0% and 24.9% reduction in comparison to control group for GAL-Am

and GLC-Am, respectively, at 100 μg/mL for both compounds) In both

conditions and concentrations tested (10 and 100 μg/mL), GAL-Am and

GLC-Am were also able to significantly reduce colonies mean area when

compared to control group (Fig 7C and E)

3.5 B16-F10 cell adhesion is not modulated by GAL-Am nor GLC-Am

B16-F10 treatment with GAL-Am or GLC-AM at 10 or 100 μg/mL did

not alter its capacity to adhere neither in a plastic substrate nor in

Matrigel® coated wells, as showed in Fig 8 The absence of adhesion

modulation reinforces the previous results, as it shows that both

treat-ments did not reduce the number of adherent cells, but rather reduced

their capacity to proliferate both in normal culture conditions and in low

confluence

Although not fully understood, it is clearly that polysaccharide

structural features, such as molecular weight, monosaccharide

compo-sition and type of glycosidic linkage, play an important role in

poly-saccharide biological activity (Xu et al., 2017) Several β-glucans

obtained from mushrooms with similar structure to GLC-Am are

described to present direct and indirect cytotoxic and anti-proliferative

activity against tumor cells (Pandya et al., 2019; Ren et al., 2012) One

example is lentinan, a β-glucan (1 → 3)-linked with β-(1 → 6)-glucose

branches, from the mushroom Lentinus edodes with a M w ranging from

400 to 800 kDa (Pandya et al., 2019) In a similar range of

concentra-tions as used in the present paper, a purified lentinan (605.4 kDa) was

able to significantly reduce the proliferation of murine breast cancer cell

line MCF-7 and murine sarcoma cell line S180 (Zhang et al., 2015)

Another similar β-glucan from L edodes presented a very similar

selec-tive cytotoxicity as demonstrated by GLC-Am against liver cancer cell

line H22 but not affecting human liver normal cell line HL7702 (Wang

et al., 2017) Interestingly, as GLC-Am presents similar structure to these

two polysaccharides but significantly smaller molecular weight, we

suggest that the β-(1 → 3)-glucose backbone and (1 → 6)-β-glucose

branches play an important role in the biological activity of β-glucans

Moreover, GLC-Am glucose linkage and its water solubility are chemical

characteristics already described as important traits eliciting mushroom polysaccharides as anticancer compounds (Lemieszek & Rzeski, 2012) Mushrooms heterogalactans have already shown antitumor activities (Ruthes et al., 2016) Some of these polysaccharides present a degree of similarity to GAL-Am, with α-D-galactose (1 → 6)-linked in the backbone

and inducing in vitro cytotoxicity and proliferation reduction against

cancer cells lines (Pires et al., 2017) A fucogalactan obtained from

Macrocybe titans mushroom, which presented a (1 → 6)-linked α-D-

galactose main chain partially substituted at O-2 by α-L-Fucp units with

a Mw of 14.2 kDa did not induce cytotoxicity nor proliferation changes in B16-F10 cell line, however it was able to reduce cell migration in a similar concentration to GAL-Am colony formation reduction effects (100 μg/mL) (Milhorini et al., 2018) A fucomannogalactan with a Mw of

17.1 kDa obtained from the mushroom Hypsizygus marmoreus with (1 →

6)-linked α-D-galactose main chain partially substituted at O-2 by α-L-

Fucp and β-D-Manp was non-cytotoxic to B16-F10 cell line at

concen-trations from 1 to 100 μg/mL, but was able to reduce cell colony area – similar to GAL-Am - at the highest concentration (Oliveira et al., 2019) Both previously cited polysaccharides containing a (1 → 6)-linked α-D- galactose main chain present a relative low Mw close to GAL-Am (9.05 kDa), which could contribute to its biological effects, as already described for other polysaccharides as a relevant structural feature to-wards higher biological activity (Cho, Lee, & You, 2011; Choi & Kim,

2013)

Acute and long-term adverse effects related to cancer treatment are still one of the major concerns and barriers in drug development Most of the induced adverse effects revolve around the consequences of target-ing cell molecules and mechanics common to cancer and normal cells (Nurgali et al., 2018) Hence, GAL-Am and GLC-Am selective reduction

of B16-F10 proliferation indicates a promising feature of these poly-saccharides that could potentially be translated to direct antitumor ac-tivity without compromising non-tumor cells

CRediT authorship contribution statement

Matheus Zavadinack, Jessica Loren da Rocha Bertage, Shayane da Silva Milhorini and Guilherme Lanzi Sassaki: Carbohydrate Chemistry Experiments, Writing - Original Draft; Daniel de Lima Bellan, Edvaldo da Silva Trindade and Fernanda Fogagnoli Simas: Biological Experiments, Writing - Original Draft, Lucimara M C Cordeiro and Marcello Iaco-mini: Writing - Review & Editing, Supervision, Funding acquisition

Acknowledgements

The authors would like to thank the Brazilian funding agencies: CAPES (Funding Code 001) and CNPq (404717/2016-0, 301719/2016-

0, 307314/2018-9) for financial support, the Chemical Technicians Thiago J dos Santos and Arquimedes Paix˜ao de Santana Filho for GC–MS and NMR analyses, respectively The authors have declared no conflicts of interest

Appendix A Supplementary data

Supplementary data to this article can be found online at https://doi org/10.1016/j.carbpol.2021.118647

References

Abreu, H., Simas, F F., Smiderle, F R., Sovrani, V., Dallazen, J L., Maria-Ferreira, D., … Iacomini, M (2019) Gelling functional property, anti-inflammatory and

antinociceptive bioactivities of β-D-glucan from the edible mushroom Pholiota

nameko International Journal of Biological Macromolecules, 122, 1128–1135 https:// doi.org/10.1016/j.ijbiomac.2018.09.062

Bhanja, S K., Rout, D., Patra, P., Sen, I K., Nandan, C K., & Islam, S S (2014) Water-

insoluble glucans from the edible fungus Ramaria botrytis In Bioactive Carbohydrates

and Dietary Fibre https://doi.org/10.1016/j.bcdf.2014.01.004