Structure-function relationships of pectic polysaccharides from broccoli by-products with in vitro B lymphocyte stimulatory activity

To study structure-function relationships of pectic polysaccharides with their immunostimulatory activity, broccoli by-products were used. Pectic polysaccharides composed by 64 mol% uronic acids, 18 mol% Ara, and 10 mol% Gal, obtained by hot water extraction, activated B lymphocytes in vitro (25–250 μg/mL).

Available online 7 December 2022

0144-8617/© 2022 The Author(s) Published by Elsevier Ltd This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/)

Structure-function relationships of pectic polysaccharides from broccoli

by-products with in vitro B lymphocyte stimulatory activity

S´onia S Ferreiraa,b,*, Alexandra Correiac,d,e, Artur M.S Silvaa, Dulcineia Ferreira Wessela,f,g,

Susana M Cardosoa, Manuel Vilanovac,d,h, Manuel A Coimbraa,**

aLAQV-REQUIMTE, Department of Chemistry, University of Aveiro, 3810-193 Aveiro, Portugal

bCICECO - Aveiro Institute of Materials, Department of Chemistry, University of Aveiro, 3810-193 Aveiro, Portugal

ci3S, Instituto de Investigaç˜ao e Inovaç˜ao em Saúde, Universidade do Porto, 4200-135 Porto, Portugal

dIBMC, Instituto de Biologia Molecular e Celular, Universidade do Porto, 4150-180 Porto, Portugal

eDGAOT, Faculdade de Ciˆencias, Universidade do Porto, Rua do Campo Alegre, 4169-007 Porto, Portugal

fSchool of Agriculture, Polytechnic Institute of Viseu, 3500-606 Viseu, Portugal

gCITAB, University of Tr´as-os-Montes e Alto Douro, 5001-801 Vila Real, Portugal

hICBAS, Instituto de Ciˆencias Biom´edicas de Abel Salazar, Universidade do Porto, 4050-313 Porto, Portugal

A R T I C L E I N F O

Keywords:

Brassica by-products

Sulfated pectic polysaccharides

Arabinogalactans

NMR

Methylation analysis

Enzyme modified pectin

Structure-function relationships

A B S T R A C T

To study structure-function relationships of pectic polysaccharides with their immunostimulatory activity, broccoli by-products were used Pectic polysaccharides composed by 64 mol% uronic acids, 18 mol% Ara, and

10 mol% Gal, obtained by hot water extraction, activated B lymphocytes in vitro (25–250 μg/mL) To disclose active structural features, combinations of ethanol and chromatographic fractionation and modification of the polysaccharides were performed Polysaccharides insoluble in 80 % ethanol (Et80) showed higher immunosti-mulatory activity than the pristine mixture, which was independent of molecular weight range (12–400 kDa) and removal of terminal or short Ara side chains Chemical sulfation did not promote B lymphocyte activation

However, the action of pectin methylesterase and endo-polygalacturonase on hot water extracted polysaccharides

produced an acidic fraction with a high immunostimulatory activity The de-esterified homogalacturonan region seem to be an important core to confer pectic polysaccharides immunostimulatory activity Therefore, agri-food by-products are a source of pectic polysaccharide functional food ingredients

1 Introduction

Broccoli by-products are a source of pectic polysaccharides (

Petko-wicz & Williams, 2020; Sch¨afer et al., 2017), which information about

their structure and potential bioactivities is scarce (Busato et al., 2020;

Xu et al., 2015; Zhang et al., 2017) In general, pectic polysaccharides

are complex heteropolysaccharides with a high proportion of GalpA

(Waldron & Faulds, 2007) They include various fragments of linear and

ramified regions covalently connected It is well established that the

linear region consists of (α1→4)-D-GalpA residues (homogalacturonan

region) carrying methyl ester groups and that can also be acetylated A

backbone of alternating (α1→4)-D-GalpA and (α1→2)-L

-rhamnopyr-anosyl (L-Rhap) residues, ramified in the Rha by galactans,

arabinoga-lactans, and/or arabinans of varying structure is named type I

rhamnogalacturonan (Vincken et al., 2003; Yapo, 2011) The diversity

of pectic polysaccharide structures is dependent on plant source, stages

of maturity, plant part, or processing

Pectic polysaccharides found in several plants have been related to health effects, from anticancer to immunomodulatory activities (Jin

et al., 2021; Ramberg et al., 2010; Yin et al., 2019) Immunomodulatory polysaccharides may strengthen innate and adaptive immunity by directly interacting with distinct cellular and humoral components of the immune system, or indirectly through complex reaction cascades between immune components (Ferreira et al., 2015) The immunosti-mulatory activity of pectic polysaccharides has been associated to the

branched regions (e.g arabinans (Dourado et al., 2004; Dourado et al.,

2006; Popov & Ovodov, 2013), arabinogalactans (Hamed et al., 2022;

Zou et al., 2017), or their combination (Westereng et al., 2008;

* Correspondence to: S.S Ferreira, LAQV-REQUIMTE, Department of Chemistry, University of Aveiro, 3810-193 Aveiro, Portugal

** Corresponding author

E-mail addresses: soniasferreira@ua.pt (S.S Ferreira), mac@ua.pt (M.A Coimbra)

Contents lists available at ScienceDirect Carbohydrate Polymers

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

https://doi.org/10.1016/j.carbpol.2022.120432

Received 19 July 2022; Received in revised form 18 November 2022; Accepted 1 December 2022

Westereng et al., 2009) In fact, the removal of the linear regions of

pectic polysaccharides by endo-polygalacturonase has been shown to

improve their complement fixing activity (Togola et al., 2008) whereas

no effect or the opposite effect was observed by the removal of the

branching regions with exo-α-L-arabinofuranosidase and galactosidases

(Kiyohara et al., 2010; Nergard et al., 2005) Decrement of complement

fixing activity or splenocytes proliferation was also observed by removal

of arabinofuranosyl (Araf) residues with weak acid hydrolysis (Diallo,

Paulsen, Liljeb¨ack, & Michaelsen, 2001, 2003; Wang, Liu, & Fang, 2005;

Duan et al., 2010; Inngjerdingen et al., 2007) However, the removal of

Araf residues after enzymatic degradation of pectic polysaccharides

from Glinus oppositifolius (L.) Aug DC (Aizoaceae) aerial parts did not

affect their complement fixing activity or their ability to stimulate

Peyer’s patches cells to produce secreted factors able to induce bone

marrow cell proliferation (Inngjerdingen et al., 2007) On the other

hand, the removal of homogalacturonan after endo-polygalacturonase

treatment of pectic polysaccharides from Pterospartum tridentatum (L.)

Willk inflorescences decreased the nitric oxide production by

macro-phages (Martins et al., 2017) These results show that different structural

features of pectic polysaccharides can be involved in the triggering or

modulation of immune responses In addition, chemical sulfation of

polysaccharides has been shown to modulate their immunostimulatory

activity (Ferreira et al., 2015) However, considering pectic

poly-saccharides, contradictory data have been reported with studies

showing improvements (Du et al., 2010) or no effect on spleen cells

proliferation (Wang et al., 2018) Therefore, the establishment of

structure-function relationships of immunomodulatory pectic

poly-saccharides is far from being fully revealed

Following the current state of knowledge about polysaccharides

immunostimulatory activity, the pectic polysaccharides from broccoli

by-products were extracted, fractionated and/or modified with 80 %

ethanol solution, size exclusion chromatography, enzymatic treatments,

chemical sulfation, and anion-exchange chromatography to study in

vitro the potential immunostimulatory activity and to establish

structure-function relationships

2 Material and methods

2.1 Broccoli by-products samples and materials

Broccoli (Brassica oleracea var Parthenon) by-product samples from

the frozen food industry were provided by Monliz SA, Alpiarça,

Portugal They were kept frozen under − 20 ◦C until use All reagents

used were of analytical grade or higher available purity

2.2 Preparation of alcohol insoluble residue (AIR)

To isolate polysaccharides from broccoli by-product cell walls,

inactivate enzymes, and remove low-molecular-weight compounds

(Knee, 1973), AIR was prepared Briefly, ethanol was added to 150 g of

frozen material, previously blended, to obtain a final concentration of

80 % (v/v) After boiling for 10 min, the mixture was cooled in a cold-

water bath and filtered in fritted funnel (G-1) The filtration residue

was dispersed again in 80 % (v/v) ethanol, boiled for 10 min, filtered,

washed with 500 mL of ethanol and 500 mL of acetone, and allowed to

dry at room temperature The AIR obtained was further milled in a

coffee grinder and the moisture content was determined by freeze-

drying

2.3 Extraction of hot water (HW) soluble polysaccharides from AIR

To extract pectic polysaccharides with green solvents, AIR was

submitted to a hot water extraction Briefly, 1 g of AIR was hydrated

with 100 mL of distilled water for 2 h, under constant stirring

After-wards, it was boiled for 1 h, using a condenser to avoid solvent loss The

mixture was cooled in a cold-water bath and the soluble material was

separated from un-extracted residue by centrifugation (24,652g, 4 ◦C, for 30 min) followed by filtration in a fritted funnel (G-3) The soluble material was concentrated in a rotary evaporator and dialysed using membranes with 12 kDa cut-off to obtain HW This sample and the residue were freeze-dried for yield determination HW carbohydrates were analysed and further purified, fractionated, and derivatized (Fig 1)

2.4 Ethanol precipitation of HW

Polysaccharides from HW were fractionated according to their sol-ubility in ethanol solutions HW (100 mg) was hydrated by the addition

of 20 mL of water, heated at 40 ◦C, for 5 min, and vortexed Some

insoluble particles, <1 % of HW dry weight, were removed by centri-fugation at 2500 g, for 10 min Absolute ethanol was added to the soluble

material to obtain a solution of 50 % ethanol After 2 h, at 4 ◦C, as no precipitate was observed, more absolute ethanol was added until reaching a concentration of 80 % ethanol This solution was kept for 2 h

at 4 ◦C, allowing to obtain the compounds that precipitate in the 80 % ethanol solution (Et80), which were separated by centrifugation

(24,652 g, 10 min, 4 ◦C) from the soluble material (SnEt) (Fig 1) Both fractions were evaporated under reduced pressure and freeze-dried for yield determination and carbohydrate analysis

2.5 Size-exclusion chromatography of fraction Et80

To unveil the molecular weight of polysaccharides found in fraction Et80, this was fractionated by a size-exclusion chromatography (SEC) column (XK 16/100, Pharmacia) of Sephacryl S400 HR (10–2000 kDa fractionation range for dextrans, Pharmacia–Biotech) The chromatog-raphy was run using 200 mM phosphate buffer, pH 7, 200 mM NaCl, and 0.02 % sodium azide as eluent, with a flow rate of 1.0 mL/min, at room temperature The sample (10 mg) was dissolved in 1 mL of the same elution buffer and then passed through the column Fractions of 2 mL were collected and assayed for total sugars according to phenol- concentrated sulfuric acid method (absorption at 490 nm) (Dubois

et al., 1956) and at 280 nm for UV-absorbing noncarbohydrate com-pounds The appropriate fractions were combined, dialysed and freeze- dried The average molecular weight of each fraction was established after a calibration of the column with dextrans of average molecular weight 12 kDa (2.8 mg, Supelco-31418–100 mg Sigma), 80 kDa (2.45

mg, Supelco-31421–100 mg Sigma), and 270 kDa (2.84 mg, Supelco-

31423 Sigma) The void volume (V0) was determined with blue dextran (2000 kDa, 5 mg) and the inner volume (Vt) of the column was determined with Glc (5 mg)

2.6 α -L-Arabinofuranosidase treatment of fraction Et80

To study the influence of the terminal Ara of pectic polysaccharides

for the in vitro immunostimulatory activity, the Et80 fraction (30 mg) was hydrolysed with 2 U of Clostridium thermocellum α-L

-arabinofur-anosidase 51 A (EC 3.2.1.55, purified from a recombinant Escherichia coli

strain, Nzytech), for 48 h at 37 ◦C with continuous gentle stirring in 5 mL

of 100 mM sodium acetate buffer, pH 5.5, containing 0.02 % sodium azide (Ferreira, Passos, Cepeda, et al., 2018) After enzymatic treatment and enzyme denaturation by placing in boiling water for 5 min, the hydrolysed material was dialysed and freeze-dried to get sample Et80_Ara

2.7 Chemical sulfation of fraction Et80

The Et80 fraction was sulfated to study the influence of this func-tional group in pectic polysaccharide immunostimulatory activity Sul-fation was performed using a methodology with chlorosulfonic acid and pyridine, according to Du et al (2010) Et80 (100 mg) was suspended in formamide (10 mL) under agitation at room temperature for 15 min,

followed by dropwise addition of sulfating reagent (286 μL

chlor-osulfonic acid and 1714 μL of pyridine) The reaction was stirred at room

temperature for 2 h and maintained at 30 ◦C for 5 h with continuous

stirring After cooling to room temperature, the solution was neutralized

with 15 % (w/v) NaOH solution, dialyzed (Mw cut-off 12 kDa) against

distilled water, concentrated, and freeze-dried to obtain the sulfated

polysaccharide (Sulf_Et80)

2.8 Enzymatic treatment of HW

Ramified regions of pectic polysaccharides from HW were purified

by enzymatic treatment to remove protein and starch, and to cleave the

homogalacturonan backbone of the pectic polysaccharides First, HW

(20 mg) was solubilized in 200 mL of 50 mM phosphate buffer, pH 7,

with 0.1 % sodium azide Pronase, a mixture of proteases (2 mg, from

Streptomyces griseus, ROCHE), was added to this solution and incubated

for 24 h, at 37 ◦C, under constant stirring Enzymes were denatured by

placing in boiling water for 5 min After cooling down, 10 μL of

α-amylase (1210 U/mg protein, 27 mg protein/mL, A-2643, Sigma), 10

μL of endo-polygalacturonase (5000 U/mL, M2, Megazyme), and 1 mL of

pectin methylesterase (75.2 U/mL, from Streptomyces avermitilis,

PRO-ZOMIX) were added and incubated for 24 h, at 37 ◦C, also under

con-stant stirring Enzymes were denatured as previous described and the

solution was dialysed, using membranes with 12 kDa cut-off, before

freeze-drying to obtain Enz_HW

2.9 Anion-exchange chromatography of Enz_HW

Anion-exchange chromatography on diethylaminoethyl cellulose

(DEAE)-Trisacryl M (1 mL of resin/7.5 μmol of uronic acids, Sigma) was

performed on a 40 × 16 mm column (XK 16/20, Pharmacia), at a flow

rate of 0.4 mL/min Enz_AIR (10 mg) was dissolved in 1 mL of 50 mM

Tris-HCl buffer, pH 7.4, 0.02 % sodium azide without formation of

insoluble material after centrifugation (2490 g, 15 min) The sample was

eluted with 40 mL of Tris-HCl buffer, and then with 20 mL solutions of

buffer containing 0.125, 0.250, 0.500, and 1.00 M NaCl, successively,

and washed with 2.00 M and 4.00 M NaCl Fractions of 1.9 mL were

collected and assayed for total sugars according to phenol-concentrated

sulfuric acid method (absorption at 490 nm) (Dubois et al., 1956) and at

280 nm for phenolics The fractions of interest were combined, dialysed,

and freeze-dried

2.10 Carbohydrate analyses

Samples were analysed for their sugar content after acid hydrolysis,

derivatisation to alditol acetates, and analysis of individual neutral

sugars by gas chromatography with flame ionization detector (GC-FID)

(Bastos et al., 2015; Blakeney et al., 1983; Selvendran et al., 1979) and quantification of uronic acids (UA) by the 3-phenylphenol colorimetric method (Blumenkrantz & Asboe-Hansen, 1973) The total sugars were determined by the sum of the amount of the individual sugars Glycosidic-linkage composition was determined by methylation of polysaccharides and analysis of partially methylated alditol acetates (PMAA) (Ciucanu & Kerek, 1984; Coimbra et al., 1996; Harris et al.,

1984; Isogai et al., 1985; Passos & Coimbra, 2013) UA linkages were detected after carboxyl reduction of methylated polysaccharides (Dourado et al., 2004; Lindberg & L¨onngren, 1978) The PMAA were separated and analysed by gas chromatography quadrupole mass spec-trometry (GCqMS) (GC-2010 Plus, Shimadzu, Japan) using a non-polar column HT5 (30 m length, 0.25 mm internal diameter and 0.10 μm stationary phase, Trajan, Australia) as described by Hamed et al (2022)

2.11 NMR experiments

NMR experiments were performed as described by Fernandes et al (2019) 1H and 13C NMR spectra of samples Et80 (25 mg), Enz_HWB (10 mg), and Et80_Ara (25 mg) was recorded in D2O (500 μL), at 60 ◦C, on a Bruker DRX 500 spectrometer operating at 500.13 and 125.77 MHz,

respectively; the chemical shifts were expressed in δ (ppm) values

relative to TSS as external reference 2D (1H,13C) gHSQC (heteronuclear single quantum coherence, using gradient pulses for selection), and 2D (1H,13C) gHMBC (heteronuclear multiple quantum coherence, using

gradient pulses for selection; delays for one-bond and long-range JC/H

couplings were optimised for 145 and 7 Hz, respectively) spectra of samples Et80 and Enz_HWB were also recorded

2.12 In vitro immunostimulatory activity assays 2.12.1 Mice

Female BALB/c mice were obtained from Charles River (Barcelona, Spain) and were kept at i3S animal facilities (University of Porto) Ex-periments followed the Portuguese rules (DL 113/2013), the European Convention for the Protection of Vertebrate Animals used for Experi-mental and Other Scientific Purposes (ETS 123), the directive 2010/63/

EU of the European parliament, and the council of 22 September 2010

on the protection of the animals used for scientific purposes Experi-ments were approved by the institutional board responsible for animal welfare (ORBEA) at i3S (code 2019-5) and by the competent national authority (Direç˜ao Geral de Alimentaç˜ao e Veterin´aria), reference number 014036/2019-07-24

2.12.2 In vitro lymphocyte stimulating effect by flow cytometry analysis

The preparation of spleen lymphocytes for stimulation tests and flow cytometry analysis was performed according to Ferreira et al (2020)

Fig 1 Scheme of fractionation and modification of hot water extract (HW) from broccoli by-product AIR

Cells were stimulated with RPMI (negative control), 2.5 μg/mL of

li-popolysaccharides (LPS) from Escherichia coli O111:B4 (Sigma, St Louis,

B cells positive control), 2.5 μg/mL of concanavalin A (T cell positive

control), or 25, 100 and/or 250 μg/mL of samples Cell viability was

assessed by fluorescent staining with propidium iodide before flow

cytometry Propidium iodide is a live cell impermeant red-fluorescent

nuclear and chromosome counterstain Since propidium iodide does

not permeate live cells, it was used as a fluorescent staining to exclude

dead cells and assess cells viability by flow cytometry The collected data

files were analysed using FlowJo v10.3 software (Tree Star Inc.,

Ash-land, OR, USA)

2.12.3 Detection of LPS

To evaluate possible endotoxin contamination in 5 mg/mL of sample,

LPS presence was evaluated after methanolysis, hexane extractions, and

acetylation of fatty acids for GC–MS analysis as 3-O-acetyl fatty acid

methyl esters (FAME), according to de Santana-Filho et al (2012)

2.12.4 Statistical analysis

Statistical analysis was performed by Student’s t-test using GraphPad

prism, Version 6.0 (GraphPad Software, Inc La Jolla, CA, USA) Results

were represented as mean ± SD from at least three experiments with

different animals (* p < 0.01)

3 Results and discussion

3.1 Broccoli hot water extractable pectic polysaccharides and evaluation

of their immunostimulatory activity

To evaluate the in vitro potential immunostimulatory activity of

pectic polysaccharides from broccoli, an alcohol insoluble residue (AIR) was prepared (60 % yield on a dry weight basis) using broccoli by- products, constituted by the remains of inflorescences, leaves, and stalks, with a carbohydrate composition similar to the broccoli available commercially (Ferreira, Passos, Cardoso, Wessel and Coimbra, 2018) A fraction rich in pectic polysaccharides was obtained by extraction of AIR with hot water, mimicking the boiling procedure when cooking vegetables

Hot water (HW) extraction yielded 12 % of high-molecular-weight material It was composed by 69 % of polysaccharides, rich in uronic acids (UA, 64 mol%), Ara (18 mol%), and Gal (10 mol%), with a small amount of Rha (1 mol%), together with Glc, Xyl, Man, and Fuc (Table 1), indicating that pectic polysaccharides were the main polysaccharides

The presence of (1→4)-GalpA residues in HW, determined by

carboxyl reduction with LiAlD4 of methylated polysaccharides, and the

identification of (1→2)-Rhap and (1→2,4)-Rhap residues (Table 2) confirmed the presence of pectic polysaccharides The high amount of

(1→5)-Araf indicates the existence of arabinans in the ramified domain

of pectic polysaccharides, as found in broccoli stems and other Brassica

(Sch¨afer et al., 2017; Stevens & Selvendran, 1980) The presence of

(1→3)- and (1→3,6)-Galp residues, as well as t-Galp, (1→6)-Galp, and t- GlcpA residues are diagnostic of type II (arabino)galactans On the other

Table 1

Yield, carbohydrate composition, and total carbohydrates of samples obtained by hot water extraction, fractionation, and enzymatic treatments of broccoli by-product alcohol insoluble residue (AIR)

Hot water (HW) extraction

80 % ethanol fractionation of HW

α- L -Arabinofuranosidase treatment of Et80

Size-exclusion chromatography of Et80

Chemical sulfation of Et80

Enzymatic treatments of HW*

Anion-exchange chromatography of Enz_HW

AIR: alcohol insoluble residue; tr traces; *treatments with pronase, α-amylase, pectin methylesterase and endo-polygalacturonase Data are expressed as mean ±

standard deviation of 3 replicates

hand, (1→4)-Galp residues are diagnostic of type I arabinogalactans, and

t-Glcp, (1→4)-Glcp, and (1→4,6)-Glcp may be related with the

occur-rence of residual starch (Femenia et al., 2000), although t-Glcp and

(1→4)-Glcp may be components of pectic polysaccharides, as cellobiose

was reported in plum pectic polysaccharides (Nunes et al., 2012) Also, t-

Xylp and (1→4)-Xylp may derive from pectic polysaccharides and/or

polysaccharides tightly associated to pectic polysaccharides (Schols

et al., 1990)

The potential immunostimulatory activity of HW extract, was

assessed in vitro in murine splenocyte cultures Cells viability was not

affected by incubation with HW extract (100 mg/L) In addition, CD3+

cells (T cells) did not show significant up-regulated expression of the

early activation marker CD69 on their surface, indicating that the

ex-tracts did not stimulate T cells In contrast, higher proportions of CD19+

cells (B cells) expressed CD69 upon stimulation by HW extract, as

compared to non-stimulated control (16 % vs 7.7 % of control, Fig 2) As

no lipid A was detected in 5 mg/mL of HW (limit of quantification of

100 ng/mL), it was excluded a possible lipopolysaccharide

contamina-tion (de Santana-Filho et al., 2012) These results show that pectic

polysaccharides from broccoli have potential immunostimulatory

ac-tivity, as observed for pectic polysaccharides from other sources

(Dourado et al., 2004; Ferreira et al., 2015; Martins et al., 2017; Popov &

Ovodov, 2013; Togola et al., 2008; Westereng et al., 2008, 2009; Zou

et al., 2017) B cells can be found along the intestinal tract in Peyer’s

patches and can be reached by pectic polysaccharides through M cells or

dendritic cells (Huang et al., 2017; Komban et al., 2019) This could

constitute a route for direct B cell activation, despite the low absorption

of pectic polysaccharides Nevertheless, indirect B cell activation may

occur through cytokines produced by polysaccharide-stimulated

enter-ocytes or phagenter-ocytes Recognition of pectic polysaccharides may occur

through pattern recognition receptors (Beukema et al., 2020), expressed

on enterocytes and on dendritic cells and macrophages located in Pey-er’s Patches or in the lamina propria (Farache et al., 2013) The detected

B cell stimulatory effect suggests a potential effect of these poly-saccharides in strengthening antibody-mediated immunity in the intes-tinal tract and their suitability to be used in functional food ingredients

3.2 Purification and modification of broccoli hot water extractable pectic polysaccharides

In order to evaluate the characteristics of pectic polysaccharides from HW extract that may be involved in their immunostimulatory

ac-tivity, two strategies were defined: 1 Fractionation by ethanol

precip-itation followed by size-exclusion chromatography, α-L- arabinofuranosidase treatment or chemical sulfation Ethanol precipi-tation allows to obtain fractions with different proportions of neutral sugars to uronic acids and size-exclusion chromatography allows to obtain fractions with different molecular weight The α-L -arabinofur-anosidase treatment and chemical sulfation allow to modify the poly-saccharide side chains by removal of terminally-linked arabinose

residues and incorporation of sulfate groups in the structure; 2

Purifi-cation by enzymatic treatment with pronase and α-amylase followed by

pectin methylesterase and endo-polygalacturonase treatments This

strategy allows to remove proteins, residual starch, and part of the homogalacturonan backbone, resulting in fractions richer in pectic polysaccharide side chains This enzymatic treated fraction was sub-mitted to an anion-exchange chromatography to separate the different domains according to their charge density

Table 2

Heatmap of glycosidic linkage analysis of samples obtained by hot water extraction (HW) from AIR of broccoli by-products and following purification, fractionation, sulfation, and enzymatic treatments mol% of glycosidic linkages were balanced with the amount of UA determined by the colorimetric method

Deduced

linkages (mol%) HW

HW - 80% ethanol fractionation

Et80 Ara removal Et80 - Size-exclusion

chromatography

Et80 sulfation

HW enzymatic treated

Enz - Anion-exchange chromatography

mol%

3.2.1 Ethanol precipitation of HW extract

To find the immunostimulatory active structural motifs of HW

polysaccharides, a fractionation according to the solubility of the

poly-saccharides in ethanol solutions was followed The HW extract was

dissolved in cold water, centrifuged, and ethanol was added to the

su-pernatant The extract was soluble in 50 % ethanol but precipitated

when the solution reached 80 % ethanol, yielding a fraction insoluble in

80 % ethanol (Et80) that was separated from the 80 % ethanol soluble

material (SnEt) The Et80 fraction accounted for 85 % of HW and

con-tained 64 % of polysaccharides (Table 1) This fraction had UA (65 mol

%) as major sugar, followed by Ara (17 mol%), Gal (10 mol%), and Rha

(1 mol%), similar in carbohydrate content and composition with the

pristine HW extract The SnEt fraction accounted for 14 % of HW mass

and had only 34 % of carbohydrates, mainly Ara (43 mol%) and UA (36

mol%), allowing to infer the presence of highly branched pectic

poly-saccharides, rich in Ara, in agreement with their high solubility in

ethanol (Fernandes et al., 2019), together with non-carbohydrate

com-pounds that are usually adsorbed to the polysaccharides and are

solu-bilized with ethanol (Fernandes et al., 2020; Gonçalves et al., 2018)

When incubated with splenocytes, fraction Et80 stimulated 11 % to

37 % of B cells in a dose-response relationship from 25 to 250 mg/L, respectively (Fig 2), showing a higher immunostimulatory activity than

100 mg/L of initial HW As the sugars and the glycosidic linkage composition of fraction Et80 were comparable to those of the initial HW, the higher immunostimulatory activity of fraction Et80 seems to be due

to the removal of non-carbohydrate compounds in the EtSn Even in small amounts, the EtSn compounds may prevent the expression of polysaccharide immunostimulatory activity, similarly to the effect of chlorogenic acids mixed with coffee arabinogalactans (Passos et al.,

2021)

3.2.1.1 Size-exclusion chromatography of Et80 fraction Polysaccharides

from Et80 fraction were submitted to size-exclusion chromatography on

a Sephacryl 400-HR to characterize their molecular weight distribution (Fig 3.a) and evaluate the effect of this parameter in their immunosti-mulatory activity Although it was not possible to resolve poly-saccharides by size, eluted compounds were separated in four fractions: Et80-I consisted on the material with molecular weight higher than 151

Fig 2 Percentage of B cells activated (CD69+) by treatment with polysaccharides from broccoli by-products at the concentrations of 25, 100, and 250 mg/L Culture medium alone (RPMI) was used as negative control and lipopolysaccharides (LPS) were used as positive control Statistical differences between RPMI (negative

control) and different stimulus are indicated above bars (* p < 0.01) Statistical differences among samples are also highlighted (* p < 0.01)

kDa, accounting for 20 % of the eluted material, and had 37 % of

polysaccharides; Et80-II consisted on the material with molecular

weight between 41 and 151 kDa, accounting for 34 % of the eluted

material, and was the richest fraction in polysaccharides (73 %); Et80-III

consisted on the material with molecular weight between 12 and 41

kDa, accounting for 25 % of the eluted material, which contained 59 %

of sugars and represented the peak of material that absorbs at 280 nm,

ie, diagnostic of the presence of phenolic compounds; Et80-IV consisted

on the material with molecular weight lower than 12 kDa, yielded 11 %,

and had vestigial amounts of carbohydrates, only 2 %

Carbohydrate composition and linkage analysis of fractions Et80-I,

Et80-II, and Et80-III (Tables 1 and 2) revealed their resemblance with

initial Et80 The higher molecular weight polysaccharides (Et80-I) were

richer in Ara (26 mol%) whereas the lower molecular weight

poly-saccharides (Et80-III) had higher amount of UA (60 mol%) Linkage

analysis showed that the branching degree of arabinans, estimated by

the ratio between the branching units (sum of (1→2,5)-, (1→3,5)-, and

twice (1→2,3,5)-Araf linkages) and total Ara, of fraction Et80-II (0.29)

was similar to the initial Et80 (0.27), lower in Et80-I and Et80-III (0.21

and 0.13, respectively) The branching degree of galactans, estimated by

the ratio between the branching units ((1→3,6)-Galp linkages) and total

Gal, was also lower for fractions Et80-I and Et80-III (0.37 and 0.24,

respectively) than for fraction Et80-II and initial Et80 (0.45 and 0.42,

respectively), indicating that the branching degree of arabinans and

galactans was correlated

To evaluate the contribution of molecular weight to fraction Et80

immunostimulatory activity, fractions Et80 I, Et80 II, and Et80 III were

incubated with spleen cells All fractions presented immunostimulatory

activity comparable to Et80 (Fig 2), showing a lack of molecular weight-

immunostimulatory activity relationship in Et80, despite the variation

of arabinans and galactans branching degree

3.2.1.2 α -L-Arabinofuranosidase treatment To evaluate the influence of

terminally-linked Ara of arabinans and arabinogalactans on the

immu-nostimulatory activity of Et80 fraction, it was treated with α-L

-arabi-nofuranosidase This enzymatic treatment allowed to recover 79 % of

the mass of Et80 fraction and 58 % of Ara in Et80_Ara subfraction

Et80_Ara had 61 % of total sugars and was mainly composed by UA (64

mol%), Gal (14 mol%) and Ara (10 mol%) Linkage analysis revealed

that the α-L-arabinofuranosidase treatment promoted the decrease of the

relative content of terminally-, 2,5-, and the 3,5-linked Ara residues, as

well as 3,6-linked Gal residues This treatment allowed to observe an

increase of the relative content of terminally- and 6-linked Gal residues These results show that the Ara residues were removed from positions 2,

3, and 5 of arabinans and 3 of galactans, in accordance with the data reported for other arabinans and arabinogalactans (Ferreira, Passos, Cepeda, et al., 2018; Peng et al., 2016; Taylor et al., 2006) The α-L- arabinofuranosidase treatment resulted in arabinans and galactans with

a branching degree of 0.21 and 0.22, respectively, lower than the 0.27 and 0.42 of the initial Et80 fraction The remaining 58 % Ara residues in Et80_Ara should be part of longer arabinan chains, as α-L -arabinofur-anosidase preferably acts on single and short chain Ara residues (Taylor

et al., 2006)

The sample Et80_Ara was incubated with spleen cells, resulting in stimulation of B cells to the same extension of that observed for Et80 fraction These results show that the removal of terminal Ara or small Ara chains did not interfere with immunostimulatory activity of broccoli pectic polysaccharides These results were in accordance with the lower contribution of Ara residues of pectic polysaccharides with immuno-modulatory activity (Inngjerdingen et al., 2007), showing the impor-tance of the core (1→3,6)- and (1→6)-β-D-galactan for the expression of their activity, in accordance with the reported arabinogalactans anti-complement activity (Peng et al., 2016; Yamada & Kiyohara, 2007; Zou

et al., 2019)

3.2.1.3 Chemical sulfation of Et80 fraction To study the effect of

chemical modification with sulfate groups, Et80 fraction was incubated with chlorosulfonic acid and pyridine, giving origin to Et80_Sulf sub-fraction Et80_Sulf had 52 % of total sugars and a sugar composition resembling Et80 fraction The sulfation was observed in the FTIR spec-trum by the presence of a strong S––O absorption band at 1260 cm− 1

(Fig 4, Castro et al., 2006) In addition, linkage analysis showed the increase of 2,5-, 3,5-, and 2,3,5-linked Ara, indicating the presence of sulfate groups in 2, 3 and 5 positions Considering Gal residues, it was possible to observe a slight decrease in 3- and 4-Gal with an increase of all other residues, allowing to infer some extent of sulfation of the gal-actan moiety Similarly, the glucan moiety seems to have been also sulfated at position 6, inferred by the decrease of 4-Glc and the increase

of 4,6-Glc

When incubated with spleen cells, Et80_Sulf showed the same immunostimulatory activity of Et80 fraction Similar results were observed for other sulfated pectic polysaccharides that retained the same spleen cell proliferation activity (Wang et al., 2018) It is possible that the sulfated polysaccharides contribute positively to the

Fig 3 Chromatograms of a) size exclusion chromatography of Et80, with indication of recovered fractions (FI, FII, FIII, and FIV), blue dextran (void volume),

dextrans (12 kDa, 80 kDa, and 270 kDa), and Glc (inner volume of the column) elution volumes and b) anion-exchange chromatography of Enz_HW with indication of recovered fractions (A, B, and C)

immunostimulatory activity of the sample, but the modification of the

galactan moiety may hidden the structural galactans features also

important for the bioactivity of these polysaccharides, with a net zero

balance

3.2.2 Pectin methylesterase and endo-polygalacturonase treatment of

purified HW extract

To isolate the branched regions of pectic polysaccharides, the hot

water extract from broccoli was treated with pronase and α-amylase, to

remove proteins and starch-like polysaccharides, respectively This was

followed by treatment with pectin methylesterase and endo-

polygalacturonase to methyl de-esterify and hydrolyse

homogalactur-onan domains from pectic polysaccharides After these enzymatic

treatments, 32 % of HW material was recovered as high-molecular-

weight material in Enz_HW fraction, which was composed by 60 % of

sugars (Table 1) This enzymatic treatment removed a large extend of

HW Glc (78 %), in accordance with the presence of starch-like

poly-saccharides Enz_HW was composed mainly by UA (37 mol%), Ara (33

mol%), and Gal (19 mol%), showing a 84 % UA removal, in accordance

with the degradation of homogalacturonans The remaining UA in

Enz_HW is expected to belong to small homogalacturonan regions or

regions with acetyl groups (Searle-van Leeuwen et al., 1996), as well as

rhamnogalacturonans and type II arabinogalactans containing GlcA

(Vincken et al., 2003; Yapo, 2011) Ara and Gal were also lost after the

enzymatic treatment (49 % and 47 %, respectively), indicating that

these residues were constituents of small side chains, which were

removed upon dialysis This inference takes into consideration the

absence of side enzyme effects on other polysaccharides beyond the

homogalacturonan moieties

To obtain pectic polysaccharides with different charge properties,

the Enz_HW fraction was submitted to an anion-exchange

chromatog-raphy (Table 1 and Fig 3.b) As the compounds eluted mainly with

0.125 M of NaCl in the buffer, it was inferred that they constituted an

acidic fraction (Enz_HWB, 50 %) The fraction that eluted with buffer

without NaCl (Enz_HWA) was not retained in the column, accounting for

36 % of Enz_HW This non retained fraction had lower amount of UA (34

mol%) than the retained fraction (40 mol%), allowing to infer that

Enz_HWB had more exposed UA residues than Enz_HWA In addition,

although the amount of Ara and Gal was similar, Enz_HWB had an Ara:

Gal ratio of 2 whereas Enz_HWA had a ratio of 1 Linkage analysis

showed that Enz_HWB had also a slightly higher proportion of 3,5- and

2,3,5-linked Ara than Enz_HWA, allowing to infer the presence of

ara-binans with higher branching degree (0.33) than Enz_HWA (0.24)

(Table 2) Accordingly, 4- and 3,6-linked Gal were recovered mainly in

the non-retained fraction Noncarbohydrate compounds from Enz_HW

(14 %) were eluted with 0.250 M of NaCl (Enz_HWC)

The incubation of Enz_HWB with spleen cells induced a stimulation

of 46 % of B lymphocytes (Fig 2), higher than the activation observed

for the hot water extract (16 %) and even Et80 fraction (26 %),

considering the same amount of stimulus (100 μg/mL) However,

Enz_HW and Enz_HWA stimulated less B lymphocytes (13 % and 10 %,

respectively) These results allow to infer that the active structural

motives from hot water extracted broccoli pectic polysaccharides are related with lateral chains rich in branched Ara with UA residues that can confer acidity to the polymer These structural features can be ob-tained by enzymatic modification of the hot water extract by pectin

methylesterase and endo-polygalacturonase, followed by anion-

exchange chromatography Moreover, it seems that the poly-saccharides recovered in Enz_HWA had a combination of polysaccharide structures that may mask the activity of Enz_HWB, as the pristine Enz_HW, which contained the same structural features, did not present the same extent of activity

3.3 Detailing of structural features from broccoli pectic polysaccharides

by NMR spectroscopy

To detail the structural features of immunostimulatory poly-saccharides from broccoli, fraction Et80, the one with higher immu-nostimulatory activity when strategy 1 (ethanol precipitation) was used

to obtain the polysaccharides, and Enz_HWB, the fraction with higher immunostimulatory activity derived from strategy 2 (enzymatic treat-ment and anion exchange chromatography separation), were analysed

by NMR spectroscopy The 13C NMR, HSQC, and HMBC spectra are represented in Figs 5, 6, and 7 According to these 1D and 2D NMR spectra, methylation analysis (Table 2), and literature about pectic polysaccharides rich in UA, arabinans, and arabinogalactans (Cardoso

et al., 2002; Dourado et al., 2006; Fernandes et al., 2019; Hamed et al.,

2022; Makarova et al., 2013; Makarova et al., 2016; Schols et al., 1990;

Shakhmatov et al., 2014; Shakhmatov et al., 2017; Shakhmatov et al.,

2019), intense signals of the anomeric carbon in the 13C NMR spectrum were attributed to C-1 of α-Araf (106.3–109.2 ppm), α-GalpA (99.2–100.2 ppm), and β-Galp (103.1–104.1 ppm) for both Et80 and

Enz_HWB Based on literature about α-glucans (Chen et al., 2017; Guo

et al., 2015), the signal at 99.5 ppm of Et80 was attributed to C-1 of

α-Glcp Low intense signals of C-1 from β-Galp (102.9 ppm), α-Rhap (99.2 and 100.1 ppm), and β-GlcpA (102.9 ppm) were also observed for

both Et80 and Enz_HWB

Prominent signals in the region of carboxyl groups were observed in the 13C NMR at δC range of 170–171 ppm, indicating the existence of UA with different vicinities Methyl and acetyl esterification were suggested

by both one-bond 13C–1H correlation of NMR resonances, in the HSQC spectrum (Fig 6), and by three-bond 13C–1H correlation of NMR res-onances, in the HMBC spectrum (Fig 7) The signal of CH3 of methyl esterified UA at 52.6/3.69 ppm and the signal of CH3 of acetyl esterified

UA at 20.3/2.04 ppm were observed in the HSQC spectrum The signals

of the carboxyl carbon correlation with methyl groups of CH3CO- at 173.6/2.06 ppm and 172.9/2.05 ppm (176.0/2.09) ppm were observed

in the HMBC spectrum (Makarova et al., 2013; Shakhmatov et al., 2017,

2019) The methyl esterification was residual in Enz_HWB in contrast with Et80 intense signal of CH3O-group, which justify different chemical shifts for GalA anomeric carbon and C-5 (Fig 6) Some assignments were not possible for GalA due to overlapping of signals and to its lower de-gree of freedom when compared to other sugar residues (Dourado et al.,

2006; Schols et al., 1990) Overall, the data indicate the presence of a

Fig 4 FTIR spectra of samples Et80 and Et80_Sulf acquired by ATR

methyl and acetyl esterified galacturonan in Et80 fraction and an acetyl

esterified galacturonan in Enz_HWB (Table 3) Methyl esterification and

acetylation, specifically at O-2 and/or O-3 positions, were previously

observed in pectins from Brassica (Westereng et al., 2006)

In 13C and 1H NMR spectra (Fig 5), C-6/H-6 of deoxy-sugar residues

Rha were observed at 16.1–16.8/1.11–1.18 ppm (Shakhmatov et al.,

2017) These results indicated the presence of t-Rhap, usually found in

the non-reducing terminal of arabinogalactans, and the presence of

(α1→2)-Rhap and (α1→2,4)-Rhap, typical of in RG-I regions, according

with what was observed in methylation analysis

The negative carbon signals in the 13C DEPT-135 (Fig 5) found at δC

69.4, 66.3–66.8, 60.9, 60.7, and 60.5 ppm can be attributed to the -CH2-

of sugar residues, namely the C-5 of Ara and the C-6 of Gal and Glc

Other assignments of the proton and carbon signals of Ara, Gal, and Glc

were reported in Table 3 The identification of signals belonging to

(α1→5)-Araf, (α1→3,5)-Araf, (α1→3)-Araf, (α1→2,5)-Araf, (α1→2,3,5)-

Araf, and t-α-Araf indicated the presence of arabinans (Dourado et al.,

2006; Fernandes et al., 2019; Shakhmatov et al., 2014, 2017, 2019)

Downfield chemical shifts of Ara residues were also found in HSQC spectrum, indicating linkages between Ara-Gal residues, as usually found in arabinogalactans (Makarova et al., 2016; Shakhmatov et al.,

2014, 2017), and confirmed by the decreasing of these signals when Et80 fraction was treated with α-L-arabinofuranosidase (Fig 5.b), where terminally linked Ara residues were mainly removed from galactans (Leboeuf et al., 2004)

The identification of signals belonging to t-β-Galp, (β1→4)-Galp, (β1→3)-Galp, (β1→6)-Galp, and (β1→3,6)-Galp agreed with the

pres-ence of 4-linked galactans, like the ones found in type I arabinoga-lactans, and 6-linked galactans ramified at 3 like the ones found in type

II arabinogalactans Galactans are usually found in ramified regions of pectic polysaccharides and type II arabinogalactans can be found both as free polysaccharides and as side chains of rhamnogalacturonans (Makarova et al., 2016; Shakhmatov et al., 2017, 2019) In this work, no correlation was observed between these two polysaccharides and RG-I, which can be due to low intensity of these possible linkages

A methyl group signal not related with carboxyl group of UA was

Gal

L

Fig 5 Representation of the 13C NMR spectra from a Et80 and b Et80_Ara Illustration of sugar residues using the nomenclature proposed by Neelamegham et al (2019) *Identification of negative signals of DEPT-135 NMR spectrum from Et80

Fig 6 Representation of HSQC spectra of fraction Et80 and Enz_HWB The notations used are given in Table 3