Harnessing marine sulfated polysaccharides to inhibit norovirus: from seaweed to solution

LETTER

Human noroviruses are the leading cause of acute gastroenteritis outbreaks worldwide, responsible for >685 million infections annually (1). Still, there are no approved vaccines, antivirals, or specific treatments. One promising approach involves the use of compounds that block norovirus attachment to histo-blood group antigens (HBGAs), which serve as essential co-factors for viral entry (2, 3). Human milk oligosaccharides (HMOs), such as 2′-fucosyllactose (2′-FL), have already demonstrated this capability, where the primary fucose moiety of HMOs bound precisely at the HBGA pocket on the capsid (4–10). Building on this concept, we further investigated the antiviral potential of polysaccharide (fucose-based) fucoidans and ulvan extracted from various seaweed species (Fig. 1), which have shown promising inhibition capacities for noroviruses (11–14) and other viruses (15).

Fig 1.

Hayworth projections of an HBGA A-trisaccharide, HMO 2′-FL, and representative sections of Fucus vesiculosus fucoidan and ulvan. A-trisaccharide is 6-deoxy-a-L-galactopyranosyl-(1→2)-[2-acetamido-2-deoxy-a-D-galactopyranosyl-(1→3)]-D-galactose (GalNAca1,3(Fuc-a1,2)Gal). The 2′-FL is 6-deoxy-α-L-galactopyranosyl-(1→2)-β-D-galactopyranosyl-(1→4)-D-glucose (Fucα1,2Galβ1,4Glc). The Fucus vesiculosus fucoidan (type II) has alternating α-L-fucose-(1, 3)-L-fucose-(1, 4) with sulfate groups and can also contain other monosaccharides like glucose, galactose, and xylose, along with acetyl groups and uronic acids (not shown). Ulvan is a highly complex and variable sulfated hetero polysaccharide that is primarily composed of repeating disaccharide units typically comprising rhamnose-3-sulfate, glucuronic acid, and iduronic acid with lesser amounts of xylose, mannose, arabinose, and galactose. These polysaccharides are generally a few 100 kDa, and have been reported up to 1000 kDa.

Using an established in vitro surrogate neutralization assay (4–6), we evaluated the ability of fucoidan and ulvan extracts (Table 1) to inhibit norovirus virus-like particle (VLP) binding to HBGA-containing human saliva samples. Inhibition profiles were comparable for two clinically significant norovirus genotypes, GII.4 (Fig. 2A) and GII.17 (Fig. 2B), but clear differences in inhibition capacities were observed. Maritech Fucus vesiculosus fucoidan demonstrated the most potent and consistent inhibitory activity for both genotypes with the lowest half-maximal inhibitory concentration (IC₅₀) values, where ^GII.4IC₅₀ = 0.71–3.81 mg/mL and ^GII.17IC₅₀ = 0.82–1.42 mg/mL for A, B, and O blood types. In contrast, the other extracts showed weaker and more variable inhibition profiles, where ulvan ^GII.4IC₅₀ = 12.01–48.25 mg/mL and ^GII.17IC₅₀ = 10.10–44.77 mg/mL; depyrogenated Fucus vesiculosus fucoidan ^GII.4IC₅₀ = 8.46–100.59 mg/mL and ^GII.17IC₅₀ = 31.50–99.95 mg/mL; and depyrogenated Undaria pinnatifida fucoidan ^GII.4IC₅₀ = 3.47–50.50 mg/mL and ^GII.17IC₅₀ = 18.60–67.75 mg/mL. The control 2′-FL was also slightly weaker, ^GII.4IC₅₀ = 15.68–22.28 mg/mL and ^GII.17IC₅₀ = 32.75–52.46 mg/mL, and was comparable to previous data (4).

TABLE 1.

Fucoidan and ulvan extracts

	(% w/w)				Monosaccharides (% mol)
Extract	Total sugars	Uronic acid	Sulfate	Polyphenolics	Fucose	Xylose	Mannose	Galactose	Glucose	Arabinose	Rhamnose
Maritech Fucus vesiculosus fucoidan	74.5	4.1	27.8	5.5	58.4	10.2	7.0	10.6	13.1	0.7	0
Ulvan	39.6	21.0	20.7	0.2	1.1	9.6	1.3	3.4	2.7	0.6	81.3
Depyrogenated Fucus vesiculosus fucoidan	64.6	4.0	29.8	0.2	82.0	9.5	0	8.5	0	0	0
Depyrogenated Undaria pinnatifida fucoidan	54.3	3.4	27.7	0.2	42.4	0.6	2.1	51.7	2.4	0	0.7

Fig 2.

Inhibition profiles of fucoidans, ulvan, and 2′-FL. Blocking assays were performed using a previously described method (4) using saliva samples representing blood types A, B, and O from healthy donor adults. Triplicate wells were coated with a 1:200 dilution of heat-inactivated (99°C, 15 min) human saliva sample overnight at 4°C. Serial dilutions starting at 50 mg/mL of Maritech Fucus vesiculosus fucoidan (≥85% fucoidan, FVS2024564), ulvan (≥85%, UOU2022559), depyrogenated Fucus vesiculosus fucoidan (≥90%, DSFVF2024520), depyrogenated Undaria pinnatifida fucoidan (≥90%, DPGFS2023001), and 100 mg/mL 2′-FL (98%, B08J8DTMGK, Layer Origin, USA) were mixed 1:1 with 25 µg/mL of (GII.4 or GII.17) VLPs and incubated for 1 h at 25°C. Negative control wells (no inhibitor, no Fc-NB26) were subtracted from all OD₄₅₀ values, and untreated VLPs were set as 100% binding. Inhibition (%) was calculated as: 1 – (mean OD₄₅₀ treated / mean OD₄₅₀ reference) × 100. The line graphs show results for GII.4 (A) and GII.17 (B) VLPs with one A-blood type (blue), two B-blood types (light green and green), and one O-blood type (orange). IC₅₀ values were determined using Excel and indicated by dashed lines. IC₅₀ values were estimates for ulvan, depyrogenated Fucus vesiculosus fucoidan, and Undaria pinnatifida fucoidan since two measurements above and below 50% inhibition were not available across all replicates, which is typically required (16).

To offer insight into potential molecular mechanisms by which these extracts might inhibit norovirus binding to HBGAs, we first generated three-dimensional (3D) structural models of fucoidan and ulvan with 4- and 64-sugar residues using GLYCAM-Web (17) and subsequently superpositioned into X-ray crystal structures of GII.4 (Fig. 3A) and GII.17 (Fig. 3B) P domains. The four-sugar fucoidan and ulvan were well-accommodated at the HBGA binding pockets on the P domains without any major steric clashes. In the context of the intact particle, a modeled left-handed helical conformation of the 64-sugar fucoidan (Fig. 3C) and ulvan molecules (Fig. 3D) extended outward from the surface of the VLP. This spatial arrangement may contribute to a steric barrier effect, whereby the fucoidan and ulvan molecules might prevent access of HBGA binding, viral aggregation, and/or interference with mobility and orientation during HBGA attachment.

Fig 3.

Structural modeling of fucoidan and ulvan bound to the norovirus capsid and possible mode of inhibition. (A) GII.4 P domain (light gray) and from left to right: X-ray crystal structure of the GII.4 P domain (light gray) in complex with A-trisaccharide (PDB ID: 4WZT), a model of a four-sugar fucoidan molecule (deep teal), and a model of a four-sugar ulvan molecule (purple) superimposed onto the fucose moiety of the A-trisaccharide (removed for simplicity). The GII.4 P domain residues that bind the fucose of the A-trisaccharide were shown (D374^side-chain, R345^side-chain, G344^main-chain, T444^side-chain, and G443^main-chain) and a flexible loop (residues ~391–396) (18–20). The red arrow shows fucoidan near this flexible loop. (B) GII.17 P domain (dark gray) and from left to right: the X-ray crystal structure of the GII.17 P domain (dark gray) in complex with A-trisaccharide (5LKC), a model showing a fucoidan molecule (deep teal), and a model showing a ulvan molecule (purple) superimposed onto the fucose moiety of the A-trisaccharide (removed for simplicity). The GII.17 P domain residues that bind the fucose of the A-trisaccharide were shown (D378^side-chain, R349^side-chain, G348^side-chain, T444^side-chain, and G443^side-chain) and a flexible loop (residues ~393–397). Models showing stereo views of (C) fucoidan and (D) ulvan molecules superimposed onto eight dimers on a cryo-EM structure of the GII.17 VLP (EMB-10759). In this model, the fucoidan and ulvan molecules radiate out from the virus particle.

Together, these inhibition data and structural models suggest a dual mode of inhibition—direct competition at the HBGA-binding pocket and a broader steric interference at the virion level. These findings provide a molecular framework for understanding fucoidan and ulvan antiviral action and offer a rationale for further development of fucoidan-based therapeutics targeting norovirus infection, considering ~4 g of fucoidan per day was found to be safe for humans (21, 22).

ETHICS APPROVAL

Saliva samples were collected under ethics approval 2023/937 and NLRD/012/21.