Development of a C5‐Azidoacetamide‐Modified 4‐Amino‐2,3‐Difluorosialic Acid Activity‐Based Probe for Labeling of Influenza A Neuraminidases

Abstract

Influenza A virus neuraminidase (NA) is central to the viral life cycle and a key target for studying sialoside recognition and hydrolysis and its impact on viral uptake, tropism, and pathogenesis. Here, we report the design, synthesis, and evaluation of 2,3‐difluorosialic acid‐based activity‐based probes for NA profiling. The probes carry a C4‐amino substituent to promote active‐site engagement and stabilize covalent trapping through a tyrosine‐sialosyl intermediate. A C5 azidoacetamide handle enables bioorthogonal tagging by CuAAC for detection, while a preclicked biotin variant supports one‐step labeling. We synthesized both the azide and biotin formats and assessed their inhibitory activity against recombinant influenza A NAs. A reactivation assay showed prolonged, hour‐scale recovery relative to related 2,3‐difluoro analogs, although the C5 modification reduced NA affinity and covalent half‐life compared with 4‐amino‐2,3‐difluoro‐Neu5Ac. In labeling experiments, the probes tagged multiple recombinant viral neuraminidases and NA present in virus samples. In addition, 9‐azido‐2,3‐difluoro‐Neu5Ac and its biotin preclicked counterpart proved potent activity‐based probes for NAs. Together, these four probes provide lead structures for further development of molecular tools for cellular profiling, viral NA activity detection, and diagnostics.

We report DFSA5Az/DFSA5bio, C5‐azidoacetamide 4‐amino‐2,3‐difluorosialic acid activity‐based probes that trap influenza A viral (IAV) neuraminidases (NA) via a stabilized covalent Tyr‐sialosyl intermediate with hour‐scale reactivation. A bioorthogonal azide enables CuAAC (or one‐step biotin) tagging for activity‐dependent Western‐blot detection at nanogram levels, labeling recombinant N1/N2, paramyxovirus HN, and neuraminidase on intact virions.

1. Introduction

Influenza A is a prevalent acute respiratory infection that can lead to severe pneumonia, particularly affecting young children, the elderly, and immunocompromised individuals, causing 300,000 to 500,000 deaths around the world annually [1]. The pathogen, influenza A virus (IAV), can be divided into subtypes based on eighteen hemagglutinin (HA) (H1‐H18) and eleven neuraminidase (NA) (N1‐N11) subtypes [2] of these two surface proteins. The HA protein binds to the terminal sialosides of the host cell receptors. The NA protein enzymatically hydrolyzes sialosides, facilitating traversal of the O‐sialoglycan‐presenting mucin barrier of the lung epithelium to access the cell surface for infection initiation and enabling the subsequent release of progeny virions from the cell surface. The HA proteins of the influenza A virus frequently undergo mutations that allow the virus to evade recognition and clearance by the host immune system [3]. In contrast, the neuraminidase (NA) is relatively conserved, preserving its critical role in facilitating the release of newly formed virions [4]. Consequently, NA has attracted attention in the past decades as a target for anti‐influenza drugs.

IAV NA is classified within the glycoside hydrolase family 34 (GH34) and typically exists as a tetramer, with each monomer harboring an active site [5]. It functions as a retaining glycosidase, catalyzing the hydrolysis of terminal α−2,3, or α−2,6‐linked sialosides from glycoconjugates. The enzyme achieves this via a double‐displacement mechanism [6] and, crucially, forms a covalent intermediate during the catalysis. Like many retaining sialidases [7], it is proposed that the catalytic residue tyrosine (Tyr) first acts as a nucleophile (Tyr highlighted in red in Figure 1), attacking the anomeric carbon (C2) of sialic acid, forming a covalent Tyr‐sialosyl intermediate via an oxocarbenium‐like transition state. Simultaneously, an aspartic acid (Asp) residue protonates the leaving group, facilitating bond cleavage. In the second step, a water molecule, activated by a glutamic acid (Glu), attacks the covalent intermediate, hydrolyzing it and releasing the β‐anomer of free sialic acid, thus completing the reaction with retention of stereochemistry. The nucleophilic tyrosine in the active site has been shown to be Tyr406, which forms the covalent intermediate with Neu5Ac [8].

FIGURE 1.

Neuraminidase double displacement active‐site mechanism via a covalent intermediate. R = glycoconjugate.

The inhibition of NA activity abrogates the release of progeny virions from the surface of infected cells [9], suppressing the viral population and thereby allowing time for the immune system of the host to eliminate the virus and infected cells. Reversible inhibitors of NA that mimic the oxocarbenium‐like transition state, like zanamivir and oseltamivir, have been applied in IAV treatment for decades. However, this has resulted in the increased appearance of antiviral resistant strains, due to high viral mutation rates that abolish or lower the affinity of these molecules to the active site [10, 11]. Covalent inhibitors that exploits the evolutionary conserved active site mechanism of NAs, such as difluoro sialic acids, are expected to be less sensitive to resistance development, and thus a promising lead for a next generation of antivirals (Figure 2A). In this class of molecules, the C2‐fluorine plays the role of a good leaving group to still allow for the formation of the covalent enzyme–substrate complex. The C3‐fluorine stabilizes the complex by drastically slowing down the subsequent hydrolysis step through inductive destabilization of the required oxocarbenium‐like transition state, thus inactivating the enzyme and labeling it (Figure 2B).

FIGURE 2.

Structure of neuraminidase inhibitors. (A) Structure of transition state mimicking NA inhibitors (Zanamivir and Oseltamivir) and covalent NA inhibitor (difluoro sialic acid). (B) Inhibition mechanism of DFSA. (C) Structure of some DFSA derivatives.

The first covalent sialidase inhibitor, the 2,3‐difluoro sialic acid, was reported in 2003 by the Withers group targeting the Trypanosoma cruzi trans‐sialidase (TcTS) [12]. They observed that the covalent probe‐enzyme intermediate hydrolyzed within seconds, similar to that observed for IAV NA discovered in a later study [13]. Therefore, inspired by the design of Zanamivir and Oseltamivir, they introduced an amine or guanidine at the C4 position of 2,3‐difluoro sialic acid (Figure 2Ca). This modification improved not only the reactivation half‐life, by stabilizing the covalent intermediate, but also the selectivity of influenza NA over human sialidase by creating a favorable binding interaction with the anionic pocket unique to influenza NAs [14]. They synthesized the axial 3F (ax) and equatorial 3F (eq) isomers of the probe, and found that the C3‐F_ax isomer generated a more stable probe‐enzyme intermediate, with a reactivation half‐life over 2300 min compared with the C3‐F_eq isomer of 50–250 min.

We reasoned that the selectivity of C4‐amino or guanidine modified DFSAs to influenza virus NA could be utilized to design an activity‐based) probe for profiling neuraminidases, by introducing a reporter tag. These probes have the potential application in the detection and monitoring of influenza viral NA, to better understand how its activity is regulated during the infection, which will guide the development of novel antivirals and diagnostics. So far, very few examples of ABPs for neuraminidases have been reported. The Wong group has reported a difluoro sialic acid‐based probe with a clickable terminal alkyne on the C5 acetamide position (Figure 2Cb), as a wide spectrum probe for a variety of sialidases from different origins [15], which mainly focused on human and bacterial sialidases. We here focus on the neuraminidase from influenza A virus by employing the lead structure, 4‐amine modified DFSA (DFSA4Am, Scheme 1), a known covalent inactivator of this enzyme, to design a novel NA ABP DFSA5Az, and its biotinylated version, DFSA5bio (Figure 2C). Based on the original structure of DFSA4Am, the C5‐acetamide was replaced with an azidoacetamide group to create tag for detection via a Cu(I)‐catalyzed azide–alkyne [3 + 2] cycloaddition (CuC3‐FaAAC) or strain‐promoted azide–alkyne cycloaddition (SPAAC). In this work, we report the synthesis of these two novel probes, DFSA5Az and DFSA5bio. We also determine their apparent IC₅₀ values and reactivation half‐lives for a set of recombinant viral neuraminidases, and we compare these results with two C9‐position functionalized 2,3‐difluoro‐Neu5Ac probes. We also show that both preclicked biotinylated probes can label NA in a virus sample and detect it on‐blot.

SCHEME 1.

(a) i) MsOH, MeOH, 70°C; ii) chloroacetic anhydride, pyridine; iii) NaOMe, MeOH; iv) Ac₂O, pyridine; 60% over 4 steps. (b) Selectfluor, water, MeNO₂, 50°C, 3‐Fax 33%, 3‐Feq 11%. (c) DAST, DCM, −40°C, 77%. (d) Triphenylphosphine, trifluoroacetic anhydride, TEA, DCM, 69%. (e) Sodium azide, DMF, RT to 70°C, 61%. (f) 1 M NaOH (aq), dioxane, 62%. (g) Biotin‐PEG4‐alkyne, CuSO₄•5H₂O, sodium ascorbate, DMF/water, 65%.

2. Results and Discussion

2.1. Synthesis of DFSA4Am‐Based Probes

A convenient preparation method for Neu5Ac analogs is chemoenzymatic synthesis, using a neuraminic acid aldolase (NANA), with ManNAc and pyruvate as substrates. However, due to the required amine substitution on C4 in DFSA4Am, this path would require an anomerically N‐substituted ManNAc, which is not compatible with the aldolase enzyme. Therefore, the probes were instead synthesized chemically. The synthesis route started from compound 1, the key known intermediate in the synthesis of both Relenza and DFSA4Am (Scheme 1). Removal of all acetyl esters and the acetamide in 1 was conducted with methanesulfonic acid in refluxing dry methanol. The resultant crude product with exposed hydroxyls and a free C5‐amine was treated with an excess of chloroacetic anhydride. Chloroacetyl esters were then removed with sodium methoxide, and the obtained crude C5‐chloroacetamide intermediate was re‐acetylated with acetic anhydride in pyridine to afford 2 in 60% yield over 4 steps. Installing the chloroacetamide at C5 allowed the late‐stage S _N 2 substitution to install the azide group at this position, after the C2,3‐difluorination sequence, and more crucially after reduction of the C4‐azide. Next, the C3‐fluorine was installed by an electrophilic addition with Selectfluor, followed by in situ hydrolysis of the anomeric intermediate to produce 3 as a ~3:1 mixture of axial: equatorial C3‐fluorines. It was previously shown that the C3‐F_ax isomer of DFSA derivatives result in a more stable covalent intermediate with the active site nucleophile (Tyr 406). Therefore, only the isolated C3‐F_ax isomer of 3 was anomerically fluorinated with diethylaminosulfur trifluoride (DAST) to afford 4. The C4‐azide was next reduced via a Staudinger reduction and protected as a trifluoroacetyl (TFA) amide in one pot to provide 5. The benefit of this TFA‐protection was that the trifluoroacetamide and all acetyl esters could be deprotected in the last step to yield the target probe, DFSA5Az. The chloride on the C5‐acetamide was now substituted for an azide group by treating 5 with aqueous sodium azide solution in DMF to yield 6. The target probe DFSA5Az was obtained by a treatment of 6 with aqueous sodium hydroxide solution in dioxane. DFSA5bio was obtained via a Cu(I)‐catalyzed azide–alkyne cycloaddition (CuAAC) click reaction between DFSA5Az and alkyne‐PEG₄‐biotin. The latter probe allowed for the comparison of the inactivation activity on influenza virus neuraminidases with a bulkier C5 modification and to compare 1‐step versus 2‐step ABP labeling. The known DFSA4Am was also prepared, using the published synthesis route [13], as a control for the two novel probe derivatives.

2.2. Inactivation Studies on Recombinant Influenza A Viral NAs

With the two ABP derivatives, DFSA5Az and DFSA5bio, in hand, we assessed their apparent inhibition of recombinant influenza A recombinant viral neuraminidase from two human strains: N1 (H1N1) and N2 (H3N2). Initially, a maximum inhibitor concentration of 100 µM was used. The control compound, DFSA4Am, exhibited low‐micromolar IC₅₀ values for both enzymes in our assay, consistent with previously published values. In contrast, both probe derivatives showed decreased potency, with bulkier substituents at the C5 amide correlating with a greater reduction in inactivation activity (Figure 3A). Because neither probe achieved half‐maximal inhibition at 100 µM, we increased the highest concentration to 375 µM (Figure 3B), and the resulting IC₅₀ values are presented in Table 1. The diminished inactivation activity likely stems from impaired hydrophobic interactions between the substrate's acetamide methyl group and nearby active‐site residues when sterically larger groups are introduced at the C5 position (PDB: 4WEG) [16].

FIGURE 3.

IC₅₀ curve for DFSA4Am‐based inactivators. (A) IC₅₀ curve for three inactivators with a highest concentration up to 100 µM. (B) IC₅₀ curve for probes DFSA5Az and DFSA5bio with the highest concentration of 375 μM (n = 3, mean ± SD).

TABLE 1.

IC₅₀ values of three DFSA based inactivators (95% CI: 95% confidential interval; data were determined by GraphPad Prism using nonlinear regression analysis).

	DFSA4Am		DFSA5Az		DFSA5bio
	IC50, µM	95%CI, µM	IC50, µM	95%CI, µM	IC50, µM	95%CI, µM
N1	2, 03	1, 83–2, 25	66, 8	56, 1–79, 7	382	259–618
N2	0, 918	0, 790–1, 07	52, 5	42, 4–65, 12	104	77, 9–142

2.3. A Novel Reactivation Assay for NA ABPs on Recombinant NA

The C3‐F_ax isomer of DFSA4Am is known to have a prolonged reactivation half‐life exceeding 100 h [13]. To determine whether this property is retained after C5 modification, we preincubated DFSA5Az and DFSA5bio with recombinant N1 protein (from H1N1) and after removal of the excess probe by centrifugal dialysis we monitored NA activity using the fluorescent substrate 4‐methylumbelliferyl‐N‐acetyl‐α‐d‐neuraminic acid (MUNANA). In our untreated control we, however, noticed that the enzymatic activity of NA severely decreased after centrifugation, probably due to partial denaturation of the recombinant enzyme. Despite attempts to prevent this by troubleshooting the workflow, we were not able to reliably retain NA activity in the control. Therefore, we developed a mild method to remove the excess inactivator that did not expose the NA to severe conditions, such as ultracentrifugation, via immobilization of the recombinant NA via its His‐tag on Nickel‐NTA‐coated magnetic beads (Figure 4A). This allowed facile removal of excess probe and on‐bead NA activity measurement afterwards.

FIGURE 4.

Overview and results for novel workflow for reactivation experiment of NA ABPs. (A) Workflow of the novel method to remove the excess of probes for the reactivation assay: His‐tagged N1 (avian H5N1) protein was incubated with probe DFSA5Az (400 µM) or DFSA5bio (400 µM) or DFSA9Az (100 µM) at 37°C for 30 min and immobilized onto Ni‐NTA beads; the excess probe was removed by mild centrifugation. The enzymatic activity at each time point was measured by a 10‐min MUNANA assay. (B) The reactivation curves of each probe on N1 H5N1 over an 8‐h period (n = 2, mean ± SD). (C) Structure of DFSA9Az.

We assessed enzyme reactivation by measuring the initial hydrolysis rates of probe‐treated and untreated NA at various time points (Figure 4B). Besides the two probes developed for this study, we included a known probe DFSA9Az (Figure 4C) [17] to compare their reactivation half‐life. After treatment with the probes and removal of the excess, both 4‐NH₂ probes showed maximum NA inhibition at around 70% to 80%, while DFSA9Az achieved approximately 60% maximum inhibition. However, as DFSA9Az shows higher affinity for NA and a lower IC₅₀ value (see supporting information, Table S1), we assume that rapid hydrolysis of the covalent intermediate of this probe already occurred during the sample preparation.

In the subsequent 8 h, DFSA5Az‐treated NA showed the slowest reactivation rate, while DFSA5bio‐treated NA gave a slightly faster recovery of NA activity over the same period. In contrast, DFSA9Az showed the fastest reactivation. These observations suggest that once a covalent substrate–enzyme complex forms, the C3‐fluorine impedes formation of the oxocarbenium‐like transition state, leading to a stabilized covalent intermediate. As previously reported, the C4 amine substitution seems to contribute to a prolonged reactivation half‐life. However, compared to the previously reported reactivation half‐life for DFSA4Am, >24 h, the azide and biotin modification negatively impacts this property. A ∼2–3 h half‐life of the covalent intermediate for the here reported probes, nevertheless, still allows for their use as useful tools for investigating IAV NAs.

To explore the covalent labeling of viral neuraminidases, we first conducted labeling experiments on recombinant N1 (human H1N1) used above and visualized the result on a western blot. The enzyme was incubated with DFSA5Az for 30 min at 37°C followed by mixing with a nonreducing loading buffer and denaturing by heating, as we noticed that thiols in normal loading buffer could reduce the azide tag, causing the failure of the on‐blot click reaction. A Bis‐Tris gel (pH 7) was used to electrophorese the samples and to avoid possible protein modification (alkylation of cysteine thiol groups) at higher pH [18]. To guarantee a clear band of the loading control, 2 μg of the enzyme were added to each sample. The lowest tested incubation concentration of DFSA5Az that still showed clear labeling of the NA was 50 μM (Figure S1A).

We included two controls to prove that labeling was selective for the neuraminidase. Coincubation with the known NA transition state inhibitor N‐acetyl‐2,3‐dehydro‐2‐deoxyneuraminic acid (DANA) at three different concentrations showed reduction of NA labeling in a concentration‐dependent fashion (Figure 5A), indicating that the binding of DFSA5Az to N1 was neuraminidase dependent. Coincubation of DFSA5Az with the viral neuraminidase (N1) and the N‐acetylneuraminate aldolase (NANA aldolase) resulted in selective labeling of the neuraminidase only (Figure 5B). No labeling was observed for the aldolase, a sialic acid‐active protein that does not have hydrolytic activity, even when present at a 10‐fold molar excess relative to N1. This demonstrates that DFSA5Az selectively targets enzymatically active neuraminidases.

FIGURE 5.

Visualizing DFSA5Az labeled recombinant N1 (2 µg). After incubation with the probe (50 µM), the samples were run on Bis‐Tris gels and blotted onto nitrocellulose membranes. The NA–probe complex coupled to biotin‐PEG₄‐alkyne via CuAAC and analyzed by anti‐biotin‐HRP. (A) Labeling with and without active site competitor DANA. (B) Coincubation of DFSA5Az with an increasing amount NANA aldolase as control for specificity.

Next, the ability of the DFSA5bio probe to label neuraminidase (NA) in a one‐step fashion was assessed. Consistent with previous inhibition results, which indicated reduced affinity of the preclicked DFSA5bio compared to DFSA5Az, on‐blot labeling required high probe concentrations under nonreducing conditions (lowest effective concentration ∼150 µM; Figure S1B). A key advantage of DFSA5bio is its lack of a reduction‐sensitive azide group, allowing denaturation under regular reducing conditions. This treatment fully dissociated the NA tetramer into monomers, improving transfer efficiency and detection on blot and thus facilitating the analysis of small NA quantities. Using silver staining as a loading control, DFSA5bio successfully labeled nanogram‐level amounts of NA (Figure S1C), highlighting its utility for labeling viral neuraminidases.

Because labeling efficiency is closely tied to NA activity, we employed DFSA5bio to label several viral proteins possessing neuraminidase activity, including NA from H1N1 and hemagglutinin‐neuraminidase (HN) proteins from human and avian paramyxoviruses. One microgram of each protein was treated with 200 µM of DFSA5bio and then analyzed by Western blot and PageBlue staining (Figure S2). The labeling intensity of each NA is shown in Figure 6A. As expected, the differences in enzymatic activity on MUNANA of these three enzymes (Figure 6B) were reflected in the labeling intensities by DFSA5bio, as observed on the blot.

FIGURE 6.

Visualizing DFSA5bio labeled influenza A NA and paramyxoviridae HN proteins. (A) On‐blot labeling intensity (as quantified by ImageJ from blot in Figure S2). (B) Neuraminidase activity determined by MUNANA assay.

After determining that both DFSA5Az and DFSA5bio could label several recombinant viral neuraminidases, we evaluated whether neuraminidase labeling still worked on whole virion particles. Approximately 1.2 × 10⁹ pdmH1N109 IAV particles were incubated with 200 µM of either biotinylated probe DFSA9bio (prepared by clicking DFSA9Az with alkynyl PEG₄ biotin, see Figure 7A for structure) or DFSA5bio. After virus inactivation with UV light and concentration, the probe‐treated virus samples were denatured and loaded onto a Bis‐Tris gel, followed by electrophoresis. The gel was blotted onto a nitrocellulose membrane and labeling was detected with HRP‐conjugated streptavidin. In the probe treated virus samples, two bands around 70 kDa were observed. Stripping the streptavidin and reblocking the membrane, followed by incubation with an anti‐NA monoclonal antibody [19], confirmed that the labeled band at exactly 70 kDa was the NA protein (Figure 7B). Samples treated with the DFSA9bio probe show more intense labeling than DFSA5bio. Both probes also labeled a slightly higher unannotated band. Additionally, we took one aliquot of each virus sample and conducted a MUNANA assay. Compared with nontreated virus, both probes showed significantly decreased NA activity, with DFSA9bio again proving to be the most effective inhibitor (Figure 7C).

FIGURE 7.

Visualizing DFSA5bio and DFSA9bio labeled IAV NA from pdm H1N1 virus. (A) Structure DFSA9bio. (B) Probe (200 µM) treated virus particles were denatured, loaded on Bis‐Tris gels, and blotted onto nitrocellulose membranes. The probe‐bound NA protein was detected by streptavidin‐HRP or anti‐NA. Loading control was stained by PageBlue. (C) NA activity in virus sample and probe treated virus samples determined by a final point measurement with MUNANA.

3. Conclusion

We synthesized four influenza neuraminidase (NA)‐targeting activity‐based probes (ABPs): the C5‐modified DFSA5Az, featuring an azide tag, its preclicked counterpart DFSA5bio, as well as the C9‐modified DFSA9Az and DFSA9bio. By using chlorine as the azide precursor, we successfully introduced a second azide group after reduction of the first at C4. These modified probes exhibited IC₅₀ values ranging from 50 to 400 µM—higher than those of the previously reported unmodified DFSA4Am—with decreased inhibition correlating to larger substituents at the C5 or C9 positions. NAs inactivated with both C5‐ and C9‐modified probes also showed prolonged reactivation half‐lives, as measured with a novel assay, supporting their potential utility in assays requiring extended NA inhibition. On‐blot labeling confirmed the ability of DFSA‐based probes to label NA on whole virions, with labeling intensity reflecting enzymatic activity. Future efforts will focus on structural refinements in this class of neuraminidase probes to improve potency and selectivity, with the goal of advancing their applications in influenza research, cellular profiling, imaging, and diagnostics.

4. Experimental

4.1. General Methods and Materials

Reagents and solvents were purchased from Merck, Acros, Alfa Aesar, Fisher Scientific, TCI, or Biosolve. For anhydrous reactions, solvents were dried over activated 3 Å or 4 Å molecular sieves for 24 h prior to use. Reactions were monitored using thin layer chromatography (TLC) on aluminum or glass plates precoated with Silica Gel 60 F₂₅₄ (Merck). Visualization of the spots was done by UV light at 254 nm or using H₂SO₄ in EtOH (5% v/v) with heating. Flash chromatography was performed on Silica Gel (Avantor delivered by VWR, 40–63 µm, or Silicycle, 60–200 µm, 60 Å). ¹H‐NMR and ¹³C‐NMR spectra were recorded on an Agilent 400‐MR (400 MHz) or a Bruker 600 Ultrashield (600 MHz) spectrometer. Chemical shifts (δ) are reported in parts per million (ppm), relative to TMS signal. All measurements were performed at room temperature. High resolution mass spectrometry (HRMS) was recorded on an Agilent 6560B DTIM‐QTOF ion mobility spectrometer used in QTOF mode only.

4.2. Methyl 5‐(Chloroacetyl)amino‐7,8,9‐Tri‐Acetyl‐2,6‐Anhydro‐4‐Azido‐3,4,5‐Trideoxy‐d‐Glycero‐D‐Galacto‐Non‐2‐Enonate (2)

Methanesulfonic acid (1.03 mL, 15.5 mmol) was added to a solution of 1 (1.16 g, 2.58 mmol) in methanol (50 mL) under a nitrogen atmosphere and the reaction mixture was refluxed for 24 h. After neutralizing with triethylamine, the mixture was concentrated in vacuo. The resulting syrup was dissolved in pyridine (70 mL) and cooled to 0°C. Chloroacetic anhydride (3.09 g, 18.06 mmol) was slowly added, and the solution was stirred for 30 min at 0°C. The reaction mixture was poured into ethyl acetate (300 mL) and washed successively with 1 M HCl (3 × 100 mL), saturated sodium bicarbonate (3 x 100 mL), and brine (150 mL). The organic layer was dried over anhydrous sodium sulfate, concentrated under reduced pressure, and purified with silica gel column chromatography (elution: petroleum ether/ethyl acetate; 7/3 → 6/4 → 5/5, v/v) to yield the per‐chloroacetylated product. To a solution of the resulting compound in dry methanol (20 mL), several drops of 30% sodium methoxide solution in methanol were added at room temperature. The mixture was stirred at room temperature for 30 min and neutralized with Amberlite IRC‐120 H⁺ form. The mixture was filtered and rinsed with methanol and the filtrate was concentrated in vacuo. The resulting crude product was dissolved in pyridine (10 mL) and treated with acetic anhydride (10 mL) at room temperature overnight. Afterwards, the solvent was removed under reduced pressure, and the crude product was purified by silica gel column chromatography (elution: petroleum ether/ethyl acetate; 7/3 → 6/4 → 5/5, v/v) to afford 2 (758 mg, 60% yield over 4 steps) as a yellowish white foam. R _f  = 0.63 (petroleum ether/ethyl acetate; 1/4) ¹H NMR (600 MHz, CDCl₃) δ 6.74 (d, J = 9.2 Hz, 1H), 6.02 (d, J = 2.5 Hz, 1H), 5.44 (dd, J = 5.7, 2.3 Hz, 1H), 5.34 (td, J = 6.1, 2.5 Hz, 1H), 4.60 (dd, J = 12.5, 2.7 Hz, 1H), 4.48 (dd, J = 10.1, 2.3 Hz, 1H), 4.43 (dd, J = 9.0, 2.5 Hz, 1H), 4.19 (dd, J = 12.5, 6.3 Hz, 1H), 4.06 (s, 2H), 4.02 (q, J = 9.4 Hz, 1H), 3.82 (d, J = 1.4 Hz, 3H), 2.13 (d, J = 1.5 Hz, 3H), 2.08 (d, J = 1.4 Hz, 3H), 2.06 (d, J = 1.4 Hz, 3H). ¹³C NMR (151 MHz, CDCl₃) δ 170.6, 170.4, 170.0, 166.6, 161.4, 145.4, 107.2, 75.6, 70.5, 67.4, 61.8, 57.7, 52.7, 48.5, 42.5, 20.9, 20.8, 20.8.HRMS (ESI): calcd. for C₁₈H₂₃ClN₄O₁₀ [M+Na]⁺ 513.0995; found 513.0995.

4.3. Methyl(5‐Chloroacetamido‐7,8,9‐Tri‐O‐Acetyl‐4‐Azido‐3‐Fluor‐2‐Hydroxy‐D‐Erythro‐β‐l‐Manno‐2‐Nonulopyranosid)onate (3)

To a solution of 2 (758 mg, 1.55 mmol) in a mixed solvent of nitromethane (15 mL) and water (1 mL) was added Selectfluor (2.14 g, 6.19 mmol), and the mixture was stirred at 50°C overnight. The reaction mixture was quenched by adding saturated bicarbonate (50 mL) and extracted by ethyl acetate. The water layer was washed with ethyl acetate (3 × 10 mL), and the combined organic layer was washed with brine and dried over anhydrous sodium sulfate. The resultant mixture was filtered, concentrated, and purified with silica column chromatography eluting with dichloromethane/ethyl acetate/acetone (8/1/1 → 7/1/1 → 6/1/1, v/v/v) to yield the desired products 3 (3F_ax: 270 mg, 33%; 3F_eq: 90 mg, 11%) as a white solid. Characterization was done on only 3Fax isomer. R _f  = 0.52 (petroleum ether/ethyl acetate 1/4) ¹H NMR (600 MHz, (CD₃)₂CO) δ 7.85 (d, J = 9.9 Hz, 1H), 5.47 (dd, J = 4.8, 2.4 Hz, 1H), 5.20 (ddd, J = 7.5, 4.8, 2.7 Hz, 1H), 5.09 (dd, J = 49.0, 2.2 Hz, 1H), 4.60 (dt, J = 10.8, 2.8 Hz, 2H), 4.50 (q, J = 10.4 Hz, 1H), 4.16 (dd, J = 12.3, 7.5 Hz, 1H), 4.11 – 4.00 (m, 3H), 3.77 (s, 3H), 2.07 (s, 3H), 1.97 (s, 6H). ¹³C NMR (151 MHz, (CD₃)₂CO) δ 205.4, 190.8, 169.8, 169.7, 169.6, 167.0, 166.5, 166.4, 94.2, 94.1, 90.0, 88.8, 71.1, 70.4, 68.2, 62.1, 59.6, 59.5, 52.2, 45.2, 42.3, 42.3, 20.0, 20.0, 19.8. ¹⁹F NMR (376 MHz, (CD₃)₂CO) δ −204.2. HRMS (ESI): calcd for C₁₈H₂₄ClFN₄O₁₁[M+H]⁺ 527.1187; found 527.1193.

4.4. Methyl (5‐Chloroacetamido‐7,8,9‐Tri‐O‐Acetyl‐4‐Azido‐2,3‐Difluor‐D‐Erythro‐β‐l‐Manno‐2‐Nonulopyranosid)onate (4)

Compound 3 (270 mg, 0.51 mmol) was dissolved in anhydrous dichloromethane (20 mL) and cooled to −40°C. DAST (0.10 mL, 0.765 mmol) was added dropwise under argon atmosphere, and the reaction was stirred for 30 min at this temperature. The reaction mixture was quenched by adding of saturated sodium bicarbonate (20 mL) and then warmed to room temperature and diluted with dichloromethane (20 mL). The organic layer was separated, washed with brine, and dried over anhydrous sodium sulfate. The crude product was purified by silica column chromatography eluting with petroleum ether/ethyl acetate (3/1 → 2/1 → 1/1 v/v) to afford 4 (207 mg, 77%) as a white foam. R _f  = 0.62 (petroleum ether/ ethyl acetate 1/4) ¹H NMR (600 MHz, CDCl₃) δ 7.00 (d, J = 7.7 Hz, 1H), 5.39 (ddd, J = 8.5, 4.4, 2.3 Hz, 1H), 5.25 (dt, J = 8.6, 1.7 Hz, 1H), 5.18 (dt, J = 49.9, 2.4 Hz, 1H), 4.56 (dd, J = 28.4, 11.0 Hz, 1H), 4.44 (d, J = 10.6 Hz, 1H), 4.37 – 4.25 (m, 2H), 4.06 (d, J = 5.5 Hz, 2H), 3.92 (s, 3H), 3.71 – 3.63 (m, 1H), 2.19 (s, 3H), 2.12 (s, 3H), 2.05 (s, 3H). ¹³C NMR (151 MHz, CDCl₃) δ 171.3, 170.6, 170.5, 169.7, 167.4, 86.64 (dd, J = 194.1, 19.6 Hz), 70.9, 70.9, 68.6, 67.3, 67.2, 61.6, 53.9, 47.7, 47.7, 42.4, 20.9, 20.8, 20.7. ¹⁹F NMR (376 MHz, CDCl₃) δ −123.13, −216.57. HRMS (ESI): calcd for C₁₈H₂₃ClF₂N₄O₁₀ [M+Na]⁺ 551.0963; found 551.0968.

4.5. Methyl (5‐Chloroacetamido‐7,8,9‐Tri‐O‐Acetyl‐4‐Amide‐Trifluoroacetic Acid‐2,3‐Difluor‐D‐Erythro‐β‐l‐Manno‐2‐Nonulopyranosid)onate (5)

Triphenylphosphine (95 mg, 0.362 mmol) was added to a solution of 4 (95.5 mg, 0.181 mmol) in anhydrous dichloromethane (30 mL), after which a mixture of triethylamine (0.063 mL, 0.453 mmol) and trifluoroacetic anhydride (0.051 mL, 0.362 mmol) in 0.5 mL dry dichloromethane was added. The reaction was stirred for 1 h at room temperature and quenched with an equal amount of saturated sodium bicarbonate solution. After extracting with dichloromethane (2 × 10 mL), the combined organic layer was dried over anhydrous sodium sulfate, concentrated, and purified by silica column chromatography eluting with petroleum ether/ethyl acetate (3/1 → 2/1 → 1/1 v/v) to afford 5 (75 mg, 69%) as a white foam. R _f  = 0.5 (ethyl acetate) ¹H NMR (600 MHz, CDCl₃) δ 7.16 – 7.05 (m, 1H), 6.71 (d, J = 8.9 Hz, 1H), 5.42 – 5.33 (m, 2H), 5.09 (dt, J = 49.8, 2.0 Hz, 1H), 4.68 (dt, J = 28.7, 9.5 Hz, 1H), 4.38 – 4.30 (m, 3H), 4.17 (dd, J = 12.6, 4.7 Hz, 1H), 4.07 – 3.93 (m, 2H), 3.91 (s, 3H), 2.12 (d, J = 3.4 Hz, 6H), 2.05 (s, 3H). ¹³C NMR (151 MHz, CDCl₃) δ 170.7, 170.7, 170.0, 170.0, 169.9, 168.1, 115.5 (d, J = 287.2 Hz), 104.3 (dd, J = 226.7, 16.9 Hz), 86.02 (dd, J = 190.8, 19.3 Hz), 73.0, 73.0, 68.9, 66.7, 61.7, 54.1, 51.5, 51.5, 51.4, 44.8, 44.8, 42.4, 21.0, 20.8, 20.7. ¹⁹F NMR (376 MHz, CDCl₃) δ −75.83, −123.83, −214.61. HRMS (ESI): calcd for C₂₀H₂₄ClF₅N₂O₁₁ [M+H]⁺ 559.1062; found 559.1068.

4.6. Methyl (5‐Azidoacetamido‐7,8,9‐Tri‐O‐Acetyl‐4‐Amide‐Trifluoroacetic Acid‐2,3‐Difluor‐D‐Erythro‐β‐l‐Manno‐2‐Nonulopyranosid)onate (6)

To a solution of 5 (39 mg, 0.065 mmol) in DMF (10 mL) was added aqueous sodium azide solution (423 mM, 235 μL). The reaction mixture was stirred at room temperature overnight and heated to 70°C for 2 h. Once the mass spectrometry showed the disappearance of 5, the reaction mixture was diluted with diethyl ether (50 mL) and washed with water (3 × 20 mL), brine (25 mL), and dried over anhydrous sodium sulfate. The crude product was concentrated and purified by silica column chromatography eluting with petroleum ether/ethyl acetate (3/1 → 2/1 → 1/1 v/v) to yield 6 (24 mg, 61%) as a white foam. R _f  = 0.5 (ethyl acetate) ¹H NMR (600 MHz, CDCl₃) δ 7.12 (d, J = 8.7 Hz, 1H), 6.47 (d, J = 9.6 Hz, 1H), 5.38 (dd, J = 3.3, 1.7 Hz, 2H), 5.09 (dt, J = 49.8, 1.9 Hz, 1H), 4.55 (dt, J = 28.2, 9.1 Hz, 1H), 4.39 – 4.29 (m, 3H), 4.17 (ddd, J = 12.7, 3.3, 1.2 Hz, 1H), 4.01 – 3.84 (m, 5H), 2.13 (d, J = 5.0 Hz, 6H), 2.04 (s, 3H). ¹³C NMR (151 MHz, CDCl₃) δ 170.6, 169.9, 169.7, 164.4, 164.2, 158.0, 157.8, 115.38 (d, J = 287.6 Hz), 85.77 (d, J = 210.7 Hz), 73.0, 68.7, 66.4, 61.5, 54.0, 52.4, 44.1, 20.9, 20.7, 20.6. ¹⁹F NMR (376 MHz, cdcl₃) δ −75.95, −123.97, −214.56. HRMS (ESI): calcd for C₂₀H₂₄F₅N₅O₁₁ [M+H]⁺ 606.1465; found 606.1468.

4.7. Methyl (5‐Azidoacetamido‐7,8,9‐Trihydroxy‐4‐Amide‐2,3‐Difluor‐D‐Erythro‐β‐l‐Manno‐2‐Nonulopyranosid)onate (DFSA5Az)

Aqueous solution of sodium hydroxide (1M) was added to the solution of 6 (44 mg, 0.073 mmol) in 1,4‐dioxane and distilled water (4 mL, 3:1 v/v) to adjust the pH to around 12 at 0°C. The mixture was gradually warmed to room temperature and stirred for 4 h. The reaction was neutralized, and concentrated, and purified with silica column chromatography (silica gel neutralized with triethylamine) eluting with ethyl acetate/methanol/distilled water (7/2/1 v/v). The product was concentrated, redissolved in distilled water, and lyophilized to afford DFSA5Az (13.2 mg, 40%) as a white cotton‐like solid. R _f  = 0.27 (ethyl acetate/methanol/water 7/2/1) ¹H NMR (600 MHz, D₂O) δ 5.13 (dt, J = 50.5, 2.7 Hz, 1H), 4.23 (t, J = 10.8 Hz, 1H), 4.02 (s, 2H), 3.88 (dd, J = 10.2, 1.4 Hz, 1H), 3.82 – 3.74 (m, 2H), 3.62 – 3.39 (m, 3H). ¹³C NMR (151 MHz, D₂O) δ 171.4, 169.3, 106.49 (dd, J = 219.7, 14.7 Hz), 88.41 (dd, J = 183.6, 19.9 Hz), 73.2, 70.4, 67.8, 62.9, 51.9, 46.3. ¹⁹F NMR (376 MHz, D₂O) δ −121.50, −217.50. HRMS (ESI): calcd for C₁₁H₁₇F₂N₅O₇ [M+H]⁺ 370.1169; found 370.1179.

4.8. Methyl (5‐(4‐(15‐Oxo‐19‐((3aS,6aR)‐2‐Oxohexahydro‐1H‐Thieno[3,4‐d]Imidazol‐4‐yl)‐2,5,8,11‐Tetraoxa‐14‐Azanonadecyl)‐1H‐1,2,3‐Triazol‐1‐yl))Acetamido‐7,8,9‐Tri‐Hydroxyl‐4‐Amide‐2,3‐Difluor‐D‐Erythro‐β‐l‐Manno‐2‐Nonulopyranosid)onate (DFSA5bio)

To a solution of DFSA5Az (1.4 mg, 0.00379 mmol) in water was added bition‐PEG₄‐alkyne (1.7 mg, 0.00376 mmol), 0.1 M CuSO₄ aqueous solution (7.58 µL) and 0.1 M sodium ascorbate aqueous solution (8.34 µL). The reaction was stirred under nitrogen for 24 h in the dark. After the removal of solvent, the crude product was purified with a P2 size exclusion column eluting with distilled water. The pure DFSA5bio was obtained after lyophilization as a white foam (2.13 mg, 65%). ¹H NMR (600 MHz, D₂O) δ 8.03 (s, 1H), 5.29 (s, 2H), 5.21 (dt, J = 50.4, 2.9 Hz, 1H), 4.65 (s, 3H), 4.52 (dd, J = 7.9, 4.9 Hz, 1H), 4.39 – 4.30 (m, 2H), 3.95 (d, J = 10.5 Hz, 1H), 3.86 – 3.74 (m, 3H), 3.71 – 3.48 (m, 17H), 3.28 (dt, J = 34.6, 4.9 Hz, 3H), 2.91 (ddd, J = 13.0, 5.0, 0.8 Hz, 1H), 2.70 (d, J = 13.0 Hz, 1H), 2.18 (t, J = 7.3 Hz, 2H), 1.71 – 1.41 (m, 4H), 1.33 (dtt, J = 12.8, 4.3, 2.5 Hz, 2H).HRMS calcd for C₃₂H₅₂F₂N₈O₁₃S [M+H]⁺ 827.3415; found 827.3420.

4.9. NA and Virus

Human codon‐optimized cDNA (Genscript, USA) encoding the NA ectodomains of N1 A/Wisconsin/09/2013 (H1N1) [20] (GenBank accession no. AGV29183.1), N2 A/Bilthoven/1761/1976 (H3N2) (GenBank: AFG99020.1; N2 H3N2) [19], N1 A/Common Tern/NL/26/2022 (HPAI H5N1, GISAID EPI_ISL_15069401) [21] HN of hPIV1 (GenBank: AAC23946.1), HN of NDV (GenBank: CAB51326.1) [22] were synthesized and used to recombinantly express these Nas, as previously published; the virus used for whole‐virus labeling was A/Netherlands/602/2009 (pdmH1N109).

4.10. IC₅₀ Enzyme Inhibition Assay

The IC₅₀ values of each ABP was determined by MUNANA assay. DFSA probes were dissolved in a NA reaction buffer (50 mM Tris/HCl, 4 mM CaCl₂, pH 6.0) and diluted into 11 different concentrations with the final volume of 50 µL and added into a row of a 96‐well‐plate, while keeping the last well containing only buffer. Subsequently, 25 µL of recombinant NA (0.001 µg/µL) in reaction buffer was added. The mixture was incubated at 37°C for 30 min. Once the incubation was finished, an aqueous solution of MUNANA (400 µM, 25 µL) was added, and the mixture was incubated at 37°C for 60 min. A stop solution (0.1 M glycine, 25% ethanol, pH = 10.7, 180 µL) was added, and the plate was placed in a fluorescence plate reader to measure the released MU at excitation 365 nm and emission 450 nm. An IC₅₀ curve was made by plotting percentage of hydrolyzed MUNANA to the log[inhibitor] using GraphPad.

4.11. Reactivation Assay

An enzyme solution of recombinant H1 from avian H5N1 (final concentration 10 nM) was incubated with DFSA9Az probe (final concentration 100 µM) or DFSA5Az (final concentration 400 µM) DFSA5bio probe (final concentration 400 µM) in NA reaction buffer (50 mM Tris/HCl, 4 mM CaCl₂, pH 6.0) for 30 min at room temperature. The solution was resuspended in 200 µL of Ni‐NTA bead suspension and incubated at room temperature for 30 min. The beads were washed three times with 400 µL of cold Bis‐Tris buffer. The Ni‐NTA beads were then resuspended in 800 µL of the NA reaction buffer. At specific time points, 50 μL of the Ni‐NTA suspension was mixed with 50 μL of MUNANA (400 μM final concentration) and incubated for 10 min at RT. The resulting supernatant was added to a well containing 100 μL of stop solution. Fluorescence was measured with a Fluostar reader (340 ex., 490 ems.), and the NA activity was calculated as percentage compared to nontreated NA.

4.12. General Procedure of NA Labeling with DFSA5Az Probe on Blot

For each sample, DFSA5Az and any mentioned compounds were incubated with 4 μg of recombinant N1 protein in a final volume of 20 μL with NA reaction buffer (50 mM Tris‐HCl, 4 mM CaCl₂, pH 6.0) and incubated at 37°C for 30 min. Each sample was mixed with 10 μL of 3x Laemmli buffer (without 2‐mercaptoethanol to avoid the reduction of the azide group) and boiled at 95°C for 10 min; then, 15 µL of each sample was loaded onto two 10% Bis‐Tris gels, respectively. The electrophoresis was performed at 90V for 10 min then 130V for 60 min. One of the gels was washed with distilled water (10 min, 3 times) and stained with 20 mL of Page blue for 1 h as loading control. The other one was blotted onto a nitrocellulose membrane using a TransBlot Turbo Transfer System (Biorad) for 7 min at 2.5 V. The membrane was washed with distilled water and treated with 2 mL of click mixture containing 0.5 μM of CuSO₄, 2.5 μM of ascorbic acid, and 0.1 μM of alkyne‐PEG4‐biotin for 2 h in the dark. After the click reaction, the membrane was washed with PBS for 5 min and blocked with a solution 5% skim milk in PBS contained 0.5% tween20 (PBS‐T) for 1 h. The blocking solution was removed by washing with 1% skim milk/PBS‐T for 5 min, and the membrane was incubated with α‐biotin‐HRP (Jackson ImmunoResearch) (1: 10 000 in 1% skim milk, 20 mL) at room temperature for 1 h. The antibody was removed by washing thoroughly with PBS‐T, and the membrane was treated with SuperSignal West Femto Maximum Sensitivity Substrate (Thermo Scientific) and imaged in Gel‐Doc system (Biorad).

4.13. General Procedure of NA Labeling with DFSA5bio Probe on Blot

For each sample, DFSA5bio was incubated with recombinant protein in a final volume of 20 μL with NA reaction buffer (50 mM Tris‐HCl, 4 mM CaCl₂, pH 6.0) and incubated at 37°C for 30 min. For virus labeling, 600 µL of virus suspension was incubated with the probe for 30 min before inactivated by UV for 5 min. The mixture was concentrated to 20 µL under air flow. Each sample was mixed with 10 μL of 3x Laemmli buffer and boiled at 95°C for 10 min; then, 15 µl of each sample was loaded onto two 10% MOPS gel, respectively. The electrophoresis was performed at 90V until the samples were seen in the running gel then switched to 130V for 60 min. One of the gels was washed with distilled water (10 min, 3 times) and stained with 20 mL of page blue for 1 h as loading control. For nanogram‐scale samples, the loading control was stained with a Pierce Silver Stain Kit (ThermoFisher). The other one was blotted onto a nitrocellulose membrane using a TransBlot Turbo Transfer System (Biorad) for 7 min at 2.5 V. The membrane was washed with PBS for 5 min and blocked with 5% BSA contained 0.5% Tween 20 for 1 h. The blocking solution was removed by washing with 1% BSA for 5 min, twice, and the membrane was incubated with α‐biotin‐HRP (Jackson ImmunoResearch) or streptavidin‐HRP (Sigma) (1: 10 000 in 1% BSA) at room temperature for 1h. The antibody/streptavidin was removed by washing thoroughly with PBS, and the membrane was treated with SuperSignal West Femto Maximum Sensitivity Substrate (ThermoFisher) and imaged in Gel‐Doc system (Biorad).

Supporting Information

Additional supporting information can be found online in the Supporting Information Section. Supporting Fig. A. Detection limits of DFSA5Az labelled N1 (H1N1) with various probe concentration on 2 µg of NA protein under non‐reducing condition. B. Detection limits of DFSA5bio labeled N1 (H1N1) with various probe concentration on 2 µg of NA protein under non‐reducing condition. C. Detection limits of DFSA5bio labeled N1 (H1N1) with various protein amount with 500 µM of probe under reducing condition. Supporting Fig. S2: Western blot and PageBlue‐stained gel supporting the quantification shown in Fig. 2A. Streptavidin blot signal was quantified in ImageJ. Non‐specific streptavidin binding to the Strep‐tag produced background signal, which was subtracted from the intensities reported in Fig. 2A. Whole gel images of Supporting Fig. S5A: Left: Western blot; Right: PageBlue staining. Whole gel images of Supporting Fig. S5B: Left: Western blot; Right: PageBlue staining. Whole gel images of Supporting Fig. S7: Left: Western blot detected by streptavidin; Middle: PageBlue staining. Right: Western blot detected by anti‐NA mAb. Supporting Table S1: IC50 value of 4‐OH, C9‐substituted probes.

Funding

This work was supported by the China Scholarship Council (Grant 201907720035) and the Horizon 2020 Framework Programme (Grant 814102).

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supplementary Material

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.