Structural insights into human parainfluenza virus ... - ACS Publications

the naturally unsubstituted C-3 position on the neuraminidase inhibitor template ... suitable for neuraminidase inhibition, at neutral pH binding of t...
0 downloads 0 Views 894KB Size
Subscriber access provided by AUSTRALIAN NATIONAL UNIV

Structural insights into human parainfluenza virus type-3 haemagglutinin– neuraminidase using unsaturated 3-N-substituted sialic acids as probes Mauro Pascolutti, Larissa Dirr, Patrice Guillon, Annelies Van Den Bergh, Thomas Ve, Robin J. Thomson, and Mark von Itzstein ACS Chem. Biol., Just Accepted Manuscript • DOI: 10.1021/acschembio.8b00150 • Publication Date (Web): 25 Apr 2018 Downloaded from http://pubs.acs.org on April 29, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Chemical Biology

Structural insights into human parainfluenza virus type-3 haemagglutinin–neuraminidase using unsaturated 3-N-substituted sialic acids as probes Mauro Pascolutti,#* Larissa Dirr,#* Patrice Guillon, Annelies Van Den Bergh, Thomas Ve, Robin J. Thomson and Mark von Itzstein* Institute for Glycomics, Griffith University, Gold Coast Campus, Queensland, 4222, Australia. Haemagglutinin-neuraminidase, sialidase, sialic acid, parainfluenza virus, inhibitors. ABSTRACT: A novel approach to human parainfluenza virus type-3 (hPIV-3) inhibitor design has been evaluated by targeting an unexplored pocket within the active site region of the virus’ haemagglutinin-neuraminidase (HN) that is normally occluded upon ligand engagement. To explore this opportunity, we developed a highly efficient route to introduce nitrogen-based functionalities at the naturally unsubstituted C-3 position on the neuraminidase inhibitor template N-acyl-2,3-dehydro-2-deoxy-neuraminic acid (Nacyl-Neu2en), via a regioselective 2,3-bromoazidation. Introduction of triazole substituents at C-3 on this template provided compounds with low micromolar inhibition of hPIV-3 HN neuraminidase activity, with the most potent having 48-fold improved potency over the corresponding C-3 unsubstituted analogue. However, the C-3-triazole N-acyl-Neu2en derivatives were significantly less active against the virus' haemagglutinin function, with high micromolar IC50 values determined, and showed insignificant in vitro antiviral activity. Given the different pH optima of the HN protein's neuraminidase (acidic pH) and haemagglutinin (neutral pH) functions, the influence of pH on inhibitor binding was examined using X-ray crystallography and STD NMR spectroscopy, providing novel insights into the multi-functionality of hPIV-3 HN. While the 3-phenyltriazole-Nisobutyryl-Neu2en derivative could bind HN at pH 4.6, suitable for neuraminidase inhibition, at neutral pH binding of the inhibitor was substantially reduced. Importantly, this study clearly demonstrates for the first time that potent inhibition of HN neuraminidase activity is not necessarily directly correlated with a strong antiviral activity, and suggests that strong inhibition of the haemagglutinin function of hPIV HN is crucial for potent antiviral activity. This highlights the importance of designing hPIV inhibitors that primarily target the receptor-binding function of hPIV HN.

Human parainfluenza viruses (hPIVs) are leading causes of respiratory disease in infants and young children, the immunecompromized, chronically ill, and elderly.1,2 Up to 100,000 hospitalizations per annum in the USA alone occur as a result of hPIV infections.3 However, there are currently neither vaccines nor specific antiviral therapies, to prevent or treat hPIV infections and the development of hPIV-specific antiviral agents is urgently needed.4,5 Critical aspects of hPIV replication in the human host are dependent upon the triple role of the viral surface glycoprotein haemagglutinin-neuraminidase (HN).6 Interaction of HN with cell-surface sialoglycoconjugates of the human airway epithelium through the haemagglutinin function initiates cell binding and promotes the activation of the viral fusion protein.7,8 In addition, the glycohydrolase (neuraminidase) activity of the HN, which cleaves terminal N-acetylneuraminic acid (N-acetyl-Neu; Neu5Ac) residues from host cell receptors, is essential for efficient virus elution from infected cells, allowing viral spread to uninfected cells.3,9 A limited number of inhibitors, predominantly based on the N-acyl-2,3-dehydro-2-deoxy-neuraminic acid (N-acylNeu2en) neuraminidase inhibitor template (1–6, Figure 1), have been used to target the hPIV HN protein.10–16 Over the past few years our group has developed potent Neu2en-based

inhibitors (such as 5 and 6, Figure 1)17 that contain a bulky substituent at C-4 combined with a C-5 isobutyramido functionality. These inhibitors were designed to target a unique feature of the hPIV-3 HN protein, the 216-cavity formed by movement of the flexible 216-loop.17 Furthermore, we also reported the first structural investigation into the catalytic mechanism of the hPIV-3 HN protein using the substrate-like compound, 7.18

Figure 1. Structures of reference hPIV-3 HN inhibitors (1-7) and the potential new target inhibitors (8).

ACS Paragon Plus Environment

ACS Chemical Biology 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Significant flexibility of the hPIV-3 HN protein19 has led us to now investigate functionalization of other positions around the N-acyl-Neu2en scaffold. Through inspection of the apo structure of hPIV-3 HN (PDB: 4XJQ) at pH 4.6 (appropriate for HN neuraminidase activity structural studies), and the structure of hPIV-3 HN in complex with inhibitor 1 (PDB: 1V3D) or 7 (PDB: 4XJR), an unexplored pocket was identified in the apo structure that is occluded upon ligand engagement (Figure 2). Notably, in all reported hPIV-3 HN inhibitor-complexed structures,18,20 the side-chain of R192 undergoes a significant reorientation, when compared to the apo structure at pH 4.6. This reorientation results in the formation of the well-known neuraminidase triarginyl cluster (in hPIV-3 HN R192, R424 and R502) to engage the inhibitor’s carboxylic acid moiety. We observed that the pocket in the apo structure would be situated adjacent to carbon 3 (C-3) of a bound N-acyl-Neu2en derivative such as N-acetyl-Neu2en (Neu5Ac2en, 1) (Figure 2a, C-3 pocket). Hence, we postulated that 1 appropriately modified at C-3 should be able to bind both the HN active site and the previously unexploited pocket observed at acidic pH [pH 4.6 or pH 4.26 (PDB: 4MZA)], prevent reorientation of R192, and inhibit HN activity.

Figure 2. Comparison of the active site region (at pH 4.6) in apo hPIV-3 HN with the hPIV-3 HN–inhibitor complexes of 7 and 1. Surface representations of the active site region in (a) the apo enzyme (PDB ID: 4XJQ), (b) the hPIV-3 HN–7 complex (PDB ID: 4XJR) and (c) the hPIV-3 HN–1 complex (PDB: 1V3D). RESULTS AND DISCUSSION Compound synthesis and HN neuraminidase inhibition. We have previously reported the synthesis of C-3 Csubstituted Neu5Ac2en derivatives21,22 as influenza virus neuraminidase inhibitors, however these compounds were found to be not active against hPIV-3 HN. In the present study, we report the rational design of novel C-3 N-substituted N-acyl-Neu2en derivatives as novel probes to explore the effect of this substitution on hPIV-3 HN activity. Given that C-3-alkyl substituted Neu5Ac2en derivatives, in which the side chains were relatively flexible,21 were inactive, we investigated the incorporation of a bulky and rigid cyclic system, a 1,4-disubstituted [1,2,3]-triazole, at C-3 on the Nacyl-Neu2en scaffold (Figure 1, 8). This cyclic system provides enhanced structural constraint and would be oriented towards the unexplored pocket to potentially optimize ligand binding. The logical choice of precursor for the synthesis of C-3 triazole-functionalized N-acyl-Neu2en derivatives required the introduction of a C-3 azide on the Neu2en scaffold (Scheme 1). A previously reported23 approach to introduce an azido moiety at C-3 of Neu5Ac, used azidonitration chemistry on protected Neu5Ac2en derivative 9. Several attempts, in our

Page 2 of 9

hands, to convert the resulting glycosyl nitrate23 or its corresponding 2-hydroxy derivative23 into an appropriate leaving group24,25 for subsequent β-elimination, however, failed to give the corresponding C-3-substituted Neu5Ac2en derivative. We therefore sought an efficient synthetic route to simultaneously introduce both an azide at C-3 and a leaving group at C-2. This led us to investigate the addition of a haloazide,26,27 in particular bromoazide,28,29 to the protected Neu5Ac2en derivative 9. Bromoazidation of an olefin can be directed to occur via either an ionic or radical mechanism by altering the reaction conditions (e.g. solvent) employed. We believed that an ionic-based addition mechanism would introduce the azide at C-2 through initial formation of a threemembered ring bromonium ion intermediate, which could then be opened at C-2 by the nucleophilic azide ion (similar to bromomethoxylation30 of 9 using N-bromosuccinimide). In contrast, a radical mechanism was anticipated27 to provide the essential inverse regiochemistry of addition, resulting in the introduction of azide at C-3 and bromide at the anomeric position. The regioselectivity of azide radical attack is most likely driven by C-2 radical resonance stabilization by the endocyclic oxygen.31 To introduce an azido group at C-3, appropriate reaction conditions were chosen to force the bromoazidation reaction to proceed via a radical pathway. Thus, a non-polar solvent under high dilution conditions to favour formation of the radical species and propagation of the reaction was employed. The reaction pathway began with the addition of bromoazide to protected Neu5Ac2en derivative 9 in dichloromethane (DCM) (Scheme 1), resulting in a mixture of two bromoazide products 10a,b in a 3:1 ratio. The chemical shift of H-3 in both products was consistent with addition of azide23 rather than bromide30 at C-3, as expected from the radical reaction pathway, with the major isomer (10a) bearing an equatorial azide (J3,4 10.5 Hz) and the minor component an axial azide (10b; J3,4 4.5 Hz). The mixture of 3-azido-2-bromo-2-deoxy-Neu5Ac derivatives 10a,b was treated with 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) in DCM to successfully afford the desired 3-azidoNeu5Ac2en derivative 11 in 77% yield over two steps from protected Neu5Ac2en 9. Scheme 1a AcO

H OAc R AcO H

AcO

CO2CH 3

O

Br

R AcO H

OAc

9

O

X

CO2CH 3

OAc Y

10a X = N 3, Y = H 10b X = H, Y = N 3

R = NHAc

AcO

H OAc

a

H OAc O

b R AcO H

OAc

CO2CH3 N3

AcO

c R

H OAc O R AcO H

11

OAc

CO 2CH 3 N N

N

12-15 R'

d

HO

H OH O R HO H

OH

N N

OH N

16-19 R' a (a) BrN , DCM, Ar, 0 °C to rt, 50 min; (b) DBU, DCM, 3 Ar, 0 °C to rt, 16 h, 77%; (c) alkyne, CuSO 4•5H2O, sodium ascorbate, MeOH/H 2O (1:1), MW (100 W), 80 °C, 20 min, 66–72%; (d) aq. NaOH, MeOH/H 2O (1:1), pH 13, 0 °C to rt, 16 h, 83–85%.

ACS Paragon Plus Environment

R'

CO2Na

12,16

N

13,17

OH

14,18

15,19

Page 3 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Chemical Biology To investigate inhibitors that would target the hPIV-3 HN active site C-3 pocket (Figure 2a), vinyl azide derivative 11 was reacted with a range of alkynes to introduce 1-N-linked 4substituted-[1,2,3]-triazoles at C-3 on the Neu5Ac2en scaffold. In order to probe the effect of the substituent of the triazole ring, four chemically diverse alkynes were chosen. Specifically, the synthesis of 3-triazole-Neu5Ac2en derivatives 12-15 was accomplished through a coppercatalysed 1,3-dipolar cycloaddition with the alkynes using microwave irradiation as the thermal source. Reaction under microwave irradiation (20 min) avoided significant decomposition of 11 that was seen in trial reactions at room temperature (16 h) or under conventional heating of 45-50 °C (2 h). Finally, base-catalysed deprotection led to the target C-3 triazole Neu5Ac2en derivatives 16-19 (Scheme 1). The C-3-triazole Neu5Ac2en derivatives 16-19 were evaluated for their ability to inhibit the neuraminidase activity of hPIV-3 HN (Table 1). The neuraminidase inhibition (NI) assay was performed using the previously described17 endpoint fluorescence-based enzymatic assay utilizing the substrate 2-(4'-methylumbelliferyl)-α-D-N-acetylneuraminide (MUN). In the first instance, screening was conducted at a standard concentration of 200 µM for each of the compounds. Only a very weak NI was observed for compounds 16 and 17 that carry a highly flexible substituent on the triazole ring, with NI of 8.3% and 10.5%, respectively. Introduction of a cyclohexyl substituent at C-4 of the triazole as in 18, resulted in a moderate improvement in inhibitory activity with an NI of 39.1%. However, a significant improvement in activity was observed for the more rigid C-3 phenyltriazole derivative 19, with an NI of 79.9% at an inhibitor concentration of 200 µM. Table 1. Inhibition of hPIV-3 HN neuraminidase activity (% NI at 200 µM) by C-3-triazole Neu5Ac2en derivatives. HO

Scheme 2a AcO

H OAc CO 2CH3

O R AcO H

AcO a, b

H OAc

OAc

CO 2CH 3

O R AcO H

20

N3

OAc

21

R = NHC(O)CH(CH 3)2

AcO c

H OAc CO 2CH3

O R AcO H

OAc

N N

N

HO d

H OH O R HO H

OH

22 =H 24 R' = CO2CH3

N N

N R'

R1

R'

CO 2Na

R'

23 =H 25 R ' = CO 2Na

a (a) BrN , DCM, Ar, 0 °C to rt, 50 min; (b) DBU, DCM, Ar, 0 °C to rt, 16 h, 75%; 3 (c) alkyne, CuSO 4•5H2O, sodium ascorbate, MeOH/H 2O (1:1), MW (100 W), 80 °C, 20 min, 59–73%; (d) aq. NaOH, MeOH/H 2O (1:1), pH 13, 0 °C to rt, 16 h, 88–89%.

The inhibitory activity of compound 23 was assessed in the standard NI assay, as described above, where it was found to have improved potency relative to its Neu5Ac2en analogue 19 (Figure 3). Replacement of the C-5 acetamide on the C-3 phenyltriazole derivative 19 with the isobutyramide (23), improved the hPIV-3 HN NI by more than two-fold (23, IC50 = 31.8 µM compared to the acetamido derivative 19, IC50 = 78.8 µM). This gain in inhibitory potency is comparable to that previously reported when the identical replacement was made on C-4-functionalized Neu2en compounds.17 Impressively, the introduction of the C-3 phenyltriazole substituent onto Nisobutyryl-Neu2en 3 provides a six-fold increase in potency over the C-3 unsubstituted analogue (3, IC50 = 188 µM). This also confirms that combined C-3 and C-5 substitution is well tolerated, providing an overall positive additive effect.

H OH O R HO H

OH

R = NHAc

CO 2Na N N

N

R'

R'

NI (%)a

16

17

18

19

8.3 ±14.6

10.5 ±2.8

39.1 ±1.3

79.9 ±0.4

a

Averaged values were calculated from two independent experiments performed in triplicate. 17,

32

With the knowledge that for hPIV-3, and hPIV-1, Neu2en-based inhibitor potency is improved when the C-5 acetamide is replaced by an isobutyramide, we decided to introduce this modification on the most active C-3-triazole Neu5Ac2en derivative 19. The synthetic approach began with a radical addition of BrN3 to N-isobutyryl-Neu2en derivative 2017,32 followed by β-elimination that generated the corresponding vinyl azide 21 (Scheme 2). Following the previously described method, compound 21 was reacted with phenylacetylene to give C-3-triazole derivative 22, which upon base-catalyzed deprotection gave the desired 3phenyltriazole-N-isobutyryl-Neu2en derivative 23.

Figure 3. hPIV-3 NI IC50 values for compounds 19 (circle), 23 (triangle) and the previously reported17 benchmark compound 3 (diamond). Averaged values were calculated from two independent experiments performed in triplicate and error bars correspond to the calculated standard deviation. hPIV-3 HN–23 complex - structure analysis. To understand the structural basis behind the inhibitory activity of the C-3 modified Neu2en derivatives, a crystal structure of hPIV-3 HN in complex with the most potent inhibitor in this series to this point, 3-phenyltriazole-N-isobutyryl-Neu2en (23), to a 1.83 Å resolution, was obtained by soaking crystals of the protein in a 50 mM solution of the inhibitor (at pH 4.6). The structure of the co-crystallized complex with inhibitor 23 had two protein

ACS Paragon Plus Environment

ACS Chemical Biology 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

molecules per asymmetric unit (binding site A and B); here we report coordination of 23 in binding site B. The electron density for 23 was overall weaker in binding site A, but using polder omit maps,33 inhibitor 23 was found in a nearly identical orientation when compared to binding site B (see Supplementary Figure 1a and Supplementary Results and Discussion). Overall, the core features of 23 – the dihydropyran ring, the C-6 glycerol side chain, and the C-5 isobutyramide – oriented in positions similar to those previously observed in structures of hPIV-3 HN–Neu2en-based inhibitor complexes.18,20 As predicted, the C-3 phenyltriazole moiety of 23 extends into the pocket observed in the pH 4.6 apo structure (Figure 4a). In order to accommodate the C-3 phenyltriazole moiety, the sidechains of F558 and the disulfide bond forming cysteines C190 and C214 have moved slightly away from the incoming ligand when compared to the apo structure (Figure 4b). The phenyl substituent on the triazole ring of 23 is well accommodated in the hydrophobic region that is generated by F558 and the hydrocarbon chain of R192. Binding of the C-3 phenyltriazole moiety is further accompanied by two hydrogen bonds formed with two active-site water molecules. Importantly, as we postulated, we did observe one major point of difference with the accommodation of the inhibitor’s C-3 substituent that is not usually seen in other hPIV-3 HN complexes. The amino acid residue R192, a member of the crucial triarginyl cluster, is not engaged with the carboxylic acid moiety of 23, an interaction that is typically seen in hPIV3 HN–inhibitor complexes.18,20 Upon binding of the C-3 substituent, the R192 side-chain is found in a similar orientation to that observed in the pH 4.6 apo hPIV-3 HN structure. Additionally, R192 forms a hydrogen bond with the C-4 hydroxy group of inhibitor 23. Furthermore, in the hPIV-3 HN–23 complex, the Y530 side-chain is oriented away from the binding site. This key catalytic amino acid is known to nucleophilically attack the C-2 anomeric carbon of the 2,3difluoro-N-isobutyrylneuraminic acid derivative 7, as previously described.18 From the crystallographic analysis, we observed the possibility to engage further interactions with R192 by introducing a negatively charged moiety at the ortho position of the phenyl ring of 23. The desired compound 3-(ocarboxyphenyltriazole)-N-isobutyryl-Neu2en 25 was therefore synthesized following our established chemical pathway – microwave mediated click reaction and base catalysed deprotection (Scheme 2). Compound 25 was subsequently tested for its capacity to inhibit the neuraminidase activity of hPIV-3 HN (Table 2). The introduction of the carboxylic acid group on the phenyl ring of inhibitor 23, resulted in a remarkable improvement of 8-fold increased NI potency for 25 (IC50 = 3.9 µM) compared to the ‘parent’ compound 23 (IC50 = 31.8 µM).

Figure 4. Crystal structure of hPIV-3 HN in complex with 3-phenyltriazole-N-isobutyryl-Neu2en 23. (a) Surface representations of the active site region in hPIV-3 HN–23 complex (PDB ID: 6C0M) with electrostatic potential (calculated using APB mapped to the surface. Coloring is continuous going from blue (potential +10 kT/e) through white to red (potential -10 kT/e). The inhibitor 23 is shown in stick representation (carbon green, oxygen red, nitrogen blue) and the electron density map of the inhibitor (2mFo-DFc, grey mesh) is contoured at 1σ. (b) Superimposition of the pH 4.6 apo and hPIV-3 HN–23 complex structures, displayed in cyan and green, respectively. The active site residues and inhibitor 23 are shown in stick representation. Active site water molecules in the hPIV-3 HN–23 complex are displayed as green spheres. Haemagglutination and virus growth assays. As hPIV-3 HN is a multifunctional protein, compounds 23, 25 were tested for their capacity to block the haemagglutinin function of hPIV-3 HN. For this purpose, utilising a standard haemagglutinin inhibition (HI) assay,17,34 the ability of 23 and 25 to inhibit virus-mediated agglutination of human red blood cells was evaluated. Interestingly, we observed that 23 and 25 displayed 9-fold and 212-fold weaker inhibitory potency (IC50), respectively, against the haemagglutinin, compared to their neuraminidase inhibitory potencies. In comparison, the C-3 unsubstituted analogue 3 displays only 2-fold difference between HI and NI.17 Subsequently, using an in-situ ELISA17,33 the in vitro antiviral effects of 23 and 25 in a human respiratory cell line were determined. Surprisingly, despite

ACS Paragon Plus Environment

Page 4 of 9

Page 5 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Chemical Biology their potencies in NI assays, compound 23 and 25 showed very weak antiviral activity. In fact, their IC50 values were > 1 mM in vitro (Table 2). Table 2. IC50 values (µM) of functional (NI and HI) and hPIV-3 virus growth inhibition assays. a

IC50 values (µ µM)

b

b

23

25

3

4

6

NI

31.8 ± 1.2

3.92 ± 0.29

187.7 b ± 39.5

21.5 ± 4.8

2.74 ± 0.23

HI

275.0 ± 35.3

830.0 ± 42.4

358.5 ± 52.0 b

16.1 ± 4.9

1.46 ± 0.36

Virus growth inhibition

> 1000

> 1000

> 1000

130.6 ± 13.0

10.3 ± 0.3

a

Averaged values were calculated from two independent experiments performed in triplicate. b Previously reported values.17

Interestingly, hPIV-1 HN and virus displayed the same trends when exposed, under identical conditions used for hPIV-3 HN and virus, to the C-3 substituted compounds 23 and 25; that is, similar µM level NI IC50 values to those seen for hPIV-3 HN, weaker HI than NI, and poor virus infection blockade (data not shown). This suggests that both hPIV-1 and hPIV-3 require good inhibition of the virus' haemagglutinin function to achieve good virus infection blockade. Paramyxovirus haemagglutinin-neuraminidase proteins differ from other neuraminidases (NAs), such as influenza virus NA, due to their dual functionality, not only within one protein but also within one binding site.20 Importantly, this is the first time that a significant discrepancy between hPIV-3 NI and HI potencies for any Neu2en-based inhibitor targeting hPIV-3 HN has been reported. The most potent Neu2en-based hPIV-3 HN inhibitors, derivative 617 and BCX 2798 (4),10 show similar inhibitory potencies for both NI and HI, and also display antiviral activity in the same µM range as in the functional assays (Table 2).17 As detailed above, although both compounds 23 and 25 showed low micromolar NI potency, they were found significantly less effective against the haemagglutinin function of hPIV-3 HN, and surprisingly showed insignificant antiviral activity (Table 2). These results suggest that targeting the haemagglutinin function of hPIV-3 HN could be crucial for the design of highly potent hPIV-3 HN inhibitors. This is in contrast to the development of potent inhibitors of influenza virus, where inhibition of the viral neuraminidase alone, for example by the Neu5Ac2en-based approved anti-influenza drug zanamivir, is sufficient to produce a strong antiviral effect. Zanamivir failed to show dose-dependent inhibition of the haemagglutinin protein of influenza virus.35,36 Interestingly, the recently reported 2,3difluoro-N-acylneuraminic acid analogues also showed poor inhibitory potency towards the HN haemagglutinin function.34,37 However, the antiviral potency from these mechanism-based inhibitors might result primarily from their long-lasting neuraminidase inhibition due to the fact that they form a covalent bond with the hPIV-3 HN protein. In fact, the long-lasting NI of these compounds can block the binding of

subsequent progeny virion to cellular receptors, providing antiviral activity in vitro.18 The hPIV-3 HN protein, based on the structure and kinetic analysis, has been described as primarily a neuraminidase with a high Km value in the order of 1 mM for the synthetic substrate MUN as well as for human alpha-1 acid glycoprotein (AGP) that leads to slow dissociation of the ligand and consequently binding activity.38 Importantly, recent work showed that strong haemagglutinin inhibition has been correlated with strong antiviral activity for Neu2en-based hPIV-3 HN inhibitors.10,17 Effect of pH on inhibitor binding. The neuraminidase and haemagglutinin assays are carried-out at the HN proteins' different optimal pH values, pH 4.6 and neutral pH respectively. hPIV HN neuraminidase activity is optimal at close to pH 5,39,40 with maximal neuraminidase activity measured in-house at pH 4.6 for hPIV-3 HN and 5.0 for hPIV1 HN. In contrast, hPIV-3 HN neuraminidase activity is reported,39 and was found in-house, to be negligible (