The pH-sensitive action of cholesterol-conjugated peptide inhibitors of influenza virus

Abstract

Influenza viruses are major human pathogens, responsible for respiratory diseases affecting millions of people worldwide, with high morbidity and significant mortality. Infections by influenza can be controlled by vaccines and antiviral drugs. However, this virus is constantly under mutations, limiting the effectiveness of these clinical antiviral strategies. It is therefore urgent to develop new ones. Influenza hemagglutinin (HA) is involved in receptor binding and promotes the pH-dependent fusion of viral and cell endocytic membranes. HA-targeted peptides may emerge as a novel antiviral option to block this viral entry step. In this study, we evaluated three HA-derived (lipo)peptides using fluorescence spectroscopy. Peptide membrane interaction assays were performed at neutral and acidic pH to better resemble the natural conditions in which influenza fusion occurs. We found that peptide affinity towards membranes decreases upon the acidification of the environment. Therefore, the released peptides would be able to bind their complementary domain and interfere with the six-helix bundle formation necessary for viral fusion, and thus for the infection of the target cell. Our results provide new insight into molecular interactions between HA-derived peptides and cell membranes, which may contribute to the development of new influenza virus inhibitors.

1. Introduction

The influenza virus remains a serious threat to global health, causing worldwide morbidity and mortality through pandemics and seasonal epidemics [1,2]. Annual epidemics infect between 5% and 15% of the world population [3]. Currently, influenza has higher incidence in elderly [4]. This virus is responsible for millions of hospitalizations and 250,000–500,000 deaths per year around the world [5,6].

Annual epidemics of influenza infections can be controlled by vaccination and antiviral drugs [7,8]. However, vaccines give limited protection and need annual updating due to antigenic drift [9]. The antiviral drugs available in the market for the prevention and treatment of influenza A virus infection, such as M2 ion channel (a transmembrane protein that forms proton channel in the viral envelope) blockers (amantadine and rimantadine) and neuraminidase (NA) inhibitors, have limited action, as a result of the circulating resistant strains and due to the appearance of side effects during drug use [10–12]. This illustrates the need for the development of new antiviral therapies, with different mechanisms of action, to fight influenza virus infection.

Influenza hemagglutinin (HA) is a key protein in the initial stages of infection. HA-mediated binding to the receptor triggers the internalization of the virion via endocytosis. Internalized viruses are trafficked on early endosomes, reaching the acidic late endosome, where exposure to low pH activates the M2 ion channel and causes a large conformational change in HA, leading to the exposure of the fusion peptide and, consequently, the fusion between the viral and endosomal membranes [13,14]. Initially, HA is synthesized as a fusion-inactive precursor, and then, by the acidification of the environment, it is activated in order to expose the N-terminal domain containing the fusion peptide [15,16]. HA has two subunits, HA1 and HA2. In order to achieve membrane fusion, after the binding of HA1 to the target receptor, HA2 becomes exposed and a loop-to-α-helix transition in this subunit enables the fusion peptide at the N-terminal region of HA2 to be inserted in the endosomal membrane, promoting its fusion with the viral membrane and the release of the viral ribonucleoproteins into the cytoplasm of the target cell [17]. Understanding the different HA conformational changes that take place during membrane fusion is important to design fusion inhibitor peptides that can prevent those changes. As several other viral class I fusion proteins, HA2 contains two heptad repeat (HR) regions, one closer to the N-terminal (NHR) and another closer to the C-terminal (CHR). For viral entry to occur, the NHR and CHR regions of a trimer of HA2 subunits must come together, forming a structure named six-helix bundle (6HB) [18,19]. Peptides derived from these HR regions are known to inhibit fusion, by binding to their complementary HR region, preventing NHR and CHR from refolding into the stable 6HB structure and inhibiting the fusion process [20]. HA-targeted anti-fusion peptides are expected to lead to novel anti-influenza drugs, as it was successfully achieved for the human immunodeficiency virus (HIV), with the approval for clinical use and inclusion in clinical practice of enfuvirtide [21].

The major challenge in developing fusion inhibitors for influenza virus has been the fact that the viral fusion occurs intracellularly. One strategy that has been applied is to tag the peptide with a lipid moiety, such as cholesterol. Cholesterol-tagged peptides may follow HA from the cell surface to the site of fusion [22]. Furthermore, cholesterol moieties are also able to orientate and localize the peptides on the target membrane, leading to a strategy that improves their antiviral activity, as reported in our previous studies for other viruses [23–27]. Following the same strategy, we used peptides designed based on HA2 domain, namely, an untagged peptide (Influenza-untagged), a peptide monomer covalently linked to cholesterol through a polyethylene glycol (PEG) spacer (Influenza-PEG₄-Chol) and a peptide dimer also linked through PEG spacers to a single cholesterol moiety ((Influenza-PEG₄)₂-Chol).

We used different fluorescence techniques to assess the interaction of the peptides with biomembrane model systems and human blood cells (erythrocytes and peripheral blood mononuclear cells, PBMCs), in order to understand the biophysical properties of the peptides. As in our previous studies with fusion inhibitors of HIV, human parainfluenza viruses (HPIV) and measles virus [23,24,28–30], here we were able to understand that cholesterol tagging is indeed a strategy that improves the peptide-membrane interaction. In order to better resemble the natural conditions where influenza fusion occurs, we evaluated the same interaction with biomembrane model systems and with human blood cells at pH 7.4 and 5.0. Our results showed that the peptide-membrane affinity decreases with acidic pH, suggesting that after peptide internalization in the endosome and upon acidification of the environment, the dynamic of the lipopeptide interaction with the cell membrane increases. We have shown that the ability of lipopeptides to associate dynamically with the cell membrane positively correlates with antiviral potency [26]. We propose this will be the case also for the influenza peptides.

2. Materials and methods

2.1. Reagents and sample preparation

Peptides were designed based on the hemagglutinin HA2 chain (Table 1). Influenza-untagged (P155–185) [22], Influenza-PEG₄-Chol (P155–185-chol) [22] and (Influenza-PEG₄)₂-Chol were synthesized as previously described [31]. Stock solutions of each peptide were prepared in dimethyl sulfoxide (DMSO), with a final concentration of 500 μM, and stored at −20 °C. The phospholipid 1-palmitoyl-2-oleoylsn-glycero-3-phosphocoline (POPC) was purchased from Avanti Polar Lipids (Alabaster, AL, USA). NaCl, L-tryptophan, DMSO, sodium citrate, chloroform and acrylamide were from Merck (Darmstadt, Germany). HEPES, cholesterol (Chol), Pluronic-F127, 1-anilino-8-naphthalene sulfonate (ANS), 5-doxyl-stearic acid (5NS) and 16-doxyl-stearic acid (16NS) were acquired from Sigma-Aldrich (St. Louis, MO, USA). Di-8-ANEPPS was purchased from Invitrogen – Molecular Probes (Eugene, OR, USA). Lymphoprep was obtained from Stemcell Technologies (Vancouver, BC, Canada).

Table 1.

Influenza peptides sequence, molecular weight and antiviral activity (IC₅₀) against influenza H3N2 live virus in a plaque reduction assay. IC₅₀ is the inhibitor concentration necessary to achieve a 50% inhibition.

Name	Sequence	MW (Da)	IC50 (μM)
Influenza-untagged	GTYDHDVYRDEALNNRFQIKGVELKSGYKDW-GSGSG-C-(CH2CONH2)-NH2	4882	>10a
Influenza-PEG4-Chol	GTYDHDVYRDEALNNRFQIKGVELKSGYKDW-GSGSG-C-(PEG4-Chol)-NH2	4265	0.4a
(Influenza-PEG4)2-Chol	[GTYDHDVYRDEALNNRFQIKGVELKSGYKDW-GSGSG-C-(PEG4)]2-Chol-NH2	9164	–

^aIC₅₀ values were obtained from [22].

The working buffers used throughout the studies were HEPES 10 mM pH 7.4 with NaCl 150 mM and sodium citrate 10 mM pH 5.0 with NaCl 150 mM. L-Tryptophan 500 μM stock solution was prepared in buffer, while ANS 2 mM and acrylamide 2 M were prepared in ultrapure H₂O. Stock solutions of the fluorescent probe di-8-ANEPPS 1 mg/mL and the lipophilic quenchers 5NS and 16NS 70 mM were prepared in ethanol. All stock solutions were stored at 4 °C, except 5NS and 16NS that were stored at −20 °C. Large unilamellar vesicles (LUV) were prepared by extrusion methods, as previously described elsewhere [32,33].

2.2. Fluorescence spectroscopy

The influenza peptides under evaluation include on their primary structure tryptophan residues, intrinsically fluorescent. This makes fluorescence techniques valuable tools to provide information on the interaction of these molecules with membranes. Membrane partition and fluorescence quenching studies using acrylamide were carried out in a Varian Cary Eclipse fluorescence spectrophotometer (Mulgrave, Australia), while the time-resolved fluorescence spectroscopy studies were performed in a LifeSpec II fluorescence lifetime spectrometer (Edinburgh Instruments, Livingston, UK). Peptide aggregation measurements were performed in an Edinburgh Instruments FLS920 Series fluorescence spectrophotometer (Livingston, UK).

Intrinsic fluorescence measurements of Influenza-untagged, Influenza-PEG₄-Chol and (Influenza-PEG₄)₂-Chol were performed with an excitation wavelength of 280 nm, except for acrylamide quenching experiments, where the excitation was done at 290 nm. For the quenching experiments, fluorescence emission was collected at a fixed wavelength of 350 nm, while for the partition studies spectra integrated from 310 to 450 nm were used. Typical spectral bandwidths were 5 nm for excitation and 10 nm for emission. All fluorescence measurements were conducted at room temperature.

Time-resolved intensity decays were obtained at 280 nm (vertically polarized) and fluorescence was acquired at 350 nm (20 nm bandwidth), using a 20 ns time span and 1024 channels in a multichannel analyzer. Fluorescence lifetimes were obtained from intensity decay fits with a sum of exponentials and using a non-linear least-squares method based on the Marquardt algorithm [34]. The quality of the fit was evaluated from χ2 values and distributions of the residuals of the autocorrelation plots.

2.3. Partition coefficient determination

Membrane partition studies were performed by successive addition of small aliquots of LUV suspensions of either POPC or POPC:Chol (2:1) to a 5 μM peptide solution, with 10 min incubation time between each addition. Assays were carried out at pH 7.4 and 5.0. Membrane partition usually yields a hyperbolic fluorescence intensity (I) vs. lipid concentration ([L]) variation profile, with I increasing as a function of [L], tending to a horizontal asymptotic line. However, in the case of our study, there is mostly a decrease in I upon increasing [L], with an additional deviation from the normal hyperbolic behavior at higher lipid concentrations. This is usually associated with self-quenching upon insertion in the membrane (for a review on different possible deviations from the most common membrane partition behavior, vd. [35]). The partition coefficient (K_p) was calculated using the equation [35]:

$$I=\frac{{\gamma }_L{K}_P\left[L\right]I_L}{1+K_P{\gamma }_L\left[L\right]+k_2{K}_P{I}_L}+\frac{I_W}{1+{\gamma }_L{K}_P\left[L\right]}$$ (1)

where,

$$k_2=\frac{k_q}{\text{ε}l{k}_f}$$ (2)

I_W and I_L represent the fluorescence intensities in aqueous solution and in the volume of the lipid membrane, respectively. [L] is the concentration of the lipid and γ_L is its molar volume. K_q is the kinetic constant of the quenching process, l is the optical path, k_f is the radiative fluorescence constant and ε is the molar absorptivity.

2.4. Peptide aggregation

In order to assess peptide aggregation, the process was followed by ANS fluorescence emission, with excitation at 369 nm and emission collected between 400 and 600 nm, using bandwidths of 5 and 10 nm for excitation and emission, respectively [36]. A solution containing 12.8 μM of ANS in HEPES buffer was titrated with a stock solution of the peptide to yield a final peptide concentration in the range of 0–8 μM. Two way-ANOVA was applied on the statistical analysis of the aggregation data.

2.5. Fluorescence quenching

Quenching of 5 μM Influenza-untagged, Influenza-PEG₄-Chol and (Influenza-PEG₄)₂-Chol intrinsic fluorescence by acrylamide (0–60 mM) [37] was performed in buffer and in the presence of LUVs containing 2 mM POPC, by successive additions of small volumes of acrylamide. For every addition, a minimum of 10 min incubation time was allowed before measurement. Data were analyzed by using the Stern-Volmer equation [34],

$$\frac{I_0}{I}=1+K_{sv}\left[Q\right]$$ (3)

where I and I₀ are the fluorescence intensities in the presence and absence of quencher, respectively, K_SV is the Stern-Volmer constant and [Q] the quencher concentration.

Quenching by the lipophilic probes 5NS and 16NS was performed by time-resolved fluorescence spectroscopy. In these assays, a 5 μM solution of each peptide was incubated with 3 mM POPC LUV. Then, successive additions of small volumes of 5NS or 16NS (in ethanol) were made, keeping the ethanol concentration below 2% (v/v) [38]. The effective concentration of the quencher at the membrane level, [Q]_L, was determined from the partition coefficient of both lipophilic molecules to biomembranes [39], and ranged from 0 to approximately 0.6 M. For every addition, a 10 min incubation time was allowed before measurement. Quenching data were analyzed by using the Stern–Volmer equation and the SIMEXDA [40] method was applied to obtain the depth of the tryptophan insertion in membranes.

2.6. Membrane dipole potential assays by di-8-ANEPPS

Human blood samples were obtained from healthy donors, with their written informed consent, at Instituto Português do Sangue e da Transplantação (Lisbon, Portugal), as approved by the Joint Ethics Committee of Faculdade de Medicina da Universidade de Lisboa and Centro Hospitalar Lisboa Norte (Lisbon, Portugal). Samples were collected to K₃EDTA anticoagulant tubes (Vacuette, Greiner Bio-One, Kremsmünster, Austria). Isolation of erythrocytes and PBMCs and labeling of these cells with di-8-ANEPPS were performed as previously described [21,41]. For erythrocytes isolation, in order to remove plasma and buffy-coat, blood sample was centrifuged for 10 min at 1200 ×g. Afterwards, isolated erythrocytes were washed three times with working buffer. They were incubated at 1% hematocrit in buffer supplemented with 0.05% (m/v) Pluronic F-127 and 10 μM di-8-ANEPPS. PBMCs were isolated with a density gradient using Lymphoprep. Cells were counted in a Moxi Z Mini automated Cell Counter (ORFLO Technologies, Ketchum, ID, USA). A suspension was prepared with a final concentration of 3000 cells/mL, in 0.05% Pluronic F-127 supplemented HEPES buffer with di-8-ANEPPS 3.3 μM. PBMC were incubated with the fluorescent probe suspension at room temperature, with gentle agitation and protected from the light. The unbound probe was removed by centrifugations at 1500 × g for 5 min. To test the membrane dipole potential of erythrocytes and PBMC at pH 5.0, sodium citrate buffer was used in the final centrifugations, after suspension incubation. The peptides were incubated with erythrocytes at 0.02% hematocrit and with PBMCs at 100 cells/μL, for 1 h, with gentle agitation, before fluorescence measurements. For LUVs labeling, to ensure the maximum incorporation of the probe, suspensions with 500 mM of total lipid were incubated overnight with 10 μM di-8-ANEPPS. The maximum concentration of DMSO in the suspensions was 2.4% (v/v) at 6 μM of peptide.

Excitation spectra and the ratio of fluorescence intensities at the excitation wavelengths of 455 and 525 nm (R = I₄₅₅/I₅₂₅) were obtained with emission set at 670 nm to avoid membrane fluidity artifacts [42,43]. Excitation and emission bandwidths for these measurements were set to 5 and 10 nm, respectively. The variation in R with the peptide concentration was analyzed with a single binding site model [44]:

$$\frac{R}{R_0}=1+\frac{R_{\min }/R_0\left[peptide\right]}{K_d+\left[peptide\right]}$$ (4)

with the R values normalized for R₀, the value in the absence of peptide. R_min defines the asymptotic minimum value of R.

3. Results and discussion

3.1. Peptide-membrane partition

In order to quantify the extent of interaction of peptides with large unilamellar vesicles (LUV), the partition coefficient (K_p) between the lipid and aqueous phases was calculated by using Eq. (1) to fit the experimental data (Table 2). The peptides tagged with cholesterol had a decrease in the fluorescence intensity in the presence of LUV of POPC and POPC:Chol (2:1) at pH 7.4, showing a higher K_p, indicating more interaction with the membranes than the untagged peptide. In the case of the Influenza-untagged peptide, no significant changes were detected, demonstrating an absence of significant peptide-membrane interaction (Fig. 1A–B).

Table 2.

Peptide-membrane partition coefficients (K_p). Parameters obtained from the fitting of the fluorescence data of the partition assays of Influenza-untagged, Influenza-PEG₄-Chol and (Influenza-PEG₄)₂-Chol using Eq. (1). All measurements were made at least in triplicate. Values are presented as mean ± standard deviation (SD).

Peptide	Lipid	pH 7.4		pH 5.0
		Kp	IL/IW	Kp	IL/IW
Influenza-untagged	POPC	≈0a	–a	≈0a	–a
	POPC:Chol (2:1)	≈0a	–a	≈0a	–a
Influenza-PEG4-Chol	POPC	2377 ± 370.3	1.17 ± 0.12	≈0a	–a
	POPC:Chol (2:1)	1938 ± 493.9	1.16 ± 0.33	≈0a	–a
(Influenza-PEG4)2-Chol	POPC	2326 ± 708.2	1.10 ± 0.11	3672 ± 940.1	0.80 ± 0.07
	POPC:Chol (2:1)	2252 ± 1657.4	1.02 ± 0.10	5154 ± 510.6	0.83 ± 0.02

^aThe type of partition curve obtained impairs the use of Eq. (1).

Fig. 1.

Partition of the peptides to lipid vesicles. Fluorescence intensity variation upon titration with LUVs of POPC and POPC:Chol (2:1) of 5 μM Influenza-untagged (circles), Influenza-PEG₄-Chol (squares) and (Influenza-PEG₄)₂-Chol (triangles). Membrane partition in HEPES 10 mM NaCl 150 mM buffer pH 7.4 (A, B) and in sodium citrate buffer pH 5.0 (C, D). Solid lines are fittings of Eq. (1) to the experimental data. Results correspond to the average of three independent replicates. Error bars represent the standard error of the mean (SEM).

As influenza virus fusion is triggered by a change to acidic conditions in the endosome, we also performed the peptide-membrane partition assays at pH 5.0. The dimer (Influenza-PEG₄)₂-Chol was the only molecule that showed a partition to both POPC and POPC:Chol lipid vesicles. In this case, not only the untagged peptide but also the monomer Influenza-PEG₄-Chol do not seem to interact with the membrane. For the conjugated peptides, the equilibrium between the aqueous and the lipid phase was reached above a lipid concentration 0.5 mM (Fig. 1C–D and Table 2).

(Influenza-PEG₄)₂-Chol was the only molecule to partition into membranes at acidic conditions, which indicates that it would remain attached to the membrane upon the acidification of the endosome during influenza virus infection. (Influenza-PEG₄)₂-Chol demonstrated a higher membrane affinity, as indicated in different assays, namely partition data and interaction with PBMC. This is in agreement with previous studies, where it was reported that viral fusion inhibitor peptide dimerization leads to a higher potency and activity [31]. In our previous studies, we suggested that dimerization combined with cholesterol tagging may be a general strategy for increasing fusion inhibitor peptides antiviral potency and for extending their in vivo half-life [24].

3.2. Peptide aggregation

Peptides conjugated with lipid moieties are prone to self-aggregation when in aqueous solutions [25,26,29]. To address this aspect, we incubated the peptide with the probe ANS and followed the interaction by fluorescence intensity measurements. ANS has a very low fluorescence quantum yield in water, but a higher one in the presence of hydrophobic pockets. This enables the assessment of the formation of peptide aggregates [36]. In this case, both the monomer Influenza-PEG₄-Chol and the dimer (Influenza-PEG₄)₂-Chol caused an increase in the fluorescence intensity of ANS (Fig. 2A) and a significant blue shift of the maximum wavelength of ANS emission (Fig. 2B), which indicates the formation of aggregates.

Fig. 2.

Aggregation of the influenza peptides evaluated by ANS fluorescence properties. ANS (12.8 μM) was titrated with small volumes of each peptide up to final concentrations. (A) Dependence of fluorescence intensity with peptide concentration. (B) Normalized fluorescence emission spectra, highlighting the ANS spectral shifts upon interacting with Influenza-untagged, Influenza-PEG₄-Chol and (Influenza-PEG₄)₂-Chol. A significant blue shift upon increasing peptide concentration and a concomitant increase in ANS quantum yield indicate that this peptide is forming aggregates. Results correspond to the average of three independent replicates. Error bars represent the SEM. *** p < 0.001.

3.3. Localization in lipid bilayers

Due to the peptide partition to lipid membranes, it is important to address whether the peptide is merely adsorbed on the membrane surface or inserted between the lipids. To test the accessibility of the tryptophan residues of the peptides to the aqueous environment, we used acrylamide as a fluorescence quencher, due to its low capacity for penetration into lipid bilayers. The fluorescence of the untagged peptide was more efficiently quenched in buffer than in the presence of vesicles of POPC (Fig. 3 and Table 3). This might suggest that in the presence of membranes the peptide changes its conformation and the tryptophan residue becomes less prone to suffer quenching by acrylamide, or it is somehow protected by the membrane.

Fig. 3.

Fluorescence quenching by acrylamide of Influenza-untagged (A), Influenza-PEG₄-Chol (B) and (Influenza-PEG₄)₂-Chol (C), with the peptide intrinsic fluorescence with (I) and without (I₀) quencher measured in the absence (squares) and presence (circles) of POPC LUVs. 5 μM peptide and 3 mM total lipid were used in these experiments. Solid lines are fittings of the Stern-Volmer equation (Eq. (3)). Using time-resolved fluorescence measurements to assess the fluorescence lifetime with (τ) and without (τ₀) quencher, we obtained Stern-Volmer plots of the quenching of Influenza-untagged, Influenza-PEG₄-Chol and (Influenza-PEG₄)₂-Chol fluorescence by 5NS (D) or 16NS (E), in POPC LUVs. [Q]_L is the local concentration of the quencher in the strict bilayer volume [43]. Depth of insertion of influenza peptides in the membrane (F), determined using the SIMEXDA method [40], yielding an average location 17.2 Å, 12.4 Å and 13.2 Å away from the center of the bilayer for Influenza-untagged, Influenza-PEG₄-Chol and (Influenza-PEG₄)₂-Chol, respectively. Distributions’ half-width at half-height were 1.4 Å for Influenza-untagged, 2.8 Å for Influenza-PEG₄-Chol and 4.0 Å for (Influenza-PEG₄)₂-Chol. Results correspond to the average of three independent replicates.

Table 3.

Quenching parameters. Stern-Volmer constants (K_SV) were obtained using Eq. (3) to fit the acrylamide quenching data of Influenza-untagged, Influenza-PEG₄-Chol and (Influenza-PEG₄)₂-Chol. All measurements were made at least in triplicate. Values are presented as mean ± SD.

System		KSV (mM−1)
	Influenza-untagged	Influenza-PEG4-Chol	(Influenza-PEG4)2-Chol
Buffer	15.70 ± 0.73	3.58 ± 0.30	8.48 ± 0.28
POPC	10.08 ± 0.38	4.36 ± 0.36	6.35 ± 0.20

It is noticeable that the quenching of the conjugated peptides in buffer is less efficient than for the untagged peptide (Table 3), which indicates that the tryptophan residues of the cholesterol-tagged peptides are less exposed to the aqueous environment. The in-depth location of the tryptophan residues inserted in lipid membranes was evaluated by using the lipophilic quenchers 5NS and 16NS. The quencher doxyl group of these molecules, either on the carbon 5 or 16 of the fatty acyl chain, when inserted in the membrane, has distinct locations. The quencher group of 5NS preferentially quenches fluorophores located at a shallower position in the membrane, whereas 16NS quenches more efficiently fluorophores located closer to the center the bilayer. Accurate determination of the membrane penetration depth is an important step in characterizing membrane interactions of proteins and peptides [45].

5NS and 16NS were able to access the tryptophan residues, indicating that those residues are located in the membrane (Fig. 3D,E). In order to obtain the in-depth distribution of the fluorophores, the SIMEXDA method [40] was applied (Fig. 3F). This method takes into account the possibility of static quenching, by using a sphere-of-action methodology [40]. Using this method, it was possible to estimate that, on average, the tryptophan residue of Influenza-untagged is located 17.2 Å away from the center of the bilayer, while those from Influenza-PEG₄-Chol and (Influenza-PEG₄)₂-Chol are located at 12.4 Å and 13.2 Å, respectively. Influenza-untagged is therefore located at a shallower position, close to the membrane surface, while the cholesterol-tagged peptides are deeply inserted in the membrane. This comparison should be made not only looking at the maxima of the distribution, but also at their half-widths at half height [40], which are 1.4 Å for Influenza-untagged, 2.8 Å for Influenza-PEG₄-Chol and 4.0 Å for (Influenza-PEG₄)₂-Chol.

Cholesterol-conjugated peptides (or, more precisely, their intrinsically fluorescent tryptophan residues) are located at the level of the phospholipid’s acyl chains, on average 12–13 Å away from the center of the bilayer, while the untagged peptide is adsorbed on the membrane surface. These results demonstrate that the tagging with cholesterol promotes peptide insertion in the membrane of the cells.

3.4. Membrane dipole potential

Peptide-membrane interactions were also evaluated using the lipophilic probe di-8-ANEPPS. As di-8-ANEPPS is an indirect reporter of the membrane dipole potential, any changes caused by the insertion or adsorption of the peptides in/on the membrane can be translated in shifts in the probe’s excitation spectra [35]. It should be noticed that only the local changes on the tryptophan residues microenvironment are monitored in the partition assays, whereas dipole potential changes are able to assess the interaction of the entire fusion inhibitor with the membrane. Making this evaluation for POPC LUV, we found that only the cholesterol-tagged peptides were able to change the membrane dipole potential (Fig. 4).

Fig. 4.

Interaction of hemagglutinin-derived peptides with di-8-ANEPPs-labeled LUVs and human blood cells. Binding profiles of Influenza-untagged, Influenza-PEG₄-Chol and (Influenza-PEG₄)₂-Chol to LUVs, erythrocytes and PBMC at pH 7.4 and erythrocytes and PBMC at pH 5.0, obtained by plotting the di-8-ANEPPS excitation ratio, R (I₄₅₅/I₅₂₅, normalized to the initial value), as a function of the peptide concentration. Cholesterol and DMSO were tested as control and no effect was observed. Results correspond to the average of three independent replicates.

In order to have a better understanding of what may happen in the bloodstream, we also studied the interaction of these peptides with human blood cells (erythrocytes and PBMC). Cell membranes were labeled with di-8-ANEPPS, as a reporter of peptide-membrane interaction [21,23–26]. Only lipid-tagged peptides induced a large change on the membrane dipole potential assessed in the differential spectra of di-8-ANEPPS at both pH values (Fig. 4). In the case of the Influenza-untagged peptide, no significant perturbation of the membrane dipole was noticed. The affinity for both cell membranes was always higher at a neutral pH; however, the peptides seemed to bind also at pH 5.0, which can be a strong indicator of their potential to target the influenza virus in the endosomes. However, it is important to point out that (Influenza-PEG₄)₂-Chol showed weak affinity to erythrocytes at pH 5.0 in comparison to pH 7.4 (K_d = 0.65 μM) (Table 4).

Table 4.

Apparent dissociation constants (K_d) and R_min values of the influenza peptides for erythrocytes and PBMC, at different pH values, obtained from the fitting of the experimental data with Eq. (4).

Peptides	pH	Erythrocytes		PBMC
		Kd (μM)	Rmin	Kd (μM)	Rmin
Influenza-untagged	7.4	–a	–a	–a	–a
	5.0	–a	–a	–a	–a
Influenza-PEG4-Chol	7.4	0.29 ± 0.05	−0.48 ± 0.02	0.56 ± 0.06	−0.56 ± 0.01
	5.0	2.33 ± 0.45	−0.68 ± 0.05	0.86 ± 0,15	−0.60 ± 0.03
(Influenza-PEG4)2-Chol	7.4	0.65 ± 0.13	−0.62 ± 0.03	0.41 ± 0.13	−0.53 ± 0.03
	5.0	–a	–a	2.23 ± 0.63	−0.65 ± 0.07

^aThe type of interaction spectra obtained impairs the use of Eq. (4).

Influenza-PEG₄-Chol has a higher affinity (lower K_d) to erythrocytes membranes at neutral pH, while (Influenza-PEG₄)₂-Chol showed a higher affinity to PBMC membranes at the same pH. In acidic environment, their affinity to blood cells membranes decrease (increased K_d). Only Influenza-PEG4-Chol presents just a slight decrease on the affinity for the PBMC membrane at pH 5.0.

Mammalian cells, such as these erythrocytes and PBMC, have a considerable amount of cholesterol in their membrane composition [46,47]. The cholesterol-conjugated peptides seem to have a higher affinity for these cell membranes in comparison with a simple model membrane of pure POPC. This reinforces that the peptides may interact preferentially with cholesterol-containing membranes.

It has been suggested that the cholesterol conjugation can be useful against viruses that fuse with the cellular membrane in endosomes, by endowing the peptides with the ability to be transfected along with the virus to intracellular sites of membrane fusion [22]. The conjugated peptides can therefore be included along the virus in the endosome, bound to the outer leaflet of the cell membrane, which later becomes the inner leaflet of the endosome membrane. Their pH-dependent membrane interaction may be an additional advantage: our membrane dipole potential variation data show that the membrane affinity of these molecules decreases upon changing the pH from 7.4 to 5.0. Based on this finding, one may propose that after endocytosis, upon the progressive acidification of the endosome, the conjugated peptides will progressively be released from the membrane, becoming confined on the volume of the endocytic pool. This release of the fusion inhibitor will occur together with the (also) pH-driven acquiring by hemagglutinin of its fusogenic and fusion inhibitor-targetable extended conformation. Therefore, the peptides would be progressively released from the membrane with the ideal timing to be able to bind the complementary heptad repeat domain of HA, impairing 6HB formation, blocking viral fusion and the entry of the viral content into the cytosol and, therefore, preventing the infection of the target cell.

4. Conclusions

In the last few years, the need to overcome resistance has greatly fueled the search for new anti-influenza drugs. The targeting of cellular factors involved in the influenza virus replication has received much attention, as such an antiviral approach could reduce viral drug resistance [8]. Towards this end, we used three different HA-derived fusion inhibitor peptides (or lipopeptides) and studied their biophysical properties, relating those values with antiviral activity data and mechanism of action.

Here, we described biophysical analysis that can be used to guide the development of effective antiviral fusion inhibitory peptides, like peptide specific lipid conjugation, dimerization, membrane binding and self-assembly properties. By adding a lipid moiety, we improved peptide antiviral potency (Table 1 [22]) through lipid affinity and helped to stabilize the peptide, and we suggest that the dimerization can also be used to regulate in vivo biodistribution, since it increases peptides availability at the membrane level, facilitating the interaction with the final target. These results correlate well with our previous studies with fusion inhibitors against other enveloped viruses (HIV, paramyxoviruses and measles virus), where we showed that specific biophysical properties of the fusion inhibitors can modulate their interaction with biomembranes, affecting their antiviral potency [24,25,30,31,48,49]. The most relevant discovery in the present work was the pH-sensitive membrane interaction of the cholesterol-conjugated peptides. This fact is of great importance, as the entry and the fusion processes occurs via an endocytic pathway, and highlights the importance of the cholesterol moiety in the composition of these inhibitors, providing a molecular level explanation for the improved antiviral activity. Developing these fusion inhibitors would set the stage suitable not only for influenza but also for other enveloped viruses with similar entry pathways. At this level, it is worth of notice that one of us has very recently reported the success of a similarly designed fusion inhibitory lipopeptide on the fight against SARS-CoV-2, both in vitro and in vivo [50].