**1. Introduction**

Influenza virus entry to the cells requires fusion of viral and cellular membranes. It is mediated by a homotrimeric viral protein hemagglutinin (HA), whose HA1 subunits bind to cellular receptors and HA2 subunits are responsible for membrane remodeling [1,2]. A critical, ye<sup>t</sup> still not exactly understood role in this latter respect is played by amphipatic N-terminal HA2 fragments, so-called fusion peptides (HAfp), which are directly inserted into target lipid bilayer and initiate the fusion process [3]. Their high sequence conservation and the existence of nonfunctional mutants, which nevertheless were demonstrated to embed within lipid bilayers, indicate that their role is not merely limited to act as inert anchors, but must also involve specific interplay with bilayer structure [4,5]. Synthetic peptides corresponding to only the first 20 N-terminal amino acids of HA2 have long been known to initiate the fusion of lipid vesicles and content mixing in vitro [6,7], without the mechanical action provided by the rest of the protein. Since the effect of fusion-abolishing mutations within HAfp is generally reproduced by the loss of function in complete fusion proteins [5], they have been extensively used as model systems to study basic aspects of viral fusion [8].

Experimental data concerning synthetic HAfp indicate that, whereas disordered in aqueous solution, they adopt mostly helical forms once bound to lipid structures such as membranes or micelles [9]. The N-terminal, strictly conserved part (residues 1–11) folds

**Citation:** Worch, R.; Dudek, A.; Borkowska, P.; Setny, P. Transient Excursions to Membrane Core as Determinants of Influenza Virus Fusion Peptide Activity. *Int. J. Mol. Sci.* **2021**, *22*, 5301. https:// doi.org/10.3390/ijms22105301

Academic Editor: Masoud Jelokhani-Niaraki

Received: 15 April 2021 Accepted: 13 May 2021 Published: 18 May 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

into a stable *α*-helix, followed by a kink region (residues 12–14) and a second, malleable C-terminal *α*-helix that starts from residue 15. Early studies, focussed on possibly shortest fusiogenic peptides composed of only 20 residues (HAfp1−20), suggested that both helices form a boomerang like structure, with the N-terminal arm partially inserted into hydrophobic lipid core, followed by a solvent-exposed kink, and the C-terminal segmen<sup>t</sup> located at lipid–water interface [10]. This view has been challenged by subsequent, nuclear magnetic resonance (NMR)-based investigation of a longer and more active 23 amino acid peptide (HAfp1−23), which was shown to adopt a tight helical hairpin structure in micelles, supposedly remaining at the lipid–water interface [11]. The hairpin structure is apparently not affected by further sequence elongation. According to NMR data, residues 24–28 form solvent exposed random coil [12], which suggests that they belong to a flexible tether that links the fusion peptide to the rigid coiled coil stem in a fusion ready, complete HA [13].

A sharply bent hairpin seems to be the actual fusion-active form of HAfp. Indeed, this explains well a strict conservation of several glycine residues aligned in a ridge that enables tight packing of helical arms as well as the fact that many amino acid substitutions that are known to abrogate fusion activity [8] are sterically inconsistent with the hairpin structure [11]. The activity of boomerang-like HAfp1−<sup>20</sup> has been attributed to the fact that it actually exists in an ensemble of states with residual ( ∼11%) closed hairpin population [12], thus also explaining its lower activity compared to HAfp1−23. Intriguingly, however, fusion-inactive W14A mutant, with complete helices and no steric hindrance that would preclude the hairpin formation, appears to exist as an overly flexible boomerang structure [14]. It indicates that hairpin stabilization may depend not only on intra-peptide contacts but also on its interaction with environment, since the W14 side chain, apparently needed to maintain the hairpin conformation, is exposed to solvent and does not participate in interhelical interactions.

Unambiguous interpretation of the above observations is hampered by the lack of clear evidence concerning the actual peptide placement within lipid bilayer and its impact on membrane structure. The length of the N-terminal helix (11 residues) is much shorter compared to typical transmembrane helical protein domains (around 20 residues [15]). Given this notion, several NMR-based studies conclude that fusion peptides orient parallel to membrane surface and stay at the lipid–water interface [11,16]. On the other hand, measurements based on attenuated total reflection Fourier-transform infrared spectroscopy (ATR-FTIR) or spin-label electron paramagnetic resonance (EPR) sugges<sup>t</sup> that peptides insert obliquely into the external membrane leaflet with the N-helix tilted on average ∼50 degrees with respect to bilayer normal [9,10,17,18], which results in buried N-terminal group and solvent-facing kink region. Moreover, recent molecular dynamics (MD) simulations indicated the possibility of fully transmembrane peptide placement which can be achieved owing to local membrane thinning caused by aqueous solvent attraction towards hydrophilic residues on both hairpin poles [19–22].

Such mixed results regarding HAfp structure and location within membrane entail the lack of consensus concerning its actual fusogenic mechanism. In this respect, multiple modes of action have been proposed, such as the promotion of lipid tails protrusions [23], lipid heads intrusions [24], membrane thinning [25,26], stabilization of positive membrane curvature [27], stabilization of negative membrane curvature [16], local membrane dehydration [28], or membrane rupture by transmembrane HAfp bundle with subsequent formation of *π* shaped intermediates [29]. Of the above, the likelihood of lipid tails protrusions [30] has been demonstrated via MD simulations as a viable predictor of membrane fusogenic propensity, and the extent to which it was increased by various peptide mutants was shown to correlate with their experimental activity [23,24]. In a recent simulation study of complete HA-mediated fusion [31], a particularly important role in inducing lipid tails protrusions was demonstrated for peptide configurations in which the charged HAfp N-terminus was buried within membrane core in agreemen<sup>t</sup> with oblique insertion mode suggested by spectroscopic measurements.

In this study, to thoroughly characterize HAfp1−<sup>23</sup> configuration landscape within POPC membranes and to assess the determinants of its fusogenic activity, we combine MD simulations with spectroscopic methods. To this end, we consider wild type (wt) peptide and two mutants, E11A and W14A, that are known to have reduced or none activity, respectively [14]. We demonstrate that the peptides are rather dynamic in membrane environment. While they preferentially remain on bilayer surface, as previously indicated, it seems that the key to explain their function is to assume transient excursions into the membrane core. According to our analysis, in contrast to generally similar characteristics of surface conformations among the considered mutants, the ability to adopt deep configuration is affected by amino acid substitutions and significantly contributes to explaining the observed, mutation-related differences in activity.

#### **2. Materials and Methods**
