Membrane Environment Sets the Functional pKa of Ionizable Lipids

Abstract

Ionizable aminolipids enable lipid nanoparticles (LNPs) to encapsulate nucleic acids at neutral pH and to release their cargo upon endosomal acidification. The discrepancy between this effective, acidic LNP pK_a and the basic intrinsic pK_a of aminolipids, however, remains poorly understood.

Here, we performed microsecond constant-pH molecular dynamics simulations of five widely used aminolipids (DODAP, DLin-MC3-DMA, DLin-KC2-DMA, ALC-0315, and SM-102) embedded in different LNP-relevant ternary DOPC/D-SPC–cholesterol membranes to quantify how aminolipid structure and membrane composition jointly govern aminolipid protonation and the associated pH-dependent membrane remodeling.

Across all systems, membrane embedding lowers the apparent aminolipid pK_a, yielding physiologically relevant values of 6–7.5 corresponding to shifts by up to 3.5 pK_a units or approx. 20 kJ mol⁻¹ with respect to the intrinsic pK_a. Strikingly, the magnitude of the pK_a shift correlates with pH-driven membrane remodeling upon deprotonation: polyunsaturated aminolipids undergo surface-to-core translocation, branched aminolipids preferentially form laterally segregated surface domains, and DODAP remains interfacially anchored through sustained hydration and hydrogen bonding. Saturated helper lipids (DSPC) systematically enhance segregation and amplify pK_a shifts relative to DOPC.

Together, these results identify membrane phase behavior as a primary regulator of aminolipid protonation equilibria and establish quantitative design principles for tuning LNP composition toward desired pK_a, membrane remodeling, and delivery performance.

Introduction

The delivery of nucleic acids using lipid nanoparticle (LNP) formulations has opened new avenues for the development of innovative therapies and vaccines targeting a wide spectrum of diseases [1]. Structurally, LNPs consist of a hydrophobic core surrounded by a lipid monolayer [1, 2]. The negatively charged nucleic acids can be efficiently encapsulated and stabilized within LNPs through the incorporation of cationic lipids [3].

LNPstypically comprise four different lipid components: Aminolipids, helper lipids, cholesterol, and PEG-lipids, each fulfilling a distinct functional role. These components are subject to diverse optimization strategies, resulting in a broad and highly tunable LNP compositional space. Aminolipids are the predominant component (usually 40 50%) of LNPs [4] and their pH-responsiveness is central to loading or encapsulation efficiency, their bio distribution, safety, and intracellular payload delivery [5]. Cholesterol is the second most abundant component (≈ 40% in currently reported LNP formulations) and provides support to surface-core organization of LNPs, con ferring fluidity while keeping a tight lipid packing preventing payload escape. The substitution of cholesterol with alternative sterols may affect delivery efficiency and favor different tissue tropism, although the precise mechanism is not clear [6]. Notably, chimeric molecules joining the two major LNP components (aminolipid and sterol) achieved satisfactory safety and efficacy profiles both in-vitro and in-vivo [7] highlighting the untapped potential for exploring novel LNP components. Helper lipids are usually zwitterionic and bilayer-forming phospholipids. While DSPC is included in all three FDA-approved LNP compositions, sphingomyelin (SM) and ceramides have been proposed as alternative helper lipids in several patents [8, 9]. Importantly, SM lipids have been shown to confer peculiar bio-distribution proper ties [10], highlighting helper lipid type and molar ratio as critical parameters for further LNP optimization.

PEG-lipids are typically included at low molar fractions (1-2%) to prevent LNP aggregation and prolong their half life in vivo [11]. The lengths of the PEG chain and of the hydrophobic anchor govern PEG desorption kinetics, thereby influencing LNP biodistribution [12]: Shorter hydrophobic anchors favor faster desorption and early formation of the protein corona promoting liver tropism, while bulkier anchors prolong LNP circulation time and tumor accumulation [13] but may cause anti-PEG immunogenicity [14]. Recent developments in the design of novel LNP formulations have removed PEG-lipids entirely from the composition [15]. Early aminolipids were based on quaternary amine headgroups, rendering them permanently cationic. This inherent cationicity facilitates strong electrostatic interactions with nucleic acids, yielding overall positively charged LNP formulations. While such formulations are generally considered beneficial for cellular uptake by promoting interactions with the negatively charged cell surface and glycocalyx [3, 16], permanently cationic lipids have been shown to induce hepatotoxicity, lung toxicity, and result in rapid clearance of LNPs from the bloodstream due to pronounced first pass effects [17].

These limitations were largely overcome by the development of cationic ionizable aminolipids (CILs), which contain secondary or tertiary amine groups capable of undergoing dynamic (de)protonation [3]. CILs are characterized by the following pH-dependent behavior:

At acidic pH (≈ 4), CILs are protonated, enabling efficient nucleic acid encapsulation [3].
Atphysiological pH (7.4), CILs adopt a deprotonated, neutral state, resulting in the formation of LNPs with a lipidic core composed of deprotonated aminolipids and cholesterol, surrounded by a lipid monolayer. In mRNA-loaded LNPs, the negative charge of the nucleic acids is shielded by locally protonated aminolipids [2, 18].
Upon endosomal acidification (≈ 6), CILs re-protonate at the LNP surface [18], enhancing electrostatic interactions with anionic endosomal lipids [19]. In addition, the bilayer-destabilizing properties of CILs likely promote membrane fusion and thereby nucleic acid release [1].

Accordingly, the protonation behavior of aminolipids is central to LNP function. It is determined by the intrinsic pKa of the aminolipid defined as the pKa of the functional ionizable group in infinite dilution conditions (pKALa). Experimentally, pKALa values are typically determined by potentiometric titration of water-soluble aminolipid analogs and can be tuned by modifying the distance between the ionizable nitrogen and the electron withdrawing groups (Fig. 1) [5].

In contrast, the apparent pKa of full LNPs (pKLNP a sic pKAL) depends not only on the intrin a , but additionally on the local LNP composition, structure, and molecular organization, and on the particular method employed to measure the pKLNP a reported intrinsic pKAL. While a values range from 7 to 10 (Fig. 1c), effective LNP formulations consistently exhibit pKLNP values between 6 and 7, i.e., consistently slightly below physiological pH.

A widely used approach to determine pKLNPa is the TNS binding assay [20], which monitors fluorescence changes upon TNS partitioning into LNP membranes. Because TNS binding increases with membrane positive charge– and thus aminolipid protonation– this method provides an indirect measure of membrane charge. We previously demonstrated that TNS localizes predominantly within the LNP monolayershell; consequently, pKLNP a. values obtained from TNS assays primarily reflect amino lipid protonation in the LNP shell rather than in the core [18]. Accordingly, TNS assays do not allow to sense the protonation states of aminolipids residing in the LNP core [5, 21]. This limitation has recently been addressed using DNA-based fluorescent probes, which revealed that the LNP core is highly permeable to protons [22].

Download the full article as PDF here Membrane Environment Sets the Functional pKa of Ionizable Lipids

or continue reading here

Membrane Environment Sets the Functional pK_a of Ionizable Lipids, Marius F.W. Trollmann, Paolo Rossetti, Rainer A. Böckmann, bioRxiv 2026.02.18.706567; doi: https://doi.org/10.64898/2026.02.18.706567