Recharging your fats: CHARMM36 parameters for neutral lipids triacylglycerol and diacylglycerol

Abstract

Neutral lipids (NLs) are an abundant class of cellular lipids. They are characterized by the total lack of charged chemical groups in their structure, and, as a consequence, they play a major role in intracellular lipid storage. NLs that carry a glycerol backbone, such as triacylglycerols (TGs) and diacylglycerols (DGs), are also involved in the biosynthetic pathway of cellular phospholipids, and they have recently been the subject of numerous structural investigations by means of atomistic molecular dynamics simulations. However, conflicting results on the physicochemical behavior of NLs were observed depending on the nature of the atomistic force field used. Here, we show that current phospholipid-derived CHARMM36 parameters for DGs and TGs cannot adequately reproduce interfacial properties of these NLs because of excessive hydrophilicity at the glycerol-ester region. By following a CHARMM36-consistent parameterization strategy, we develop improved parameters for both TGs and DGs that are compatible with both cutoff-based and particle mesh Ewald schemes for the treatment of Lennard-Jones interactions. We show that our improved parameters can reproduce interfacial properties of NLs and their behavior in more complex lipid assemblies. We discuss the implications of our findings in the context of intracellular lipid storage and NLs’ cellular activity.

Molecular Condensate in a Membrane: A Tugging Game between Hydrophobicity and Polarity with Its Biological Significance

“…Campomanes et al found that the CHARMM36 parameters for TAG and DAG molecules were too hydrophilic in the glycerol−ester region and reparameterized the charges. 35 This agrees with our finding that, with the current CHARMM36, TAG starts to aggregate at a much higher concentration than the ones reported in experiments. Therefore, we note that the nonbonded parameters can impact the simulation results, and a careful treatment of the force field parameters will further pave the road for development in this field.…”

“…Their MK9 parametrization has more polar carbonyl groups and a smaller van der Waals radius for hydrogen atoms than the CHARMM36 FF. Campomanes et al found that the CHARMM36 parameters for TAG and DAG molecules were too hydrophilic in the glycerol–ester region and reparameterized the charges . This agrees with our finding that, with the current CHARMM36, TAG starts to aggregate at a much higher concentration than the ones reported in experiments.…”

“…The CHARMM-GUI website has extensively modeled important biological membrane systems and provides 18 models of these systems (). Currently UQ, MK, TAG, and DAG lipids are all available in CHARMM-GUI Membrane Builder. − We plan to provide the modified DAG/TAG parameters based on Campomanes et al…”

Capturing the liquid-crystalline phase transformation: Implications for protein targeting to sterol ester-rich lipid droplets

“…14 It is important to note that these magnitudes of hydration are likely overestimated since the standard CHARMM36 parameters used in our model for TG are not ideal for bulk hydration properties and interfacial (water/lipid) surface tension. 33,34 The central limitation in these, and all parameters developed to date, is lack of polarizability, which is essential to capture proper neutral lipid interactions in both nonpolar core and polar interfacial environments. For example, bulk TG is hydrophobic, but at a water interface the glycerol moieties should have larger partial charges in response to proximal water.…”

“…10,36,37 Updates to the additive forcefields of TG that correct bulk hydration and interfacial water/lipid surface tension do not correctly represent the APL of a LD monolayer, likely due to increased hydrophobicity that retains TG in the LD core. 38,39…”