Surfactant Proteins SP-B and SP-C in Pulmonary Surfactant Monolayers: Physical Properties Controlled by Specific Protein–Lipid Interactions

The lining of the alveoli is covered by pulmonary surfactant, a complex mixture of surface-active lipids and proteins that enables efficient gas exchange between inhaled air and the circulation. Despite decades of advancements in the study of the pulmonary surfactant, the molecular scale behavior of the surfactant and the inherent role of the number of different lipids and proteins in surfactant behavior are not fully understood. The most important proteins in this complex system are the surfactant proteins SP-B and SP-C. Given this, in this work we performed nonequilibrium all-atom molecular dynamics simulations to study the interplay of SP-B and SP-C with multicomponent lipid monolayers mimicking the pulmonary surfactant in composition. The simulations were complemented by z-scan fluorescence correlation spectroscopy and atomic force microscopy measurements. Our state-of-the-art simulation model reproduces experimental pressure–area isotherms and lateral diffusion coefficients. In agreement with previous research, the inclusion of either SP-B and SP-C increases surface pressure, and our simulations provide a molecular scale explanation for this effect: The proteins display preferential lipid interactions with phosphatidylglycerol, they reside predominantly in the lipid acyl chain region, and they partition into the liquid expanded phase or even induce it in an otherwise packed monolayer. The latter effect is also visible in our atomic force microscopy images. The research done contributes to a better understanding of the roles of specific lipids and proteins in surfactant function, thus helping to develop better synthetic products for surfactant replacement therapy used in the treatment of many fatal lung-related injuries and diseases.


SP-C Tilt Angle
The tilt angle of SP-C was calculated as the the angle between the z axis (normal to the monolayer plane) and a vector joining residues Lys11 and Leu32 C α atoms using the gmx gangle tool provided with GROMACS. 1 The tilt angles were averaged over the four repeats and over the length of the simulations, and the error estimate is given as the standard error.
Lipid Acyl Chain Tilt Angle The phospholipid acyl chain tilt angle was calculated as the angle between the z axis (normal to the monolayer plane) and a vector joining the 1st and 16th carbon (C21 and C216) in the sn-1 fatty acid chain using the GROMACS tool gmx gangle. The lipid acyl chain tilt angles were averaged per lipid type over the four repeats for the SPB and SPC systems and over the two repeats for the NoP systems. The distance of a lipid from the protein was calculated with the GROMACS tool gmx mindist every 1 ns as the minimum distance of a lipid phosphorus atom from the protein.
Surface Pressure-Area Isotherm Monolayer surface pressure Π as a function of APL A was calculated as Π(A) = γ 0 − γ(A), where γ and γ 0 are the surface tensions of the surfactant-covered and surfactant-free air-water interfaces, respectively. The surface tension γ was calculated as γ = L z × (P N − P L ) /2, where L z is the length of the simulation box z axis normal to the interface, P N = P zz and P L = (P xx + P yy ) /2 are the normal and planar components of the pressure, respectively, and the factor 1/2 accounts for the two interfaces in each simulation system. The values were obtained with the GROMACS tool gmx energy.
The area of the monolayers was extracted from the size of the simulation box as L x × L y , where L x and L y are the lengths of the simulation box x and y axes, respectively. The trajectories were analyzed in 100 ns sections and the values were calculated as an average over the section. The results were then averaged over the repetitions and the error estimate was calculated as standard error.

Lipid Diffusion Coefficients
The lateral diffusion of lipids was quantified using displacement distributions. To this end, we extracted the displacements of lipid centers of mass over 10 ns intervals. The values extracted from different surface pressure regimes (0-15, 15-30, 30-45, and >45 mN/m) were histogrammed. For simplicity, displacements across periodic boundaries were discarded, yet this has no effect on the shape of the distribution. The distributions were normalized and fitted by 2 where r is the length of the displacement, ∆ the time interval, and D the diffusion coefficient.
We chose ∆=10 ns to avoid probing anomalous diffusion at short timescales, 3 while also probing the desired surface pressure interval; Using larger ∆ value would mean that the displacements start and end at very different surface pressures in the dynamic simulation.
The diffusion coefficients were measured for two (NoP system, two monolayers) or four (SPB and SPC systems, two replicas with two monolayers) monolayers and averaged. Standard error of these independent samples was used as the error estimate. The SP-PG contacts were mapped onto the protein structure (BETA field in a PDB file), and rendered using VMD to visualize the spatial distribution of the favourable contacts.