New Insights into the Molecular Structure of Tear Film Lipids Revealed by Surface X-ray Scattering

The tear film lipid layer (TFLL) is a unique biological membrane that serves a pivotal role in the maintenance of ocular surface health. Reaching an overarching understanding of the functional principle of the TFLL has been hampered by a lack of insights into the structural and functional roles played by individual lipid classes. To bridge this knowledge gap, we herein focus on studying films formed by principal lipid classes by surface scattering methods. Through grazing incidence X-ray diffraction and X-ray reflectivity studies, we reveal quantitative data about the lattice distances, molecular tilt angles, and mono/multilayer thickness and density profiles for central TFLL lipid classes under close to simulated physiological conditions. In addition, we discuss the correlation of the results to those obtained previously with the natural lipid composition of meibum.


Grazing incidence x-ray diffraction
Instrumentation: GIXD data were collected at the ID10 beamline of the European Synchrotron Radiation Facility (ESRF) in Grenoble, France.The incident monochromatic beam energy was 22.0 keV.The beam incidence angle was 0.045 degrees (approx.0.79 mrad).A Dectris Mythen 2K detector was used with a resolution of 200 pixels per degree.The beam size was 10 µm in the vertical dimension and 34 µm in the horizontal dimension.Attenuators were employed as sample damage was observed when exposed to the full beam flux.Oxidation was minimized by continuously pumping helium into the Langmuir trough environment during measurements.GIXD data were also collected at the SIRIUS beamline of the SOLEIL Synchrotron in Paris, France.The incident monochromatic beam energy was 8.0 keV.The beam incidence angle was 2.03 mrad.The beam size was 110 µm in the vertical dimension and 2 mm in the horizontal dimension.Attenuators were employed as sample damage was observed when exposed to the full beam flux.Oxidation was minimized by continuously pumping helium into the Langmuir trough environment during measurements.A 2D Pilatus3 1M detector (Dectris, Switzerland) was used associated with a Soller collimator leading to a resolution of about 0.07nm -1 .

Langmuir trough
ID10 is equipped with a Langmuir trough measuring 170 mm x 438 mm and with a depth of 3 mm.A single moveable barrier controls surface pressure which is measured using a Wilhelmy balance (Surface Pressure Sensor Model PS4, Nima Technology Ltd).When in use, the trough is sealed with a cover and filled with helium.Temperature is control by a thermostated bath and an active vibration isolation system (MOD-2 S, Halcyonics) was used.Standard PBS buffer solution was used as the substrate.
SIRIUS is equipped with a homemade Langmuir trough of 700 cm 2 , and the pressure is measured using a Wilhelmy balance (Riegler and Kirstein GmbH).The mean molecular area is controlled by a single moveable barrier with a maximum compression ratio of 3.8.

Data analysis:
Observed features in the GIXD scans were fitted with Gaussian functions and the peak centroids were used to calculate the coherence length, tilt angle and lattice parameters (a,b) of each of our samples.The lattice type (i.e., NN or NNN) was identified by careful inference through consideration of the position of observed features.
The Scherrer equation can be used to calculate the coherence length, that is related to the average size of the crystalline areas in the film: For a sample with a rectangular lattice, parameters a and b can be calculated with the following equations: where qx,y 20 is the peak centroid of the non-degenerate peak and qx,y 11 is the peak centroid of the degenerate peak in the qx,y plane.
Tilt angle can be calculated in the NNN case with: where n and d refer to the non-degenerate and degenerate peaks, respectively; and in the NN case with:

X-ray reflectivity
Instrumentation: XRR measurements were conducted at beamline ID10 at ESRF.The sample environment was identical as described in the GIXD experimental section.A Maxipix detector (2560 x 256 pixels) was used with a distance of 891.7mm between sample and detector leading to a resolution of 1 degree per 283 pixels.

XRR data analysis:
The XRR data were analysed using the Refl1d program 2 .The lipid film structure was modelled using a slab model, i.e., the lipid heads and tails were assumed to form layers with different electron densities.The scattering length density (SLD) profiles were modelled using three parameters per each layer implemented in the model: layer thickness, electron density and roughness, which described the roughness of the interface between the layers.

Supporting results
Supporting Table 1.variations in fitted model parameters as a function of temperature and pressure.Tilt angles were not able to be calculated for some samples as the Gaussian peak was beyond the measurement region.