*2.3. Physicochemical Characterization*

To image the ligand distribution, fluorescent ligand streptavidin Oregon Green 488 was conjugated to the biotinylated microbubbles as described previously [29]. Briefly, microbubbles were first washed by flotation: 0.9 mL microbubble suspension was placed

in a 3 mL syringe and topped with 2.1 mL saline solution saturated with C4F10. After 45 min, the subnatant was drained, and the microbubbles were resuspended in 0.3 mL saline solution saturated with C4F10. Then, 22.5 µL of streptavidin-Oregon Green 488 (2 mg/mL) was allowed to incubate with 0.7–1.0 <sup>×</sup> <sup>10</sup><sup>8</sup> microbubbles for 30 min on ice. The excess of streptavidin was washed away by flotation as described above, with resuspension of the microbubbles in 0.2 mL saline solution.

To measure the microbubble size distribution and concentration, a Coulter Counter Multisizer 3 (Beckman Coulter, Mijdrecht, The Netherlands) was used. To quantify particles between 1 and 30 µm, a 50 µm aperture tube was used. To evaluate the polydispersity of the samples, the span value was calculated, defined as (*d*90 − *d*10%)/*d*50%, where *d*90, *d*10 and *d*50% are the microbubble diameters below which 90, 10 and 50% of the cumulative number of microbubbles was found. Samples were measured after the first flotation wash and again after conjugation with streptavidin Oregon Green 488.

The streptavidin-conjugated microbubbles were imaged by microscopy as described by Langeveld et al. [16]. In short, the microbubbles were placed between quartz glass in 87% glycerol (*v*/*v* in phosphate-buffered saline) to reduce Brownian motion and imaged with a Leica TCS 4Pi confocal laser-scanning microscope [30]. An axial resolution up to 90 nm was achieved with a matched pair of aligned opposing 100× glycerol HCX PL APO objective lenses (numerical aperture 1.35). For excitation of Oregon Green 488, a 488 nm laser was used, and for excitation of rhodamine-DHPE, a 561 nm laser was used. Images were recorded in 3D as *y*-stacked *xz*-scans in a green (500−550 nm) and red (580−640 nm) spectral channel. The "voltex" function was used to volume-render the image stacks with AMIRA (Version 2020.2, FEI, Mérignac Cedex, France).

Quantitative analysis was performed on the 4Pi microscopy data using customdeveloped image analysis software in MATLAB (Mathworks, Natick, MA, USA), based on the method described by Langeveld et al. [16]. The microbubble coating was subdivided into 32 parts, of which the mean fluorescence pixel intensity (*I*part for the green channel and *I*part-rhod for the red channel) was calculated. The median intensity of all parts (*I*median for the green channel and *I*median-rhod for the red channel) was calculated per microbubble. To evaluate the ligand distribution, parts were classified as inhomogeneous when the absolute difference between *I*part and *I*median was more than two-thirds times the value of *I*median (i.e., |*I*part − *I*median| > 2/3 × *I*median), and the percentage of inhomogeneous parts was calculated per microbubble. To evaluate the lipid phase distribution, parts were classified as LC phase when the value of *I*part-rhod was less than one-third of *I*median-rhod (i.e., *<sup>I</sup>*part-rhod < 1/3 <sup>×</sup> *<sup>I</sup>*median-rhod). The LC phase surface area was first calculated in <sup>µ</sup>m2, and then a percentage of the total analyzed surface area per microbubble. Before evaluating the ligand distribution or the lipid phase distribution, an additional normalization step was included in the image analysis. This step corrected for a difference in fluorescence intensity between the center and the top or bottom of the microbubbles, likely caused by attenuation of the laser light leading to a lower fluorescence signal at the center of the sample. The normalization factor was calculated based on the median *I*part (for the green channel) or the median *I*part-rhod (for the red channel) per angular part from all microbubbles (Supplemental Figure S1). To determine the number of microbubbles with buckles, the microbubble coating was manually scored for fluorescent signal outside and attached to the microbubble coating, based on the red channel (rhodamine-DHPE signal). Only bright spots with 1 µm diameter or larger were classified as a buckle.
