The spatial protein organization, topography and physical mechanics of native carboxysomes, the self-assembling carbon-fixation machinery in cyanobacteria, were characterized.

Nanomechanical measurements of cells and single molecules with atomic force microscopy (AFM) require accurate calibration of two parameters: the spring constant of the cantilever (k) and the inverse of the optical lever sensitivity (InvOLS). The most established calibration approach in liquid involves determining the InvOLS by acquiring force-distance curves on a stiff surface, k is then calculated using the thermal spectrum (PSD) of the cantilever via the equipartition theorem. Recent studies have proposed using cantilevers with calibrated k and then determining the InvOLS from the thermal spectrum. These non-contact approaches improve the precision of nanomechanical measurements compared to conventional contact-based approaches. The Sader method or the recent global calibration initiative (GCI) are accurate approaches and do not require knowledge of the InvOLS to determine k, thus they would allow one-step calibration of AFM in liquid. However, both methods assume high quality factor cantilevers, not the case for most cantilevers in liquid. Here we assess the accuracy and precision of the Sader and GCI methods in liquid on two types of cantilevers with low Q-factor using two different PSD fitting models (SHO and Pirzer). We evaluate the two approaches using only the thermal spectrum in liquid to calibrate both k and the InvOLS. While both methods led to similar results, the GCI approach is less prone to systematic uncertainties and, using the SHO model, provides higher accuracy in k and the InvOLS. Therefore, the proposed SHO, GCI-based approach utilizing only the thermal spectrum in liquid is precise and accurate and allows one-step calibration of AFM.

The volume of a protein can be estimated from its molecular weight. This approach has also been applied in force microscopy experiments. Two factors contribute to the determination of the volume from a force microscope image, the applied force and the tip radius. Those factors act in opposite directions. Here, we demonstrate that in the optimum conditions to image a protein, the apparent volume deduced from an AFM image overestimates the real protein volume. The lateral broadening due to the tip finite size, makes the simulated volume to exceed the real protein volume value, while the force applied by the tip tends to decrease the measured volume. The measured volume could coincide with the real volume for either a point-size tip at zero force or when the compression exerted by the tip compensates its dilation effects. The interplay between the above factors make unsuitable to apply the molecular weight method to determine the volume of a protein from AFM data.

The use of atomic force microscopy on nanomechanical measurements requires accurate calibration of 7 the cantilever's spring constant (k c ) and the optical lever sensitivity (OLS). The thermal method, based 8 on the cantilever's thermal fluctuations in fluid, allows estimating k c in a fast, non-invasive mode. How-9