Insulin secretory granules labelled with phogrin-fluorescent proteins show alterations in size, mobility and responsiveness to glucose stimulation in living β-cells

The intracellular life of insulin secretory granules (ISGs) from biogenesis to secretion depends on their structural (e.g. size) and dynamic (e.g. diffusivity, mode of motion) properties. Thus, it would be useful to have rapid and robust measurements of such parameters in living β-cells. To provide such measurements, we have developed a fast spatiotemporal fluctuation spectroscopy. We calculate an imaging-derived Mean Squared Displacement (iMSD), which simultaneously provides the size, average diffusivity, and anomalous coefficient of ISGs, without the need to extract individual trajectories. Clustering of structural and dynamic quantities in a multidimensional parametric space defines the ISGs’ properties for different conditions. First, we create a reference using INS-1E cells expressing proinsulin fused to a fluorescent protein (FP) under basal culture conditions and validate our analysis by testing well-established stimuli, such as glucose intake, cytoskeleton disruption, or cholesterol overload. After, we investigate the effect of FP-tagged ISG protein markers on the structural and dynamic properties of the granule. While iMSD analysis produces similar results for most of the lumenal markers, the transmembrane marker phogrin-FP shows a clearly altered result. Phogrin overexpression induces a substantial granule enlargement and higher mobility, together with a partial de-polymerization of the actin cytoskeleton, and reduced cell responsiveness to glucose stimulation. Our data suggest a more careful interpretation of many previous ISG-based reports in living β-cells. The presented data pave the way to high-throughput cell-based screening of ISG structure and dynamics under various physiological and pathological conditions.

Where and represent the waist of the PSF along radial and axial direction, respectively. Fig. 1 shows (left panel) the section of the PSF over the plane = 0 and (right panel) the profile ( , 0,0). The radial waist is indicated as a red segment and can be regarded as the "size" of the PSF on the focal plane.
Intercept of the iMSD curve g(ξ, η, ζ, τ) is defined as the spatiotemporal image correlation, as function of the spatial lag-variable , , and the temporal lag-variable . We can define the position vector ⃗ ≡ ( , , ) andfor point-like particles-we can express ( ⃗, ) as [(1), Eq. 2.5]: (2) Where ( ⃗, ) is the probability density function of particle displacement, it depends on the investigated dynamics and determines the time evolution of g. ( ⃗) is the PSF and its Gaussian approximation is given in Eq. 1.
Under this assumption, ( ⃗, ) regulates the trend of 2 ( ), whose boundary value 2 (0) is fixed by ( ⃗) through the radial waist of the PSF. In other words, Eq. 3 can be evaluated by combining Eq. 2 and Eq. 1 and by considering that where is the Dirac delta function. Hence, the intercept of the iMSD curve is 2 (0) = 2 .
Contribution of particle size.
When the particle under study has more than one fluorophore and its size is not negligible, g(ξ, η, τ) must include also the spatial extension of the particle itself. If we define ( ⃗) as the spatial distribution of the fluorophores on the particle surface, Where And is the PSF. For 2D Motion, we can focus on a radial profile, then generalize the solution over the ( , )-plane. In this regard, by assuming a uniform distribution of the fluorophores, ( , 0) can be written as: Where is the particle radius and is a constant normalization factor. By combining Eq. 1, 5 and 6, we obtain: Where the integral arises from the convolution of the involved functions and its analytic solution involves the error function. Eq. 7 can be approximated by the Gaussian function: In this regard, Fig. 2 shows the analytic solution of 1 (black lines) and the corresponding Gaussian fitting curve (blue line) for different R values. Under this approximation, the waist 1 is 1 2 = 2 + 2 2 Finally, by combining Eq. 1, 8, 9 and evaluating at zero time-lag, follows that the intercept of the iMSD curve reads 2 (0) = 1 2 , thus the particle radius can be estimated as: custom made Matlab algorhythm (as described in Methods). F) iMSD curves obtained for movie in A). Fitting procedures lead to reported values, in agreement to tracking-based retrieved ones. G) Zoomed region of experimental iMSD with relative fitting curve. The square root of the y-axis intercept retrieved by fitting is used to estimate the average size (diameter) of imaged ISGs. H) size (diameter) distribution of imaged ISGs. Values were calculated from the first frame of the movie for which iMSD curve in (G) was reported. I) Example of size calculation for imaged ISGs. The intensity profile along the yellow line was fitted with a Gaussian function to estimate of spot size.