3D Bayesian cluster analysis of super-resolution data reveals LAT recruitment to the T cell synapse

Single-molecule localisation microscopy (SMLM) allows the localisation of fluorophores with a precision of 10–30 nm, revealing the cell’s nanoscale architecture at the molecular level. Recently, SMLM has been extended to 3D, providing a unique insight into cellular machinery. Although cluster analysis techniques have been developed for 2D SMLM data sets, few have been applied to 3D. This lack of quantification tools can be explained by the relative novelty of imaging techniques such as interferometric photo-activated localisation microscopy (iPALM). Also, existing methods that could be extended to 3D SMLM are usually subject to user defined analysis parameters, which remains a major drawback. Here, we present a new open source cluster analysis method for 3D SMLM data, free of user definable parameters, relying on a model-based Bayesian approach which takes full account of the individual localisation precisions in all three dimensions. The accuracy and reliability of the method is validated using simulated data sets. This tool is then deployed on novel experimental data as a proof of concept, illustrating the recruitment of LAT to the T-cell immunological synapse in data acquired by iPALM providing ~10 nm isotropic resolution.


Analytical tool methodology
As explained in the main text, we assign to all localisations a density estimate, L 3D (r), as an indication of the local clustering around each point. A localisation is defined as being a local maximum if it has been assigned the highest L 3D (r) value in a sphere of the same radius r, centred on it. The choice of which specific maxima to consider as being associated with true cluster positions is based on their topographic prominence (TP) value rather than their absolute L 3D (r) values, previously shown to improve performance 25 . We define the topographic prominence of a maximum in the context of pointillist data as the interval between its L 3D (r) value and a base value (the key col value) which allows connectivity to higher maximum. The key col is defined as the lowest value of L 3D (r) for which two maxima are connected. Two maxima are considered connected if a path exits between them using a step size smaller than the mean CSR nearest neighbour distance stepping only between points with L 3D (r) value above the key col value. We next apply a threshold T to the calculated TP values of each local clustering maximum. Only maxima with a TP value above T are seen as identifying clusters, hence by varying the TP threshold, different cluster proposals are generated. Finally, once a maximum above the threshold has been identified, the attribution of localisations to the associated cluster relies only on connectivity to the selected maximum.

Cell culture and transfection:
Jurkat T E6.1 cells were cultured in RPMI supplemented with Glutamax, 10% Foetal Calf Serum (FCS), 100 mg/ml Penicillin and 100 g/ml streptomycin, incubated at 37˚C in a 5% CO 2 incubator. Cells were electroporated with Amaxa Nucleofector Kit V (Lonza, Germany). Plasmid DNA contained LAT fused to mEos3.2 under the control of the CMV promoter. Once transfected, the cells were incubated in medium for 36 hours at 37˚C to allow expression of tagged LAT.

Synapse formation and fixation:
The coverslip was coated with anti-CD3 (2 L/mL) and anti-CD28 (5 L/mL) in Hank's Balanced Salt Solution (HBSS) before being incubated for one hour. The coverslip was then washed three times with HBSS before being used for synapse formation.
The transfected Jurkat T cells were washed in HBSS and spun down twice before being suspended at a concentration of 1 × 10 6 cells per mL for imaging. The cells were allowed to settle on the coverslip (t=0 s for the activation time course). The coverslip was placed back in the incubator for 4 or 8 minutes depending on the condition. For the control case, the same steps as the 8 minute case were followed, with uncoated coverslips. Cells were then fixed at room temperature using a 2-step protocol. First, cells were fixed using 3% paraformaldehyde (PFA) in 80 mM kPIPES at a pH of 6.8 for 5 minutes then washed twice with HBSS. The coverslip was then placed in 3% PFA in 100 mM NaB 4 O 2 at a pH of 11 for 10 minutes and washed three times with HBSS before imaging. Finally, samples were mounted between a lower 25 mm diameter #1.5 coverslip (also containing the fiducial markers) and an upper 18 mm diameter #1.5 coverslip 10 . Coverslips were bonded together using epoxy (3 M) and sealed to prevent buffer evaporation using Vaseline (Unilever).

Imaging:
The imaging was performed on an iPALM microscope at the Advanced Imaging Center at the Howard Hughes Medical Institute's Janelia Research Campus 10 . 30,000 frames were acquired, using an integration time of 50 ms each. Two lasers were used during the acquisition: a pulsed 405 nm laser, delivering 5-30 W/cm 2 to photoconvert the fluorescent protein, as well a continuous wave (CW), 561 nm laser at 1 kW/cm 2 for fluorescence readout. The 405 nm laser pulse width was continuously increased during acquisition to maintain an approximately constant number of visible molecules in each frame. Fluorescence was collected in the range 570 -615 nm (FF01-593/40-25, Semrock).

Image processing:
The raw data from the acquisition was processed by the Peak Selector software package 5 . Lists of localisations were extracted from the raw data, which was, in turn, filtered by uncertainty (to less than 30 nm in all dimensions). Drift correction in x,y and z and tilt correction was performed using fiducials. Finally, the data was cropped into 2 x 2 x 0.6 m ROIs.

Statistics:
P values were calculated using an unpaired, two-tailed, Mann-Whitney U-test in Prism software.

Code availability:
All code related to this analysis is be available at: GitHub public "OwenlabKCL" page.