The TurboID biotinylation labelling radius is sufficiently small to allow streptavidin-based imaging of target proteins by light microscopy

A) Schematic representation of the trypanosome NPC (Obado et al., 2016) including selected transport factors and their enrichment observed in TurboID experiments followed by streptavidin affinity capture and LC MS/MS analysis. Proteins quantified are filled in shades of green representative of the corresponding t -test difference increments. The respective bait protein is drawn in pink and undetected proteins are in gray. NUP96, NUP76 and NUP110 experiments were analysed with TurboID-HA fusions at both respective termini (as indicated). The nuclear envelope is drawn in sand and nucleoporins and transport factors are numbered in the legend (right) according to (Obado et al., 2016) (NMD3, Tb927.7.970; RanBP1, Tb927.11.3380; RanBPL, Tb927.10.8650; exportin 1, Tb927.11.14340; GAP, Tb927.10.7680),

B) Schematics illustrating the imaging concept for a protein of interest (POI) by either Cy3- streptavidin or anti-HA coupled with secondary Alexa488-labelled antibodies.

(C-D) Nuclear pore proteins of Trypanosoma brucei (C) and HeLa cells (D) were expressed fused to the streptavidin-HA tandem tag and detected with streptavidin-cy3 (pink) and anti- HA (green). Representative single plane images of a Z-stack is shown, as raw data.

Streptavidin can detect targets within phase-separated regions, while most antibodies fail

(A-B) T. brucei MEX67 (A) and T. brucei nucleolar protein NOG1 (B) were expressed fused to a C-terminal TurboID-HA tandem tag and cells probed with streptavidin-cy3 (pink) and by anti- HA immunofluorescence (green). Representative single plane images of an unprocessed Z- stack series are shown.

(C-E) Cells expressing the stress granule marker protein PABP2 fused to TurboID-HA were starved (2 hours PBS) and starvation stress granules detected by streptavidin (pink, Cy3) and anti-HA (green, Alexa488). The starvation experiment was performed in biological triplicates. One representative image of starved cells is shown as Z-stack projection (72 slices a 140 nm, sum slices) in C. For each replicate, intensity profiles across one of the larger granules of the cell were measured for 25 cells in both fluorescence channels. The profiles for replicate 1 are shown in D. For each granule, the granule diameter was calculated from the profiles at 50% fluorescence and the difference in diameter between the HA- and streptavidin stain is presented in E for each replicate, as quotient of granule diameters. Note that despite differences between the three replicates, likely arisen from starvation conditions being not 100% reproducible, the HA stain consistently delivered a larger granule diameter than the streptavidin stain, consistent with preferentially peripheral staining of the granule by anti- HA. For replicate 2, the fluorophores were switched, with essentially the same result (Figure S2 in Supplementary Materials).

(F) Human NUP54 fused to TurboID-4HA was expressed in HeLa cells and cells were probed with both anti-HA (green, Alexa488) and streptavidin (Cy3, shown in pink). Streptavidin, but not anti-HA detects NUP54 at the nuclear pores. A single plane image of a Z-stack is shown as raw data.

(G and H) T. brucei MEX67 (G) and NOG1 (H) were expressed as TurboID-Ty1 fusion proteins and detected with streptavidin (Cy3, shown in pink) and anti-Ty1 (BB2, green). Representative single plane images of unprocessed Z-stack images are shown.

(I) Trypanosome wild type (WT) cells were probed for MEX67 with polyclonal antiserum (kind gift of Mark Carrington, University of Cambridge; secondary antibody Alexa 488, shown in pink). One representative single plane image of an unprocessed Z-stack image is shown. (J) T. brucei MEX67 (left) and NOG1 (right) were expressed as eGFP fusion proteins and detected with Cy5 labelled eGFP nanobodies. Representative single plane images of unprocessed Z-stack images are shown.

Streptavidin imaging yields higher signal intensities than immunofluorescence

(A-C) Enhanced signal in standard light microscopy

Trypanosome cells expressing NUP158-TurboID-HA were labelled with combinations of either streptavidin-Cy3 and anti-HA/Alexa488 secondary (A) or with streptavidin-Alexa488 and anti-HA//Alexa 594 secondary (B). Z-stack images were recorded (48 slices a 140 nm). Representative, unprocessed single plane images are shown (A and B). The maximum Alexa488 fluorescence was quantified from Z-stack projections (sum slices) from 60 cells probed with anti-HA or streptavidin; the data are presented as a dot blot (waist is median; box is IQR; whiskers are ± 1.5 IQR) (C)

(D) Improved signal in expansion microscopy

Trypanosome cells expressing NUP76-TurboID-HA or MEX67-TurboID-HA were imaged using Pro-expansion or Ultra-expansion microscopy. Single plane and Z-stack projections (sum slices) of the streptavidin and anti-HA signal are shown for one representative nucleus. All images were deconvolved in proExM, except for NUP76 anti-HA.

(E-F) Improved signal in CLEM

Trypanosome cells expressing NUP158-TurboID-HA were embedded in LR-White resin. Slices were probed with streptavidin and anti-HA and imaged by light microscopy (E) followed by electron microscopy (CLEM) (F).

Visualisation of protein interactions with TurboID

(A) Trypanosome cells expressing PABP2-TurboID-HA were probed with streptavidin and anti-HA. A representative image of the posterior part of a cell is shown (single plane of a Z- stack processed by computational clearing). The position of posterior pole granule is indicated by arrows.

(B) Trypanosome cells were transformed to co-express PABP2-mChFP and ALPH1-eYFP. A representative image of the posterior part of a cell is shown (single plane of a Z-stack processed by deconvolution). The position of posterior pole granule is indicated by arrows. (C) Trypanosome cells expressing TurboID-HA-MLP2 were probed with streptavidin and anti- HA. One representative image of an interphase, prophase, metaphase and an anaphase cell are shown. All images are unprocessed single plane images, with the exception of the DAPI image that is a Z-stack projection (max intensity of 48 slices a 140 nm).

(D and E) Trypanosome cells expressing NUP65 (D) or NUP75 (E) fused to a C-terminal TurboID-HA tag were probed with streptavidin. Single plane and Z-stack projection (sum slices of 48 slices a 140 m) of unprocessed images are shown.

A refined map of the T. brucei nuclear pore complex

Each known T. brucei nuclear pore protein was expressed fused to TurboID-HA, each both at the N- and C-terminus. Cells were labelled with anti-HA and with streptavidin.

(A) For each NUP, a representative image of the nucleus is shown as unprocessed, single plane of a Z-stack. We evaluated the extent of anti-HA stain in colour increments (shown as a bar above images).

(B) The HA-signal at the nuclear pores was mapped onto a schematic representation of the trypanosome NPC (modified from (Obado et al., 2016)) using the same colour increments. (C) Scheme of the T. brucei nuclear pore, with known FG NUPs (Obado et al., 2016) shown in black on the left side and prediction of phase separation (Chu et al., 2022) mapped on the right.