Development of the SPEx protocol

(A) SPEx workflow. Cultured HeLa cells are fixed, and immunofluorescence is used to label the subcellular compartment of interest with fluorescent antibodies. Next, the cells are embedded in a hydrogel. To allow for later expansion, proteins are denatured overnight at 95 °C. NHS staining is performed to non-specifically label proteins before expansion of the gel in H₂O. Expanded gels are dried onto PPS membrane slides, and laser microdissection of the areas of interest is performed. LC–MS of the collected areas and subsequent data analysis are used to identify proteins enriched in specific subcellular compartments. Created with BioRender.com. (B) Cells stained for total protein content with NHS 488 and immunostained with a GM130 antibody, with and without expansion using a modified TREx protocol. Scale bar: 100 µm. (C) Cells stained for total protein content with NHS 594, with and without expansion using a modified TREx protocol. Scale bar: 100 µm. (D) Measured nuclear diameters for expanded and non-expanded cells. Data are plotted as mean ± SD (non-expanded: 20.53 ± 1.94 µm; expanded (hydrated): 179.70 ± 17.26 µm) (n = 10 cells; N = 2 independent experiments indicated by different colors). (E) Average expansion factor calculated based on the measurements in (D). Data are plotted as mean ± SD (expanded (hydrated): 8.934 ± 0.40) (N = 2 independent experiments indicated by different colors).

Optimization of the SPEx protocol for laser microdissection

(A) Cells stained for total protein with NHS 488 and immunostained for GM130, with and without expansion using a modified TREx protocol. Expanded samples were dried onto PPS membrane slides prior to imaging. Scale bar, 100 µm. (B) Cells stained for total protein with NHS 594, with and without expansion using a modified TREx protocol. Expanded samples were dried onto PPS membrane slides prior to imaging. Scale bar, 100 µm. (C) Nuclear diameters in expanded (dried) and non-expanded cells. Data are mean ± s.d. (non-expanded: 20.53 ± 1.94 µm; expanded (dried): 153.30 ± 21.16 µm; n = 10 cells, N = 2 independent experiments indicated by different colors). (D) Expansion factor calculated from measurements in (C). Data are mean ± s.d. (7.61 ± 0.42; N = 2 independent experiments indicated by different colors). (E) Pre– and post-cut images of all isolated sample types. Dissected cells were stained, expanded and dried according to the SPEx protocol. Images are not to scale (magnification of the objective used for dissection is indicated). Scale bar, 100 µm. (F) UpSet plot showing shared and unique proteins identified across the different regions of interest (ROIs) shown in (E). (G) Principal component analysis (PCA) of proteomics samples from the same ROIs shown in (E).

Application of SPEx to the nucleus

(A) Volcano plot of pairwise proteomic comparison between Nuclei and CellsMinusNuclei samples. The dotted square indicates proteins significantly enriched in the nucleus (adjusted P < 0.05, log₂FC > 0). Red dots represent proteins annotated as nuclear according to Gene Ontology (GO). FC, fold change. (B) Number of proteins significantly enriched in the nucleus in (A) (dotted square) that are annotated as nuclear (Yes) or not nuclear (No) according to GO. (C) HeLa cells immunostained for C7orf50 and stained with DAPI and NHS 594. DAPI is omitted from the merged image for clarity. Scale bar, 10 µm. (D) Functional enrichment analysis of all proteins shown in (A) for cellular components using Gene Ontology (GO) via STRING (string-db.org). The six most significantly enriched terms (ranked by FDR) are shown. All significant terms (FDR < 1%) are listed in Supplementary Table S7. (E) Same analysis as in (D), showing the six most significantly depleted GO terms. (F) Volcano plot of pairwise proteomic comparison between WholeCells and CellsMinusNucleus samples. The dotted square indicates proteins significantly enriched in WholeCells (adjusted P < 0.05, log₂FC > 0). Red dots represent proteins annotated as nuclear according to GO. (G) Same analysis as in (B) for proteins significantly enriched in WholeCells (F, dotted square).

Application of SPEx to nucleoli

(A) HeLa cells immunostained for NCL (nucleolar marker) or SON (nuclear speckle marker) and stained with DAPI and NHS 594. DAPI is omitted from the merged image for clarity. Scale bar, 10 µm. (B) Volcano plot of pairwise proteomic comparison between Nuclei and NucleiMinusNucleoli samples. The dotted square indicates proteins significantly enriched in nuclei (adjusted P < 0.05, log₂FC > 0). Red dots represent proteins annotated to localize to the nucleolus according to Gene Ontology (GO). (C) Number of proteins significantly enriched in (B) that are annotated as nucleolar (Yes) or not (No) according to GO. (D) Functional enrichment analysis of all quantified proteins shown in (B) for cellular components using Gene Ontology (GO) via STRING (string-db.org). The seven most enriched significant terms (ranked by enrichment score) are shown. All significant terms (FDR < 1%) are listed in Supplementary Table S11. (E) Same analysis as in (D), showing the two significantly depleted GO terms. (F) HeLa cells immunostained for SPOUT1 and stained with DAPI and NHS 594. DAPI is omitted from the merged image for clarity. Scale bar, 10 µm.

Application of SPEx to the Golgi

(A) Volcano plot of pairwise proteomic comparison between Golgis and CellsMinusGolgi. The dotted square indicates proteins significantly enriched (adjusted P < 0.05, log₂FC > 0) in the Golgi. Red dots represent proteins annotated to localize to the Golgi and/or to be extracellular according to GO. (B) Number of proteins significantly enriched in the Golgi in (A) that are annotated as Golgi and/or extracellular (Yes) or not (No) according to GO. (C) Functional enrichment analysis of all quantified proteins shown in (A) for cellular components using Gene Ontology (GO) via STRING (string-db.org). The six most enriched terms (ranked by enrichment score, FDR < 1%) are shown. All significant terms (FDR < 1%) are listed in Supplementary Table S12. (D) Same analysis as in (C), showing the six most depleted GO terms. (E) HeLa cells immunostained with antibodies for KIAA2013 and GBF1 (Golgi marker) and stained with DAPI. DAPI is not shown in the merged image for clarity. Scale bar, 10 µm. (F) Bar plot comparing the number of proteins assigned to the Golgi (blue) in our SPEx dataset (this study, A) with other published spatial proteomics datasets (Fasimoye et al., 2023; Geladaki et al., 2019; Hein et al., 2025; Orre et al., 2019; Schessner et al., 2023). Proteins overlapping with GO Golgi/extracellular annotations are shown in red. (G) Specificity of GO Golgi/extracellular annotation in the different proteomics datasets, determined from (F).

Key resource table

(A) Number of identified proteins for different input cell numbers. Data are mean ± SD: 50 cells: 2964; 10 cells: 2374; 5 cells: 2417; 1 cell: 982.5 ± 220.6 (n = 1 for 50, 10, and 5 cells; n = 6 for 1 cell; N = 1 experiment). (B) Volcano plot of pairwise proteomic comparison between nuclei and whole cells. Red dots represent proteins annotated as nuclear according to GO. (C) Volcano plot of pairwise proteomic comparison between nuclei and CellsMinusNuclei (same data as Figure 3A). Red dots represent proteins annotated as mitochondrial and not nuclear according to GO. (D) Schematic illustrating the SPEx “standard” approach and the SPEx “subtraction” approach. In the standard approach, the proteome of the organelle of interest (e.g., nuclei) is compared to the proteome of whole control cells from which the organelle was removed. In the subtraction approach, the proteome of a whole cell is compared to the proteome of a cell lacking the organelle of interest. In both approaches, proteins localized to the organelle of interest are enriched, and absent proteins are depleted. Created in BioRender.com. (E) Venn diagram depicting the overlap of identified nuclear proteins between the standard SPEx (Figure 3A) and subtraction SPEx approach (Figure 3F). (F) Bar chart illustrating the number of proteins in the overlap shown in (E) with (Yes) and without (No) nuclear GO annotation. (G) Bar plot comparing the number of proteins assigned to the nucleus (blue) in our SPEx dataset (this study, A) with other published spatial proteomics datasets (Geladaki et al., 2019; Hein et al., 2025; Orre et al., 2019). Proteins overlapping with nuclear GO annotation are shown in red. (H) Specificity of nuclear GO annotation in the different proteomics datasets, determined from (G).

(A) Volcano plot of pairwise proteomic comparison between nucleoli and NucleiMinusNucleoli. The dotted square indicates proteins significantly enriched in nucleoli (adjusted P < 0.05, log₂FC > 0). Red dots represent proteins annotated to localize to the nucleolus according to GO. (B) Number of proteins significantly enriched in nucleoli in (A, dotted square) that are annotated as nucleolar (Yes) or not (No) according to GO. (C) A-431 cells immunostained for RBM5 and tubulin. Image adapted from Protein Atlas (www.proteinatlas.org). Scale bar, 10 µm. (D) A-431 cells immunostained for SNW1 and tubulin. Image adapted from Protein Atlas (www.proteinatlas.org). Scale bar, 10 µm. (E) Bar plot comparing the number of proteins assigned to the nucleolus (blue) in our SPEx dataset (this study, A) with the published proteomics dataset (Orre et al., 2019). Proteins overlapping with nucleolar GO annotation are shown in red. (F) Specificity of nucleolar GO annotation in the different proteomics datasets, determined from (E).

(A) Volcano plot of pairwise proteomic comparison between WholeCells and CellsMinusGolgi. The dotted square indicates proteins significantly enriched (adjusted P < 0.05, log₂FC > 0) in the WholeCells. Red dots represent proteins annotated to localize to the Golgi and/or to be extracellular according to GO. (B) Number of proteins significantly enriched in the Golgi in (A) that are annotated as Golgi and/or extracellular (Yes) or not (No) according to GO.