Heterogeneity in endocytic capacity promotes Mtb phenotypic diversity.

A) Schematic of experiment to assess Mtb redox states in cells of different endocytic capacities. RAW 264.7 macrophages were infected with Mtb Mrx1-roGFP2 H37Rv and pulsed with dextran-TMR for 10 minutes. Cells were then fixed and analyzed by flow cytometry. Based on cargo fluorescence intensity, infected cells were categorized into ‘high’ or ‘low’ endocytic capacity subpopulations, and the redox status of intracellular Mtb was assessed within each group. A total of 10,000 infected cells were analyzed, with the lower and upper 30% of the dextran-TMR distribution used to define the ‘low’ and ‘high’ endocytic subpopulations, respectively. The proportions of bacterial subpopulations within each host subpopulation were calculated as described in the Methods. B) Infected cell subpopulations defined based on dextran uptake harbour Mtb of different redox states denoted as EMSH-reduced, basal and oxidised. Percentage of infected cells across these host subpopulations are indicated. C,D) EMSH-reduced, basal and oxidised Mtb in THP-1 monocyte derived macrophages (C) or mouse bone marrow derived macrophage (D) subpopulations defined by dextran uptake. Data the representative of at least three independent biological experiments. Error bars represent standard deviation from the mean of technical replicates. *denote p-values from Student’s t-test (paired):**p < 0.01, ***p < 0.001; ns, not significant.

Characterization of variance in the endocytic capacity across cell types.

A) RAW 264.7 macrophages were pulsed with dextran conjugated to fluorophore for 10 minutes to assess endocytic activity. Cells were then sorted based on fluorescence intensity from the two ends of the distribution : the lowest and highest 10–15 % were designated as ‘Low’ and ‘High’ endocytic capacity subpopulations, respectively. B) EEA1 and Rab5 immunostaining in Dextran-TMR sorted Parent, ‘Low’ and ‘High’ subpopulation of RAW 264.7 macrophages, fixed immediately after sorting. Density plots from 800-2500 cells shows the distribution of the number of puncta and total intensity of Dextran, EEA1 and Rab5 endosomes between the subpopulations. C) Distribution of dextran uptake from 10,000 cells in parental and sorted RAW 264.7 macrophage subpopulations. Fluorescence intensity profiles of dextran-AF647 in the parental population, as well as in the ‘Low’ and ‘High’ endocytic capacity subpopulations, were assessed either immediately after sorting (i, Day 0) or after three days in culture post-sorting (ii, Day 3). Bar graphs represent binning analysis of the percentage of cells from each group falling into defined fluorescence bins, based on the dextran distribution of the corresponding parental population at each time point. Error bars indicate standard deviation from the mean of three independent biological replicates. D,E) Dextran or transferrin uptake assay was performed independently in the different cell lines indicated. The coefficient of variation (CV) for dextran-TMR and transferrin-488 was calculated from the fluorescence intensity distributions after background subtraction. Background subtraction was performed by subtracting the mean fluorescence intensity of the unstained control from each event’s scale value. Error bars indicate standard deviation from mean between three biological replicates. F) HeLa GFP-Rab5 cells are sorted into ‘low’ and ‘high’ subpopulations based on GFP-Rab5 intensity, fixed and immunostained for EEA1. Representative images show GFP-Rab5 and EEA1 endosomes from parental, ‘low’ and ‘high’ endocytic subpopulations. Density plots from 60-90 cells show the distribution of the number and intensities of the endosomes between the subpopulations. G) Distribution of GFP-Rab5 intensity from 10,000 cells ‘low’ and ‘high’ sorted HeLa GFP-Rab5 subpopulations either immediately (i, day 0) or three days (ii, day 3) post-sorting. Black dots denote the parental distribution. Bar graphs represent binning analysis of the percentage of cells from each group falling into defined fluorescence bins, based on the GFP-Rab5 distribution of the corresponding parental population at each time point. Error bars indicate standard deviation from the mean of three independent biological replicates. Scale bar =10 µm

Time-lapse imaging reveals association between endocytic capacity and the cell cycle.

A) Representative frames from time-lapse imaging of HeLa GFP-Rab5 cells. Movies were recorded every 30 minutes for 72 hours. Scale bar = 10 µm. B) Distributions of (i) total GFP-Rab5 intensity and (ii) mean GFP-Rab5 intensity normalized to cell area were quantified from the initial populations of 51 ‘low’ and 39 ‘high’ subpopulation cells at the indicated timepoints. C) Trajectory of total GFP-Rab5 intensity for representative cells across multiple rounds of cell division. D) Distribution of total GFP-Rab5 intensity in ‘low’ and ‘high’ subpopulations across multiple cell generations. E) Probability of asymmetric division between two daughter cells in parent, ‘low’, and ‘high’ subpopulations.F) Violin plot showing the inter-mitotic time for cells in parent, ‘low’, and ‘high’ subpopulations across different generations. Dots indicate population medians. p-values were computed using Wilcoxon Rank Sum (statistical test). Results are representative of two biological replicates.

Cell cycle progression through interphase explains substantial variation in endocytic capacity.

A) HeLa BAC GFP-Rab5 cells were pulsed with fluorescently labeled transferrin, followed by Hoechst staining and analyzed by flow cytometry. Contour plots show positive correlation between GFP-Rab5 (i) or transferrin (ii) and Hoechst intensity. B) HeLa GFP-Rab5 cells were fixed, immunostained for EEA1, and imaged. Contour density plots show relationships between DAPI intensity and total GFP-Rab5 (i) or EEA1 (ii) intensities. Data are representative of three independent experiments. For A, B, numbers denote mean ± SD of the correlation coefficients from three independent biological replicates.C) HeLa BAC GFP-Rab5 cells were pulsed with fluorescent transferrin, stained with Hoechst, and analyzed by flow cytometry. Cells were gated into G1, S, and G2/M phases based on Hoechst intensity. Bar graphs show mean total GFP-Rab5 and transferrin intensities within each gate. Error bars denote SD from four technical replicates. Data are representative of at least three independent biological replicates.D, E) HeLa BAC GFP-Rab5 cells were treated with 2.5 mM thymidine (Thy) for 18 hours to arrest the cell cycle in late G1/S phase. Cells at 0 hrs and 8 hrs post-treatment are labeled as block and release, respectively.D) Cells were pulsed with transferrin, fixed, and stained with Hoechst. Distributions of Hoechst, GFP-Rab5, and transferrin intensities are shown. Bar graphs display mean total GFP-Rab5 and transferrin intensities in each condition. Error bars indicate SD from two technical replicates. E)HeLa BAC GFP-Rab5 cells after thymidine block or release were fixed and immunostained for EEA1 to assess early endosomal content. Density plots from 450-1700 cells show distributions of nuclear, GFP-Rab5, and EEA1 intensities per cell. Data are representative of three independent experiments. (Scale bar = 10 µm).F) RAW 264.7 macrophages were pulsed with fluorescent transferrin, stained with Hoechst, and analyzed by flow cytometry. Cells were gated into G1, S, and G2/M phases. Bar graphs show mean transferrin intensity in each gate. Error bars represent SD from three technical replicates. G) RAW 264.7 macrophages were treated with 2 mM thymidine for 12 hours to arrest cells in late G1/S. Cells at 0 hrs and 4 hrs post-treatment (block and release) were pulsed with dextran, fixed, and stained with Hoechst. Distributions of Hoechst and dextran-TMR intensities are shown. Bar graphs depict mean total dextran-TMR intensity in each condition. Error bars represent SD from three technical replicates. For C-G, data are representative of at least three independent biological replicates.*denote p-values from Student’s t-test:*p < 0.05,**p < 0.01, ***p < 0.001; ns, not significant.

Cell cycle is a major contributing factor to endocytic heterogeneity in a population.

A) Schematic of the single-cell RNA sequencing experiment. RAW 264.7 macrophages were pulsed with Dextran-TMR for 10 minutes and sorted into ‘low’ and ‘high’ endocytic subpopulations.B) Analysis pipeline used for scRNA-seq data.C,D) Pathway analysis from GO TERM enrichment showing biological processes enriched in genes upregulated in the ‘high’ (C) or ‘low’ (D) subpopulations. E) i) Venn diagram showing overlap of differentially expressed genes (DEGs) between ‘high’ vs ‘low’ endocytic capacity subpopulations, and DEGs from thymidine release (4 hrs) versus parent populations. ii) Pathway enrichment analysis of genes common to both. iii) Pathway enrichment of genes distinct to endocytic capacity.F) HeLa BAC GFP-Rab5 cells were imaged during thymidine (Thy) treatment or low FBS condition (1% FBS). Slopes were extracted from GFP-Rab5 intensity across time using a linear fit model from tracks with atleast 7 datapoints. i) Representative single cell tracks from the different conditions (dots) with the fits (solid lines). ii, iii) Comparison of slopes of total GFP-Rab5 cellular intensity between the indicated conditions. For panel iii, experimental conditions of thymidine block and release duration are shown in the figure. Slopes are calculated from 281, 56, 179, and 256 tracks for the experimental conditions denoted as Unt (untreated), L-FBS (low FBS), Thy bl (Thymidine Block) and release, respectively. *** and n.s. denote p-values < 0.0001, and > 0.1 respectively, from Student’s t -test.G) Variance decomposition analysis to quantify the contribution of cell cycle progression to endocytic heterogeneity. Error bars represent the standard deviation from the mean across three biological replicates.

Cell cycle progression promotes phenotypic diversity of Mtb

A) Experimental scheme to assess the effect of cell cycle progression on intracellular Mtb phenotypic states. B) RAW 264.7 macrophages were treated with 2mM thymidine for 12 hrs for arresting the cell cycle in late G1/S phase. The cells were infected with Mtb Mrx1-roGFP2 H37Rv at 3.5 hrs post release, fixed and stained with Draq5. Graph shows the proportion of infected cells containing EMSH-reduced, basal and oxidised Mtb subpopulations between untreated or synchronised conditions (T-release) as indicated. Numbers denote the percentages of total and infected cells in each host cell subpopulations. C) Experimental scheme to assess plasticity of Mtb phenotypic states as cells move through the cell cycle. RAW 264.7 cells were infected with Mtb Mrx1-roGFP2 and pulsed with dextran for 10 mins in serum free media and sorted into ‘low’ or ‘high’ subpopulations based on the dextran fluorescence. D) Intracellular Mtb EMSH subpopulations were analysed from 10,000 infected cells in the ‘low’ and ‘high’ endocytic subpopulations either immediately (i), or 2 days (ii) after sorting. Arrows indicate that the respective sorted subpopulations are cultured separately from each other. Error bars represents standard deviation of mean from 3 technical replicates, numbers denote p-value. Results are representative of atleast two independent biological replicates)(*denote p-values from Student’s t-test:**p < 0.01, ***p < 0.001.)

Interaction between bacterial and host heterogeneity in post differentiated macrophages.

A) Schematic detailing the experiment to assess the relationship between endocytic capacity and cell cycle stages in macrophages post-differentiation. B) 2D FACS contour plot showing the endocytic capacity and cell cycle stages of THP-1 monocytes and the macrophages derived from them. To generate these plots, cells were pulsed with dextran for 10mins, fixed and stained with Hoechst 33342 (2 µg/mL for 20 min). Cells were gated into G1/S and G2/M subpopulations based on the Hoechst intensity. Proportions of cells in these gates are indicated. The subpopulations of the lower and upper 30% bounds of dextran intensity were defined as ‘low’ and ‘high’ dextran subpopulations, respectively. Numbers inside the plots indicate the proportion of G1/S and G2/M cells within the ‘low’ and ‘high’ dextran gates, and are expressed as mean ± standard deviation from three independent experiments. C, D, E, F) Proportion of different phenotypic subpopulations of Mtb Mrx1-roGFP2 in THP-1 macrophage (C,D) or BMDM (E,F) binned based on host cell endocytic capacity (C,E) or cell cycle stages (D,F). For these experiments, cells were infected with Mtb Mrx1-roGFP2 followed by pulsing with dextran for 10 mins, fixing, staining with DRAQ5 and measuring the endocytic capacity, cell cycle stages and redox of status of intracellular Mtb in the same cell population by flow cytometry. Numbers indicate the mean +/- SD of percentage infected cells from atleast three technical replicates. G,H, I, J) Among the EMSH-oxidized and reduced subpopulations, the percentage of bacteria in ‘low’ and ‘high’ dextran populations from THP-macrophages (G) or bone marrow derived macrophages (I), or cells in G1/G0 and G2 stages from THP-1 (H) or bone marrow derived macrophages (J) was quantified and represented as bar graphs. All data represents mean ± standard deviation of atleast three technical replicates in each subpopulation, and are representative of 3 independent biological experiments.

Heterogeneity in endocytic capacity promotes Mtb phenotypic diversity.

(A) RAW 264.7 macrophages were infected with Mtb Mrx1-roGFP2 H37Rv for 2 hours at an MOI of 10 and pulsed with transferrin-AF647 for 10 minutes. Cells were fixed and analyzed by flow cytometry. Based on transferrin fluorescence intensity, infected cells were categorized into ‘high’ and ‘low’ endocytic capacity subpopulations. The redox status of intracellular Mtb was then assessed within these subpopulations. 10,000 infected cells were analyzed, and the top and bottom 30% of the transferrin fluorescence distribution were gated as ‘high’ and ‘low’ endocytic cells, respectively. The proportion of bacterial redox states (EMSH reduced, basal, and oxidized) was determined as described in the Methods. The percentage of infected cells across these host subpopulations is indicated. Data shown are representative of at least three independent biological experiments. Error bars represent the standard deviation of technical replicates. *denote p-values from Student’s t-test:**p < 0.01, ***p < 0.001; ns, not significant.

Characterization of variance in endocytic capacity across cell types.

(A) Dextran and transferrin uptake assays were performed independently in the indicated cell lines. Correlation coefficients between dextran-TMR and transferrin-488 intensities were calculated from single-cell measurements. Error bars represent the standard deviation from the mean of three biological replicates.(B) Hela BAC GFP-Rab5 single-cell colonies were analysed after uptake of fluorescently tagged dextran and transferrin using flow cytometry. Heatmap shows the correlation coefficient between GFP-Rab5, dextran and transferrin. Data from three colnies are shown with error bars between colonies.(C) HeLa GFP-Rab5 single-cell colonies were sorted into ‘low’ and ‘high’ subpopulations based on GFP-Rab5 intensity. Shown are the distributions of GFP-Rab5 intensities from 10,000 cells in the ‘low’ and ‘high’ sorted HeLa GFP-Rab5 subpopulations either immediately after sorting (i, day 0) or four days post-sorting (ii, day 4). Red dots indicate the parental population. Distributions from five independent colonies are shown.

Time-lapse imaging reveals an association between endocytic capacity and the cell cycle.

(A) Representative frames from time-lapse imaging of HeLa GFP-Rab5 cells used for segmentation and lineage tracking via Cellpose2 and ilastik, respectively. (B) Cells were binned into ‘high’ and ‘low’ groups (21 cells each) based on total GFP-Rab5 intensity of 84 cells from parental population at the start of imaging. Distribution analysis of these cell lineages across the indicated timepoints confirms that observed effects are not artifacts of the cell sorting procedure.

Cell cycle progression through interphase explains substantial variation in endocytic capacity.

(A) HeLa BAC GFP-Rab5 cells were treated with 2.5 mM thymidine for 18 hours to arrest the cell cycle in late G1/S phase. Cells at 0 hours and 8 hours post-treatment are labeled as “block” and “release,” respectively. Cells were gated into G1, S, and G2/M phases based on Hoechst intensity. Bar graphs show the percentage of cells in each phase. Error bars denote standard deviation (SD) from four technical replicates. Data are representative of at least three independent biological replicates. (B, C) HeLa BAC GFP-Rab5 colonies K4 and J10 were subjected to a similar experiment. Cells were pulsed with transferrin, fixed, and stained with Hoechst. Distributions of Hoechst, GFP-Rab5, and transferrin intensities are shown. Bar graphs display the percentage of cells in different cell cycle stages. Error bars represent SD from three technical replicates. (D) RAW 264.7 macrophages were treated with 2 mM thymidine for 12 hours to arrest cells in late G1/S. Cells at 0 hours and 4 hours post-treatment (“block” and “release”) were pulsed with dextran, fixed, and stained with Hoechst. Bar graphs depict the percentage of cells in different cell cycle stages. Error bars represent SD from three technical replicates. Data are representative of at least three independent biological replicates. *denote p-values from Student’s t-test:***p < 0.001.

Cell cycle is a major contributing factor to endocytic heterogeneity.

(A,B) Volcano plots showing differentially expressed genes between ‘low’ vs ‘high’ and parent vs thymidine release (4 h). (C) Venn diagram showing overlap of differentially expressed genes across conditions.

Cell cycle analysis using DRAQ5 in RAW cell synchronization.

RAW 264.7 macrophages were treated with 2 mM thymidine for 12 h to arrest cells in late G1/S phase. Cells were then released into fresh medium for 3.5 hrs, followed by infection with Mtb Mrx1-roGFP2 H37Rv (MOI 20) for 30 min. After fixation and DRAQ5 staining, cell cycle stages were analyzed by flow cytometry. G1/S was gated based on the first sharp DRAQ5 intensity peak, while G2/M was gated at approximately twice the median intensity of G1.

Cell cycle analysis using DRAQ5 in THP-1 and BMDM cells.

THP-1 and BMDM cells were infected with Mtb Mrx1-roGFP2 at an MOI of 10, followed by a 10-minute dextran pulse, fixation, and staining with DRAQ5. Endocytic capacity and cell cycle stages were assessed by flow cytometry. Cell cycle phases were gated as indicated: the G1 gate was drawn based on the first sharp peak in DRAQ5 intensity; the G2 peak was gated using a threshold approximately twice the median intensity of G1.