Rab11b ^−/− mice were developed by us through CRISPR-CAS9 genome editing previously described in (Joseph et al., 2023; D'Agostino et al., 2019). Mice were housed and bred under a specific pathogen free condition. The comparison of wild type and Rab11b ^−/− mice were made in both male and female mice, and experiments were repeated using mice from different mating pairs. All animal experiments were conducted in accordance with NIH guidelines and approved by the Institutional Animal Care and Use Committee (IACUC) at Rutgers University.

The procedure for transmission electron microscopy was described previously (Yu et al., 2014a; Gao and Kaestner, 2010; Balasubramanian et al., 2023; Das et al., 2015; Flores et al., 2021; Yu et al., 2020). Briefly, intestinal tissue samples, approximately 1 mm³ in size, were collected and fixed overnight at 4°C in TEM fixative (2.5% Glut and 2.0%PFA in 0.1M Sodium Cacodylate (NaCac) pH 7.4). Following fixation, the samples were washed twice for 10 min each with 0.1M NaCac pH 7.4. The tissue samples were then incubated in 1% OsO4 in 0.1M NaCac pH 7.4 solution for 1 h in the dark. Subsequently, the samples were washed twice with 0.1M NaCac for 10 min each time and then rinsed twice with milli-Q for 10 min each time. The tissue samples were then incubated with 1% Uranyl Acetate in H₂O for 30 min. After incubation, the samples were washed twice with ddH2O for 10 min each time. The samples were then dehydrated through 50%, 70%, 80%, 90%, and 100% ethanol(2x). The samples were further dehydrated using 1:1 and 3:1 Propylene Oxide:Ethanol and 100% Propylene Oxide for 10 min each. For resin embedding, the samples were left overnight in 1:1 Propylene Oxide: EMBed-812 followed by overnight incubation in 100% EMBed-812. Final embedding was in EMBed-812 activated with DMP at 65°C for 48 h. Ultrathin sections were post-stained using lead citrate and uranyl acetate. Three independent mice were used for each genotype with 3 grids prepared for each mouse. To assess mitochondrial integrity, the morphology of mitochondria was categorized into four grades based on structural and membrane characteristics: normal (intact cristae and preserved membrane integrity), grade 1 (mild abnormalities with minor disruptions in cristae structure), grade 2 (pronounced abnormalities, including swelling and loss of cristae definition), and grade 3 (severe damage with extensive cristae disruption and compromised membrane integrity). Percentage of mitochondria in each grade was calculated for each genotype. The proportion of mitochondrial grades was visualized as a stacking bar graph. The analysis was performed based on a total of 50–70 microscopic images from the 3 mice per genotype.

RNA extraction and bulk RNA sequencing were performed according to the protocols described in ((Joseph et al., 2023) Total RNA was extracted from 2 to 3 mm sections of jejunum tissue using the Qiagen RNeasy Plus Mini Kit (#74134). Each tissue sample was immersed in 800 μL of RLT lysis buffer (provided by manufacturer) supplemented with 10 μM β-Mercaptoethanol. Tissues were homogenized at 4°C by multiple passages through a 20-gauge needle (BD #305176). Following homogenization, RNA isolation was performed according to Qiagen RNeasy Mini protocol by applying to a spin column and RNA elution by RNase-free water. The quality of the extracted RNA was validated by Bioanalyzer and subjected to sequencing described in (Joseph et al., 2023).

Gene Set Enrichment Analysis (GSEA) (Subramanian et al., 2005) was conducted as detailed in our earlier work (Joseph et al., 2023). Heatmaps and leading-edge enrichment plots were generated using the Molecular Signature Database (MSigDB). One thousand permutations were performed on each gene list tested, and the normalized enrichment score (NES) and nominal P-value were reported for each analysis.

3×FLAG-RAB11A-pQXCIP or 3×FLAG-RAB11B-pQXCIP was transiently transfected into HEK293T cells. After 24 h, cells were lysed, protein concentrations were measured by Bradford assay, and 2 mg of lysate was immunoprecipitated overnight with anti-FLAG M2 affinity gel beads at 4°C. Proteins were resolved by SDS-PAGE, stained with Ruby Red, and bands ≥25 kDa were analyzed by mass spectrometry. Proteomic analysis of transfected HEK293T cells was described in previous study (Joseph et al., 2023).

For immunofluorescence staining, formalin-fixed paraffin-embedded tissue sections were deparaffinized and rehydrated. Antigen retrieval was performed in monohydrate citric acid buffer (pH 6) by boiling slides for 15 min, followed by cooling to 40°C, washing in deionized water for 5 min, and PBS for 5 min. Slides were blocked for 1 h in blocking solution (0.1% Triton-X100, 2% BSA, 2% normal donkey or goat serum in PBS), then incubated overnight at 4°C with primary antibodies diluted in blocking solution. The primary antibodies used included anti-Tom20 (Thermo Fisher, Cat# PA5-52843) which was applied at a 1:500 dilution. After three 10-minute PBS washes, slides were incubated with secondary antibodies diluted in blocking solution for 1 h at room temperature, washed twice in PBS for 10 min each, counterstained with DAPI for 15 min, rinsed in PBS, and mounted using Prolong Gold Antifade (Invitrogen P36930). Imaging was performed using a Zeiss LSM 980 microscope equipped with Airyscan2. Image processing was performed using ImageJ software.

HEK293T cells were grown in Dulbecco’s Modified Eagle’s Medium (DMEM) (Corning, CM10013) supplemented with 10% fetal bovine serum (FBS) (GIBCO, A3160501). Caco2 BBE cells were grown in DMEM, supplemented with 20% FBS. Cells were transiently transfected with mCherry blank, mCherry-tagged RAB11A or RAB11B plasmid using Lipofectamine 3000 (Invitrogen, L3000008). After 12–16 h post-transfection, the cells were incubated with MitoTracker Green (Fisher Scientific, M46750) at a 1:1,000 dilution for 1 h (for HEK293T); 1:1,500 dilution for 15 min (for Caco2 BBE cells). Live cell culture conditions were maintained at 37°C throughout image acquisition with 5% CO2. Single plane confocal or Z-stack imaging were performed using a Zeiss LSM 980 microscope equipped with Airyscan2. 3D Volume Rendering was used to visualize the depth perception of the colocalization, in XY-planar, and Surface Rendering was used to assess the colocalization in XZ-planar. The number of microscopic fields were randomly selected for quantification and was reported in the graph. Number of colocalization points per field were recorded and analyzed with ImageJ, and the statistical result was quantified from at least 3 independent experiments using GraphPad Prism software, employing one-way ANOVA analysis.

Intestinal tissues were harvested from WT and Rab11b knockout mice. Tissues were lysed in RIPA buffer (Thermo Fisher #89901) with protease inhibitor (Sigma, #11836170001). Total concentration of proteins was determined using the Pierce™ BCA Protein Assay Kit (Thermo Fisher, #23225). 20 μg proteins were resolved on 15% acrylamide-Tris-BIS-0.1% SDS gel, and 6.16% acrylamide-Tris-BIS gel was used as stacking. Proteins were transferred to nitrocellulose membrane (Cytiva, 10600004). Primary antibodies were diluted with 1:1,000 for detection of GAPDH, GeneTex GT239; BCL2, Santa Cruz, sc-7382; NDUFS3, Santa Cruz, sc-374282; SLC25A3, Santa Cruz, sc-376742; and Tom20, Thermo Fisher PA5-52843. Proteins were detected by ECL Prime Western Blotting Detection Reagents (Cytiva, RPN2232).

The ileum was harvested from each mouse and immediately placed in an ice-cold petri dish containing 1X PBS (Fisher Scientific, SH30256LS). The tissues were cut open longitudinally, and fecal contents were removed by shaking the opened tissue in the PBS-filled petri dish. The tissue was then cut into 5 mm pieces and transferred to a 50 mL Falcon tube containing 30 mL of ice-cold 1X PBS. The pieces were washed by inverting the tube 10–15 times. Washing step was repeated three times, each time transferring the tissue pieces to a new tube with fresh ice-cold 1X PBS. Tissue pieces were transferred to a tube containing 30 mL of crypt isolation buffer, consisting of 1X HBSS without calcium and magnesium (ThermoFisher, 14170112), supplemented with 600 µL of 0.5 M EDTA (Invitrogen, AM9260G). After 15-min, the solution was replaced with fresh crypt isolation buffer, and the tissues were shaken for an additional 30 min at 4°C. The tubes were then vigorously shaken to dislodge epithelial cells, and the turbid solution was passed through a 70 μm cell strainer (Fisher Scientific Cat# 22363548), which was rinsed with 10 mL of PBS that also passed through the cell strainer. Tissue disruption and cell isolation were monitored by brightfield microscopy. The cell suspensions were centrifuged at 500 g for 5 min at 4°C. The resulting pellet was resuspended in 2 mL of TrypLE buffer (ThermoFisher Cat# 12605–028) and split into two 15 mL Falcon tubes, each containing 1 mL of cell suspension in 9 mL of TrypLE buffer. 10 μL of reconstituted DNase (Qiagen Cat#79254) was added to each tube, and the samples were incubated in a 37°C water bath for 20 min, with regular inversions and observations every 3 min to monitor dissociation. When the suspension consisted of greater than 80% single cells, 1 mL of FBS was added to stop the reaction. The solutions were then passed through a 40 μm cell strainer (Fisher Scientific Cat# 22–363–547) into a 50 mL Falcon tube and centrifuged at 700 g for 5 min at 4°C.

For flow analysis of different epithelial cell populations, the cell pellet was resuspended in 500 µL of FACS buffer (F12/DMEM, ThermoFisher, Cat# 12634–010) containing 2 µM ROCK inhibitor (Sigma, Cat# Y0503), 2% BSA, 2% FBS (Sigma, Cat# F2442), and 2 mM EDTA, and then transferred to an Eppendorf tube. The cells were centrifuged at 700 g for 5 min at 4°C and resuspended in 100 µL of FACS buffer containing anti-mouse Fc (1:100, BioLegend, Cat# 101330; RRID: AB_312801) for 10 min on ice. After centrifugation at 700 g for 5 min at 4°C, the pellet was resuspended in 200 µL of a surface staining solution. This solution was prepared by adding EpCAM-APC (CD326) at a 1:200 dilution, CD24-APC/Fire 750 at a 1:200 dilution, and Fc at a 1:100 dilution in FACS buffer. Additionally, DAPI was added at a 1:500 dilution. Alternatively, a different surface staining solution was prepared using CD24-APC/Fire 750 (BioLegend, Cat# 101839) at a 1:200 dilution, EpCAM-FITC (BioLegend, Cat# 118207) at a 1:200 dilution, and Live/Dead 405 (ThermoFisher, Cat# L34963) at a 1:200 dilution.

For MitoTracker and MitoSox analysis, the pellet was resuspended in a staining solution composed of MitoTracker Green (ThermoFisher, Cat# M46750) at a 1:1,000 dilution, MitoSox Red (ThermoFisher, Cat# M36008) at a 1:5,000 dilution, and HBSS with Ca²⁺ and Mg²⁺ (ThermoFisher, Cat# 24020117). The cells were then incubated at 37°C for 30 min. Following the incubation period, 1 mL of FACS buffer was added, and the cells were centrifuged at 700 g for 5 min at 4°C. The resulting cell pellet was resuspended in 500 µL of FACS buffer, and 1 µL of DAPI was added before flow cytometry sorting. Alternatively, the cells were incubated at 37°C for 30 min with MitoTracker Orange (Invitrogen, Cat# M7510) at a 1:1,000 dilution. Cells were then sorted using the BD Biosciences Aria II Flow Cytometer (BD FACSAria II). For all fluorophores, a cell pellet was stained with a single antibody corresponding to each primary antibody used, and a combination of all antibodies was utilized to establish compensation controls. Single viable epithelial cells, Paneth cells, and stem cells were gated based on forward scatter, side scatter, and negative staining for DAPI or Live/Dead marker. Subpopulations were further gated using EpCAM, CD24, MitoTracker, or MitoSox CellROX.

Immunostaining and live cell imaging results were quantified on 10 to 12 independent microscopic fields and imaged using a ZEISS LSM 980 microscope. All flow cytometry statistical data was obtained using FlowJo software (TreeStar, version 10.7.1). Statistical analyses were performed using GraphPad Prism software, employing unpaired t-tests for two-group comparisons and one-way ANOVA for multigroup analysis.

The RNA‐Seq datasets are available in Gene Expression Omnibus, accession number: GSE232493 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE232493). The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifier PXD042300 and PXD042335.

To profile the protein interactome associated with RAB11A or RAB11B, we conducted proteomic analyses employing mass spectrometry on 3× Flag-tagged RAB11A and RAB11B co-immunoprecipitants from HEK293T cells (Joseph et al., 2023). The mass spectrometry data was sorted to identify interacting partners that were unique to RAB11A but not RAB11B (Figure 1A) and interactors unique to RAB11B but not RAB11A (Figure 1B). Notably, the RAB11B proteomic network revealed a significant association with mitochondrial proteins, including the RAB11FIP5 (RAB11 family interacting protein 5) (Figures 1C,D). The top 10 mitochondrial protein targets, ranked by spectrum counts and unique peptide numbers, in RAB11B-mitochondria network were proteins involved in metabolism and mitochondrial regulation (Figure 1C). String analysis of these RAB11B mitochondrial proteome were strongly related to the mitochondrial matrix (Figure 1D).

To appreciate the cellular localization of RAB11A, RAB11B, and potential intersection with mitochondria, HEK293T cells were transiently transfected with either mCherry-tagged RAB11B or RAB11A and subjected to live cell fluorescence confocal microscopy (Figure 1E). The RAB11B and RAB11A fluorescent signals appeared to be structures that are typically identified as vesicular puncta (Figure 1E). In comparison of MitoTrackerGreen and RAB11 A/B fluorescence signals, there were distinct contacting “spots” where the fluorescence signals for mitochondria and RAB11B appear to be co-localized at this level of resolution (see Figure 1F). Such colocalization spots were more readily observed with RAB11B as compared to RAB11A in experiments conducted in parallel (Figure 1G). We performed similar live cell imaging experiments in Caco2 cells to determine if the observation holds in this intestinal epithelial cell line where endogenous RAB11 is abundant. Single plane confocal, Z-stack imaging and 3D surface rendering revealed RAB11-mitochondrial contact sites in Caco2 cells (Figure 2A), while RAB11B has significantly more of these contact sites than RAB11A or mCherry alone (Figure 2B).

To examine if there is general mitochondrial defect in vivo, we performed immunostaining for Tom20, a translocase of outer mitochondrial membrane, in wild type (WT) and Rab11b knockout (KO) intestinal tissues. Tom20 localization was observed throughout the entire cell, with prominent staining observed in the subapical cytoplasm region (white box in Figure 2C). In contrast, Tom20 staining appeared to be disrupted in the Rab11b KO epithelial cells and became localized to the entire cell body in the crypt-villus junction region (white box in Figure 2D). The signal became weaker in Rab11b KO crypt cells (red box in Figure 2D).

We then performed Western Blotting analysis to assess the mitochondrial protein abundances in WT and Rab11b KO intestinal tissues. While the total Tom20 abundances remained unchanged, SLC25A3, a mitochondrial phosphate carrier, and NDUFS3, a vital part of mitochondrial complex I of the respiratory chain, were both increased in Rab11b KO intestines (Figures 2E,F). Interestingly, BCL2, a key regulator of the permeabilization of the outer mitochondrial membrane in programmed cell death, was also increased in Rab11b KO intestinal tissues (Figures 2E,F). These biochemical and immunohistochemical results suggest that mitochondrial localizations and functions are altered in the absence of Rab11b.

To examine if Rab11b deletion impacted the intestinal transcriptomic readouts of mitochondrial function, we analyzed the intestinal bulk-RNA sequencing from WT and Rab11b KO littermates. Gene set enrichment and heat map analysis revealed a significant increase of glucose metabolism in Rab11b KO intestines compared to WT (p < 0.05, Figures 3A,B). Specifically, PPP1R3E that is involved in positive regulation of glycogen biosynthetic process and PFKFB22 that regulates the level of F26BP for control of glycolysis process were robustly increased in Rab11b KO intestines (Figures 3C,D). Interestingly, genes related to superoxide metabolic processes are reduced in Rab11b KO intestinal tissue (Figures 3E,F). Reductions in genes such as CD177, FANCC, TYROBP and ITGB2 (Figures 3F–I), which regulate inflammation and superoxide production, suggest that Rab11b ablation potentially disrupted the trafficking of these receptors and mitochondrial superoxide production.

Gene set enrichment analysis identified that genes related to mitochondrial protein targeting are increased in RAB11B KO intestines (Figure 3J). Figure 3K provides heat maps for the top 10 hits and three of these, HSPA1L, PDCD5 and BAG3, are linked to maintaining mitochondrial homeostasis and integrity (Tahrir et al., 2017; Bock et al., 2015; Ni et al. 215). Interestingly, RhoU GTPase is significantly elevated in Rab11b KO tissues (Figure 3L) (Hall, 1998) and this Rho family member appears to play a role in epithelial cell development, actin cytoskeletal dynamics, and motor protein activity. Overall, these results suggest that genetic ablation of Rab11b led to an enhanced mitochondrial protein targeting gene signature possibly in response to an altered mitochondrial function in tissue homeostasis.

To examine if Rab11b was required for mitochondrial activity, we used fluorescence-activated cell sorting (FACS) of intestinal epithelial cells isolated from WT and Rab11b KO crypts prepared from small intestine. The fluorescent probes MitoTracker Green (green fluorescence) and MitoSox (red fluorescence) were used to measure mitochondrial quantity and superoxide for mitochondrial activity, respectively. EpCAM⁺DAPI⁻ cells were used to gate for live epithelial cells (Figure 4A). When gating for MitoTracker Green/MitoSox signals, live epithelial cells with the highest mitochondrial number and mitochondrial activity were positioned in Quadrant 2 (Figure 4B). Notably, the number of Rab11b KO epithelial cells in Quadrant 2 were significantly reduced relative to the number of WT (Figures 4B,C). Rab11b KO epithelial cells were most abundant in Quadrant 4 that represented cells with low mitochondrial content and activity (Figures 4B,C). These results echo the transcriptomic results indicating an altered mitochondrial function upon Rab11b deletion.

By transmission electron microscopy (EM), WT enterocytes, readily identified by their brush borders, contained abundant mitochondria with well-defined inner and outer membranes, transverse cristae, and electron dense matrix. Consistent with fluorescence localization (Figure 2C), EM revealed that mitochondria were largely positioned in the apical domain of WT enterocytes (Figure 4D). Rab11b knockout enterocytes still elaborated apical brush borders, however their mitochondria lacked organization and electron density seen for WT mitochondria (Figure 4E). Mitochondria in Rab11b KO epithelial cells showed a dramatic reduction in cristae, associated with breakage and loss of transverse cristae, loss of matrix electron density, and discontinuous outer membrane in some cases (red arrowheads in Figure 4E). Events of mitophagy or autophagosomes appeared more visible in Rab11b KO cells (arrows in Figure 4E), potentially reflecting an early reported observation in HeLa cells (Grimsey et al., 2016). Mitochondrial structural integrity was categorized into mild (grade 1), moderate (grade 2) or severe (grade 3) loss of integrity (Figure 4F). Rab11b KO enterocytes showed an overall 60% reduction in normal-appearing mitochondria, with over 20% mitochondria exhibiting moderate to severe structural or membrane damage (Figure 4G). These ultrastructural findings were validated in different KO mice that were analyzed in parallel with WT mice, confirming that Rab11b is required for maintaining mitochondrial morphology in enterocytes.

To examine if the observed mitochondrial defects were also present in non-absorptive cells, such as intestinal stem cells and Paneth cells in crypts, we performed additional flow cytometry analysis using gating strategy (Sato et al., 2011), specific for CD24⁺ non-Paneth cells, Paneth cells, and enteroendocrine cells (EECs) (Figure 5A). Mitotracker Orange (CMTMRos), a cationic fluorescent probe for negative mitochondrial membrane potential, showed that WT Paneth cells have a relatively higher signal accumulation than Rab11b KO Paneth cells, indicating a reduced mitochondrial activity in the absence of Rab11b (Figure 5B). This reduction was not observed in non-Paneth crypt epithelial cells (Figure 5C). However, mitochondrial superoxide labeled by MitoSox Red did not significantly change or showed an insignificant increase in Rab11b KO Paneth cells (Figure 5D), which could potentially result from electron leakage from damaged mitochondria. These results collectively suggest that mitochondrial membrane potential and metabolic activity are changed in response to Rab11b loss, with Paneth cells potentially being more vulnerable.

Electron microscopy of mitochondria in Paneth (Figure 5G) and stem cells (Figure 5H) confirmed the disrupted mitochondrial inner membranes and cristae structures described in Rab11b KO enterocytes. Quantitative analysis of mitochondrial structure revealed that approximately 80%–90% mitochondria in Rab11b deficient Paneth cells and stem cells are abnormal (Figures 5E,F). Thus, the loss of electron density within the mitochondrial matrix induced by Rab11b loss appear to be consistent across different intestinal epithelial cell types. These findings collectively suggest that Rab11b is crucial for maintaining mitochondrial structure and function in gut epithelial cells.

Several prior studies have indicated that Rab proteins may play a role in regulating mitochondrial dynamics. For instance, it has been shown that depolarized mitochondria can be sequestered in Rab5-positive endosomes, which later mature into Rab7-positive late endosomes before being delivered to lysosomes for degradation (Hammerling et al., 2017). Furthermore, Rab8a has been identified as a mitochondrial receptor for lipid droplets (Ouyang et al., 2023), while Rab32 has been shown to coordinate mitochondrial fission (Alto et al., 2002).

Notably, RAB11A, which primarily localizes to the recycling endosomal compartment, appears to regulate the mitochondrial recruitment of the fission protein Drp1 through its effector FIP1C/RCP, which ultimately modulates mitochondrial fission and fusion events (Landry et al., 2014). However, the role of the closely related small GTPase RAB11B in mitochondrial function remains largely unexplored. Our collective observations from Rab11b genetic ablation and ultrastructural characterization confirmed a distinct mitochondrial defect in different IEC cell types. Our proteomic and live cell imaging analysis pointed to a potential physical intersection of Rab11b vesicles with mitochondrial protein complex. Loss of Rab11b may affect mitochondrial movement, assembly or fission, as reflected by the drastic disruption of mitochondrial membrane integrity and functions related to membrane potential and ROS production. The transcriptomic results from Rab11b knockout intestines suggested a potential tissue compensation in response to the impaired mitochondrial functions. Taken together, we propose a previously unappreciated contribution of RAB11B to the maintenance of mitochondrial structure and homeostasis in the gut epithelial cells.