Abstract

   In the liver, mitochondria are exposed to different concentrations of
   nutrients due to their spatial positioning across the periportal and
   pericentral axis. How the mitochondria sense and integrate these
   signals to respond and maintain homeostasis is not known. Here, we
   combine intravital microscopy, spatial proteomics, and functional
   assessment to investigate mitochondrial heterogeneity in the context of
   liver zonation. We find that periportal and pericentral mitochondria
   are morphologically and functionally distinct; beta-oxidation is
   elevated in periportal regions, while lipid synthesis is predominant in
   the pericentral mitochondria. In addition, comparative
   phosphoproteomics reveals spatially distinct patterns of mitochondrial
   composition and potential regulation via phosphorylation. Acute
   pharmacological modulation of nutrient sensing through AMPK and mTOR
   shifts mitochondrial phenotypes in the periportal and pericentral
   regions, linking nutrient gradients across the lobule and mitochondrial
   heterogeneity. This study highlights the role of protein
   phosphorylation in mitochondrial structure, function, and overall
   homeostasis in hepatic metabolic zonation. These findings have
   important implications for liver physiology and disease.

   Subject terms: Nutrient signalling, Metabolism, Mitochondria
     __________________________________________________________________

   Kang et al. reveal structural and functional differences in
   mitochondria across the hepatic lobule. Mitochondrial distinct
   phosphoproteome influences their functions highlighting how nutrient
   availability helps to shape mitochondria zonation.

Introduction

   Mitochondria are highly dynamic organelles that play critical roles in
   cell physiology, including energy production through oxidative
   phosphorylation (OXPHOS), metabolic signaling pathways, and
   biosynthesis^[47]1–[48]5. In response to changes in the environment,
   such as nutrient availability, mitochondria undergo remodeling through
   fission, fusion, and mitophagy^[49]2,[50]4. These dynamic
   rearrangements in mitochondrial architecture are fundamental to
   mitochondrial function, morphology, and homeostasis^[51]4,[52]5.
   Further, these dynamic changes are modulated, in part, through
   post-translational modifications, like phosphorylation, allowing
   mitochondria to rapidly and reversibly adjust their metabolic
   output^[53]6–[54]8. In vivo, mitochondria in cells within organs are
   exposed to varying levels of nutrients due to their spatial positioning
   with respect to the blood supply. How this affects and/or regulates
   mitochondrial dynamics is not known.

   This is particularly important in the liver, a central metabolic organ
   that balances whole-body nutrient availability. Within the liver,
   hepatocytes are organized into polygonal units called lobules.
   Nutrient-rich blood flows unidirectionally into the lobule, entering
   through the hepatic artery and portal vein (which will hence be
   referred to as periportal; PP) and drains into a single central vein
   (pericentral; PC). Consequently, the hepatocytes are exposed to
   different levels of regional metabolites and metabolic burdens
   depending on their relative location within the lobule^[55]9,[56]10. It
   has been previously shown that these gradients direct hepatocytes in
   different parts of the lobule to express different genes, a phenomenon
   known as liver zonation^[57]10,[58]11. Wnt ligands secreted by PC
   endothelial cells are a major driver of zonal gene
   expression^[59]12–[60]14. Although single-cell RNA sequencing has
   enhanced our understanding of liver zonation^[61]15,[62]16, the
   functional consequences are often inferred solely from gene expression.

   Electron microscopy studies have revealed notable differences in
   mitochondrial morphology between PP and PC hepatocytes^[63]17–[64]20.
   These observations suggest that cells on the PP–PC axis, separated by
   up to a mere 300 μm, possess mitochondria with distinct functions.
   However, it is not known how spatial separation affects mitochondria
   functions in vivo. Furthermore, whether mitochondrial variations are
   determined genetically or continuously adjusted metabolically via
   dynamic nutrient gradients requires elucidation.

   In the present study, we aimed to define the relationship between
   mitochondrial function, structure, and spatial positioning in the
   hepatic lobule. We also sought to identify mechanisms that regulate
   mitochondrial diversity during homeostasis. Our results reveal hepatic
   mitochondrial zonation, combining structural and functional features
   with specialized mitochondrial subpopulations within the liver lobule.
   Further, they highlight the role of protein phosphorylation and
   nutrient sensing in dynamically tuning zonated mitochondrial functions.

Results

A comparative mitochondrial proteome of sorted hepatocytes

   To investigate how the spatial positioning of cells in the liver
   affects mitochondrial functions, PP and PC hepatocytes from the livers
   of four ad-lib-fed mice were enriched using unique surface markers and
   fluorescence-activated cell sorting (FACS; Fig. [65]1A–C). E-cadherin
   and CD73 antibodies were used to enrich PP and PC hepatocytes,
   respectively, and Western blots were performed for validation
   (Fig. [66]1D).

Fig. 1. Spatial enrichment of PP and PC hepatocytes.

   [67]Fig. 1
   [68]Open in a new tab

   A Schematic diagram depicting the workflow. Hepatocytes were isolated
   from the murine liver using two-step collagenase perfusion, after which
   unique surface markers were applied to label cells. Labeled cells were
   then enriched using fluorescence-activated cell sorting (FACS) to
   obtain hepatocytes from different zones. Illustration created using
   BioRender. B Immunofluorescence staining of a liver section showing the
   zonal distribution of CD73 (cyan) and E-cadherin (magenta).
   Representative image from three independent experiments. C
   Representative two-dimensional scatter plot of hepatocytes labeled with
   CD73 and E-cadherin. D Western blot of spatially sorted hepatocytes.
   Representative blot from n = 3 independent experiments. Source data is
   provided as a Source Data file for (D). PP periportal, PC pericentral,
   GS glutamine synthetase.

   Next, PP and PC hepatocytes were subjected to tandem mass tag
   (TMT)-based quantitative mass spectrometry for total proteome analysis.
   The goal was to gain insight into mitochondrial functions by
   establishing a quantitative map of mitochondrial protein abundance
   along the PP–PC axis. Principal component and hierarchical clustering
   analyses showed sample grouping based on spatial origin (PP or PC)
   (Supplementary Fig. [69]1A, B). Of 5018 proteins identified, 46% were
   zonated, meaning they had a biased expression toward PP or PC
   hepatocytes (Fig. S[70]1C, D; Supplementary Data [71]1). Pathway
   analysis highlighted PP and PC-restricted processes consistent with a
   previous study describing gene expression^[72]16 (Supplementary
   Fig. [73]1E, F).

   The list of quantified proteins was compared with the murine MitoCarta
   3.0 database, which consists of 1,140 proteins. We identified 829
   mitochondrial proteins, 422 of which were enriched in PP mitochondria
   and 113 in PC mitochondria (Fig. [74]2A, B). To gain insight into the
   functions of PP and PC mitochondria, the top 25 unique proteins were
   selected, and their location and pathway within the mitochondria were
   determined using MitoCarta 3.0 database (Fig. [75]2C, D). Selected
   mitochondrial proteins were also validated by immunofluorescence
   (Supplementary Fig. [76]2).

Fig. 2. Comparative mitochondrial proteome of spatially sorted hepatocytes.

   [77]Fig. 2
   [78]Open in a new tab

   A Total number of non-mitochondrial proteins (gray) and mitochondrial
   proteins (green) was detected by mass spectrometry (left; n = 4 mice).
   Percentage of periportal (PP), pericentral (PC), and unzonated (UZ)
   mitochondrial proteins (right); zonated expression based on a p-value
   (0.05). B Volcano plot shows the log[2] PC/PP fold-change (x-axis) and
   the −log[10] p-value (y-axis) for mitochondrial proteins. Proteomics
   data was analyzed with Limma R package (v3.40.6). C, D Spatial
   distribution of the top 25 PP or PC mitochondrial proteins. Proteins
   were color-coded to reflect their cellular function. Pathways were
   listed based on the frequency at which they appeared. E Abundance of
   representative nuclear and mitochondria-encoded respiratory chain
   proteins are shown with floating bar graphs. The center represents the
   mean, top and bottom represent maxima and minima. The p-values were
   calculated with Limma R package (v3.40.6). F
   Bioinformatic STRING analysis of the PC mitochondria proteomic data.
   The interaction map illustrates the functional association of PC
   mitochondrial metabolism with cytosolic lipid synthesis. Data presented
   as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

   Proteins enriched in PP mitochondria were primarily localized to the
   inner membrane or the mitochondrial matrix and were involved in amino
   acid metabolism or OXPHOS. Thus, we examined the spatial expression of
   selected OXPHOS proteins and found that PP mitochondria expressed
   higher levels of nuclear and mitochondrial-encoded OXPHOS components
   (Fig. [79]2E). In contrast, proteins enriched in PC mitochondria
   localized to the outer and inner mitochondrial membranes and matrix. In
   addition to regulating mitochondrial structure, dynamics, and contact
   with other organelles, many of the identified PC proteins are involved
   in lipid metabolism, detoxification, and carbohydrate metabolism
   (Fig. [80]2D). Notably, citrate synthase (CS), a tricarboxylic acid
   cycle (TCA cycle) enzyme, was highly expressed in PC mitochondria. In
   addition to the TCA cycle, when transported into the cytosol, citrate
   can be a precursor for lipid synthesis. Indeed, STRING analysis
   suggested a functional link between PC mitochondrial proteins involved
   in pyruvate metabolism and phospholipid membrane biosynthesis and lipid
   synthesis occurring in the cytosol (Fig. [81]2F).

PP mitochondria display enhanced bioenergetic capacity

   We next performed a functional evaluation of the PP and PC mitochondria
   to determine their bioenergetic capacity. Intravital microscopy^[82]21
   was used to examine mitochondrial membrane potential in the intact
   liver of anesthetized mice. Mitochondria were labeled with MitoTracker
   green and TMRE, with the former labeling mitochondria matrix and the
   latter indicating membrane potential. TMRE labeled mitochondria in PP
   regions more intensely, indicating higher membrane potential
   (Fig. [83]3A, B and Supplementary Fig. [84]3A, B). Isolated hepatocytes
   were labeled with Anti-CD73, Anti-E-cadherin antibodies, and JC1, a
   ratiometric fluorescent reporter of mitochondrial membrane potential.
   Consistent with the in vivo data, PP cells labeled in suspension had
   higher mitochondrial membrane potential than PC cells (Fig. [85]3C, D).
   Together, these data show that PP mitochondria have a higher membrane
   potential that is not disrupted by spatial sorting, allowing the use of
   these cells for further physiological evaluation.

Fig. 3. PP mitochondria display higher bioenergetic capacity.

   [86]Fig. 3
   [87]Open in a new tab

   A Intravital microscopy of the hepatic lobule labeled with
   tetramethylrhodamine ethyl ester (TMRE), and MitoTracker Green to
   evaluate the relative mitochondrial membrane potential. The scale bar
   is 100 μm. Similar trends were observed in three independent
   experiments. B Mitochondrial membrane potential was evaluated by
   measuring fluorescence intensity (AU) along a PP–PC axis line. C, D
   Measurement of mitochondrial membrane potential using JC1 and flow
   cytometry in spatially sorted hepatocytes. Dot plots of a
   representative experiment is shown. The bar graph shows similar trends
   in three independent experiments. Stacked bar plot of relative
   percentages of green or red positive cells from n = 3. Two-way ANOVA,
   Šidák multiple comparisons test p < 0.0001. E Oxygen consumption rate
   (OCR) in spatially sorted hepatocytes using the XF Mito Stress Test Kit
   and Seahorse XF96 Analyzer. Samples were normalized to cell number.
   Similar trends were observed in four independent experiments. F Maximum
   respiration capacity in spatially sorted hepatocytes expressed relative
   to PP. Data presented as mean ± SD from n = 4 independent experiments.
   Statistical significance was calculated with two-tailed unpaired
   Student’s t-test ***p = 0.0007. G, H Substrate dependency assay in
   spatially sorted hepatocytes using the MitoFuel Flex test. ATP
   production rate relative to PP is shown in cells treated with etomoxir,
   UK5099, or BPTES. Data presented as mean ± SD from n = 4 independent
   experiments. Statistical significance was calculated using one-way
   ANOVA, Dunnett’s multiple comparisons test; ns not significant;
   **p = 0.003. I ATP content relative to PP in spatially sorted
   hepatocytes using the colorimetric luciferase assay from three
   independent experiments. J Citrate synthase activity relative to PP in
   spatially sorted hepatocytes. Data presented as mean ± SD from n = 5
   independent experiments. Statistical significance was calculated using
   two-tailed unpaired Student’s t-test *p = 0.04. K Intracellular
   triglyceride (TG) concentration relative to PP in spatially sorted
   hepatocytes. Data presented as mean ± SD from n = 4 independent
   experiments. Statistical significance was calculated using two-tailed
   unpaired Student’s t-test **p = 0.002. Periportal (PP) is labeled in
   red; pericentral (PC) in blue.

   Subsequently, mitochondrial oxygen consumption rate and substrate
   preferences were evaluated using the Seahorse XF Analyzer in spatially