henhaobio
Sign in

bioRxivCell BiologyPreprint

Spatial distribution of blood-brain barrier membrane proteins is controlled by sorting motifs and physiological signals in vivo

Joseph AmickiD, Shamika BhandarkariD, Douglas R. WilcoxiD, Solenn Boloker, Yi Daniel LuiD, Chenghua GuiD

PDFDOICode

In-depth summary

Claude Opus 5.5 · xhigh effort · 7 Nov 2026 · from the full text of version 2 (22 Sept 2026) · CC BY 4.0

STED imaging quantifies luminal versus abluminal localization of blood-brain barrier proteins in vivo; tyrosine motifs drive abluminal sorting of OAT3 and MHC I, and CD31 shifts day to night.

Where blood-brain barrier membrane proteins sit on the luminal (blood-facing) versus abluminal (brain-facing) endothelial surface has been hard to measure, because prior immuno-electron microscopy has few compatible antibodies and sparse labeling. This paper adds a quantitative, super-resolution, in vivo readout.

Key results

  1. ShownA STED readout around the endothelial nucleus recovers the expected polarity of reference proteins, validating the abluminal enrichment score (mean abluminal over mean luminal intensity).

    PODXL 96.1% luminal, enrichment 0.03 (5 mice); HA-OAT3 76.9% abluminal, enrichment 2.77 (4 mice); the non-targeted membrane control Alfa-PM 55.5% abluminal, enrichment 0.86 (3 mice). · Development of an approach to image and measure…; Quantification and statistical analysis

  2. ShownMouse brain endothelial proteins span a continuum from strongly luminal to strongly abluminal; some are non-polarized relative to the control.

    P-gp 98.2% luminal (0.02); ICAM2 86.5% luminal (0.11); MHC I 3.07; integrin α6 3.97; integrin β1 4.97; active integrin β1 16.76 (93.7% abluminal). Mfsd2a, GLUT1, TfR and Cav1 did not differ from Alfa-PM (nested ANOVA, Dunnett, log-transformed). · Spatial distribution of BEC membrane proteins in…; Quantification and statistical analysis

  3. ShownTyrosine residues in cytoplasmic domains are needed for abluminal enrichment of OAT3 (Y437A/V440E) and MHC I (Y342A); both mutants became roughly equally distributed between membranes.

    4 mice per construct; nested t-tests plus a linear mixed-effects model with animal as random effect. HA-MHC I matched endogenous MHC I, and mutant patterns differed from a perinuclear ER marker, arguing against ER retention. · Protein-intrinsic motifs control BBB membrane…; Quantification and statistical analysis

  4. SuggestedA computational screen of mouse BEC-expressed proteins produced a candidate list of cytoplasmic sorting motifs (YXXØ and [DE]XXXL[LI]), of which only two proteins were tested experimentally.

    15,813 BEC-expressed genes were filtered to 1,861 transmembrane, cell-membrane genes; 1,061 had YXXØ and 335 had [DE]XXXL[LI]. After human-ortholog and expression filters: 883, 237 and 218 with both. · Protein-intrinsic motifs control BBB membrane…; Computational search of potential soring motifs…

  5. ShownCD31 localization changes with time of day: nearly equal on both membranes at ZT2 and luminally enriched at ZT14. GLUT1, ICAM2, PODXL and Cav1 did not change significantly.

    6 mice per time point and protein, 10 cells per mouse; nested one-way ANOVA with Bonferroni correction; image analysis blinded to harvest time. · Extrinsic factors modulate membrane protein…; Day versus night measurements; Quantification and statistical analysis

Methods, data and reproducibility

2D STED imaging of immunostained mouse (and human postmortem) brain sections, tracing luminal and abluminal membranes around the endothelial nucleus in ImageJ/Fiji to compute percent distribution and an abluminal enrichment score. Motif function was tested with tagged wild-type and mutant proteins expressed in brain endothelial cells using the BI30 AAV capsid.

  • Mouse brain sections stained for endogenous membrane proteins or tagged constructs, measured per cell · mouse · n = 3 to 5 mice per protein; 6 mice per group for day/night
  • Human postmortem mesial temporal cortex (adults aged 36-72) stained for GLUT1, Cav1 and PODXL · human · n = 3 samples per protein
  • scRNA-seq dataset of brain endothelial cell expression used for the motif search, mapped to UniProt annotations · mouse · n = 15,813 genes
Methods & tools
Abberior Facility Line STED microscope, Abberior Lightbox software, ImageJ/Fiji, GraphPad Prism 9.4.0, MATLAB fitlme, Python (numpy, pandas, regular expressions), UniProt, BI30 AAV capsid, Alfa epitope tag, Transmission electron microscopy
Software & environment
GraphPad Prism 9.4.0; Abberior Lightbox software 2024.48.21878-gc86bbd647c; Python with numpy and pandas for the motif search.
Reproducibility gaps
No public repository for images is stated; gene fragment sequences, antibodies, motif lists and summary statistics are in Supplemental Tables 1 to 4 only.
Limitations (5)
  • Measurements are restricted to the area around the endothelial nucleus, and it has not been formally shown that this region represents the whole cell surface. · stated by the authors
  • The downstream trafficking machinery that reads the tyrosine motifs is not identified, and the mechanism and physiological relevance of the circadian CD31 shift remain unknown. · stated by the authors
  • The polarity categories are a continuum: Mfsd2a (p=0.0639) and CD31 (p=0.0320) sit on either side of the significance threshold despite similar enrichments. · stated by the authors
  • Motif dependence was tested for only two proteins, using virally expressed tagged constructs, and the day/night comparison covers five proteins at two time points. · noted in this summary
  • Experimenters were not blinded to group or protein identity outside the day/night experiment, and the human data are limited to three proteins in 3 samples each. · noted in this summary

Read this if you study polarized membrane protein trafficking in endothelium or want to judge luminal versus abluminal localization of candidate blood-brain barrier delivery targets.

Citations
2

Abstract

The blood-brain barrier (BBB), formed by brain endothelial cells (BECs), creates a safe and homeostatic environment for proper brain function. Together with pericytes and astrocytes, the BBB controls substance influx and efflux into and out of the brain. While BECs are extraordinarily thin, their luminal and abluminal plasma membranes face markedly different environments: the blood and brain parenchyma, respectively. How BBB membrane proteins are spatially distributed between these membranes and the factors regulating their localization are not clear. Here, we establish a method for measuring polarized protein sorting at the BBB in vivo. We characterize the distribution of transporters and receptors on the luminal and abluminal plasma membranes and identify protein-intrinsic motifs that control polarized sorting. Finally, we observe a change in membrane protein distribution that aligns with circadian rhythms, revealing an under-appreciated dimension of BBB dynamics. This method can be broadly used for evaluating membrane protein landscape changes during development and disease and for choosing optimal molecular targets for BBB-crossing therapeutics.

Reproducibility ledger

Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.

Sign in to file a reproducibility report.

Related papers

Similarity Map
SelectedSame fieldCross-field
THISPfau '21Wang '22Yousofvand '25Zhou '25Tian '20Shih '18
63% semantic matchNeuroscience
Sarah J. Pfau, Urs H. Langen, Theodore M. Fisher+6 more·bioRxiv·Apr 26, 2021· 13 citations