An mTurq2-Col4a1 mouse model allows for live visualization of mammalian basement membrane development

Abstract

Basement membranes (BMs) are specialized sheets of extracellular matrix that underlie epithelial and endothelial tissues. BMs regulate the traffic of cells and molecules between compartments, and participate in signaling, cell migration, and organogenesis. The dynamics of mammalian BMs, however, are poorly understood, largely due to a lack of models in which core BM components are endogenously labeled. Here, we describe the mTurquoise2-Col4a1 mouse in which we fluorescently tag collagen IV, the main component of BMs. Using an innovative planar-sagittal live imaging technique to visualize the BM of developing skin, we directly observe BM deformation during hair follicle budding and basal progenitor cell divisions. The BM's inherent pliability enables dividing cells to remain attached to and deform the BM, rather than lose adhesion as generally thought. Using FRAP, we show BM collagen IV is extremely stable, even during periods of rapid epidermal growth. These findings demonstrate the utility of the mTurq2-Col4a1 mouse to shed new light on mammalian BM developmental dynamics.

Figure 1.

mTurq2-Col4a1 endogenously tagged reporter design and basement membrane localization in adult mouse tissues. (A) Schematic of the COL4A1 subunit and type IV collagen trimeric protomer with an N-terminal mTurq2 fluorescent tag. (B) Schematic of the Mus musculus Col4a1 genomic locus and mTurq2 insertion site. The mTurq2 gene was inserted into Exon 2 between the coding sequences for the signal peptide and 7S domain. (C) Sagittal section of dorsal skin from mTurq2-Col4a1/+ adult mouse labeled with ColIV antibodies (magenta) and Hoechst to mark nuclei (blue; grayscale in C’ and C’’). (C’) mTurq2-COL4A1 (cyan) localizes to BM surrounding the hair follicle, where it overlaps with ColIV staining (C’’, magenta). Scale bars, 100 µm. (D–D’’) High magnification view of mTurq2-COL4A1 (cyan) localization in the interfollicular epidermis from mTurq2-Col4a1/+ adult mouse labeled with ColIV (magenta) and E-Cadherin (E-Cad; grayscale). Scale bar, 20 µm. (E–E’’) High magnification view of mTurq2-COL4A1 (cyan) localization in the interfollicular epidermis from mTurq2-Col4a1/+ adult mouse labeled with Perlecan (magenta) and E-Cadherin (E-Cad; grayscale). Scale bar, 20 µm. (F) Kidney section from mTurq2-Col4a1/+ adult mouse labeled with ColIV (magenta) and Hoechst to mark nuclei (blue). Scale bar, 100 µm. (G–G’’) High magnification view of mTurq2-COL4A1 (cyan) localization in the kidney of mTurq2-Col4a1/+ adult mouse labeled with ColIV (magenta in G; grayscale in G’) Hoechst (blue). mTurq2-COL4A1 expression can be seen in renal tubules (arrows), Bowman’s capsule (arrowheads), and the mesangial matrix (asterisks). Scale bar, 100 µm. (H) Kidney section from mTurq2-Col4a1/+ adult mouse labeled with Perlecan (magenta) and Hoechst (blue). Scale bar, 100 µm. (I–I’’) High magnification view of mTurq2-COL4A1 (cyan) localization in the kidney of mTurq2-Col4a1/+ adult mouse labeled with Perlecan (magenta in G; grayscale in G’) and Hoechst (blue). mTurq2-COL4A1 expression can be seen in renal tubules (arrows), Bowman’s capsule (arrowheads), the mesangial matrix (asterisks), and the collecting duct (CD). Scale bar, 100 µm.

Figure S1.

Related to Fig. 1. mTurq2-Col4a1 basement membrane localization in adult epidermis and kidney. (A–F) Additional representative images from a second biological replicate mTurq2-Col4a1/+ adult mouse. (A and B) Dorsal skin sections from mTurq2-Col4a1/+ adult mouse labeled with ColIV (A and A’; magenta) or Perlecan (B and B’, magenta) and E-Cadherin (grayscale). mTurq2-COL4A1 (cyan) localizes to the basement membrane underlying the interfollicular epidermis and the dermal vasculature. Scale bars, 20 µm. (C) Additional biological replicate of kidney sections from mTurq2-Col4a1/+ adult mouse labeled with ColIV antibodies (magenta) and Hoechst to mark nuclei (blue). Scale bar, 100 µm. (D–D’’) High magnification view of mTurq2-COL4A1 (cyan) localization in the kidney of second biological replicate mTurq2-Col4a1/+ adult mouse labeled with ColIV (magenta in D; grayscale in D’) Hoechst (blue). Scale bar, 100 µm. (E) Kidney section from mTurq2-Col4a1/+ adult mouse labeled with Perlecan antibodies (magenta) and Hoechst (blue). Scale bar, 100 µm. (F–F’’) High magnification view of mTurq2-COL4A1 (cyan) localization in the kidney of mTurq2-Col4a1/+ adult mouse labeled with Perlecan (magenta in F; grayscale in F’) Hoechst (blue). Scale bar, 100 µm.

Figure 2.

E18.5 mTurq2-Col4a1 mouse backskin shows correct and robust expression of mTurq2-COL4A1 in the dermal–epidermal junction basement membrane. (A) Average intensity projection of backskin from WT E18.5 embryo labeled with Hoechst (merge, top), anti-laminin β-1 antibody (A’), and anti-collagen IV antibody (A’’). No signal was observed in the mTurq2 channel (A’’’). (B) As in A, except backskin from a mTurq2-Col4a1/+ embryo. Robust mTurquoise2 signal recapitulates the labeling pattern of the collagen IV antibody. (C) Pearson’s correlation coefficient (r) measuring the correlation between laminin β-1 and collagen IV expression (antibody labeling) in WT and mTurq2-Col4a1/+ E18.5 backskins as labeled. Larger datapoints—biological replicates (n = 3 embryos), smaller datapoints—individual images per replicate (n = 10 per embryo). (D) Average intensity projection of backskin from WT E18.5 embryo labeled with Hoechst (merge, top), anti-laminin β-1 antibody (D’), and anti-perlecan antibody (D’’). (E) As in D, except backskin from a mTurq2-Col4a1/+ embryo. (F) As in C, except measuring the correlation between laminin β-1 and perlecan. Scale bar, 20 µm. ns = not significant, Mann–Whitney U test.

Figure 3.

Normal ultrastructural organization of the epidermis and dermal-epidermal junction in mTurq2-Col4a1/+ embryonic skin. Transmission electron micrographs of ultrathin skin sections from E18.5 WT control (A, C, E, and G) and mTurq2-Col4a1/+ (B, D, F, and H) embryos. (A and B) Ultrastructural overview of skin architecture. Der = dermis, BL = basal layer, SL = spinous layer, GL = granular layer, SC = stratum corneum, HF = hair follicle. The dotted line denotes dermal–epidermal boundary. Scale bars, 10 µm. (C and D) Zoomed in view of the boundary between the basal epithelial layer and dermis. Dermal–epidermal junction (DEJ) is highlighted in pink. Dotted boxes denote regions magnified in E and F. Scale bars, 2 µm. (E and F) Two representative examples of DEJ region of the skin from WT (E and E’) and mTurq2-Col4a1/+ (F and F’) embryos. Nuc = nucleus. Scale bars, 500 nm. (G and H) Zoomed in views of individual hemidesmosomes at the DEJ. HD = hemidesmosome, PM = plasma membrane, LL = lamina lucida, LD = lamina densa. Scale bars, 100 nm.

Figure S2.

Related to Fig. 3. Ultrastructural organization of the epidermis and dermal–epidermal junction in mTurq2-Col4a1/+ embryonic skin. (A–D) Representative TEM images from additional experimental replicates of E18.5 WT control (A and B) and mTurq2-Col4a1/+ (C and D) embryonic skin. Der = dermis, BL = basal layer, SL = spinous layer, GL = granular layer, SC = stratum corneum. The dotted line denotes dermal–epidermal boundary. Scale bars, 10 µm. (E–H) Representative images of individual hemidesmosomes at the dermal–epidermal junction region of WT control (E and F) and mTurq2-Col4a1/+ (G and H) embryonic skin. HD = hemidesmosome, PM = plasma membrane, BM = basement membrane. Scale bars, 250 nm.

Figure 4.

Spatial distribution and correlation between key basement membrane components in mTurq2-Col4a1/+ embryonic backskins resembles that of WT littermates. (A) Average intensity projection (XZ optical reconstruction) of backskin from WT E15.5 embryo labeled with Hoechst (merge, top), anti-laminin β-1 antibody (A’), and anti-collagen IV antibody (A’’). No signal is observed in the mTurq2 channel (A’’’). (B) As in A, except backskin from a mTurq2-Col4a1/+ embryo. Robust mTurq2-COL4A1 signal recapitulates the labeling pattern of the collagen IV antibody. (C) Pearson’s correlation coefficient (r) measuring the correlation between laminin β-1 and collagen IV expression (antibody labeling) in WT and mTurq2-Col4a1/+ E15.5 backskins as labeled. Larger datapoints—biological replicate means (n = 3 embryos), smaller datapoints—individual images per replicate (n = 10 per embryo). (D) Fraction of laminin β-1 contribution to the laminin β-1:collagen IV overlap in E15.5 backskins; refer to the main text for methodology. (E) As in D, except collagen IV contribution to laminin β-1:collagen IV overlap. Dots represent individual images (n = 10 per embryo) and boxes denote the minimum and maximum mean per biological replicate (n = 3 embryos). (F) Average intensity projection (XZ optical reconstruction) of backskin from WT E15.5 embryo labeled with Hoechst (merge, top), anti-laminin β-1 antibody (F’), and anti-perlecan antibody (F’’). (G) As in F, except backskin from an mTurq2-Col4a1/+ embryo. (H–J) As in C–E, except measuring the correlation and overlap fractions between laminin β-1 and perlecan. (K–T) As in A–J, except backskins are from E12.5 embryos. Scale bars, 20 µm. ns = not significant, Mann–Whitney U test.

Figure 5.

PS Multiview imaging of mTurq2-Col4a1 mouse model. (A) Schematic of established live imaging method. The skin explant is mounted on an agarose pad and imaged using spinning disc microscopy. To visualize the tissue in Z, the image is software-reconstructed. Scale bar, 50 µm. (B) Schematic of PS Multiview live imaging. The skin explant is mounted as previous, but over an agarose ridge. Both planar (XY) and sagittal (“XZ”) focal planes are then imaged. Scale bar, 50 µm. (C) Snapshots taken from a time-lapse live imaging movie of hair follicle invagination in E14.5 mTurq2-Col4a1/+;mTmG/mTmG backskin. (D) Representative image showing mean fluorescence intensity calculation methodology. Orange brackets highlight the hair follicle base region where signal intensity was measured. Magenta brackets highlight the IFE region where the signal intensity was measured. Scale bar, 50 µm. (E) Quantification of mTurq2-COL4A1 signal in IFE, placode, and hair germ stage. Each dot represents an individual fluorescence intensity measurement as shown in D. Larger dots show means of independent experiments. Error bars = SEM. ***P = <0.001, Mann–Whitney U test. ns = not significant. (F) Snapshots taken from a time-lapse live imaging movie of basal cell division in E15.5 mTurq2-Col4a1/+;mTmG/mTmG backskin. Interphase, metaphase, and post-cytokinesis stages are shown (left to right). The bottom panel shows the same images with membrane curvature measurements. Scale bar, 10 µm. (G) Quantification of the average angle of BM deformation as shown in F. Each dot represents an average of two angles. n = 23 dividing cells. Error bars = mean + SD. *P = <0.05 (0.0366), **P = <0.01, ***P = <0.001, one-way ANOVA with Dunn’s multiple comparisons test. (H) Snapshots taken from time-lapse live imaging movies of placode, late-placode, and hair germ stages as labeled, in E14.5 and E15.5 Col4a1/+;mTmG/mTmG backskins. The bottom panel shows separate channels, note clear mTurq2-COL4A1 “baskets” surrounding the developing dermal condensate. Scale bar, 50 µm.

Figure S3.

Related to Fig. 5. The epidermal BM deforms around dividing basal progenitor cells. (A) Additional example of snapshots taken from time-lapse live imaging movie of basal cell division in E15.5 mTurq2-Col4a1/+;mTmG/mTmG backskin. Interphase, metaphase, and post-cytokinesis stages are shown (left to right). Bottom panel shows same images with membrane curvature measurements shown. Scale bar, 10 µm. (B) Additional example as A. (C) Quantification of the average angle of BM deformation as shown in Fig. 5 F. Data is normalized to the interphase angle for each individual dividing cell. Each line represents one dividing cell. Dark blue line represents the average of all dividing cell lines. n = 23 dividing cells. Error bars = mean + SD. (D) Average intensity projection of backskin from mTmG/mTmG E16.5 embryo labeled with Hoechst (merge, D’) and anti-β4 integrin antibody (merge, D”). (E–E”) Average intensity projection (XZ optical reconstruction) of the image shown in D. (F and G) As D and E, except showing an image taken at 4× zoom. Scale bars, 10 µm. (H–J) Representative TEM images from E18.5 mTurq2-Col4a1/+ embryonic skin showing a dividing basal progenitor cell maintaining hemidesmosome adhesions with the BM during mitotic rounding. BL = basal layer, DEJ = dermal-epidermal junction, HD = hemidesmosome, PM = plasma membrane. Dotted boxes (‘) and (“) in H denote regions magnified in I and J, respectively. Scale bars, 2 µm (H), 100 nm (I), 50 nm (J).

Figure 6.

mTurq2-COL4A1 is highly stable within the embryonic epidermal basement membrane. (A) Planar overview of mTurq2-COL4A1 localization in E15.5 whole mount epidermis. White boxes indicate BM lying at the dermal–epidermal junction of the interfollicular epidermis (IFE) and along the rim of a budding hair follicle (HF). Scale bar, 25 µm. See also Video 3. (B) Still images from representative fluorescence recovery after photobleaching (FRAP) experiments of mTurq2-COL4A1 within the IFE (top panels) and rim of a hair follicle (bottom panels). Scale bars, 10 µm. (C) FRAP recovery curves of mTurq2-COL4A1 within the IFE at E13.5 (grey circles; n = 36), E15.5 (blue squares; n = 32), and at the rim of E15.5 hair follicles (magenta triangles; n = 30). Gray dotted box indicates the region rescaled in C’. Each point is the mean normalized intensity. Values were acquired across n > 3 biological replicates. Error bars = SD.