High-resolution imaging and quantification of plasma membrane cholesterol by NanoSIMS

Abstract

Cholesterol is a crucial lipid within the plasma membrane of mammalian cells. Recent biochemical studies showed that one pool of cholesterol in the plasma membrane is "accessible" to binding by a modified version of the cytolysin perfringolysin O (PFO*), whereas another pool is sequestered by sphingomyelin and cannot be bound by PFO* unless the sphingomyelin is destroyed with sphingomyelinase (SMase). Thus far, it has been unclear whether PFO* and related cholesterol-binding proteins bind uniformly to the plasma membrane or bind preferentially to specific domains or morphologic features on the plasma membrane. Here, we used nanoscale secondary ion mass spectrometry (NanoSIMS) imaging, in combination with ¹⁵N-labeled cholesterol-binding proteins (PFO* and ALO-D4, a modified anthrolysin O), to generate high-resolution images of cholesterol distribution in the plasma membrane of Chinese hamster ovary (CHO) cells. The NanoSIMS images revealed preferential binding of PFO* and ALO-D4 to microvilli on the plasma membrane; lower amounts of binding were detectable in regions of the plasma membrane lacking microvilli. The binding of ALO-D4 to the plasma membrane was virtually eliminated when cholesterol stores were depleted with methyl-β-cyclodextrin. When cells were treated with SMase, the binding of ALO-D4 to cells increased, largely due to increased binding to microvilli. Remarkably, lysenin (a sphingomyelin-binding protein) also bound preferentially to microvilli. Thus, high-resolution images of lipid-binding proteins on CHO cells can be acquired with NanoSIMS imaging. These images demonstrate that accessible cholesterol, as judged by PFO* or ALO-D4 binding, is not evenly distributed over the entire plasma membrane but instead is highly enriched on microvilli.

Keywords: NanoSIMS; anthrolysin O; cholesterol; microvilli; perfringolysin O.

Fig. 1.

NanoSIMS imaging of cholesterol-binding proteins on the plasma membrane of CHO-KI cells. CHO-K1 cells were plated on silicon wafers and grown in Ham’s F-12 medium containing 10% lipoprotein-deficient serum (Materials and Methods) for 5 d. The cells were then loaded with cholesterol as described in Materials and Methods. The cells were subsequently grown in medium lacking supplemental cholesterol for 44 h. Next, the cells were incubated with 20 μg/mL [¹⁵N]PFO* (A) or [¹⁵N]ALO-D4 (B) for 2 h at 4 °C. NanoSIMS images were generated based on secondary electrons (SEs); other images were created based on the ¹⁵N/¹⁴N ratio. (Scale bar, 10 μm.) The color scale shows the range of ¹⁵N/¹⁴N ratios. (C) High-magnification image of the cell shown in B. (Scale bar, 3 μm.) (D) Line scan demonstrating the ¹⁵N/¹⁴N isotope ratio across microvilli on the surface of the plasma membrane (white lines in C and D). Pixel, ∼19.5 nm. (E) SEM image of a CHO-K1 cell grown on a silicon wafer and fixed with 4% (vol/vol) PFA plus 1% glutaraldehyde followed by 2.5% (vol/vol) glutaraldehyde. A higher-magnification image of the boxed area on the Left is shown in the image on the Right. Red arrow shows a microvillus at the perimeter of the cell; white arrows show microvilli on the surface of the cell. (F) SEM image of CHO-K1 cells grown on a silicon wafer and fixed with 2.5% (vol/vol) glutaraldehyde in 0.1 M sodium cacodylate followed by 1% osmium tetroxide in 0.1 M sodium cacodylate. (Scale bar, 10 μm.)

Fig. 2.

NanoSIMS analysis of [¹⁵N]ALO-D4 binding to the plasma membrane of CHO-KI cells grown under standard conditions. CHO-K1 cells (imaged in A–C) were plated on silicon wafers and grown overnight and given either no treatment (A); treated with 10 mM methyl-β-cyclodextrin (+MβCD) for 15 min at 37 °C (B); or treated with 100 milliunits/mL of sphingomyelinase for 30 min at 37 °C (+SMase) (C). The cells were then washed and incubated with 20 μg/mL [¹⁵N]ALO-D4 for 2 h at 4 °C. NanoSIMS images were generated based on ¹²C¹⁴N^– ions (to visualize cell morphology) and on the ¹⁵N/¹⁴N ratio. (Scale bar, 10 μm.) The color scale shows the range of ¹⁵N/¹⁴N ratios. (D) ¹⁵N/¹⁴N ratios in microvilli (black solid circles) and nonmicrovilli regions (red solid circles) (n = 30) of the plasma membrane of two nontreated (NT) and SMase-treated (SMase) CHO-K1 cells. (E and F) Bar graphs depicting ¹⁵N/¹⁴N ratios in microvilli and nonmicrovilli regions of the plasma membrane in NT and SMase-treated cells. Data were analyzed with an unpaired Student’s t test with Welch’s correction.

Fig. 3.

NanoSIMS imaging of [¹⁵N]ALO-D4 binding to the plasma membrane of CHO-KI cells that had been loaded with cholesterol. CHO-K1 cells were plated on silicon wafers and grown for 5 d. The cells were then loaded with cholesterol by incubating the cells for 2 d in medium containing 300 μM cholesterol (Materials and Methods). The cells were then washed and grown without supplemental cholesterol for 44 h. Next, the cells were plated on silicon wafers, grown overnight, and then given no treatment (A); treated with 10 mM methyl-β-cyclodextrin for 15 min at 37 °C (+MβCD) (B); or treated with 100 milliunits/mL of sphingomyelinase for 30 min at 37 °C (+SMase) (C). The cells were then washed and incubated with 20 μg/mL [¹⁵N]ALO-D4 for 2 h at 4 °C. NanoSIMS images were generated based on secondary electrons (SEs) and on the ratio of ¹²C¹⁵N^– to ¹²C¹⁴N^– secondary ions (¹⁵N/¹⁴N). (Scale bar, 10 μm.) The color scale shows the range of ¹⁵N/¹⁴N ratios. (D) ¹⁵N/¹⁴N ratios in microvilli (black solid circles) and nonmicrovilli regions (red solid circles) (n = 30) of two nontreated (NT) and SMase-treated (SMase) cells. (E and F) Bar graphs of ¹⁵N/¹⁴N ratios in microvilli and nonmicrovilli regions on the plasma membrane of NT and SMase-treated cells. Data were analyzed with an unpaired Student’s t test with Welch’s correction.

Fig. S1.

NanoSIMS imaging of [¹⁵N]ALO-D4 binding to the polylysine-coated silicon wafer substrate. (A) Secondary electron (SE) image of a cholesterol-loaded CHO-K1 cell that had been treated with MβCD, showing the cell body (Upper Right) and microvilli projecting from the edge of the cell (Lower Left). (B) A ¹⁵N/¹⁴N ratio NanoSIMS image reveals low levels of ¹⁵N enrichment on the polylysine-coated silicon wafer substrate between microvilli projections at the edge of a cell. (C) Graph showing the ¹⁵N/¹⁴N ratio in different regions of the cell body of CHO cells and in the adjacent cell-free polylysine-coated silicon wafer substrate. We calculated the ¹⁵N/¹⁴N ratio in 20 regions of the cell body (circles) and 20 cell-free regions of the silicon wafer substrate (diamonds) surrounding two MβCD-treated cells (black) and two nontreated cells (NT; red). The ¹⁵N/¹⁴N ratio in MβCD-treated cells fell to levels that were very close to the natural abundance of ¹⁵N (which is 0.0037), reflecting the known capacity of MβCD to unload cholesterol from cells. In the absence of MβCD, the ¹⁵N/¹⁴N ratio in the cell body was ∼0.033. MβCD treatment had little or no effect on nonspecific binding of [¹⁵N]ALO-D4 to the polylysine-silicon wafer substrate. The biochemical basis for ALO-D4 binding to the polylysine-coated silicon wafer substrate is unknown but presumably involves electrostatic interactions that are insensitive to MβCD.

Fig. 4.

NanoSIMS analysis of [¹⁵N]ALO-D4 binding to CHO-K1 cells. CHO-K1 cells were plated on Thermanox plastic coverslips and grown for 5 d. The cells were then loaded with cholesterol as described in Materials and Methods. Coverslips received no treatment (A); treatment with 10 mM methyl-β-cyclodextrin for 15 min at 37 °C (+MβCD) (B); or treatment with 100 milliunits/mL of sphingomyelinase for 30 min at 37 °C (+SMase) (C). The cells were then washed and incubated with 20 μg/mL [¹⁵N]ALO-D4 at 4 °C for 2 h. Next, the cells were fixed, dehydrated, resin embedded, and sectioned. NanoSIMS images were generated based on ¹²C¹⁴N⁻ secondary ions (to define cell morphology) and ¹⁵N/¹⁴N ratios {to visualize binding of [¹⁵N]ALO-D4}. Peaks in ¹⁵N/¹⁴N ratios on the line graphs are centered above the plasma membrane. (Scale bar, 10 μm.) (D–F) Line graphs showing ¹⁵N/¹⁴N ratios across cells.

Fig. 5.

NanoSIMS imaging of [¹⁵N]lysenin binding to CHO-K1 cells. CHO-K1 cells were plated on silicon wafers. (A) The cells received either no treatment (NT); treatment with 100 milliunits/mL of sphingomyelinase for 30 min at 37 °C (+SMase); or treatment with 10 mM MβCD for 15 min (+MβCD) at 37 °C. The cells were then washed and incubated with 20 μg/mL [¹⁵N]lysenin for 1 h at 4 °C. NanoSIMS images were generated based on ¹²C¹⁴N⁻ secondary ions (to define cell morphology) and on the ¹⁵N/¹⁴N ratio {to visualize [¹⁵N]lysenin binding}. (Scale bar, 10 μm.) (B) ¹⁵N/¹⁴N ratios in microvilli (black solid circles) and nonmicrovilli regions (red solid circles) (n = 60) of nontreated and MβCD-treated cells. (C and D) Bar graphs depicting ¹⁵N/¹⁴N ratios in microvilli and in nonmicrovilli regions of NT cells and MβCD-treated cells. Data were analyzed with a Student’s t test with Welch’s correction and with a Mann–Whitney test. Both tests yielded the same level of statistical significance.

Fig. 6.

Immunofluorescence microscopy to assess the binding of ALO-D4 and mCherry-lysenin to CHO-K1 cells. CHO-K1 cells were plated on glass coverslips and grown overnight before receiving one of three treatments: incubating cells in medium alone at 37 °C for 30 min (i.e., no treatment); incubating cells with medium containing 100 milliunits/mL of SMase at 37 °C for 30 min (+SMase); or incubating cells with medium containing 10 mM MβCD at 37 °C for 15 min (+MβCD). Binding of ALO-D4 and lysenin to the surface of cells was assessed by confocal microscopy as described in SI Materials and Methods. Cell nuclei were visualized with DAPI (blue). (Scale bar, 20 μm.)