Subscriber access provided by TUFTS UNIV
Communication
Spectrally Resolved, Functional Super-resolution Microscopy Reveals Nanoscale Compositional Heterogeneity in Live-cell Membranes Seonah Moon, Rui Yan, Samuel Kenny, Yennie Shyu, Limin Xiang, Wan Li, and Ke Xu J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.7b03846 • Publication Date (Web): 04 Aug 2017 Downloaded from http://pubs.acs.org on August 4, 2017
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of the American Chemical Society is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 5
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Journal of the American Chemical Society
Spectrally Resolved, Functional Super-resolution Microscopy Reveals Nanoscale Compositional Heterogeneity in Live-Cell Membranes Seonah Moon,†,∥ Rui Yan,†,∥ Samuel J Kenny,† Yennie Shyu,† Limin Xiang,† Wan Li,† Ke Xu*,†,‡ †Department of Chemistry, University of California, Berkeley, California 94720, United States ‡Division of Molecular Biophysics and Integrated Bioimaging, Lawrence Berkeley National Laboratory, Berkeley, California 94720, United States
Supporting Information Placeholder ABSTRACT: By recording the full fluorescence spectra and superresolved positions of ~106 individual polarity-sensing solvatochromic molecules, we reveal compositional heterogeneity in the membranes of live mammalian cells with single-molecule sensitivity and ~30 nm spatial resolution. This allowed us to unveil distinct polarity characteristics of the plasma membrane and the membranes of nanoscale intracellular organelles, a result we found to be due to differences in cholesterol levels. Within the plasma membrane, we observed the formation of low-polarity, raft-like nanodomains upon cholesterol addition or cholera-toxin treatment, but found this nanoscale phase separation absent in native cells. The ultimate sensitivity achieved through examining the spectra of individual molecules thus opens the door to functional interrogations of intracellular physicochemical parameters at the nanoscale.
Being chemically interconnected by diffusion and vesicular transport, cellular membranes are nevertheless compositionally and functionally heterogeneous.1 Biochemical studies indicate remarkable compositional differences between the isolated plasma membrane and organelle membranes,1a,2 but their native organization in living cells is difficult to visualize3 given the nanoscale dimensions of organelles. For the plasma membrane per se, research over the past two decades has debated the possible coexistence of liquid-ordered (Lo) and liquiddisordered (Ld) membrane domains, but this proposed phase separation at the nanoscale remains indeterminate.4 The difficulty of probing nanoscale heterogeneity in live-cell membranes arises from a lack of means to visualize it in an unbiased manner.4c,4d,5 Labeling specific membrane components or phases with fluorescent probes is subject to labeling specificity and may shift the native equilibrium by stabilizing or disrupting the target being labeled. Environment-sensitive fluorescent probes,6 in particular solvatochromic fluorophores that exhibit spectral shifts in media of varied chemical polarity,7 provide a possibility to sense membrane heterogeneity without the need to label (and so potentially disturb) a specific target. Instead, the fluorophore may indiscriminately sample the membrane and reports local polarity through spectral changes: a lower local polarity corresponds to less membrane hydration and thus more orderly packed lipids.5b,8 With conventional detection methods, however, the
diffraction of light limits spatial resolution to ~300 nm, and it is still difficult to obtain the full fluorescence spectra for every pixel. The low spatial resolution also limits polarity sensitivity as local differences in spectrum are averaged over all probes in the diffraction-limited volume. The ultimate sensitivity of environment-sensitive probes may be reached if each probe molecule is individually examined, thus avoiding the averaging of potentially distinct spectra of different molecules. Taking advantage of stochastic fluorescence blinking, super-resolution microscopy (SRM) techniques9 like STORM (stochastic optical reconstruction microscopy)10 and PAINT (points accumulation for imaging in nanoscale topography)11 localize single molecules with high spatial resolution. Recently, SRM has been extended to also capture spectral information.12 In particular, by recording the emission spectra and positions of ~106 blinking single molecules, with spectrally resolved STORM (SR-STORM) we demonstrated in fixed cells excellent spectral separation and spatial resolution for four fluorescent probes 10 nm apart in emission wavelength.12a A recent study12d combined Nile Red, a solvatochromic dye, with SR-PAINT to enable surface hydrophobicity mapping. By achieving Nile Red-based SR-STORM and SR-PAINT for livecell membranes, we here detected local variations in membrane chemical polarity with single-molecule sensitivity and ~30 nm spatial resolution. This functional (as opposed to conventional, shape-only) SRM approach enabled us to directly visualize nanoscale compositional heterogeneity in the membranes of live mammalian cells. We originally realized SR-STORM with a dual-objective design in which the sample was sandwiched between two opposing objective lenses.12a For compatibility with live cells, in this study we built a new system based on an inverted microscope (Figure 1a and Supplementary Information). Fluorescence blinking of single Nile Red molecules was achieved with a 561 nm excitation laser through photoswitching (STORM)13 or reversible binding to membranes (PAINT).11 Obtained single-molecule fluorescence was split into two light paths. Path 1 provided the position of each molecule, and a dispersive prism was inserted into Path 2 to generate spectra of the same molecules. The resultant single-molecule images and spectra in the wide-field were concurrently recorded on two different areas of an EM-CCD at 110220 frames per second (4.5-9 ms integration per frame; Figure 1bc and
ACS Paragon Plus Environment
Journal of the American Chemical Society
1.0
Path 1
M
L
EMCCD
M
L
Path 2 Slit L
BS
Prism
M
Path 2
1 2 µm
1
2 3
0.0
e
t1
2 µm
Molecule 1 Molecule 2 Molecule 3 Average 580
600 620 640 660 Wavelength (nm)
680
t1 + 30 s
d
0.5
2 3
c
0.5
1.0
b Path 1
Fluorescence Intensity
a
0.0
DOPC Only DOPC:SM DOPC:SM:Chol 580
600 620 640 660 Wavelength (nm)
680
Figure 1. SR-STORM/SR-PAINT with Nile Red. (a) Schematic of setup. Slit is at the camera port of an inverted microscope. L, lens; BS, beam splitter; M, mirror. (b) A small region of the concurrently acquired images and spectra of single Nile Red molecules in a DOPC bilayer, obtained in a 6-ms snapshot. Crosses denote the mapped spectral positions of 590 nm for each molecule. (c) Spectra of the 3 molecules in (b), compared to that averaged from 280,898 single molecules from the same sample. (d) The averaged spectra for single Nile Red molecules labeled to supported lipid bilayers of different compositions (DOPC, DOPC:SM 1:1, and DOPC:SM:Chol 1:1:1). (e) Sequential STORM images of a Nile Red-labeled live COS-7 cell at 30 s separation. Magenta, green, and cyan arrows point to structural changes in the plasma, ER, and mitochondrial membranes, respectively. We first applied Nile Red-based SR-PAINT to supported lipid bilayers of different compositions. Averaged spectra of the measured single Nile Red molecules (Figure 1d) showed the reddest spectrum for a bilayer of the unsaturated lipid 1,2-dioleoyl-sn-glycero-3-
a
2 µm
b
2 µm
t1
phosphocholine (DOPC), a model system for the Ld phase of the plasma membrane.15 A bilayer of a 1:1 mixture of DOPC and sphingomyelin (SM), a more saturated lipid, showed a modest (~5 nm) blue-shift. Further addition of cholesterol (Chol) led to a substantial (~20 nm) blue-shift. These results are consistent with previous measurements on model lipid vesicles:5b,8,15 Nile Red exhibits fluorescence blueshifts in media of reduced polarity; the DOPC-only bilayer is the least orderly packed and so the most hydrated and polar, whereas cholesterol assists the packing of sphingolipids into more ordered and less hydrated membrane phases. We next achieved Nile Red-based STORM and PAINT for membranes in live mammalian cells. For STORM, cells were labeled with 100 nM Nile Red, and imaged in a buffer containing ascorbic acid to assist photoswitching. For PAINT, unlabeled cells were imaged in a buffer containing 3 nM Nile Red for reversible binding to the membrane during imaging. For both approaches, the cell plasma membrane and the membranes of intracellular organelles were well labeled and visualized at the nanoscale [Figure 1e; arrows point to example endoplasmic reticulum (ER) (elongated tubular structures), mitochondria (thicker lumps), and plasma membrane]. With 50% fluorescence split to the image channel, we detected ~800 photons per molecule. This value is comparable to that of several known membrane STORM dyes, and translates to a spatial resolution of ~30 nm (localization precision in full width at half maximum).16 Consistent with this resolution, the thinner ER tubules appeared 50-100 nm in width in our images, in agreement with previous results.16 Three-dimensional STORM/PAINT indicated that for the plasma membrane, it was typically the top (apical) membrane that was imaged, which went out of the focal range for thicker parts of the cell (Figure S1). STORM/PAINT image sequences were obtained at ~30 s time resolution, which allowed us to track the morphological evolution of cellular membranes far beyond the diffraction limit (Figure 1e and Movie S2). Integrating Nile Red-based cell-membrane imaging with SRSTORM/SR-PAINT, we next asked whether the lipid compositionsensing capability we demonstrated for supported bilayers could reveal possible heterogeneities in the membrane composition of live mammalian cells. Remarkably, we observed significant spectral differences between Nile Red molecules at the plasma membrane and at the intracellular membranes of nanoscale organelles. To present both the spectral and spatial information of every detected Nile Red molecule (~106 total for a typical image), we calculated the spectral mean of each molecule as the intensity-weighted average of
c
t1 + 1 min
1.0
Ave. single-molecule spectrum
Movie S1 for representative raw data at a supported lipid bilayer and a live cell, respectively). With ~30 molecules detected in each frame (probe density 106 single Nile Red molecules in a few minutes, thus enabling the reconstruction of super-resolution SRSTORM/SR-PAINT images that carried functional information on local chemical polarity. Concurrent positional and spectral recording may also be achieved via the zeroth and first diffraction orders of a grating,12c,12d,14 but gratings achieve low light efficiency, and the strong dispersion makes it difficult to probe densely labeled two-dimensional structures like cellular membranes.
Ave. single-molecule spectra
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 2 of 5
d PM MT ER DOPC DOPC:SM DOPC:SM:Chol
0.5
0.0
580 600 620 640 660 680 Wavelength (nm)
Spectral position of single molecules 612 nm 648 nm Low
Polarity
High
Figure 2. Spectrally resolved, functional SRM visualizes polarity differences between organelle and plasma membranes in live cells. (a) True-color SR-PAINT image of a Nile Red-labeled live PtK2 cell. Each detected single molecule is color-coded according to its spectral mean [color bar below (d)]. (b,c) Sequential true-color SR-STORM images of a Nile Red-labeled live COS-7 cell at 1 min separation. Arrows point to notable structural changes in ER. (d) Averaged spectra of single Nile Red molecules from different nanoscale regions in live cells at the plasma membrane (PM), mitochondria (MT), and ER, compared to that at model supported lipid bilayers of different compositions.
ACS Paragon Plus Environment
wavelength,12a,17 and used this value to assign a color on a continuous scale (612-648 nm) as we plotted the position of each molecule. The resultant “true-color”12a SRM images showed strikingly different colors for the plasma membrane (blue) and organelle membranes (yellow) (Figure 2a-c). Image sequences further showed that as the plasma and organelle membranes underwent dynamic structural rearrangements at the nanoscale, their respective spectral characteristics were maintained (Figure 2bc, Movie S3, and Figure S2). Highly similar results were obtained from SR-PAINT and SR-STORM (Figure 2a-c and Figure S3), and across different cell types (Figure 2a: PtK2, rat kangaroo epithelial cell; Figure 2bc: COS-7, monkey fibroblast). Fixed cells exhibited similar spectral characteristics as live cells (Figure S4), indicating that the membrane compositions are stable upon chemical fixation. The locally averaged single-molecule spectra, as computed from nanoscale subareas of the SR-STORM/SR-PAINT data, were nearly identical for mitochondrial and ER membranes, but showed a strong blueshift of ~20 nm for the plasma membrane (Figure 2d), with results from different cells being identical (Figure S5). Distribution of the measured single-molecule spectral means (Figure S6) showed similar standard deviations for the mitochondrial and plasma membranes over a DOPC bilayer (~6 nm), likely limited by our spectral precision. A slightly larger standard deviation of 7 nm was observed for the ER membrane. Together, these results reveal fundamental differences between the organelle and plasma membranes: the redder spectra of organelle membranes suggest that they are physically more polar, structurally less ordered, and functionally more flexible and permeable to water, consistent with their intracellular functions.1a,2a
a
c
d
and high (~30-40%) cholesterol levels,1a,2 respectively. As cholesterol assists the packing of lipid bilayers into more ordered and less hydrated membrane phases,1a,5b,8 it may explain the significant differences in single-molecule spectra we observed between organelle and plasma membranes. Indeed, a comparison of our SR-STORM/SR-PAINT results on live-cell membranes and supported lipid bilayers showed that the organelle membranes are spectrally similar to cholesterol-free bilayers, whereas the plasma membrane is spectrally similar to the DOPC:SM:Chol (1:1:1) bilayer (Figure 2d). To understand whether cholesterol is indeed the driving force behind the membrane polarity differences we observed, we next combined Nile Red-based SR-STORM with cholesterol manipulation via methyl-β-cyclodextrin (MβCD).18 Depleting cholesterol with MβCD led to a strong redshift of Nile Red single-molecule spectra at the plasma membrane but little change at organelle membranes, as evidenced by both true-color SRM images and locally averaged single-molecule spectra (Figure 3ab and Figure S7 for fixed and live cells, respectively). In contrast, upon addition of 1 mM water-soluble cholesterol (cholesterol–MβCD), the organelle membrane spectra blue-shifted markedly to become closer to that of the plasma membrane as the latter remained spectrally unchanged (Figure 3ce and Figure S8 for fixed and live cells, respectively). While live cells were resistant to further addition of cholesterol (Figure S8), for fixed cells substantial blue-shifts were observed for both the plasma and organelle membranes with 5 mM water-soluble cholesterol, so that both became bluer than that of the untreated plasma membrane (Figure 3de). A portion of the organelle membranes, however, appeared resistant to cholesterol addition, as indicated by nanoscale regions with redder colors in true-color SRM images and a shoulder peak in the locally averaged single-molecule spectra. Together, these results indicate that cellular cholesterol levels were responsible for the polarity differences we found for the plasma and organelle membranes. a
b
SM 612 648 nm emission nm Low Polarity High
2 µm
2 µm 612 Single-mol. 648 1.0 emission nm nm
b
Low
0.5
0.0
Polarity
600
650 Wavelength (nm)
e
High
2 µm
0.5
PM OM PM, 5 min MβCD OM, 5 min MβCD PM, 20 min MβCD OM, 20 min MβCD
700
2 µm
0.0
PM OM PM, 1 mM Chol OM, 1 mM Chol PM, 5 mM Chol OM, 5 mM Chol Low-polarity nanodomains
600
650 Wavelength (nm)
700
Figure 3. The observed heterogeneity in cellular membrane polarity is driven by cholesterol. (a) True-color SR-STORM image of a Nile Redlabeled fixed COS-7 cell after depletion of cholesterol with 5 mM MβCD for 20 min. (b) Averaged single-molecule spectra at the plasma membrane (PM) and organelle membrane (OM) after 5 and 20 min treatment of 5 mM MβCD, compared to that of untreated cells. (c,d) True-color SR-STORM images of Nile Red-labeled fixed COS-7 cells after cholesterol enrichment with 1 mM (c) and 5 mM (d) watersoluble cholesterol. Arrows in (d) point to low-polarity nanodomains. (e) Averaged single-molecule spectra at the plasma membrane (PM) and organelle membrane (OM), as well as at the low-polarity nanodomains, after cholesterol enrichment, compared to that of untreated cells. Bulk measurements on isolated organelle and plasma membranes indicate that the former and latter are characterized by low (
