Zero-Mode Waveguide Nanowells for Single-Molecule Detection in Living Cells

Abstract

Single-molecule fluorescence imaging experiments generally require sub-nanomolar protein concentrations to isolate single protein molecules, which makes such experiments challenging in live cells due to high intracellular protein concentrations. Here, we show that single-molecule observations can be achieved in live cells through a drastic reduction in the observation volume using overmilled zero-mode waveguides (ZMWs- subwavelength-size holes in a metal film). Overmilling of the ZMW in a palladium film creates a nanowell of tunable size in the glass layer below the aperture, which cells can penetrate. We present a thorough theoretical and experimental characterization of the optical properties of these nanowells over a wide range of ZMW diameters and overmilling depths, showing an excellent signal confinement and a 5-fold fluorescence enhancement of fluorescent molecules inside nanowells. ZMW nanowells facilitate live-cell imaging as cells form stable protrusions into the nanowells. Importantly, the nanowells greatly reduce the cytoplasmic background fluorescence, enabling the detection of individual membrane-bound fluorophores in the presence of high cytoplasmic expression levels, which could not be achieved with TIRF microscopy. Zero-mode waveguide nanowells thus provide great potential to study individual proteins in living cells.

Keywords: fluorescence correlation spectroscopy; fluorescence enhancement; fluorescence microscopy; live-cell imaging; palladium; single-molecule fluorescence; zero-mode waveguide.

Figure 1

Schematic of the experiment and fabrication of overmilled ZMWs. (A) Schematic of a cell on top of an array of overmilled ZMWs. Nanowells below Pd ZMWs allow for the observation of single membrane bound fluorophores despite a high abundance of cytoplasmicfluorophores. (B) Pd is evaporated onto a glass coverslip, and ZWMs are created by focused ion beam milling. Pore diameters used in the study ranged between 100 nm and 280 nm, and overmilling depths ranged between 0 nm and 200 nm. (C) SEM image showing ZMWs with different pore diameters. (D) The depth of milling was measured by cutting through the pores with a focused ion beam and measuring the height when imaging under an angle of 52°.

Figure 2

FDTD simulations of the excitation field and fluorescence emission within a nanowell underneath a ZMW. (A) Schematic of the simulation setup. A dipole was placed at varying depths within the aperture and excited by a plane wave incident from the bottom (wide-field, WF) or under an angle of 70° resembling conditions used in TIRF microscopy. (B, C) Resulting distributions of the excitation field intensity for wide-field (B) or TIRF (C) excitation for a ZMW diameter of 200 nm and an overmilling depth of 200 nm. The electric field is polarized along the x-axis. (D, E) Computed quantum yield and detection efficiency of the dye Alexa488 as a function of the z-position along the central pore axis (dashed line in B, C). (F–I) Z-profiles of the excitation intensity along the central pore axis (F, H) and the detected signal S(z) (G, I) under wide-field (F, G) and TIRF (H, I) excitation. The position of the metal membrane is indicated as a gray-shaded area.

Figure 3

Experimental characterization of fluorescence properties in ZMWs. (A) Schematic of a ZMW with freely diffusing Alexa488 dye. (B) SEM (top) and confocal fluorescence (middle) images of a pore array with a milling depth of 150 nm. The fluorescence image was acquired at a 1 μM concentration of Alexa488. The intensity profile of the fluorescence image is shown below. The scale bar corresponds to 1 μm. (C) Example fluorescence time trace (binning: 1 ms) acquired at a concentration of 500 nM Alexa488 for a ZMW with a diameter of 280 nm and no overmilling (h = 0 nm). (D–F) Heatmaps of the average number of particles in the observation volume N_FCS, fluorescence lifetime τ, and signal enhancement factor defined as the ratio of the counts per molecule in the ZMW compared to free diffusion, ε_ZMW/ε₀, acquired for 500 nM of Alexa488. Data marked as N/A could not be quantified due to insufficient signal. (G) Comparison of measured and predicted fluorescence lifetimes from FDTD simulations for an overmilling depth of 200 nm. The lifetimes of the free dyes are shown as dashed lines. (H) Predicted signal enhancement compared to a free-diffusion experiment as a function of the z position obtained from FDTD simulations (see Supplementary Figure 19 for details). (I) Linear regression of the measured versus the predicted signal enhancement. (J) Comparison of measured and predicted enhancement factors as a function of the overmilling depth.

Figure 4

The cell membrane protrudes into nanowells. (A) U2OS cells expressing CD40TM-BFP on a ZMW array of version 1. Bright-field (left) and BFP fluorescence (right) images are shown. White outline represents an estimated outline of a cell on the surface. (B) Intensity profiles of BFP fluorescence along different pore diameters (100 nm to 280 nm) for a milling depth of 50 nm (left) and 100 nm (right). (C) Fraction of pores with detectable BFP signal for each pore size, as determined from 7367 pores potentially covered by cells.

Figure 5

Imaging of single fluorophores in live cells that protrude into nanowells. (A) U2OS cells expressing cytoplasmic BFP were grown on the ZMW arrays of version 2 and imaged using transmission light (left) or BFP fluorescence (right). Scale bar: 10 μm. (B) Intensity profiles of BFP and transmission light for pores in the yellow box in (A). (C) BFP fluorescence time traces of the pores are indicated in B. (D) Representative time traces of BFP intensity for pores with stable high (blue line) or stable low (green line) BFP signal intensity. The pores correspond to the pores denoted by the red box in A. Images were acquired every 7.5 s. (E) Each dot represents the average BFP intensity of a time trace for an individual pore showing a stable signal. There are two distinct populations for each pore size. The number of measurements per pore size ranges from 113 to 413. (F) Average BFP intensity (mean ± SD from 3 independent experiments) of the high BFP signals in E. (G) BFP intensity profiles under wide-field and TIRF illumination. Under TIRF illumination, peak intensities are increased, while background levels are reduced. (H) Average BFP intensities of high- and low-intensity pores under TIRF or WF illumination. Error bars represent the standard deviation.

Figure 6

Suppression of cytoplasmic background signal using Pd ZMWs. (A) Schematic of U2OS cells expressing cytoplasmic BFP and HaloTag-TfR localized in the plasma membrane. (B) Representative TIRF image of JFX650-labeled HaloTag-TfR. Graph on the right represents an intensity profile along the yellow line. (C, D) Representative images of bright-field (BF), BFP, and JFX650-HaloTag acquired through nanopores. Fluorescence time trace of a single pore, representing the yellow boxed area in the images. Pore diameter d = 220 nm; milling depth d = 50 nm (C) and d = 200 nm (D). Time interval, 5 s for BFP, 500 ms for JFX650-Halo. (E) Expression of cytoplasmic HaloTag in the same cell line as in A. (F) Representative TIRF image of JFX650-labeled HaloTag-TfR. Graph on the right represents an intensity profile through the yellow line. Scale bar: 2 μm. (G, H) Representative images of BF, BFP, and JFX650-HaloTag acquired under the same imaging condition as C and D, but from the cell line additionally expressing cytoplasmic HaloTag. Fluorescence time trace of a single pore representing the yellow boxed area in the images. Pore diameter d = 220 nm; milling depth h = 50 nm (G) and h = 200 nm (H). Time interval, 5 s for BFP, 500 ms for JFX650-Halo. (I) Analysis of the Halo intensity time trace using a hidden Markov model to determine the Halo peak intensity and background noise of each trace. The background noise is defined as the standard deviation (σ) of the background intensity. (J) Distribution of Halo peak intensities for pores with different milling depths and a constant pore diameter of 220 nm, obtained from the cell line without cytoplasmic HaloTag described in A. (K) Background noise σ in the red HaloTag channel from individual pores. Each dotted line represents a single pore. The mean is indicated by black bars. Number of pores: n = 86, 133, 116, 177, 72, 113, 14, 69 (in the same order as the graph).