Spatial high resolution of actin filament organization by PeakForce atomic force microscopy

Abstract Objectives To investigate the heterogeneous feature of actin filaments (ACFs) associated with the cellular membrane in HeLa and HCT‐116 cells at the nanoscale level. Materials and Methods Fluorescence microscopy coupled with atomic force microscopy (AFM) was used to identify and characterize ACFs of cells. The distribution of ACFs was detected by Fluor‐488‐phalloidin–labelled actin. The morphology of the ACFs was probed by AFM images. The spatial correlation of the microvilli and ACFs was explored with different forces of AFM loading on cells. Results Intricate but ordered structures of the actin cytoskeletons associated with cellular membrane were characterized and revealed. Two different layers of ACFs with distinct structural organizations were directly observed in HCT‐116 and HeLa cells. Bundle‐shaped ACFs protruding the cellular membrane forming the microvilli, and the network ACFs underneath the cellular membrane were resolved with high resolution under near‐physiological conditions. Approximately 14 nm lateral resolution was achieved when imaging single ACF beneath the cellular membrane. On the basis of the observed spatial distribution of the ultrastructure of the ACF organization, a model for this organization of ACFs was proposed. Conclusions We revealed the two layers of the ACF organization in Hela and HCT‐116 cells. The resolved heterogeneous structures at the nanoscale level provide a spatial view of the ACFs, which would contribute to the understanding of the essential biological functions of the actin cytoskeleton.

and the network ACFs underneath the cellular membrane were resolved with high resolution under near-physiological conditions. Approximately 14 nm lateral resolution was achieved when imaging single ACF beneath the cellular membrane. On the basis of the observed spatial distribution of the ultrastructure of the ACF organization, a model for this organization of ACFs was proposed.

Conclusions:
We revealed the two layers of the ACF organization in Hela and HCT-116 cells. The resolved heterogeneous structures at the nanoscale level provide a spatial view of the ACFs, which would contribute to the understanding of the essential biological functions of the actin cytoskeleton.

| INTRODUC TI ON
Actin filaments (ACFs) play a crucial role in maintaining the morphology of the cytoplasmic membrane [1][2][3][4] and assist the membrane in performing fundamental physiological functions of the cell such as immune surveillance, signal transduction, endocytosis and exocytosis. [5][6][7][8][9] The ACF possess a well-organized architecture, which is mainly constructed by the helical assembly of actin molecules. Most of these helices are organized to form either a cortex ACF, the intricate network beneath the cellular membrane, or a finger-like bundle ACF, highly specialized structure emanating from the cellular membrane to form the core of microvilli and filopodia. 10,11 In fact, ACFs provide a highly ordered scaffolding not only for the cellular membrane and organelles but also for the membrane-associated proteins. 4,12,13 The requirements of the extensive and intimate interactions with the cellular membrane, organelles and proteins make the ACF an extraordinary hierarchical multi-scale architecture. 1,14,15 For a better understanding of the fundamental role of ACFs in the physiological processes of cells, it is critical to characterize the spatial distribution of ACFs at the nanoscale level.
Advanced techniques have been developed to characterize the ACFs, and new functions have been proposed. 16,17 However, describing the morphology and ultrastructure of the ACFs remains a challenge because of the absence of appropriate techniques. Optical microscopy can be applied to study the ACFs using specific cytoskeleton-labelling molecules, but its resolution is restricted by the diffraction limit of the light. 18,19 In contrast, electron microscopy (EM) is a powerful tool for characterizing the ACFs at the nanoscale level, 20 but it provides information on the geometry alone and not on the chemical and physical properties. Moreover, the structures observed under EM cannot be directly verified in situ by other imaging techniques like fluorescence microscopy (FM). In addition, it is time-consuming, and the samples cannot be observed under physiological conditions, which is of paramount importance when working with biological samples. These limitations of the currently available methods restrict their use for the characterization of the natural structure of the ACF and limit our understanding of its intricate structures and fundamental functions. Filling up the gap between the nanometre and micrometre detection levels to explore the inherent structural relationship of the ACFs under physiological conditions is very important.
The vertical organization of ACFs in this thin region has not been observed directly.
In this study, by coupling FM with modern PeakForce AFM, we uncovered the detailed location and structure of the actin cytoskeleton relative to the cell membrane. We directly observed multiple types of ACF structures at different depths associated with the cell membrane and obtained ~14 nm lateral resolution when imaging cortex actin beneath the cellular membrane, thus providing a spatial view of the ACFs at the nanoscale level.

| Actin filament staining
The HCT-116 cells were seeded in confocal plates at a density of 4 × 10 4 cells per well, respectively, in 500 μL complete cell culture medium for 24 h. To observe the ACF and microvilli structures using optical microscopy, the fixed cells were stained using Oregon Green™ 488 phalloidin (5-10 units/well, 30 minutes) (Invitrogen) according to the manufacturer's instructions and then washed three times with PBS.

| Fluorescence image acquisition
A laser scanning confocal microscope (Leica TCS SP8) was used for observing the multi-layered cell structure. Focal planes were adjusted to obtain the image at different depths. Other fluorescent images were collected using a fluorescence microscope (Leica DMI 3000 B) by oil immersion 63× (NA1.4, Leica) with 480 nm laser light.

| AFM data acquisition
Atomic force microscopy studies were carried out using a BioScope Resolve, AFM (Bruker), coupled with an inverted fluorescence microscope. AFM imaging was performed using a PeakForce-QNM mode.

| Data analysis
All probe calibrations were conducted using Bruker NanoScope 9.3 software. Indentation value was directly extracted from the PeakForce-QNM deformation channel. All data represent values from at least three independent experiments and were analysed by a custom NanoScope analysis program V1.8 (Bruker). Subsequent analyses of the AFM data were done using GraphPad software.

| FM images showing the different layers of ACFs in a HCT-116 cell
We chose HCT-116 cell line because there were plentiful microvilli at the apical surface which would facilitate the morphological observation of the cellular surface in FM images. 38 To provide a comprehensive view of the ACF in a cell, we characterized the ACFs at different locations of individual cells with FM. A small organic molecule, Oregon Green TM 488 phalloidin which binds to the ACF with high specificity, was used to label the ACF of fixed HCT-116 cells.
In Figure 1A, clear fluorescent signals were presented at the apical centre of the cell. Various dots and short lines were clearly observed, while at the periphery it was blurry. In Figure 1B, the outline of the cell appeared and few dots scattered were clearly shown. In Figure 1C, the whole cell presents with high fluorescence intensity both at the centre and around the periphery. A large number of long filament structures were shown. The observed architectures under FM agreed well with previously reported results, 39 suggesting that these structures were the ACF organization in HCT-116 cell.

| High-resolution AFM images resolving the ultrastructure of the ACFs associated with the cellular membrane in a HCT-116 cell
To depict the detailed morphology of the ACFs, we imaged the surface of the cell with high resolution using AFM ( Figure 3A). Tadpoleshaped structures on the cellular surface were observed ( Figure 3A, black). Unlike the root which had a larger diameter, the top had a smaller diameter due to its relatively larger freedom ( Figure 3B, left).
The measured length of this structure ranged up to ~680 nm, and its diameter was ~180 nm in the middle ( Figure 3B, middle and right).
Considering the widening effect of the AFM tip and the reported size of the microvillus, we speculate that the observed structure correlates with that of a single microvillus of ~100 nm in diameter. 9 We noticed that the topography of the single microvillus was slightly different from previously reported structures. 33 This may be due to the difference in the loading force during imaging and the largely increased length of the microvillus in different cells.
Additionally, numerous undulating lamellar structures of various sizes were observed ( Figure 3C, left), whose length and width  1,30,31,42 The ridges of these networks had an average diameter of ~14 nm ( Figure 3E,F, bottom), and the smallest ridge measured was ~12 nm. Considering that the widening effect of the AFM tip usually affects the apparent width of the probed objects as mentioned above, we presumed that the observed structure in Figure 3E was single ACF, although its size is larger than the diameter of single ACF, which is estimated at 5-9 nm in theory. 43

| AFM images revealing the spatial heterogeneous structures of the ACFs comprising two different layers of ACF organization in HCT-116 cells
To explore the spatial correlation between the observed microvilli and cortex ACFs, we tried to use a continuously increasing peak force to map the cellular surface ( Figure 4A). In a series of AFM images ( Figure 4B), we observed two different layers of ACF organization. In the first layer, we observed the bundle-shaped ACFs covered with cellular membrane-the microvilli. The topography of the microvillus in the isolated (red) and assembly forms (yellow) was remarkably clear in high resolution when the peak forces were <2.0 nN. The topography images displayed sudden variations at certain loading forces, and most microvillus structures became dim or disappeared when the loading force was >8.0 nN, but some network structures were observed with blurred architectures (green).
Deformation analyses indicated that the indentation of the tip had reached >200 nm ( Figure 4C), which is deeper than that of the diameter of a single isolated microvillus, and nearly one third of the width of the lamellar structures, indicating that the image was obtained by the tip mostly touching the membrane tightly. A further increase in the force revealed the coexistence of thick and thin network patterns of the ACFs. Further deformation analyses demonstrated that the tip indentation was >300 nm (Figure 4C), implying the presence of another layer network structure which is beneath the cellular membrane, that is in the second layer of ACFs.
Interestingly, by adjusting the peak force to the previously used 0.5 nN, the high-fidelity morphology on the first layer of cellular membrane reappeared; the single microvillus and the assembly of brush-border structure were observed in their previous shape, orientation and size. This suggested that the microvilli on the cellular surface were highly soft and flexible. Statistical data showed that for the HCT-116 cells, an indentation >300 nm was required to reveal the network structure and an addition of ~100 nm indentation still showed a noticeable similar dense network of the ACF ( Figure 4D), implying that the cortex ACFs are also highly flexible.
To test whether the similar feature ACF structure exist in the other epithelial cells, we imaged HeLa cells, a commonly used cell line in laboratories, using the same way. As expected, similar

| D ISCUSS I ON
ACFs are highly ordered and specialized hierarchical organizations which contribute to the fundamental functions of cells; therefore, it is vital to unravel the spatial distribution of the ACFs. AFM has been proven to be a powerful tool for precisely measuring topography of ACFs of cells. So far, the ACFs around the marginal position of cultured cells were visualized clearly. Since the detected ACFs were covered by cellular membrane, the ACFs are generally exhibited as a thick and straight paralleled linear conformation. [44][45][46][47] As a result, the detailed structure information was hard to obtain. In fact, there is little research on the distribution and morphology of ACFs at the nanoscale level, especially in the context of the apical surface of cells, due to its small size, complex topography and soft property at this position. The vertical organization of ACFs in this thin region has been poorly studied under physiological conditions. In this study, using two different yet highly complemen- The undulating lamellar topography of the assembled microvillus structure was directly detected in near-physiological condition for F I G U R E 5 Spatial heterogeneous structure of the ACFs comprises two layers of actin-based filaments in a HeLa cell. A, Topography AFM image of a HeLa cell. B, Representative morphology AFM images of the cellular membrane with a continuously increasing Peak Force of 1-10 nN. Single microvillus (red), lamellar structure (yellow) and cortex actin (green) F I G U R E 6 A two-layer mode for the spatial heterogeneous distribution of the ACF organization the first time. Moreover, the fine lattice structure made of the thin and dense ACFs beneath the cellular membrane at the apical surface of cells showed distinct morphology from that of the thick ACF organization around the periphery of cells. 45 Interestingly, the multiple-layer structure at the cellular surface was routinely resolved at the nanoscale level by simply adjusting the peak force of AFM. We proposed a model of ACFs. The ACFs are made of two layers; the first layer is made of the bundle ACFs forming the single and assembled microvillus emanating from the cell membrane; the second layer is an intricate network of thin ACFs beneath the cell membrane.
In summary, we have demonstrated that improved methodology for imaging ACFs allows more precise characterization of the ACF organization of cellular membrane. Specifically, the ACF organized into an intricate but ordered structure in this thin region with cellular membrane have been explored at the nanoscale level. This heterogeneous distribution of cytoskeletal structures provides new insights into the relationship between the structure and function of the ACFs with the cellular membrane and its associated organelles.

CO N FLI C T O F I NTE R E S T
The authors confirm that there are no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.