Real-Time Visualization of Microtubule and Protofilament-Scale Dynamics in Multi-Microtubule Arrays by Atomic Force Microscopy

Abstract

Microtubules, polymers of α, β-tubulin heterodimers, are organized into dynamic multi-microtubule arrays for diverse cellular functions. The dynamic properties of microtubule arrays govern their structural and functional properties. While numerous insights into the biophysical mechanisms underlying microtubule organization have been gleaned from in vitro reconstitution studies, the assays are largely restricted to visualization of single or pairs of microtubules. Thus, the dynamic processes underlying the remodeling of multi-microtubule arrays remain poorly understood. Recent work shows that Atomic Force Microscopy (AFM) enables the visualization of nanoscale dynamics within multi-microtubule 2D arrays. In this assay, electrostatic interactions permit the non-specific adsorption of microtubule arrays to mica. AFM imaging in Tapping Mode, a gentle method of imaging, allows the visualization of microtubules and protofilaments without sample damage. The height information captured by AFM imaging enables the tracking of structural changes in microtubules and protofilaments within multi-microtubule arrays over time. The experimental data from the method described here reveal previously unseen modes of nanoscale dynamics in microtubule bundles formed by the microtubule-crosslinking protein PRC1 in the presence of the depolymerase MCAK. The observations demonstrate the potential of AFM imaging in transforming our understanding of the fundamental cellular process by which multi-microtubule arrays are dynamically assembled and disassembled.

Basic Protocol 1: Sample preparation and real-time visualization of microtubule arrays by Atomic Force Microscopy

INTRODUCTION:

Microtubule polymers, composed of linear protofilaments of α, β-tubulin heterodimers, form the backbone of diverse cellular structures such as the spindle in dividing cells and axonemes in the cilia. In these structures, individual microtubules are arranged in arrays of well-defined size and geometry (Mani, Wijeratne, & Subramanian, 2021). For instance, anti-parallel microtubule arrays form the midzone of the mitotic spindle, and parallel arrays of microtubule doublets comprise the axoneme. The size and stability of these multi-microtubule arrays depend on the dynamics of individual polymers within these arrays. Numerous cell biological and biochemical studies have established that microtubule dynamics is tightly regulated by associated proteins to for proper organization and function of structures such as the spindle and the cilium.

Insights into microtubule dynamics have come from in vitro reconstitution of microtubule assembly and disassembly using light microscopy or electron microscopy-based imaging methods (Gudimchuk & McIntosh, 2021). However, light microscopy assays are limited to single or pairs of microtubules due to the challenges in resolving individual polymers (25 nm diameter) within a larger bundle of closely spaced microtubules. Even in the case of single microtubules, the dynamics of individual protofilaments (4 nm diameter) are challenging to image in real-time. While electron microscopy provides a high-resolution view of protofilaments and microtubules, it does not permit the direct visualization of the dynamics of individual microtubules in real-time. Recent work shows that Atomic Force Microscopy (AFM) is a promising method to overcome these technical barriers. By bridging the spatial-temporal resolution gap between light and electron microscopy, this technique permitted the visualization of previously unobserved modes of disassembly of multi-microtubule arrays by microtubule depolymerizing enzymes (Wijeratne, Marchan, Tresback, & Subramanian, 2022). The different modes of disassembly provide insights into how the action of depolymerases may result in either the large-scale remodeling or the precise length-regulation of microtubule arrays during processes such as cell division and cell migration.

This protocol describes an AFM-based assay for immobilizing and imaging microtubule arrays organized by the mitotic spindle-associated protein, Protein Regulator of Cytokinesis-1 (PRC1), which preferentially crosslinks microtubules in an antiparallel orientation (Bieling, Telley, & Surrey, 2010; Subramanian, Wilson-Kubalek, Arthur, Bick et al., 2010). PRC1-crosslinked bundles are formed by mixing polymerized microtubules with recombinant protein. The bundles adhere non-specifically on mica via electrostatic interactions by Mg²⁺ cations included in the buffer (Hamon, Curmi, & Pastre, 2010). Tapping Mode AFM enables the stable imaging of micron-sized arrays for over 30 mins. Upon examining AFM images and their corresponding height information, individual microtubules (25 nm) and protofilaments (4 nm) within microtubule arrays become apparent. After introducing a depolymerizing enzyme, MCAK (Helenius, Brouhard, Kalaidzidis, Diez, & Howard, 2006), changes in height profile at each time point, provide a direct readout of depolymerization at individual microtubule and protofilament levels within multi-microtubule arrays. These AFM-based assays provide a viewpoint, not captured using other imaging techniques like light and electron microscopy, and thus have the potential to transform our understanding of the fundamental processes underlying the remodeling of microtubule arrays.

Basic Protocol 1 describes how to prepare single microtubules and microtubule array samples, how to visualize microtubule arrays for AFM imaging, and how to analyze AFM data.

BASIC PROTOCOL 1

Basic protocol title:

Sample preparation and real-time visualization of microtubule arrays by Atomic Force Microscopy

Introductory paragraph:

This protocol describes a method for preparing and imaging samples with single microtubules and microtubule arrays by AFM. Before imaging, mica discs are glued to specimen discs. An assay mixture of either single microtubules or microtubule arrays in BRB80 buffer supplemented with 5 mM MgCl₂ is prepared. Next, the liquid probe holder is also cleaned and dried. The assay mixture is deposited on mica where the surface adsorption of negatively charged microtubules is mediated by electrostatic interactions between the Mg²⁺ cations on mica. During imaging, the frequency response of the cantilever is calibrated, and the scanning parameters are chosen. Static or time-lapse images are acquired of the microtubule samples. After imaging, the raw data are processed and analyzed.

Materials:

Reagents, solutions, starting samples

1X BRB80 (see Reagents and Solutions)

Magnesium chloride (Sigma Aldrich)

Fluorescent microtubules (see Ref. Mani, Marchan, & Subramanian, 2022)

Protein of interests (PRC1 microtubule crosslinker, MCAK microtubule depolymerase

(see Refs. Subramanian, Wilson-Kubalek, Arthur, Bick et al., 2010; Helenius, Brouhard, Kalaidzidis, Diez, & Howard, 2006)

Isopropanol

Sterile water

Hardware and instruments (see Fig. 1)

Figure 1.

Hardware for sample preparation and imaging.

5 Minute^® epoxy, colorless to light yellow (Devcon, #230-14250)

Highest grade V1 AFM mica discs, 10 mm (Ted Pella, #50)

AFM/STM metal specimen discs, 15mm (Ted Pella, #16218 )

Aven Technik Tweezers 7-SA tweezer (Aven Tools, # 18072USA)

BL-AC40TS tips (radius: 8 nm; spring constant: 0.09 N/m; Oxford Instruments, # 803.OLY.BL-AC40TS)

Mini Centrifuge (Argos Flexifuge, Model C1000)

Liquid probe holder (Oxford Instruments)

Optical stereo microscope, M60 (Leica)

AFM Cypher S/ES instrument (Oxford Instruments)

Scotch tape

Protocol steps with step annotations:

BRB80 buffer preparation

• 1Refer to the Reagents and Solutions section for BRB80 buffer contents.
• 2Prepare 1 mL of AFM imaging buffer by supplementing BRB80 buffer with an additional 5 mM MgCl₂ (final concentration).

  Keep buffer at room temperature until use.

Microtubule and proteins preparation

• 3Prepare fluorescent GMPCPP microtubules as described previously (Mani, Marchan, & Subramanian, 2022).

  Fluorescent microtubules are not necessary. However, prior to AFM experiments, fluorescent microtubules are helpful to quickly assess if microtubules are present in the sample by fluorescence microscopy, where in a 100 × 100 μm field of view, a high microtubule density (>100 microtubules) is preferred. This increases the chance of locating microtubules in a smaller scanning region (20 × 20 μm) during AFM imaging.

• 4Prepare PRC1 reagents (Subramanian, Wilson-Kubalek, Arthur, Bick et al., 2010).

  PRC1 crosslinks microtubules into antiparallel arrays.

  Doublet microtubules from purified axonemes could also be studied using this method (Linck, 1976; Waterman-Storer, 2001).

• 5Prepare depolymerase reagents.

  In this assay, MCAK was used, which depolymerizes the microtubule from both plus and minus ends (Helenius, Brouhard, Kalaidzidis, Diez, & Howard, 2006).

Mica-specimen discs attachment

• 6Open 5 Minute^® epoxy tube, squeeze a drop of both the resin and the hardener in equal amounts into a plastic container and mix thoroughly for 1 min until the mixture looks cloudy.
• 7Apply a small amount of the epoxy glue mixture to the center of the specimen disc.
• 8Lift a mica disc with a tweezer and place it on the specimen disc.

  Quickly perform Steps 6–8 to prevent the glue mixture from drying out.

• 9Press the mica down gently with the wooden end of a cotton swab to spread the glue evenly between the mica and specimen disc. The mica disc should be glued flat and not tilted on the specimen disc.
• 10Leave the mica-specimen discs to dry overnight.

Liquid probe holder preparation

• 11Clean the AFM’s probe holder by squirting 70% isopropanol on the probe holder and rinse thoroughly with sterile water.
• 12Dry the probe holder with nitrogen air.
• 13Place a BL-AC40TS cantilever until it sits flat in the cantilever pocket in the probe holder using an Aven Technik Tweezers 7-SA tweezer. Tighten the screw until snug so that the cantilever is secure in the pocket of the probe holder.
• 14Nudge the tip on either side with a 7-SA tweezer to ensure that the tip is secure and not loose.
• 15Examine if the tip is still intact under an optical microscope. The BL-AC40TS tip should be visible under the optical microscope with a magnification of 60x.

Microtubule assay mixture preparation

For single microtubules:

• 16Mix 10 μL of imaging buffer and 3–4 μL of 2 μM microtubules in a tube.

  Microtubules should not be put on ice. Keep microtubules at or above room temperature to prevent depolymerization.

For microtubule arrays:

• 17Mix 6 μL of imaging buffer, 3 μL microtubules, and 1 μL of 0.1–1 μM of PRC1 in a tube.

Pre-imaging preparation

All AFM experiments are conducted at room temperature, utilizing Tapping Mode in liquid with a silicon tip.

• 18Cleave the mica with scotch tape to expose an even and fresh layer and place the cleaved mica disc on the sample stage.
• 19If imaging microtubule arrays, centrifuge the microtubule assay mixture at 6000 rpm (from Step 18) on a table-top centrifuge for 5 minutes. If imaging single microtubules, this step is not needed.
• 20Deposit 10 μL of the microtubule assay mixture on the middle of the mica disc.
• 21After 5 minutes of incubation, add 10 μL of additional imaging buffer to the center of the mica disc.
• 22Add a 10 μL droplet of the imaging buffer on the probe holder to pre-wet the tip before mounting the cantilever holder on the AFM cell body.

  This step helps to reduce air bubbles in the liquid during imaging.

• 23Mount the cantilever holder on the AFM cell body.
• 24Decrease the gap between the cantilever and probe holder until droplets on the tip and the mica make contact. The tip should be incubated in liquid for 10–15 min before imaging.

Imaging

Engaging the tip on the surface

• 25Set the focus position of the cantilever and the surface. Center the laser spot on the cantilever.
• 26Capture a thermal plot of the cantilever.

  In liquid, the drive frequency of the BL-AC40TS-tip is ~25–30 kHz.

• 27Tune the cantilever manually. Set the setpoint to ~8 nm.
• 28Increase the drive amplitude until the free amplitude is above 8 nm. Approach and engage the tip to the substrate.
• 29Set the initial scan parameters to obtain an image. Set the scan size to 1 × 1 μm and integral gain just before the image becomes slightly noisy.
• 30Do an initial scan to determine if the tip is engaged on the surface. Increase the drive amplitude, until the trace and retrace profiles overlap.

  Steps 26–31 describe the procedure before engaging the tip on the surface with the Asylum Cypher ES/S AFM. The order of these steps may differ depending on the brand of AFM. Refer to the instrument user manual for calibration and steps on how to land the tip on the surface for liquid imaging.

Acquiring an initial AFM image

• 31Set the scan size to 20 × 20 μm and acquire an image to visualize if microtubules are present in the scanning region.

  The 20 × 20 μm scan image helps in locating clean regions for imaging. Usually, if a microtubule sample is dense via fluorescence microscopy (>100 microtubules within a 100 × 100 μm field-of-view), microtubules should be observed in the AFM images.

• 32Locate regions on the 20 × 20 μm scan, which shows flat microtubule arrays or single microtubules. Zoom into these regions for stable imaging.
• 33To visualize individual microtubules within an array, decrease the scan size to ~<3 μm, and set the sample resolution to 256 × 256 pixels.
• 34To visualize individual protofilaments within a microtubule, decrease the scan size to ~<1 μm, and set the sample resolution to 256 × 256 pixels.

Time-lapse experiments

• 35Zoom into a 1–2 μm region of interest, preferably with microtubule ends visible in the scan area.
• 36Acquire this t=0 image and pause the scanning.
• 37Increase the distance between tip and sample by 100–200 nm.

  The tip-sample distance needs to be optimized based on the AFM experimental setup. The tip-sample distance should be increased while maintaining the liquid droplet between the tip and sample.

• 38Inject 10 μL of the 7 nM depolymerase and 1 mM ATP with a pipette without touching the mica disc to prevent losing the region of interest.

  The amount of liquid on the mica was monitored to ensure that the sample did not dehydrate. To prevent the liquid from drying out, buffer solution can be exchanged to the mica disc. After testing a range of enzyme concentrations, 7 nM was found to be the preferred concentration. The final conditions were empirically determined because some protein can be non-specifically adsorbed on the mica surface and cantilever in these experiments.

• 39After adding the depolymerase, start imaging at a frame rate of ~3 mins/frame (256×256 pixels at ~1.5 Hz).

  In our AFM experiments, the long axis (y-axis) of the microtubule is perpendicular to the direction of AFM scanning (x-axis or fast-axis) in single microtubules or microtubule arrays. To align the microtubules at the desired orientation, the scanning angle can be changed. Data could be collected on microtubules in various orientations due to random binding to the surface. It is important to ensure that the observations from AFM imaging are not specific to microtubule orientation with respect to the scan direction. A troubleshooting guide for resolving typical issues during AFM imaging of microtubules are discussed in Table 2.

• 40Image up to ~30 mins while constantly maintaining the drive amplitude of the tip throughout the experiment.

  The drive amplitude was maintained slightly above a point where the tip starts to contact the surface.

• 41If starting a new experiment, change the tip to avoid the contamination and ensure that the same quality of AFM images is obtained as determined by the spatial resolution or the width of the sample.

  Because of sample contamination and tip damage, using the same tip for all the experiments is not possible.

• 42After the experiment, transfer the data to a computer for data analysis.

  Control experiments without depolymerases, without ATP or different surfaces should also be performed to ensure that the features are not arising from the tip damage or the surface. During these experiments, microtubules should remain at >20 nm for the entire duration of the experiments. See Table 1 for a list of suggested control experiments.

Table 2.

Troubleshooting Guide for AFM imaging of microtubules and microtubule arrays

Problem	Possible Cause	Solution
Microtubules absent during AFM imaging	Microtubules absent in the sample Microtubules are depolymerized	Use fluorescent microtubules to check if microtubules are present Do not mix microtubules with cold buffers should be kept at room temperature
A few microtubules present during AFM imaging	Not enough microtubule density	Double the concentration of microtubules
Microtubules disappear after the first scan	Microtubules not sticking to the surface, or scanning velocity is too fast	Reduce scan rate
Double microtubule effect	AFM tip cantilever split into two	Use new AFM tip
Microtubules appear fuzzy	Microtubules are near tall structures or AFM tip is not sharp, causing the tip to press against the microtubules and not follow their contours.	Scan only in flat microtubule regions Change AFM tip – a sharper tip can better conform to microtubule surface Increase scan size – imaging a larger area can compensate for the loss of resolution Increase integral gain – help can improve signal to noise ratio
Microtubule protofilaments not visible	AFM tip is not sharp	Use new AFM tip
Microtubules disappear after flowing in protein solution or buffer	Flow buffer solution too cold or too much liquid added to the surface	Gently flow liquid into the sample and ensure the flow solution is not cold
Aggregates on the surface	Microtubule sample contains tubulin aggregates	Spin the microtubule sample down, pipette the liquid from the bottom of the tube
Microtubules disappearing with AFM scanning	AFM tip causing damage	Lower the drive amplitude Use a new AFM tip (resonant frequency should be ~25 kHz in liquid) Reduce scan points, and scan lines Use fresh reagents

Table 1:

List of control experiments

Control Experiment	Surface	Reason
Microtubule arrays + BRB80 + 5 mM MgCl2	Mica	Buffer control – depolymerization is not arising from adding 5 mM MgCl2
Microtubule arrays + BRB80 + 5 mM MgCl2 + 1 mM ATP	Mica	ATP control - depolymerization is not arising from adding 1 mM ATP
Microtubule arrays + BRB80 + 5 mM MgCl2 + 1 mM ATP + 7 nM MCAK Image every 5–10 minutes	Mica	Tip control - depolymerization is not arising from AFM tip
Microtubule arrays + BRB80 + 5 mM MgCl2 + 1 mM ATP + 7 nM Kip3p	Mica	Motor control – depolymerization features are specific to depolymerase. Kip3p is a motor protein which depolymerizes microtubules from the plus-end.
Microtubule arrays + BRB80 + 5 mM MgCl2 + 1 mM ATP + 7 nM MCAK	Glass	Surface control - depolymerization not arising from surface

Data analysis

Raw AFM data processing

• 43Open AFM data processing software

  Generally, data analysis could be performed by any AFM analysis software. In this protocol, Gwiddyon (http://gwyddion.net/) was used. Only the cleanest AFM images are used for obtaining height profiles.

• 44Assess all the AFM images (height, phase, and amplitude) and determine which images should be analyzed.

  Typically, the height images are processed. However, if the phase and amplitude images are clear and provide additional information that is not present in the height images, such as striations from protofilaments, they should also be processed.

• 45Flatten images (height, phase, and amplitude).
• 46After AFM data processing of static images, classify the structures in the images as microtubules (~25 nm) or protofilaments (~4 nm) by obtaining the height profiles.

  Based on the microtubule’s EM structure, the topography of the microtubule could be characterized by 1) its cross-sectional height (~25 nm) and 2) striations from protofilaments depending on the scan size. Similarly, the cross-sectional height of a protofilament should be 4 nm based on the EM structure.

• 47Save the images as figure files (.png, jpegs, or tiffs). The height profiles could be saved as distance (x) versus height (z) values (.dat or .xls).
• 48To compile the images into movies, open the time-lapse images separately in ImageJ (Rasband, W.S., ImageJ, U. S. National Institutes of Health, Bethesda, Maryland, USA, https://imagej.nih.gov/ij/, 1997–2018) and make a z-stack of the image files in order of data acquisition. Save the z-stack as a .tiff or .avi file.

Calculation of microtubule depolymerization rates from AFM images

• 49To calculate depolymerization rates of a microtubule, calculate the average length change along the microtubule between the first frame and the last frame from the time-lapse images.
• 50To determine the effect of neighboring microtubules on depolymerization rates, count the number of microtubules that are physically contacting an individual microtubule in parallel within a large bundle. For example, a microtubule with no neighbors nearby has N=0, a microtubule in contact with another microtubule has N=1, and a microtubule in contact with two microtubules has N=2.
• 51In addition to end-depolymerization, we observed the propagation of defects, which are defined as discontinuities in the microtubule lattice, in the presence of MCAK. To calculate the rate of defect propagation, after the appearance of a defect, measure the average diameter or length changes from both edges of a defect. The ‘diameter’ is the change in width over time around the diameter of the microtubule, and ‘length’ refers to the change in length over time along the length of the microtubule.

ALTERNATE PROTOCOL 1. Protocol for Coating Surface with Poly-L-lysine and Immobilizing Microtubules

Materials:

Poly-L-lysine solution (1 mg/ml; Sigma Aldrich; # P4832-50ML)

Glass coverslips (10 mm; Ted Pella)

Imaging buffer (see Basic Protocol 1; Step 2)

Fluorescent microtubules (see Basic Protocol 1; Step 3)

Sterile water

Protocol steps with step annotations:

1. Prepare a 0.2 mg/mL Poly-L-lysine solution by adding sterile water.

2. Coat the glass coverslip with 40 μL of the Poly-L-lysine solution.

3. Incubate the coated coverslip for 45 minutes at room temperature.

4. Rinse the coated coverslip with sterile water.

5. Rinse the coated coverslip with imaging buffer.

6. Add 3 μL of microtubules and 10 μL of imaging buffer to the coated coverslip.

7. Incubate the coated coverslip with the microtubule and imaging solution for 10 minutes.

Note: All steps should be performed under sterile conditions to avoid contamination. It is important to handle the coverslip carefully to avoid scratching the coated surface. Store the coated coverslip in the imaging buffer until imaging.

REAGENTS AND SOLUTIONS:

• BRB80, 1X (pH adjusted to 6.8 with KOH)

  • 80 mM PIPES
  • 1.5 mM MgCl₂
  • 0.5 mM EGTA

• Store up to 2 years at 4°C

• 1 mM Dithiothreitol (DTT)

• 1 mM Adenosine Triphosphate (ATP)

COMMENTARY:

Background Information:

AFM is a powerful technique that spans a spatial-temporal resolution range not readily accessible by other visualization methods. Its high spatial resolution has allowed for imaging of protofilaments within single microtubules (Hamon, Curmi, & Pastre, 2010; Muller-Reichert, Chretien, Severin, & Hyman, 1998; Schaap, Carrasco, de Pablo, & Schmidt, 2011). In addition, its high-temporal resolution has enabled the visualization of nanoscale-level organization on single microtubules and actin filaments (Kodera, Yamamoto, Ishikawa, & Ando, 2010; Nasrin, Ganser, Nishikawa, Kabir et al., 2021; Owa, Uchihashi, Yanagisawa, Yamano et al., 2019; Schaap, Carrasco, de Pablo, & Schmidt, 2011). In our recent study, we adapted AFM to visualize real-time depolymerization of multi-microtubule arrays (Wijeratne, Marchan, Tresback, & Subramanian, 2022). We describe the protocol here, which allows for the stable imaging and remodeling of multi-microtubule arrays at the nanoscale over time. Mica polysilicate, a negatively charged surface, is used for microtubule immobilization and it is a common substrate for AFM imaging of biological molecules (Hamon, Curmi, & Pastre, 2010; Hansma & Laney, 1996; Heenan & Perkins, 2019). The surface adsorption of negatively charged microtubules is mediated by the electrostatic interactions between the Mg²⁺ cations on mica (Pastre, Pietrement, Fusil, Landousy et al., 2003). Tapping Mode allows the visualization of microtubule arrays at the single microtubule and protofilament levels. In addition, height information captured by AFM imaging helps to track changes in microtubule and protofilament dynamics over time. This protocol was empirically optimized through numerous trials and errors and heavily borrowed from a long-standing interest in studying microtubule dynamics using TIRF microscopy.

Critical Parameters:

AFM imaging mode

During AFM imaging of single microtubules or microtubule arrays, a gentle AFM mode is crucial to minimize sample damage. To achieve this, it is recommended to conduct AFM experiments in liquid using Tapping Mode, along with a small radius tip and spring constant. For instance, in this protocol, we used the Asylum Cypher S/ES with a silicon tip (BL-AC40TS; radius: 8 nm; spring constant: 0.09 N/m; Oxford Instruments).

AFM experiments

Injecting depolymerase solution

Before adding a depolymerase, the tip-sample distance should be increased just enough so that solution could be gently pipetted onto the mica surface without moving it. If the droplet formed between the cantilever and surface disappears, the liquid probe holder should be taken out and mounted again. If there is no way to inject a solution directly onto the surface without taking the probe holder out, two options are available: a) Unmount the liquid holder, add liquid on top of the mica, quickly mount the holder and start imaging. However, if the enzyme’s reaction is slow, changes to the microtubules could still be captured. In addition, the use of stabilizers, such as GMPCPP microtubules, PRC1-crosslinked microtubule arrays, or axonemes should help slow down the depolymerization reaction. The scanned area will likely be lost here, and the t=0 image would not be acquired. b) Use inlets commonly found in liquid probe holders for flowing in liquid through a syringe with tubing. This is the preferred method for injecting solution as it preserves the position of the scanned area and allows for immediate imaging to capture early depolymerization events.

Obtaining high-quality images

After adding a depolymerase, the amount of liquid on the mica should be monitored to ensure that the sample does not dehydrate. Additional buffer solution should be added as needed to maintain solution volume during the experiments. For stable imaging, regions only showing flat microtubule arrays should be scanned. During each scan, the drive amplitude should be monitored and maintained slightly above a point where the tip starts to make contact with the surface. To avoid sample contamination and tip damage, a new AFM probe should be utilized while maintaining the same quality of AFM images as determined by the spatial resolution of the sample.

Control experiments

After a time-lapse experiment, additional experiments should be performed to ensure the observations of microtubule remodeling are specific to the depolymerase. To ensure that depolymerization is occurring at the expected time scales, corresponding TIRF experiments should be performed with the same batch of proteins to confirm the depolymerization timescales are consistent between the two imaging modalities. To exclude the possibility that depolymerization arises from the AFM tip or the surface, relevant control experiments should be performed to match the duration of each experiment (Table 1).

Troubleshooting:

This table is a troubleshooting guide for resolving typical problems encountered while AFM imaging of single microtubules and microtubule arrays.

Understanding Results:

This protocol will result in the immobilization of microtubule samples on the substrate and in static or real-time visualization of protofilament-level organization of single microtubules or microtubule arrays by AFM. The following sections describe a discussion of understanding important protocol steps and interpreting AFM data.

Microtubule immobilization

Microtubule samples are immobilized non-specifically on mica in the presence of BRB80 buffer with 5 mM MgCl₂ (Fig. 2). Mica (polysilicate), a negatively charged surface, is a common substrate for AFM imaging of biological molecules (Hamon, Curmi, & Pastre, 2010; Hansma & Laney, 1996; Heenan & Perkins, 2019). The surface adsorption of negatively charged microtubules is mediated by the electrostatic interactions with the Mg²⁺ cations on mica (Pastre, Pietrement, Fusil, Landousy et al., 2003). MgCl₂ at 5 mM concentration is a common component of microtubule assays and does not depolymerize microtubules significantly without enzymes. In addition to mica, we tested microtubule depolymerization experiments by MCAK using a glass surface. Qualitatively similar results were obtained on glass indicating that the features of PRC1-crosslinked microtubule bundles and their depolymerization by MCAK were not influenced by immobilization on mica. However, pretreatment of glass with polyL-lysine was needed to immobilize microtubules prior to AFM imaging and the image quality was not as good as that achieved when imaging on mica.

Figure 2.

Steps for sample preparation and imaging of microtubules. Created with BioRender.com.

In addition to using Mg2+ cations and poly-L-Lysine, other methods can immobilize microtubules for AFM imaging. APTES can covalently link microtubules to the substrate by reacting with the hydroxyl group of the substrate and the carboxyl groups on the microtubules. Similarly, lipid bilayers functionalized with antibodies or ligands can also immobilize microtubules. Streptavidin is another method for immobilizing microtubules, which binds to biotinylated microtubules with high affinity.

AFM imaging technique

Tapping mode, a gentle technique for imaging biological materials, is used for AFM imaging of microtubule arrays (Wijeratne, Marchan, Tresback, & Subramanian, 2022). In this mode, the AFM cantilever oscillates near its resonant frequency and makes intermittent contact with the surface while scanning (Fig. 3A). The resulting AFM image, color-coded by height, is obtained by the z-piezo movement, maintaining the setpoint of amplitude interactions. When the AFM tip encounters changes in topography, the amplitude feedback changes the z-piezo height to re-establish the amplitude setpoint. The lateral (x,y) and vertical parameters (z) determine the resolution of the AFM image (Fig. 3B). The high z-resolution could be exploited for differentiating microtubules and protofilaments in arrays.

Figure 3.

Schematic of AFM experimental set-up, description of the AFM output image and an example AFM image of DNA. A) Schematic of AFM experiment operated in tapping mode. Laser light is reflected off the back of the cantilever to the photodiode. The deflection of the cantilever is measured as it records the surface by gentle tapping of the tip, allowing an image of the topography to be acquired. In liquid, the resonant frequency of the cantilever is 25 kHz and the tip radius is 8 nm. B) Output AFM image of a hypothetical sphere. The AFM image shows the topography in x and y and it is color-coded by height (z), by dark to light (brown to cream) C) An example AFM image of DNA imaged by our AFM system (Cypher ES) in liquid (20 mM Tris + 1 mM EDTA + 10 mM MgCl₂). The x-y scale bar is 200 nm and the z-scale bar is 0 to 3 nm (brown to cream). Adapted from Ref. (Wijeratne, Marchan, Tresback, & Subramanian, 2022).

Resolution

Two parameters determine the x, y resolution of an AFM image: tip shape and pixel size: (1) Tip shape: In the AFM tip, the radius of curvature and the aspect ratio (radius/length) control the highest available lateral resolution with a certain tip. The smaller the radius of curvature, the higher is the ability to resolve small structural features. In this protocol, a tip of an 8 nm radius was used. (2) Pixel size: In the AFM image, the pixel size determines the x,y resolution where small features than a pixel size could not be resolved (Fig. 3B). The x, y resolution in our images are 256 × 256 pixels, where the pixel size is the scan size divided by 256. For example an AFM image with a 1 × 1 μm scan size (x,y) is 1 μm/256 pixels = ~4 nm/pixel. In this protocol, the pixel size is ~ 3–4 nm for the high-resolution imaging of microtubules and ~10–20 nm for the microtubule bundles.

The z-resolution is determined by the vertical piezo scanner movement, having less than 1 Å sensitivity and a baseline system noise < 40 pm. The high resolution of AFM in z-height permits visualization of biological molecules that have a small height, such as double-stranded DNA (Fig. 3C), by AFM. Similarly, the cylindrical shape of the microtubule and the high resolution in z compared to the x-y are advantageous to distinguish between singles, doubles, or components of microtubules. The z-height is also advantageous because the lateral x-y measurements are impacted by the mobility of the microtubules, particularly in scan directions, dilating lateral measurement values. The loss of protofilaments results in a change in height along the short axis of the microtubule, leading to changes in the topography pattern. This unique feature of AFM can be exploited to image depolymerization of individual protofilaments in microtubules and to reliably image individual microtubules in larger arrays by monitoring the height along with the array.

Interpreting AFM images

Estimating topography and analyzing an AFM image of a single microtubule or protofilament

The topography of a single microtubule or protofilament can be estimated from reported EM images. Based on the EM structure of a microtubule, its topography will be characterized by 1) its cross-sectional height (~25 nm) and 2) striations from protofilaments depending on the scan size. Figure 4 shows EM images of a microtubule and the corresponding estimated topography. As predicted, the maximum cross-sectional height of microtubules is ~25 nm at all scan sizes. However, the measured width of the microtubule during AFM imaging does not accurately reflect the actual width of the microtubule due to tip dilation, an apparent increase in the lateral dimension during imaging objects with a small width (Fig. 4). This occurs when the tip cannot perfectly follow the contours of the sample, and instead the tip presses against the sample, causing a broadening of the image due to limitation in AFM tip sizes. Therefore, in this protocol, microtubules in arrays and protofilaments in microtubules could be distinguished by taking advantage of the high z-resolution of the AFM. Similarly, based on the EM structure, the cross-sectional height of the protofilament should be 4 nm. As shown in Fig. 4, the predicted height agrees with our experimental observation. While tip dilation doesn’t accurately reflect the width of the sample, it can help increase the lateral resolution of the AFM image. This is because the apparent broadening of the sample can be used to extract topographical features of the sample. For instance, at 1 × 1 μm scan size, striations along the microtubule length, corresponding to protofilaments, are clear, allowing the identification of protofilaments within microtubules and arrays.

Figure 4.

EM structure, cartoon schematic, and high-resolution AFM image of an A) microtubule and B) protofilament. The height profiles in the inset of the AFM images were drawn along the dotted line. The x-y scale bar in A) is 100 nm and B) is 20 nm. The z-scale bar in A) is 30 nm and B) 4 nm. From Ref. (Wijeratne, Marchan, Tresback, & Subramanian, 2022).

Analyzing an AFM image of a depolymerizing microtubule

Since the z resolution of AFM is in sub-nanometers, structures on the order of 4 nm diameter could readily be discerned, even when the lateral resolution is lower at larger scan sizes. This feature is exploited in our assay to observe protofilament loss within microtubules in large arrays where it is necessary to image a larger field of view, as illustrated in Fig. 5. A partially depolymerized section of a single microtubule has segments where the topography corresponds to 1–2 or more protofilaments, which could be distinguished based on the topographical height changes. Due to microtubule depolymerization, the change in height along the z-axis enables the observation of protofilament loss by measuring differences in the surface topography over time. These height differences indicate the type of protofilament-level activity of the depolymerase. The loss of protofilaments could be characterized as synchronous or asynchronous, where protofilaments depolymerize at the same rates or different rates, respectively. In the presence of MCAK, the depolymerization of protofilaments within a microtubule is asynchronous. In addition, loss of protofilaments could also occur within microtubule breaks, where these defects are propagated over time.

Figure 5.

AFM imaging of a depolymerizing microtubule. High-resolution AFM image of a depolymerizing microtubule and its corresponding height profiles along the indicated lines on the different microtubule sections. Peaks with heights less than 20 nm indicate protofilaments. The x-y scale bar is 30 nm. The z scale bar is 30 nm. From Ref. (Wijeratne, Marchan, Tresback, & Subramanian, 2022).

Another feature of depolymerization is the appearance of curved structures as observed by EM. However, the AFM visualization of curved structures peeling away from the microtubule after MCAK depolymerization is limited for reasons. First, during depolymerization, these structures form transiently and many of them may be missed in the timescale of the imaging. Second, only the curved protofilaments that form along the mica surface can be read out by the AFM tip. Therefore, only a small fraction of flared ends could be captured. However, it is possible to capture the transient appearance of these structures [see Ref. (Wijeratne, Marchan, Tresback, & Subramanian, 2022); SI Appendix, Fig. S2B].

Analyzing an AFM image of microtubule arrays

Imaging microtubule arrays requires scanning a large field of view. While this lowers the lateral resolution (nm/pixel), the spacing between microtubules in an array causes a height change between neighboring microtubules, which is greater than the z-axis resolution of AFM. As in Fig. 6, each microtubule in a bundle with nine microtubules could be identified (bold numbers). The change in height indicates the spacing between the microtubules, enabling the ability to distinguish microtubules in a bundle. In a microtubule array, the crosslinking distance determines the distance between neighboring microtubules in an array and suggests the change in height while moving from one microtubule to the next one. In PRC1, the crosslinking distance (~35 nm) is larger than the diameter of a microtubule, resulting in a change in the height of 4 nm or more, which readily allows to distinguish neighboring microtubules in a bundle.

Figure 6.

AFM imaging of microtubule bundles. Comparison of A) single microtubule and B) microtubule bundle and corresponding height profiles. The x-y scale bar in A) is 40 nm and B) is 60 nm. The z scale bar in A) is 30 nm and B) is 40 nm. From Ref. (Wijeratne, Marchan, Tresback, & Subramanian, 2022).

Analyzing an AFM image of a depolymerizing microtubule array

The binding of proteins such as PRC1 to microtubules could change the surface topography corresponding to the molecular dimensions of the protein. This feature and the larger field of view needed to image larger arrays makes it challenging to image lone protofilaments within a single intact microtubule. However, the loss of protofilaments in a depolymerizing microtubule is readily discernible. This is because the change in topography with protofilament loss, due to the change in height is significantly greater than the z-resolution in AFM imaging, allowing for the reliable visualization of protofilament dynamics within PRC1-bound microtubules (Fig. 7).

Figure 7.

AFM imaging of depolymerizing microtubule array. AFM images of a depolymerizing microtubule bundle at A) 0 min, B) 6 min, and C) 12 min and its corresponding height profile along the indicated dotted line. The blue arrows in the plots indicate the individual depolymerizing microtubules and the change in height within the bundle over time. Peaks of heights less than 20 nm indicate protofilaments. The x-y scale bar is 50 nm. The z scale bar in is 30 nm. From Ref. (Wijeratne, Marchan, Tresback, & Subramanian, 2022).

Time Considerations:

Before the day of the experiments, a stock solution of the BRB80 buffer should be prepared and stored at 4°C until use. A 1 mL stock solution of BRB80 containing no nucleotide or reducing agent should last for several experiments and be made every few days. The microtubule and protein reagents should also be ready and tested prior to the start of the experiment (for example by fluorescence imaging). The mica-specimen discs could be glued on the day of the experiments. If preparing the discs on the same day, leave 1–2 hours to dry before use.

On the day of the experiments, microtubules should be polymerized and ideally checked with fluorescence microscopy as described previously (Mani, Marchan, & Subramanian, 2022). A 100 × 100 μm region should occupy >100 microtubules in a squash (Mani, Marchan, & Subramanian, 2022). The AFM liquid probe holder should be cleaned and dried. Take 5–10 minutes to thoroughly clean the probe holder to eliminate remnants from previous experiments and to increase the quality of the data. Finally, the assay mixture should be prepared and deposited on the mica and left to incubate for 5 minutes. The AFM cantilever should be fixed on the liquid probe holder during incubation.

During imaging experiments, if the microtubule density is sufficient on the mica, several microtubules should be present in the initial 20 × 20 μm scan. If no microtubules are present, move to another region on the mica disc and do a quick scan (low resolution at ~160 × 160 pixels to save time) or prepare a new assay mixture with double the microtubule concentration. Fluorescent microtubules are helpful for quickly checking the microtubule density with fluorescence microscopy before preparing the assay mixture for AFM experiments. The acquisition of a static high-resolution image takes >5 minutes after increasing the number of scan lines and scan points to ~256 × 256 pixels and scanning at 1 Hz. For time-lapse experiments, each image takes ~3 minutes when imaging at 256 × 256 resolution with a 1.5 Hz scan rate for up to ~30–40 min.

After AFM imaging, take 5–10 minutes to clean and dry the liquid probe holder. The mica discs should be stored and re-cleaved in the following experiments. Discard the microtubule and protein reagents after the experiments.

DATA AVAILABILITY STATEMENT:

The data that support the findings of this study are available from the corresponding author upon reasonable request.