Surprises in the skeletons of neurons: Super-resolution imaging leads to discovery of regularly spaced actin rings in axons

by Parizad M. Bilimoria

Taking the neuronal skeleton by STORM Xiaowei Zhuang, Ke Xu, and Guisheng Zhong (left to right) gathered around the STORM imaging apparatus used to discover the periodic actin-spectrin structure in axons. Xu and Zhong are co-first authors on the study published in Science. Courtesy of MCB Graphics.

Taking the neuronal skeleton by STORM
Xiaowei Zhuang, Ke Xu, and Guisheng Zhong (left to right) gathered around the STORM imaging apparatus used to discover the periodic actin-spectrin structure in axons. Xu and Zhong are co-first authors on the study published in Science. Courtesy of MCB Graphics.

The skeletons of neurons have long fascinated scientists. Beyond giving these cells their unique shapes and structures, the cytoskeleton plays an active role in many processes that define neurons—from the birth of axons and dendrites to the development of synaptic sites and the dynamics of neurotransmitter release. We now know the major molecular constituents of the neuronal cytoskeleton—for example, actin, which fills the tips of growth cones and the heads of dendritic spines, and tubulin, whose dimers comprise the microtubules that serve as railroad tracks for axonal or dendritic cargo. We know how fast cytoskeletal molecules polymerize, what makes them stable or unstable, and how their ends are decorated. We can purify and manipulate them in test tubes or visualize their natural dynamics in live cells. And yet, despite this intense level of scrutiny, it appears we have some very basic facts left to learn about the organization of the neuronal cytoskeleton and its associated membrane structures.

Ke Xu and Guisheng Zhong, postdoctoral fellows in the laboratory of Xiaowei Zhuang, Professor of Chemistry and Chemical Biology and Physics, a member of the Conte Center at Harvard and an investigator of the Howard Hughes Medical Institute, recently discovered that axons contain a periodic structure of actin rings. In a study published recently in Science, the trio describes the organization of these rings—detected for the first time thanks to technical advances in super-resolution imaging. Each ring is composed of short actin filaments capped on their ends by the adducin protein. The rings wrap around the circumference of the axon shaft, spaced out at 180-190 nanometer intervals by tetramers of spectrin molecules. Spectrin, like actin, is a ubiquitous skeletal protein known to be abundant in the brain.

The structure is “amazingly beautiful” and completely unexpected, Zhuang notes. “It was pretty shocking to us that there’s such a regular, periodic structure of such a long term order in the axon.”

“It was an epiphany for me to see… just beautiful,” says Vann Bennett, the George Barth Geller Professor in the Departments of Cell Biology, Biochemistry, and Neurobiology at Duke University and also an investigator of the Howard Hughes Medical Institute, who was not involved in the research. “The actin rings are precisely in register analogous to rings of a trachea.”

Actin rings in the axon (Top) STORM image of the repeating sub-membrane actin-spectrin structure found in axon shafts. Actin is shown in green and spectrin (ẞII isoform, labeled at its C-terminus) in purple.(Right) Model for the periodic actin-spectrin structure. Actin rings, composed of short actin filaments (green) capped by the adducin protein (blue), are wrapped around the circumference of the axon and separated by spectrin tetramers (purple) at 180-190 nanometer intervals. Images courtesy of Zhuang lab.

Actin rings in the axon
(Top) STORM image of the repeating sub-membrane actin-spectrin structure found in axon shafts. Actin is shown in green and spectrin (ẞII isoform, labeled at its C-terminus) in purple.(Right) Model for the periodic actin-spectrin structure. Actin rings, composed of short actin filaments (green) capped by the adducin protein (blue), are wrapped around the circumference of the axon and separated by spectrin tetramers (purple) at 180-190 nanometer intervals. Images courtesy of Zhuang lab.

model.jpg

Actin rings in the axon

(Top) STORM image of the repeating sub-membrane actin-spectrin structure found in axon shafts. Actin is shown in green and spectrin (ẞII isoform, labeled at its C-terminus) in purple.(Right) Model for the periodic actin-spectrin structure. Actin rings, composed of short actin filaments (green) capped by the adducin protein (blue), are wrapped around the circumference of the axon and separated by spectrin tetramers (purple) at 180-190 nanometer intervals. Images courtesy of Zhuang lab.

Thus far the only other cell type in which the structure of a spectrin and actin network is known is the red blood cell, where spectrin and actin are arranged into hexagonal and pentagonal networks together with associated proteins.

“Xiaowei’s images are the first ones ever to be seen outside of red [blood] cells,” says Bennett, who originally discovered spectrin in the brain. “So this is likely going to have implications for everywhere else that spectrin is located.”

In the present study the actin-spectrin structure was detected in brain slices as well as cultured neurons from rodents. It first appeared at day 5 in vitro in neurons cultured from the hippocampus (a region of the temporal lobe famous for its role in learning and memory) and remained present in mature hippocampal neurons. Day 5 in vitro is a developmental stage when the axons of hippocampal neurons have just been established and are beginning to grow longer and branch. The structure was not present in dendritic shafts at any age studied—dendritic shafts instead had long actin filaments running along their length.

Based on co-imaging studies, the Zhuang team believes that this sub-membrane actin-spectrin structure may underlie the regular distribution of sodium channels in the axonal membrane. This could be important because sodium channels are necessary for ‘action potentials’—the spikes in electrical activity that neurons use to communicate with each other. Bennett agrees, pointing out that the spectrin partner protein ankyrin is known to be concentrated in regions of the axonal membrane where there are lots of sodium channels. The newfound structure could also underlie the organization of a variety of other membrane proteins.

Additionally, since axons are extremely long and thin—in the case of the sciatic nerve in humans, running all the way from the base of the spine to the toes—and spectrin molecules are known for their flexibility, the actin-spectrin structure might be providing an elastic form of mechanical support that allows axons to remain intact as animals move. In fact, in the worm C. elegans, loss of spectrin can lead to the breakage of axons, unless the worms are paralyzed and there is no mechanical strain on the spectrin-less axons.

An unexpected pattern

I always feel like this is a sign of a really good researcher, someone who can just notice this kind of thing when it’s not that obvious to others yet.”
— Xiaowei Zhuang on the keen initial observations made by postdoctoral fellows Ke Xu and Guisheng Zhong

Xu and Zhong first noticed hints of the actin-spectrin structure when gearing up to study actin organization in synapses with stochastic optical reconstruction microscopy (STORM). STORM, invented by the Zhuang lab in 2006, is a super-resolution light microscopy technique that allows researchers to visualize structures smaller than the wavelength of light itself by taking advantage of photo-switchable fluorescent probes. This type of probe allows a pool of labeled molecules to be imaged in small subsets over time, such that the precise spatial position of individual molecules (whose images ordinarily would overlap with that of adjacent molecules) can be determined. An image of extraordinarily high resolution can then be constructed by integrating information from numerous fluorescence sampling sessions—each for the same region of space, just with different subsets of the pool of fluorescent molecules activated.

A few years ago, Zhuang’s group started using STORM to characterize in 3D the molecular architecture of synapses in rodent brains, working in collaboration with the laboratory of Catherine Dulac, Chair and Professor of Molecular and Cellular Biology and also a member of the Conte Center. Recent advances in STORM methods have significantly improved the limits of resolution since that work began, such that it’s possible now to distinguish structures smaller than 10 nanometers in the xy-plane and 20 nanometers in the z-plane (one nanometer being one-millionth of a millimeter). Xu and Zhong were eager to use these advances to visualize the actin cytoskeleton of pre- and post-synaptic structures.

But after fluorescently labeling actin molecules and imaging neurons with the latest STORM capabilities, the two were distracted by the pattern of staining in axon shafts. While faint and hard to make out at first, it looked somewhat discrete and regular—not like the continuous long lines of actin filaments they expected to find running along the length of an axon. And it was interesting how the pattern looked different in axons versus dendrites. So although not initially sure of what they were seeing, they decided to dig deeper and optimize labeling and imaging conditions to study axonal shafts instead of synaptic sites. After many rounds of optimization, the pictures grew clearer and it became evident that there was indeed a discrete and periodic pattern of actin rings.

“I always feel like this is a sign of a really good researcher, someone who can just notice this kind of thing when it’s not that obvious to others yet,” Zhuang observes, emphasizing that the discovery could easily have been missed if it weren’t for Xu and Zhong’s keen initial observation.

The subsequent finding of spectrin tetramers linking the actin rings and adducin caps within the actin rings came from screening many actin-binding partners. Essentially, the team used published data on well-known actin-associated proteins to develop their hypotheses about spectrin and adducin. For example, studies from the Bennett lab had revealed some years ago that adducin promotes the binding of actin to spectrin.

A flood of questions

“I think Xiaowei’s structural tool is like the telescope of Galileo. She can now see all kinds of things.”
— Vann Bennett commenting on the power of STORM imaging

The discovery of this new cytoskeletal structure in axons has opened the gates to a flood of questions, including how the structure develops and what its functional roles are. It is also interesting to ponder why the structure had never been described before. Many laboratories worldwide have used light and electron microscopy to study the actin cytoskeleton in neurons. And although it is much harder to label specific proteins for detection with electron microscopy than with fluorescence-based light microscopy, like STORM electron microscopy does offer an image resolution on the nanometer scale.

Perhaps, Zhuang suggests, one key issue is that actin is the predominant cytoskeletal protein in growth cones and dendritic spines, but not in axonal or dendritic shafts. The latter are packed at an incredibly high density with microtubules and intermediate filament proteins as well, so to track a specific type of filament within axonal or dendritic shafts—especially actin filaments, the thinnest of the three main cytoskeletal elements—poses a tremendous challenge. Depending on the angle at which a cell or tissue sample is cut, the actin rings of axons might appear as dots or small lines within individual electron microscopy images. So if one didn’t suspect the existence of the regular actin-spectrin structure, it would be hard to reconstruct.

The new study also suggests a number of directions for future investigation. For instance, what happens at synaptic sites, the original focus of Xu and Zhong’s work? Does the actin-spectrin structure exist in the synapse-rich terminal arbors of axons? Does the periodicity ever change, depending on a neuron’s environment and the signals it receives from other cells? Are there nervous systems diseases in which the structure is compromised, and fixing it might be an option? Zhuang’s lab is interested in all these questions and more.

In terms of diseases, Bennett comments, it will be important to investigate whether the new skeletal structure is involved in motor neuron diseases like amyotrophic lateral sclerosis (ALS). Neuroscientists have long wondered why motor neurons are selectively susceptible to degeneration in these diseases—whether the very long axons of these neurons make them the ones most likely to suffer when mechanical support systems break down. And the Zhuang lab findings will give researchers a new, axon-specific support structure to examine.

“I think Xiaowei’s structural tool is like the telescope of Galileo,” concludes Bennett. “She can now see all kinds of things.”

But instead of the distant moons of Jupiter revealed by Galileo’s telescope, STORM is uncovering the beauties hidden in our own bodies and the species around us. One can only wonder what other surprises lie waiting in the skeletons of our brain cells.

>>> Download as a PDF