Yang Lab Research Update

Ge Yang at Histochemistry 2012
Ge shows his best pre-QSTORM imaging tracking the movement of cargo vesicles along microtubules in Drosophila, and discusses the quest to understand the possible roles of the tau protein in regulating and sometimes disrupting that traffic. We see his first efforts to transition from microinjecting quantum dots to harnessing endocytosis by including the QDs in the cell growth medium. Early QSTORM imaging of Tau looks very promising.



Karine and Carol Lynn teamed up to write this summary of the Research Update Presentation delivered by Ge Yang at the Athens, Georgia Meeting on December 11, 2011.

Ge Yang’s Lab: Neural Transport in Fruit Flies

Drosophila Melanogaster - The common fruit fly.

Ge’s QSTORM group at Carnegie-Mellon includes graduate students Yiyi Yu (doing wet lab work) and Minhua Qiu (doing both wet lab and computation). Two undergraduate students – Stacy Lee and Breanna Stillo- are also contributing. Ge is applying for supplemental funding in 2012 to hire one additional undergraduate research student.

This Carnegie-Mellon laboratory is working toward applying the QSTORM technique to studying the transport of resources through the axons of nerve cells in fruit flies. Why do we care?   Transport of  materials and resources within nerve cells is critical to their survival and proper functioning; defects in so-called “cargo transport” have been implicated in many aging-related neurodegenerative diseases like Alzheimer’s. Ge wants to use the QSTORM imaging technique to get a better understanding of how transport is regulated in neurons, and to test out a particular hypothesis that overproduction of a particular protein, Tau, is strongly linked to breakdowns in the transport system, leading to “roadblocks,” neurodegeneration, and cell death.


Our brains and nervous systems are largely made up of neuron cells which process and transmit information to and from every region of our bodies.

Anatomy of a Neuron – Protein synthesis takes place in the cell body (purple) surrounding the nucleus (green). Dendrites around the cell body (also in purple) receive signals (electrochemical charges) from the axon terminals of other nerve cells. The cell transmits these signals out through its long axon (covered in myelin sheathing –yellow) toward the axon terminals to be sent to the next neuron.  A single axon can have multiple terminals forming synapses with the dendrites of other nerve cells.

Cross-section of a Synaptic Terminal –  The synapse is a tiny gap between an axon terminal of one nerve cell and the dendrite of a receiving nerve cell. Neurotransmitters emitted by the sending cell carry the charge signal to receptor molecules on the surface of the dendrite of the receiving nerve cell.


Cartoon of axon transport (not to scale) – A single axon contains dozens of hollow 25-nm diameter microtubules running along its length from cell body to terminus (only two are shown here).

The microtubules serve as two-way highways for the transport of cargo. Most of the cargo is transported within “vesicles,” membrane wrapped containers. The vesicles carry essential cell nutrients and building materials (e.g., for neurotransmitters) to the far (distal) end of the axon, and return spent materials for recycling to the cell body. Keep in mind that in our bodies a single axon can extend more than a meter, with a diameter of less than 2 microns; thus, they are very long compared to their tiny width. In fact, if they were as wide as a single hair, they would be 30-50 meters long.  So the molecular motors that move cargo along these very long narrow highways need to be pretty robust.

Sketch of motor proteins moving a single vesicle along a microtubule – A variety of different interlocking proteins provide the mechanism for transport of the vesicle along the microtubules. Kinesin protein complexes serve as motors pulling the vesicles toward the positively charged axon terminal. Dynein protein complexes drive the cargo assembly in the reverse, or negatively charged direction, toward the cell body. It is unclear how the movement of cargo is regulated, so that the right cargo is delivered to the right place at the right time. Dozens of different kinds of microtubule-associated proteins have been identified that may play regulatory roles.

Role of Tau Protein – One of the microtubule-associated proteins, Tau, is thought to have a role helping to stabilize the microtubules and regulating cargo transport. Lying across the microtubules, a Tau protein acts as a kind of roadblock. However, there is a mechanism for removing the roadblock when cargo needs passage.  (Images from Ballatore et al, Nat. Rev. Neurosci. 2007; Morris et al, Neuron, 2011)

It seems that a phosphate group binds to the Tau protein, lifting it off the microtubule to allow cargo passage, and then loses the phosphate group and reattaches after the cargo passes. It’s still a mystery how the Tau protein “knows” when to take on a phosphate group, or “phosphorylate,” to lift it off the microtubule and allow cargo to pass on by.

Studies have shown that Tau appears in higher concentrations toward the “distal” or far ends of axons. This may be due to the greater regulation needed close to the synaptic terminals, where neurotransmitters leave the axon to transfer to neuroreceivers on the dendrite of another nerve cell.  (Perhaps it’s like the line up at a port for cargo loading and unloading.)

How Are Tau Proteins Associated with Alzheimer’s Disease? With Alzheimer’s Disease, post-mortem biopsies often show clumps or tangles of  Tau proteins in neural tissue.      In landmark research done with fruit flies by Wittmann, et al (Science, 2003), Tau tangles were not present in flies with induced Alzheimer’s, but abnormal configurations and phosphorylations of Tau were, leading the researchers to suspect that it was these earlier changes to the Tau protein and their resulting effects that led to neurodegeneration, not the resulting tangles that are found in human and other models.   Ge thinks that the nanometer-scale resolution capability of QSTORM imaging will allow him to investigate the behavior and accumulation of Tau in live fruit fly nerve cells, where active transport along the axon microtubules can be observed.  Rather than relying on post-mortem tissue analysis, QSTORM will allow him to see how changes in the Tau affect transport in action inside neurons.

Fruit Flies. Fruit flies are an excellent model organism for this type of work because they are well studied; much of their genetics, development, and anatomy is well mapped out. It is relatively easy to make and test the effects of gene mutations. Also, their larvae are transparent.


Micron-scale imaging in fruit fly larvae
Ge’s group began making micron-scale images of cargo transport in live fruit fly larvae modified by a gene mutation that allowed only a single neuron in each nerve bundle to express fluorescent green protein along the length of its axon. Since each nerve is made up of dozens of neurons, this simplified the imaging task.
Fig. 1. Fruit fly larva (3 mm) showing vertical sectioning (denticles).

Fig. 2.  Imaging revealing brain and nerves connecting to segments of the fly’s musculoskeletal system.  A single neuron within each nerve bundle  is tagged with green fluorescent protein.

Fig. 3. Schematic of fly’s nervous system. Small denticle-associated segments along each neural pathway are delineated with red boxes, A1, A2, etc.

Fig. 4. Close-up of the A3 axon segment, showing fluorescent-tagged vesicles. Imaged with wide field microscope.

YouTube Preview Image

Fig. 5. Movie showing movement of vesicles to and from distal end in segment A3.  

Analyzing Vesicle Transport

As part of his investigation of the role of Tau in regulating cargo transport in nerve cell axons, Ge wants to measure the relative speed of vesicles as they move back and forth along the microtubules in various segments of nerve pathways between the central nervous system and the synaptic terminals at the distal end of the axon.  Do the vesicles slow down as they approach the synapse, and is that slow-down associated with an accumulation of Tau protein “roadblocks?”  Is that condition reversible, or does it lead to cell death?

The next set of images demonstrate how Ge’s team measures the speed and direction of vesicle transport in each consecutive segment of a nerve pathway.   The time lapse images are layered to show the traces of fluorescence left by each individual vesicle.

Fig 1.   For each segment of a nerve pathway, the movement of individual vesicles is tracked over a period of time.


Fig. 2.  A plot of distance over time reveals the movement of individual vesicles heading both outward (anterogade) toward the synapse and inward (retrograde) toward the cell body, along the axon.  Vertical lines represent stationary vesicles. The more slope to the line, the greater speed of transport in both the anterograde and the retrograde directions.


Fig. 3  Below.  A side-by-side comparison of the speed of vesicle movements along axon segments A4 and A5 in three genetically distinct fly larvae, including a control normal larva, and two types of larvae with disease-level Tau expression.  The comparison shows higher velocity of movement in both directions in the control group.  The first mutant Tau model (hTauWT), shows many stationary vesicles, especially toward the distal end of the axon (the right end of segment A5), but still shows some vesicles continuing to proceed toward the synapse.  The second mutant Tau R401 type shows a total cessation of forward (distal) movement. All movement seems directed back toward the cell body.

These data support the hypothesis that Tau expression affects the speed and direction of vesicle movement.   However, the vesicles targeted in this experiment were associated with the transport of a beta amyloid precursor, which is already associated with Alzheimer’s Disease and neuron degeneration.  The results were not as consistent when imaging the movement of other types of vesicles.

A brief reflection on the imaging technology used in Ge’s lab

Ge’s lab does not have a STORM imaging set-up, and he took a moment to explain how he is nonetheless able to get the visual data he is showing us, using a wide-field microscope.

His lab acquired a piezo-electric crystal stage that can be moved with remarkable nanoscale precision. His lab rests on granite bedrock, common in the Pittsburgh region.

The stage is about 6 inches across.

Ge also acquired software that enabled him to obtain  location information more precise than normally allowed by the 250 nm diffraction barrier.  To see how this is done, notice below how the location of the point spread functions (PSFs) of two different sources of light can both be described as occurring in pixel 2,2.  And yet, it is clear to the eye that the first one is located toward the upper left of that pixel, and the second one is located toward the lower right of than pixel.  The software enables the use of data showing the spillover of light into neighboring pixels, using it to better approximate the center of the PSF, where the tiny source of light is located.




In order to test the stability of the new stage and the accuracy of the localization software, Ge’s team fixed fluorescent beads to a slide, and controlled the stage to move in various nanoscale increments.

The graphs below plot the directed, controlled movement of the piezo stage (in red) and the locations of a fluorescent beads as reported by the software (in green).

The first plot shows excellent correspondence between the software’s indication of the position of the bead and the controlled movement of the stage in 100 nm increments. The second plot, in which the stage is moved in 24 nm increments, shows a growing discrepancy between the known positions and the positions being reported by the software. The third plot, where the incremental movement has been reduced to 16 nm, begins to show increasingly erratic fluctuation of the software’s readings. On the other hand, it is easy to see that the piezoelectric stage itself is showing very little variation. The stage is resting on an air table which cushions it from external vibration from the room, and the building the apparatus rests in may be anchored on top of granite bedrock. Ge estimates, from these plots, that the stage (not the software) is providing single nanometer stability, which is why he is recommending the same experimental set-up to Peter.   On the other hand, these tests show that the software, although greatly improving the resolution available through the wide-field scope, is limited to about 20 nm precision.  This is why Ge is excited about being able to work with Peter and Jessica to get the resolution that QSTORM should provide.

Back to the Imaging Data – Tackling the Imaging of Tau Proteins

The mixed vesicle velocity data motivated Ge to try to identify the distribution patterns of the Tau proteins within the axon. This is very hard to do with current imaging resolution, because there are so many microtubule-associated proteins operating within the 1-2 micron diameter of an axon (perhaps 70), and many others operating near the axon membrane. Ge believes that if QSTORM is successful, biologists will have sufficient resolution to track the activities of each one of these proteins in the axon, and thus to better understand the dynamics of axon transport overall. Ge’s team began this effort to image the Tau protein using PC-12 cells (derived from rat medulla nerve cells).

The team used antibodies to attach Alexa 488 dye molecules to Tau proteins in the PC-12 cells, in a process called immunostaining.  They then used nerve growth factor (NGF) to induce the cells to grow long thin “processes” or “neurites” that are anatomically similar to axons.  The team sent the cells to Peter Kner’s lab for STORM imaging.  Below, is the STORM image Andrew sent back. The detailed localization of Tau proteins along the neurite below shows far higher resolution than previous images of Tau in nerve cells.

Ge would like to be able to compare the distribution of Tau alongside the distribution of transport blockages along microtubules in the axon – to determine if there is indeed some correlation between transport blockage points and the presence of Tau. Transport blockage areas (see below) seem to come in near-regular patterns – could these be related to the relative spatial concentration of Tau protein in axons?  Peter asked at this point whether two-color STORM imaging would be useful for the biologists, and they both heartily agreed.

First Attempts to Substitute Quantum Dots for the Alexa Dye

Ge’s group did not have success microinjecting QDs into the PC-12 cells. The QDs killed the cells. The students, Breanna and Stacy, then tried a number of things:

•   With standard QDs from Invitrogen, the team observed a lot of clumping on the uncoated tip of the pipette. Occasionally they were successful, but the cells died. (In fact, Beth Brainerd’s team at Brown advised Ge not to try coating the tip of the pipette; when they tried it, they found the coating caused quenching of the QDs.)

•   Ge’s students had success microinjecting dextran (a glucose-based dye) into individual PC-12 cells, proving they had mastered the microinjection technique. But co-injecting dextran with the QDs failed.

•    They next tried PEG-coated QDs, recommended by Gang Ruan in Jessica Winter’s lab, but that did not work either.

•    They then received micelle-coated QDs (MQDs) from Jessica’s lab. These generated less clumping, and they had success getting them into the cells. Like Beth’s team, they were surprised by the rapid diffusion of the fluorescent dots throughout the cell (they expected to be able to see individual QDs), but the cells still died. (Jessica added in here that she had partnered with another biologist to try to inject the MQDs into another kind of cell and they noticed – not rapid diffusion – but aggregation and clumping of the MQDs in the cell.)

The conclusion of the discussion that ensued next among the QSTORM PIs in the room was that Jessica would try making some DSP-Peg coated Invitrogen QDs to send to both Ge and Beth, to see if the different coating and the smaller size of a single coated QD makes a difference.

Next Steps

•   Ge is going to abandon work with PC-12 cells, which, after all, are not real nor are they live functioning neural cells. He will move on to work with live mouse hippocampal neurons.
These mouse hippocampal neurons are more difficult to culture, but ultimately more relevant to neurobiology. Ge will first work with Peter to test STORM imaging of the axons using Alexa dye immunostaining. This will require “fixing” the cells. That kills them, and so this first experiment with them will not reveal any protein or transport activity. Ge hopes to quickly move on to imaging live cells with QDs, and to writing a joint paper with Peter.

•    Meanwhile, Ge’s students will need to figure out the best way to inject QDs into the mouse hippocampal neurons. Ge suspects that one has to develop a custom protocol for microinjection of QDs into each specific type of cell line, so there was no point in continuing to try to find the best way to microinjecting various types of QDs into the PC-12 cells.
•    Acting on Beth’s advice, Ge will try the intermediate step of practicing with fixed cells. That will allow him to quickly test new QDs in the system, such as the Invitrogen QDs with DSP-PEG coatings that Jessica plans to send them both.
•    Later, Ge plans to transition to Drosophila nerve cells, his real target. However, imaging those cells live will require the use of adaptive optics, because they will be thicker samples.
•    Ge plans to write a further grant to get additional funding to study the regulation of transport in neurons; it is widely recognized as an important focus for research.