Breakthrough allows scientists to watch how molecules morph into memories (w/ video)

Watching molecules morph into memories (w/ video)
Credit: Albert Einstein College of Medicine

In two studies in the January 24 issue of Science, researchers at Albert Einstein College of Medicine of Yeshiva University used advanced imaging techniques to provide a window into how the brain makes memories. These insights into the molecular basis of memory were made possible by a technological tour de force never before achieved in animals: a mouse model developed at Einstein in which molecules crucial to making memories were given fluorescent "tags" so they could be observed traveling in real time in living brain cells.

Efforts to discover how make memories have long confronted a major roadblock: Neurons are extremely sensitive to any kind of disruption, yet only by probing their innermost workings can scientists view the molecular processes that culminate in memories. To peer deep into neurons without harming them, Einstein researchers developed a in which they fluorescently tagged all molecules of messenger RNA (mRNA) that code for beta-actin protein – an essential structural protein found in large amounts in and considered a key player in making memories. mRNA is a family of RNA molecules that copy DNA's genetic information and translate it into the proteins that make life possible.

"It's noteworthy that we were able to develop this mouse without having to use an artificial gene or other interventions that might have disrupted neurons and called our findings into question," said Robert Singer, Ph.D., the senior author of both papers and professor and co-chair of Einstein's department of anatomy & structural biology and co-director of the Gruss Lipper Biophotonics Center at Einstein. He also holds the Harold and Muriel Block Chair in Anatomy & Structural Biology at Einstein.

In the research described in the two Science papers, the Einstein researchers stimulated neurons from the mouse's hippocampus, where memories are made and stored, and then watched fluorescently glowing beta-actin mRNA molecules form in the nuclei of neurons and travel within dendrites, the neuron's branched projections. They discovered that mRNA in neurons is regulated through a novel process described as "masking" and "unmasking," which allows beta-actin protein to be synthesized at specific times and places and in specific amounts.

This video is not supported by your browser at this time.
In two papers in this week's Science, researchers in the laboratory of Albert Einstein College of Medicine's Robert Singer, Ph.D., focused on beta actin, an essential protein found abundantly in neurons (nerve cells). The Einstein researchers developed a mouse in which all molecules of messenger RNA (mRNA) that code for beta actin were labeled with green fluorescent protein. After stimulating cultured neurons and brain tissue of the mouse, they were able to observe fluorescently tagged molecules of beta actin mRNA within the living neurons. In the first study, led by Hye Yoon Park, Ph.D., the researchers observed beta-actin mRNA being transcribed by the beta actin gene and then traveling within the neuron's branched projections, or dendrites, to sites where synthesis of beta actin protein would occur. This movie shows beta-actin mRNA molecules traveling within the dendrites of a cultured live hippocampal mouse neuron. Credit: Hye Yoon Park, Ph.D.

"We know the beta-actin mRNA we observed in these two papers was 'normal' RNA, transcribed from the mouse's naturally occurring beta-actin gene," said Dr. Singer. "And attaching green fluorescent protein to mRNA molecules did not affect the mice, which were healthy and able to reproduce."

Neurons come together at synapses, where slender dendritic "spines" of neurons grasp each other, much as the fingers of one hand bind those of the other. Evidence indicates that repeated neural stimulation increases the strength of synaptic connections by changing the shape of these interlocking dendrite "fingers." Beta-actin protein appears to strengthen these synaptic connections by altering the shape of dendritic spines. Memories are thought to be encoded when stable, long-lasting synaptic connections form between neurons in contact with each other.

The first paper describes the work of Hye Yoon Park, Ph.D., a postdoctoral student in Dr. Singer's lab at the time and now an instructor at Einstein. Her research was instrumental in developing the mice containing fluorescent beta-actin mRNA—a process that took about three years.

Dr. Park stimulated individual hippocampal neurons of the mouse and observed newly formed beta-actin mRNA molecules within 10 to 15 minutes, indicating that nerve stimulation had caused rapid transcription of the beta-actin gene. Further observations suggested that these beta-actin mRNA molecules continuously assemble and disassemble into large and small particles, respectively. These mRNA particles were seen traveling to their destinations in dendrites where beta-actin protein would be synthesized.

In the second paper, lead author and graduate student Adina Buxbaum of Dr. Singer's lab showed that neurons may be unique among cells in how they control the synthesis of beta-actin protein.

"Having a long, attenuated structure means that neurons face a logistical problem," said Dr. Singer. "Their beta-actin mRNA molecules must travel throughout the cell, but neurons need to control their mRNA so that it makes beta-actin protein only in certain regions at the base of dendritic spines."

Ms. Buxbaum's research revealed the novel mechanism by which brain neurons handle this challenge. She found that as soon as beta-actin mRNA molecules form in the nucleus of hippocampal neurons and travel out to the cytoplasm, the mRNAs are packaged into granules and so become inaccessible for making protein. She then saw that stimulating the neuron caused these granules to fall apart, so that mRNA molecules became unmasked and available for synthesizing beta-actin protein.

But that observation raised a question: How do neurons prevent these newly liberated mRNAs from making more beta-actin protein than is desirable? "Ms. Buxbaum made the remarkable observation that mRNA's availability in neurons is a transient phenomenon," said Dr. Singer. "She saw that after the mRNA molecules make beta-actin protein for just a few minutes, they suddenly repackage and once again become masked. In other words, the default condition for mRNA in neurons is to be packaged and inaccessible."

These findings suggest that neurons have developed an ingenious strategy for controlling how -making proteins do their job. "This observation that neurons selectively activate protein synthesis and then shut it off fits perfectly with how we think memories are made," said Dr. Singer. "Frequent stimulation of the neuron would make mRNA available in frequent, controlled bursts, causing beta-actin protein to accumulate precisely where it's needed to strengthen the synapse."

To gain further insight into memory's molecular basis, the Singer lab is developing technologies for imaging neurons in the intact brains of living mice in collaboration with another Einstein faculty member in the same department, Vladislav Verkhusha, Ph.D. Since the hippocampus resides deep in the brain, they hope to develop infrared fluorescent proteins that emit light that can pass through tissue. Another possibility is a fiberoptic device that can be inserted into the brain to observe memory-making hippocampal neurons.

The titles of the two Einstein papers are: "Visualization of Dynamics of Single Endogenous mRNA as Labeled in Live Mouse," and "Single Beta-actin mRNA Detection in Neurons Reveals a Mechanism for Regulating Its Translatability."

More information: "Single β-Actin mRNA Detection in Neurons Reveals a Mechanism for Regulating Its Translatability", Science, 2014.
"Visualization of Dynamics of Single Endogenous mRNA as Labeled in Live Mouse", Science, 2014.

Related Stories

New research provides insight into epilepsy

Dec 20, 2013

Experiments using mice have led to new research results showing that the amount of microRNA-128 has a great impact on the musculoskeletal system. If the level of microRNA-128 is increased, it leads to lower ...

Scientists observe single gene activity in living cells

Apr 21, 2011

Researchers at Albert Einstein College of Medicine of Yeshiva University have for the first time observed the activity of a single gene in living cells. In an unprecedented study, published in the April 22 online edition ...

Recommended for you

Brain encodes time and place of taste memory

15 minutes ago

Have you ever eaten something totally new and it made you sick? Don't give up; if you try the same food in a different place, your brain will be more "forgiving" of the new attempt. In a new study conducted ...

User comments

Adjust slider to filter visible comments by rank

Display comments: newest first

tomcloyd
5 / 5 (1) Jan 23, 2014
It would be nice if this piece were NOT so misleadingly headlined.

Molecules do NOT morph into memories. Neural networks do. It's that simple.

Who writes this stuff?
JVK
not rated yet Jan 23, 2014
Experimental evidence shows that molecules morph into memories without telling us anything about the stimulus that is responsible for memory formation or how any stimulus results in morphological changes in synapses that involve messenger RNA.
tomcloyd
not rated yet Jan 23, 2014
Molecular encoding of external information? OK - demonstrate that and you have a Nobel prize.

The synaptic changes are part of the neural net formation, said neural net being where the pattern that is the memory is encoded. This is elementary neurology.
JVK
5 / 5 (1) Jan 23, 2014
Are you saying that the cause of morphological changes in different cell types is elementary neurology, or that you don't have a clue about how the molecular encoding of external information occurs in the context of Systems Biology from atoms to ecosystems?
JVK
not rated yet Jan 23, 2014

tomcloyd
5 / 5 (1) Jan 24, 2014
I'm saying that information does NOT code at the molecular level. Molecules cannot do that. What they CAN do is participate in the restructuring of neural networks which acquire the capacity for recognizing complex patterns of perception, and stimulating responses - or providing identification of a pattern so that some other part of the meta-network can choose a response to emit. What I just described is elementary neurology. Most of it's been worked out in rather primitive animals, with simple brains which permit the study of morphological changes in simple neural networks in response to environmental changes. Molecules are obvious involved in the encoding of the learning, but the learning is NOT in the molecules, but rather in the structures of which they are a part - the networks of neurons involved.

None of this is new, actually. The article headline is therefore highly inappropriate.
tomcloyd
5 / 5 (1) Jan 24, 2014
A highly authoritative source for what I'm saying - a book that should be available in any decent college library.

"Memory storage is the change in probability of activating a particular neural network pattern in the future." Obviously, before learning occurs and memory is created, that probability is zero.

Siegel, D. J. (2012). The developing mind : how relationships and the brain interact to shape who we are. New York: Guilford Press. p. 48.

Siegel is a Clinical Professor of Psychiatry at UCLA.
russell_russell
5 / 5 (1) Jan 24, 2014
@tom
You are correct...if DNA is not molecular, and has nothing to do with information, storage or memory. Congratulations.

A probability of zero implies zero interaction no matter if the variable time is considered or not.

"I'm saying that information does NOT code at the molecular level. " - Tom
What is DNA?
russell_russell
not rated yet Jan 24, 2014
"Dr. Park stimulated individual hippocampal neurons of the mouse and observed newly formed beta-actin mRNA molecules within 10 to 15 minutes."

What accounts for short-term memory? For example, repeating what you stated a second ago, two seconds ago, three seconds ago, etc., etc..
tomcloyd
3.7 / 5 (3) Jan 24, 2014
Russell - I cannot trot out a course in elementary neurology to clear this up. I think you're being argumentative. Of course DNA encodes information, but it is not learning, and learning is the subject of the research reported above. DNA is a product of evolution. Learning is a product of environmental interaction. Both obviously involve molecules, and I've never said otherwise. But even with DNA, it is groups of molecules acting together that get the job done. They aren't organized in network, as are neurons, in the brain. Neurons do THEIR work using molecules, of course, but learning - memory - is not IN the neuron, much less in the molecules of the neuron. If you understood the brain as an information PROCESSING system, I wouldn't be having to go over any of this.

By analogy, my word processing dictionary is composed of atoms and molecules in patterns, but we never deal with it at that level.

Short term memory is neural nets also; they just aren't set up to endure.
micahscopes
5 / 5 (2) Jan 24, 2014
@tomcloyd you guys are arguing the trees vs. the forest. The molecules and their situation are both completely implicated in memory.

It's like trying to understand where economic value is stored. Is it stored in social relationships? Does it exist in the physical work people do, the valuable materials we exchange? Does physical money itself embody the economy? The spreadsheets in the corporate office? Well the answer is yes to all these questions, because economic value moves through and is hinged on all of these at some point or another. And just like memory doesn't necessarily depend on biological molecules, economic value can be stored on a computer instead of a piece of paper.

Lately there have been a number of studies showing that protein formation is critical to the formation and recollection of certain kinds of memory. This is another one.

Can art exist without sculptures? Sure, but sculpture is still a medium for transmitting art.
micahscopes
4.5 / 5 (2) Jan 24, 2014
@tomcloyd: this isn't obvious elementary neurology. it's an active field of research in different fields of science and mathematics!
tomcloyd
not rated yet Jan 24, 2014
Micah, again: memory necessary involves molecules, so studying molecules will inform us about aspects of memory, for sure.

Here's another example: I have two piles of bricks and mortar. Both are equal in quality and size. One is arranged in a building, and other is not. At the level of bricks and mortar they are indistinguishable. But they clearly are not the same on some other level. The difference? It's what's called a gestalt or system property. Add the invisible factor of organization (neural net organization, in the brain), and you get a building. In the brain, environmental interactions build these "buildings", but obviously without quality brick and mortar it is to no avail.

You can look at the molecules all day and you will not find memory, any more than looking at the bricks will give the building.

I must journey on. I deal with buildings. I'm a psychotherapists. Molecules are NOT what I work with. Thanks to all for an interesting discussion.
JVK
2 / 5 (1) Jan 24, 2014
Olfactory/pheromonal input causes the epigenetic landscape to become the physical landscape of DNA via the conserved molecular mechanisms of de novo gene creation in the organized genomes of all species.

The omission of epigenetic cause and effect in these articles merely attests to the probability that the authors have not grasped Kohl's Laws of biology:

Law #1) Finding food -- and remembering how your innate response (at the molecular level) to food odors allowed you to do that -- is a 'condition of life.'

Food is metabolized to species-specific pheromones.

Law # 2) The response to pheromones (at the molecular level) controls the physiology of reproduction via the innate ability to recognize the difference between food and conspecifics.

These Biological Laws would have been obvious to Darwin if he had known anything about genetics. However, Law #1 should be obvious to anyone who has not accepted mutation-driven evolution, since it doesn't occur in species that starve to death.
Surly
5 / 5 (1) Jan 24, 2014
I'm saying that information does NOT code at the molecular level. Molecules cannot do that. What they CAN do is participate in the restructuring of neural networks which acquire the capacity for recognizing complex patterns of perception, and stimulating responses - or providing identification of a pattern so that some other part of the meta-network can choose a response to emit.

That's true for short- and intermediate-term memory. Blocking protein synthesis blocks the formation of long-term memories, which is evidence that long-term memories are stored in molecules. The evidence for the reconstructive memory hypothesis also suggests that memories are stored physically, not just in synaptic changes.
Porgie
not rated yet Jan 25, 2014
So will this allow you to implement math formulas into my brain and let me speak French?
tomcloyd
not rated yet Jan 25, 2014
I'm saying that information does NOT code at the molecular level....

That's true for short- and intermediate-term memory. Blocking protein synthesis blocks the formation of long-term memories, which is evidence that long-term memories are stored in molecules. The evidence for the reconstructive memory hypothesis also suggests that memories are stored physically, not just in synaptic changes.


That is a non sequitur. That protein synthesis is essential to formation of long term memory does NOT mean that it IS long term memory. Carbon is also essential to formation of long term, memory. So...? You have a conspicuously missing minor premise. Isn't construction of valid arguments taught any more?
tomcloyd
5 / 5 (1) Jan 25, 2014
So will this allow you to implement math formulas into my brain and let me speak French?


Your sarcasm indicates you get it, for the notion of "molecular memory", discredited quite some time ago, implies that what you desire is indeed possible. It turned out not to be, for the simple reason that memory, upon careful examination, was found to be exactly as I've described it: encoded in neural networks, NOT molecules. No neuropsychologically-informed person has proposed the molecular memory idea for a long time.
antialias_physorg
not rated yet Jan 26, 2014
I'm saying that information does NOT code at the molecular level.

Neither are they. So what's your problem?
anywallsocket
not rated yet Feb 17, 2014
It would be nice if this piece were NOT so misleadingly headlined.

Molecules do NOT morph into memories. Neural networks do. It's that simple.

Who writes this stuff?


It is misleading because it's a title and is meant to be hyperbolic and catchy so it yields more views. The writers write this stuff, and the writers write for the website.
tomcloyd
not rated yet Feb 17, 2014

It is misleading because it's a title and is meant to be hyperbolic and catchy so it yields more views. The writers write this stuff, and the writers write for the website.


Precisely. And this sort of thing is NOT responsible science reporting, due to the blatant mislead. Since there's been no change, they obvious don't give a damn. Impressive. Also very responsible. Not.

I no longer regard this site as a prime source of science reporting, and will say so to those who look to be for advice.