Skip to Main Content

NEW YORK — Andre Fenton got to design his lab when he joined the Center for Neural Science at New York University a few years ago, and he made sure that he had a lot of closet space. But his closets do not contain brooms or shoes.

Each one is lined with black curtains and has wires and cameras hanging from the ceiling. In the middle of each closet is a disk where Fenton places mice or rats. As the rodents explore the arena, they soon discover that one section delivers a shock. It’s a lesson they don’t soon forget.


And by watching them learn that lesson, Fenton can learn how memories are made.

Fenton and a long-time collaborator, Todd Sacktor, have used these arenas to come to a remarkable conclusion about these animals’ brains: a single molecule, according to the scientists, alters neurons to form long-term memories.

To be more precise, Fenton and Sacktor came to that conclusion a decade ago. In 2006, they published their report on the molecule, called PKMzeta, to immediate acclaim. Science named it one of the great scientific advances of the year. Other scientists followed their path and began studying PKMzeta as well.


The research promised to illuminate what happens in our own brains since mammalian brains are similar to each other in so many ways.

But in 2013, everything came to a screeching halt. Two teams of scientists published back-to-back papers indicating that mice didn’t need PKMzeta to make memories. Many scientists decided Fenton and Sacktor had been wrong.

Fenton and Sacktor did not surrender, though. Instead, they have been fighting to restore PKMzeta’s good name.

This May, they declared that it’s their opponents who were wrong, and presented a series of experiments to show why. PKMzeta remains a crucial memory molecule, they argue, and they’re now moving ahead with research to understand how it works — research that may someday point toward new treatments for disorders like Alzheimer’s disease.

In an age when we get a lot of our medical news in click-baity headlines and hasty tweets, it’s easy to believe that scientific research is constantly barreling forward like a jet. The saga of PKMzeta shows just how contorted the true path of science can be. Fenton and Sacktor have worked together with what some of their colleagues consider a near-obsession for 14 years on PKMzeta. And after all that work, and years of setbacks, they feel like only now are they just able to start figuring out what this molecule does to let us form memories.

“Now we know it’s important,” Fenton said in a recent interview. “But what is it actually doing? I can tell you it’s crucial biochemically, but you shouldn’t feel satisfied that you understand memory any better. Because you don’t. We still don’t know what it means for memory.”

Dr.Todd Sacktor
Todd Sacktor at his SUNY Downstate lab in Brooklyn, N.Y. Chantal Heijnen for STAT

The rodent arena

Fenton’s path toward a career studying memory started with a single university lecture on frog eyes. Until then, he had been interested in philosophy and literature, contemplating the nature of reality and how we understand it. But then one of his professors at McGill University in Montreal explained how a researcher named Jerry Lettvin studied the optic nerves of frogs in the 1960s.

Lettvin found that the nerves fired electrical impulses only in response to certain stimuli. Small, black, moving objects — a fly, for instance — caused the nerves to fire. Small, black objects that didn’t move — like a dead fly — left the nerves quiet.

“To me that was the answer to the ‘What’s real?’ question,” said Fenton. Our brains have evolved to perceive only certain things and store them in our memory. “And there’s no way a philosopher would understand that, but neurobiologists could with electrodes. So that seemed the right thing to study.”

Fenton switched to studying neurobiology, programming a computer to play sounds to crickets. After graduating from college in 1991, he headed to Prague to work for a legendary neuroscientist at the Institute of Physiology named Jan Bures.

“Bures studied every part of the brain,” said Fenton. “He said, ‘Pick a part.’”

Fenton chose to study a small region of the brain called the hippocampus, which is crucial for memory. Bures’s lab, meantime, turned out to be a good place to learn how to become a tinkerer.

“It was Eastern Europe, so they didn’t have a lot of money, but they were smart and clever,” said Fenton.

After two years in Prague, Fenton moved to Brooklyn, N.Y. to get his PhD at SUNY Downstate Medical Center. He wasn’t satisfied with the existing tests scientists used to study memory, and so he built himself a new one: his circular arena.

When he put a mouse or a rat into the arena, it quickly discovered a section of the disk where it got a mild electric shock. By inserting electrodes into the brains of the animals, Fenton and his colleagues could eavesdrop on neurons in the hippocampus as the animals committed the location to memory.

Fenton’s research revealed, among other things, that the rodents can simultaneously build different mental maps — based on vision or smell, for example. They can make themselves the center of their map, or use landmarks to construct them.

“I said, ‘I’ve found this molecule I think is important. But can you help me figure out how to study it in memory?’”

Todd Sacktor, SUNY Downstate

At Downstate, Fenton met Sacktor, who had also been studying memory for years, but in a profoundly different way. Instead of observing live animals and listening in on their brains, Sacktor was sifting through the molecules that neurons make when new memories form.

In the 1960s, scientists found that neurons go through a molecular upheaval during learning. The result of all this biochemistry is the sprouting of new projections, called synapses, that let neurons communicate with each other. As a result, when one neuron fires an electrical impulse, a linked neuron will be more likely to follow suit. This process, known as long-term potentiation, is believed to strengthen associations in the brain.

After the discovery of long-term potentiation, a number of scientists probed neurons to figure out how they were growing synapses. As a graduate student in the 1990s, Sacktor became convinced that some unknown molecule established those long-term memories and maintained them.

Studying rat hippocampus cells in a dish, he noticed a promising protein that hadn’t been observed before. He dubbed it PKMzeta. The more he studied it, the more convinced he became that it was crucial for memory. It was only produced in the brain, appearing only in the synapses of neurons during long-term potentiation.

To see just how important it was, Sacktor and his colleagues ran an experiment on the neurons. They zapped the cells with electric pulses to create strong connections between them. Then they injected a chemical called ZIP, which blocks PKMzeta. After the injection, the scientists discovered, the long-term potentiation disappeared. Sacktor became convinced that PKMzeta wasn’t just an important molecule. He decided it was necessary and sufficient, on its own, to lock in memories.

The work took years, and got Sacktor little attention. But he kept at it, convinced PKMzeta was important.

“In a day and age where people jump on bandwagons or are influenced by fashion, I find myself full of admiration for someone who works out a problem and sticks with it and carries it through either to success or the bitter end,” said Richard Morris, a neuroscientist at the University of Edinburgh. “There are not so many people in science like that.”

Memories wiped away

Sacktor knew that few scientists would be persuaded by findings from cells in a dish. He needed to show that PKMzeta mattered to the brains of living animals. As an expert on biochemistry, Sacktor didn’t know much about running an experiment on memory itself. So he approached Fenton for help.

“I said, ‘I’ve found this molecule I think is important,’” Sacktor recalled. “‘But can you help me figure out how to study it in memory?’”

The offer put Fenton outside of his own comfort zone. When he studied rats and mice, he took it for granted that some mysterious molecules were creating and maintaining long-term potentiation, thus forging new memories. But he had no idea what those molecules were. “I was going to leave that for someone else to figure out,” said Fenton.

While Fenton was skeptical that PKMzeta would be as important as Sacktor believed, he agreed to use his arena to test the idea. The two scientists began designing experiments to observe PKMzeta in living animals.

In one experiment, Fenton’s postdoctoral student, Eva Pastalkova, developed a way to study long-term potentiation in a rat. She inserted electrodes into a rat’s brain and delivered tiny pulses of electricity to excite a few neurons in the hippocampus. That triggered long-term potentiation in the cells. She and her colleagues then doused those neurons with ZIP. When the ZIP blocked PKMzeta in the rats, their long-term potentiation disappeared.

That success gave Fenton and Sacktor the confidence to turn their attention from long-term potentiation to full-blown memory. They put rats in the arena and let them learn to avoid the shock-delivering region. A day later, the scientists injected ZIP into the hippocampus of the rats to block PKMzeta. When the scientists put the rats back in the arena, they blundered back into the shock zone.

In other words, the scientists had wiped the memories away.

When Fenton, Sacktor, and their colleagues published their results in 2006, PKMzeta immediately became a hot commodity among memory researchers. Fenton and Sacktor got big grants to dive deeper into their research, and over 100 other labs started playing with PKMzeta, too.

“We were saying, ‘We could do something nobody had done before, and you could try it, too, and it’s really simple,’” said Sacktor.

The other researchers found that blocking PKMzeta caused animals to forget all sorts of memories, such as bad tastes. “The majority of types of memories could be erased, even if you erased them three months later,” said Sacktor.

Sacktor and Fenton knew they had skeptics who doubted that a single molecule could be so essential for memory. “A lot of people didn’t like it — at all,” said Fenton. But for years those grumbles remained faint and distant.

And then, the day after New Years 2013, the grumbles grew as loud as a thunderclap.

One of the many scientists who was intrigued by Fenton and Sacktor’s paper was Richard Huganir, the director of the department of neuroscience at the Johns Hopkins University School of Medicine. Huganir wanted to know exactly how PKMzeta oversaw the formation of memories.

He and his colleagues decided the best way to go about investigating that question was to create genetically engineered mice — so-called knockouts — that couldn’t make any PKMzeta at all. If Fenton and Sacktor were right, the mice should not have been able to form long-term memories.

But that wasn’t what Huganir found.

“The mice are actually fine. In fact, they’re really fine,” he said. “We were quite surprised.”

Hunagir and his colleagues then repeated the key step in Fenton and Sacktor’s procedure: After the knockout mice formed new memories, the scientists injected ZIP into their brains. If ZIP only targeted PKMzeta, then it shouldn’t have any effect on mice that couldn’t make the molecule. Once again, the experiment didn’t go the way they expected. ZIP made the mice forget.

The scientists concluded that some other molecule was being affected — perhaps a number of them. While they had no idea which molecules these were, one thing was clear: It couldn’t be PKMzeta.

Huganir and his colleagues wrote up their report and sent it to Nature. It turned out that Robert Messing of the University of California, San Francisco, and his colleagues had independently conducted a similar experiment with mouse knockouts and come to the same conclusion. Nature decided to publish both papers in the same issue in January 2013.

“We were disappointed in this, and we think there are major problems in the hypothesis,” said Huganir.

Dr.Todd Sacktor
Equipment used for the study of memory in mice. Chantal Heijnen for STAT

A scientific backlash

When Messing and Hunagir’s papers came out, Fenton and Sacktor weren’t too worried at first. They were sure the other scientists must have made some kind of blunder. After all, many other scientists had come to the same conclusion as Fenton and Sacktor.

“You’d have to throw out huge amounts of data for no reason,” said Sacktor.

It turned out the scientific community felt differently.

“Within a week of those papers coming out, pretty much all across the Western world, work on PKMzeta stopped, except for me and Andre basically,” said Sacktor.

Scientists who applied for new grants to study PKMzeta got turned down — including even Sacktor. Fortunately, he had won a merit award from the National Institutes of Health in recognition of his PKMzeta work. He had enough extra money from that to fund a direct rebuttal of Huganir and Messing.

Sacktor realized he and Fenton would have one chance to really make their case and restore PKMzeta’s reputation. They decided to investigate a hunch they had about why the knockout mice could still remember: Another molecule was stepping in to take PKMzeta’s place — one that normally was not crucial to longer-term memories.

Their suspicions turned to a molecule called PKC-iota-lambda, which Sacktor and Fenton usually refer to as just “iota.”

Iota and PKMzeta are very similar in their structure, but they have different jobs in the body. While PKMzeta is only produced in the brain, iota gets made all over the body, starting at the first moment of conception. Cells use it to move other proteins to one side or the other. Iota helps make one side of a gut cell able to absorb food, for example, and the other to deliver that food to our blood system.

In his earliest research, Sacktor had noticed that iota sometimes appeared in neurons in the first hour of memory formation, but then vanished. He didn’t know what to make of that, and soon he focused all his attention on PKMzeta.

Now, he and Fenton turned their attention back to iota. Perhaps, they thought, the lack of PKMzeta in the knockout mice allowed iota to continue building up in neurons. In short, it was taking on the job of forming memories.

To test this idea, the scientists peered into the hippocampus of the knockout mice. They didn’t see any PKMzeta, obviously, since the mice couldn’t make it. But they did detect high levels of iota — four times higher than in normal mice. The scientists then applied a chemical to the knockout mice that only blocks iota, but not PKMzeta. Now, the mice couldn’t pass their memory test.

Fenton and Sacktor explain their results as an evolutionary story stretching over hundreds of millions of years. Our distant animal ancestors started out with just iota, which they used to move proteins around their cells. Iota also proved to be good at moving proteins around in neurons. And so it helped produce a primitive form of memory. Later, in our fish ancestors, iota accidentally got duplicated. The new copy evolved into a better molecule for maintaining long-term memory — PKMzeta, in other words.

To test that scenario, the scientists gave the knockout mice a more challenging test of their memory to see how well iota compared to PKMzeta as a memory molecule.

“We made these tasks more complicated — like raising the bar on an IQ test,” said Sacktor. “They can do simple memory tasks, but as soon as you ask them to think about four things rather than just two things, the animals just can’t do it at all.”

Morris, who has followed the PKMzeta story for 25 years, thinks that the new research, published in May in the journal eLife, is compelling.

“When the other papers came out, it seemed like they had demolished the story,” said Morris. “I think they’ve come back with some very convincing evidence. It looks like a pretty powerful rejoinder. It’s not something that can be dismissed lightly.”

Huganir, on the other hand, isn’t persuaded. He questions whether the drug they used to shut down iota, for example, didn’t also block some other molecules. And he is skeptical that iota can easily leap into the breach and take over PKMzeta’s job.

“Conceptually it doesn’t make any sense,” he said. “It’s not this magic molecule that it’s been touted to be.”

Sacktor predicts that PKMzeta will gain more support as scientists start publishing the studies that they had kept in cold storage for the past three years. Meanwhile, he and Fenton are going to charge ahead with their own research.

On his monitor, Fenton showed me glowing pictures of hippocampi, the neurons making PKMzeta glowing green. “I can design experiments to observe it, instead of destroying it,” said Fenton. “Now I can ask, ‘Where does it go?’”

For now, Fenton wants to answer these questions to understand how PKMzeta actually works. But eventually, he hopes his work translates into treatments that can help people with memory-related disorders.

“If you look at the tangles in an Alzheimer’s brain,” Fenton said, “they’re enriched with PKMzeta. Why that is, I don’t know. But one way to think of it is that the tangles act like a sponge. Maybe zeta sticks to something in the tangle.”

It’s also possible that some brain disorders are caused by too much PKMzeta. It might be behind pathological memories, or epileptic seizures triggered by abnormally strong connections between neurons. Fenton speculated that a drug that could alter levels of PKMzeta might be able to treat those conditions.

But the philosopher in Fenton can’t help but wonder about the ethics of affecting memory since memories are so important to who were are. “We’d have to deal with this new person that we’ve created,” said Fenton. “I would feel nervous about that sort of thing until we learn about what we’re really playing with.”

  • I am sure conceptual reduction of the experience of memory to the action of molecules is a component of any theory of memory.

    It would be nice, however, if:
    (a) we had some (any) idea what the term memory refers to.
    (b) what the purported relation between the experience of memory and the activity of neural players consists in.
    (c) how non-sentient (apologies to al panpsychists) matter enables subjective experience.
    (d) Why reduction (which has the inevitable effect of reformulating (in true Procrustean fashion) the phenomenon to fit the theory/method is not held accountable for reducing the phenomenon out of existence.
    (e) We appreciated that remembering does not necessarily result in memory (see point “A”).
    (f) psychological and neuroscience begin to take a mature and well-reasoned approach to the topic(s) of inquiry.

    Psychological “science” is an enterprise fraught with conceptual vacuity. the sooner we come to grips with this, the more likely we may make contributions that cut nature at its natural (rather than human) joints.

  • Fascinating story. I’m not 100% convinced their inhibitors don’t have off-target effects, but I think this helps highlight how germline knockouts from birth aren’t the be-all-end-all, as there’s certainly a lot of compensation by other proteins occurring. I wonder what they would see with a CRISPR/Cas9 knockout of PKMzeta in adulthood, perhaps after the animals learned a task?

  • The story of PKMzeta from its beginning through the crisis to the later glory was very catching, simple and at the same time convincing the need for scientific thinking of alternate possibilities and hopeful patient perseverence in the pursuit of truth. Its an easily assimilable and palatable enlightener for scientists from all disciplines.

Comments are closed.