The Gates of Immortality

By Yves Barral The Gates of Immortality Did biology evolve a way to protect offspring from the ravages of aging by creating a physical barrier that separates the parent from its young? Dr. Stanley Flegler, Visuals Unlimited he idea that every organism must age was a concept that surprised many biologists. For a long time, aging was thought to be a process occurring only in multicellular organisms. The reason for this arguably odd presumption was that we knew so

Yves Barral
Oct 1, 2010

The Gates of Immortality

Did biology evolve a way to protect offspring from the ravages of aging by creating a physical barrier that separates the parent from its young?

Dr. Stanley Flegler, Visuals Unlimited

he idea that every organism must age was a concept that surprised many biologists. For a long time, aging was thought to be a process occurring only in multicellular organisms. The reason for this arguably odd presumption was that we knew somatic cells—such as those that comprise the kidney, brain, and liver—lost their functionality over time: they aged. Furthermore, those cells divided only a limited number of times, around 50, after which they reached the so-called Hayflick limit, stopped proliferating, and died.

Unicellular organisms were thought to be capable of dividing forever, as long as conditions allowed: one generation begetting the next down through time—a sort of immortality. If unicellular organisms were like somatic cells, then...

It wasn’t until the 1950s that researchers who thought about aging began to change their minds. It became clear that the daughter cells of some unicellular organisms seemed to rejuvenate, to start from scratch, while the mother cells accumulated the cellular aberrations that signaled aging. This pattern of aging was seen in such evolutionarily distant organisms like Saccharomyces cerevisiae, known as budding or baker’s yeast, and bacteria such as Caulobacter crescentus and Escherichia coli.1–3 Aging, it seems, is a universal property of all living beings.

For me, that realization begged a more fundamental question, one that as biologists, we are scarcely allowed to ponder: Why do cells allow some mistakes to accumulate? If evolution is such a powerful process—one that finds solutions to all manner of problems—how could there be processes or problems that can’t be fixed?

As I continued my research into aging and cell division, I couldn’t help but think about how to categorize these “unfixable” problems like aging. Could there be a mathematical description that might capture and explain biology’s fallibility?

The theory of evolution, as formulated by Darwin, stands in its remarkable simplicity as an example of a single set of rules applicable to all biological systems. This theory, in its original expression, was remarkably devoid of mathematics (although it has since been described with great success in mathematical models). But biology in general appears to be so subject to deviations, exceptions, and diversification, rather than convergences, that it is difficult to derive simple and general principles. Nonetheless, it was this possibility of convergence that drove me to biology in the first place.

In high school I had studied mathematics, physics, and biology, looking for the places where these disciplines overlapped. As an undergrad, I asked the geneticist Bernard Dujon, now at the Pasteur Institute in Paris, whether I could work on yeast genetics in his lab. I told him I was interested in finding a way to capture the regularity of biological processes. Bernard, in response to my excited naïveté, answered sternly: “You are welcome in the lab, but be certain that there is no theory in biology and there will never be any. Biology is a field that escapes theory,” he told me. “Research in biology has to be pragmatic, or fail.”

That was the first blow to my dreams as a 20-year-old aspiring scientist. The second was that Bernard would not allow me to work in his lab until I learned to speak English. He arranged for me to study yeast genetics for a summer with his collaborator Walton Fangman at University of Washington in Seattle. It proved to be a very productive and intense 3 months. During the day, I would learn yeast genetics and English. In the evening I immersed myself in the only connection I had to the French language: a translation of the book, Gödel, Escher, Bach, on the conceptual connections between the world-class mathematician, the illustrator, and the composer.4 It seemed fitting that as I was learning what seemed like an absurd language, riddled with exceptions, I was also exploring a mathematical theorem—Gödel’s—that focused on the exceptions in mathematics: the equations that could not be solved.

Gödel’s theorems explained that all powerful mathematical systems (like algebra) contained completely valid assertions that were neither provable as right nor as wrong, indicating that all logical systems that were consistent and useful were incomplete. Although I didn’t know what to make of this at the time, I had a strong feeling that this idea should have remarkable consequences in biology. Despite its lack of overarching rules, biology seemed to be ruled by a sort of implacable underlying logic. Most biological systems function on a set of “if, then” rules that are the cornerstone of logical arguments. For example, if a receptor is activated, then a particular signal transduction program is initiated, with each downstream molecule following its own “if, then” stipulations for activation.

How, I wondered, could the concept of unsolvable problems, of incomplete systems, be applied to biology? Somehow the idea seemed plausible—life seems to be something that is always incomplete—but how could that incompleteness be manifest as something tangible, definable, measurable? These were the questions that I turned over in my head those summer evenings in Seattle.

hen I returned to France, I buried my thoughts about Gödel and decided to get pragmatic about science, following Bernard’s advice lest I fail as a researcher. I turned my efforts to yeast, particularly budding yeast. Like many researchers, I found the ability of these cells to control their size fascinating. Were cells aware of their shape? It was at least clear that our understanding of such processes was woefully incomplete!

Could there be a mathematical description that might capture and explain biology’s fallibility?

I set up my own lab at ETH Zurich in 1999, with a continued focus on yeast cell division. Like stem cells, yeast cells divide asymmetrically; the smaller of the two resulting cells is called the bud. In stem cells, the product of asymmetric division is an immortal mother and a daughter with a finite reproductive life span. In the case of yeast, the bud is the stem cell, not the mother cell, as one might think at first. It is the bud that forms the immortal lineage, that regenerates itself and stays pluripotent, while its mother cell changes over time and ages. It is as if, in order to stay young, the buds have to leave something behind in the mother cell. This was strange, because in principle these buds are made of the same cytoplasm, the same organelles, and the same plasma membrane as their mothers. How could the buds be so profoundly and uniformly distinct from the mother? These considerations led me to a deeper investigation of how the bud separates from the mother. How did the mother cell “know” where it stopped, and where the bud began? If there was some sort of boundary, how did it work?

Age gates
A ring of septin proteins are embedded in cellular membranes at the pinch between mother and bud. These rings prevent membrane proteins, including some of the mother’s nuclear pores, from passing to the bud. While the barrier doesn’t prevent matter in the nucleoplasm from passing to the bud, the extrachromosomal rDNA circles (ERCs), a potential proxy for aging, remain attached to the pores and are not passed to the bud.

We started by looking at the septin ring that Leland Hartwell had found in yeast in the 1970s.5 John Pringle had shown that this ring of proteins separating the mother and the bud was responsible for the cleavage of the cell during cytokinesis. I was curious about whether septins might do more than squeeze off a new cell—perhaps this ring also prevented the new bud from mixing with the mother cell. We labeled proteins associated with the plasma membrane first with affinity tags, later with fluorescent markers. Although these tagged proteins diffused freely on the cell membrane, they did not cross the bud neck unless the septin ring was disrupted.6 This was particularly true for so-called polarity factors, proteins that localize to one part of the plasma membrane and contribute to polarization. In yeast, these factors travel to the bud emergence site at the plasma membrane where they drive the polarization of the cytoskeleton and growth toward that area. We discovered that the diffusion barrier was required to retain these polarity factors in the bud during bud growth.

This wasn’t the first example of protein belts separating a cell membrane7—the classic illustration was the epithelial cell, where tight junctions encircling the cell formed a lateral diffusion barrier that prevented membrane lipids and proteins from crossing from the apical to the basolateral surfaces, and vice versa. Septins offered a second example, suggesting that diffusion barriers were an important mechanism, conserved from yeast to mammalian cells.

What no one expected was that this belt of proteins might create a barrier that also impeded diffusion of the membrane proteins in cellular organelles from migrating into the bud, acting like a roadblock at the intersection of the mother cell and the bud. The membranes of the yeast cell organelles, such as the endoplasmic reticulum or the envelope of the dividing nucleus,7–9 were shared in a way that prevented a free exchange of proteins between the mother and the bud. The mother and bud, therefore, were really kept distinct and separate from each other from the onset of cell division. Complex machinery, present at the bud neck from the onset of bud emergence, was acting to keep the composition and identity of the mother and bud distinct.

t hadn’t taken very long for me to return to questions of aging. If the diffusion barriers helped the bud to keep its polarity markers—necessary for separation from the mother—could it be that it was those same barriers allowed the bud to split from its mother in a pristine, stem-cell–like state while the mother “aged”?

I had to wonder, what is aging after all? Is it something positively tangible, something that we could define otherwise than a loss: loss of fitness, loss of potential, loss of viability? There is at least one type of molecular marker that correlates well with aging, at least in yeast: extrachromosomal ribosomal DNA circles (ERCs).10 These circles, or plasmids, of DNA are excised from the ribosomal DNA (rDNA) locus on the chromosome and replicate at each division cycle. The ERCs, however, are redundant to the chromosomal ribosome genes—unnecessary elements that accumulate in the nucleus of the mother cell as it ages. Although ERCs do not represent the only mechanism of aging in yeast, their accumulation is related to aging: cells in which ERC formation is delayed live longer, whereas cells with increased ERC formation die sooner. It would follow then, that their retention by the mother cell contributes largely to the age asymmetry between mother and bud.

Mortal and immortal lineages
Most buds produced by a mother cell will also become bud-producing mothers themselves. A new mother has the capacity to produce 30 buds before she dies with a large number of extrachromosomal rDNA circles (ERCs) accumulated in her nucleus. When the barrier separating the mother from the bud was experimentally disrupted, it allowed ERCs to enter the bud in large number. As a result, the mother lived longer, while the reproductive life of her daughters was shortened proportionate to the amount of ERCs she received.

ERC retention, then, must rely on some intrinsic asymmetry of the nucleus. For example, the yeast nucleus, which does not disassemble during mitosis, as in mammalian cells, always sends its oldest spindle pole body (SPB) to the bud. The SPB, which organizes the duplicate genomes during mitosis, is embedded in the nuclear envelope, and duplicates at each cycle to form a single new SPB. If we forced the nucleus to send the old SPB to the bud only half of the time, would the ERCs then segregate to the bud the other half of the time? To our surprise, this did not happen; the ERCs remained in the mother cell. The exclusive retention of the ERC plasmids within the nucleus of the mother cell only diminished when we disturbed the septin diffusion barrier that divided the nuclear envelope.8

As a consequence, yeast cells lacking the septin diffusion barrier can pass these molecular markers of aging to their daughters.8 Without the diffusion barrier the mothers were longer lived, but their daughters behaved as if they were older at birth: in other words, they had the capacity to replicate fewer times. The fact that ERCs remained in the mother’s part of the nucleus indicated that the plasmids had to be linked to something embedded in the nuclear membrane. Since septins only blocked diffusion of molecules in the membrane (they did not, for example, create a webbing across the cytoplasm), none of the molecules freely floating in the nucleoplasm would be affected. We observed that these plasmids were associated with the nuclear envelope, more precisely with the basket of nuclear pores on the inside of the nuclear membrane, and that this association was required for their retention in the mother cell. Taken together, our data suggest the intriguing idea that aging, whatever it is, respects diffusion barriers, and that these boundaries prevent the propagation of aging-related molecules into newborn buds.

It may be that the cell’s solution to its unsolvable problems is simply to age, to compartmentalize the components that bear too much resemblance to self and slough them off.

It is still unclear at this point whether these findings have any parallel in other eukaryotes, but we think they might.6 Indeed, the process of sperm generation shows intriguing similarities with the budding process in yeast, at least in terms of the maturation of the future sperm’s nuclear envelope. The emergence of the sperm head involves the migration of the nucleus through a perinuclear ring. During this process, the nuclear envelope is combed, leaving behind its nuclear pores, which, in many cases, are then excluded from the sperm nucleus. Thus, it is tempting to speculate that we, too—like yeast—keep our sperm as young as possible each time we prepare to form a newborn.

ver the years, these observations and findings have led me back to Gödel and his ideas of unsolvable problems. Aging seemed the perfect example of a process in which the cell could not detect ambiguous molecules (either overtly damaging, or beneficial) and repair itself.

To my mind, ERCs are emblematic of objects that are ambiguous to the cell. They have the same chemical nature, the same repeating composition as the chromosome, and therefore cannot be targeted for destruction without risking damaging the chromosomes as well. They take on the characteristics of entities that are both self and nonself. Gödel was able to mathematically characterize the unsolvable problems he encountered and describe them with a universal rule. Might ERCs help define the universal properties of the unfixable errors that accumulate with age? What prevented biological systems from being complete?

At its core, the generation and accumulation of ERCs is a problem of symmetry—ERCs are generated by errors in DNA repair. When the DNA repair (or recombination) machinery resolves the Holliday Junction, it has one of two options, excision or repair. But because of the local symmetry at the Holliday Junction, the recombination machinery cannot detect a difference between the incoming strands of DNA, and therefore cannot favor one solution over the other. In order to handle such an unsolvable problem, the cell simply produces both outcomes with equal probability, with the production of ERCs and DNA repair occurring exactly 50 percent of the time.

What if structural asymmetries, such as cell polarity, might have actually emerged to counteract the logical problems that symmetric events such as DNA repair generate for the cell? If true, it implies that studying symmetric processes in biology could reveal new insights about aging.

It may be that the cell’s solution to its unsolvable problems is simply to age, to compartmentalize the components that bear too much resemblance to self and slough them off, producing a life that lacks these deformities. Although the yeast cell might not be able to distinguish ERCs from the chromosomes, it found ways to sort them out and confine them to the mother cells. Diffusion barriers could play a central role in this process. It is interesting because they are likely to simply retain in the mother cell anything that is not actively being chosen and pulled into the bud, such as chromosomes or vesicles. Thus, they offer a remarkable solution to the retention of ambiguous objects, that is, objects that the cell cannot distinguish as being right or wrong, objects that therefore remain invisible to cellular machineries.

Last, if aging is a consequence of Gödel’s theorem in biology and of the cell’s incompleteness, then aging is not a program but an inescapable fact. The quest for a cure to the aging “disease” will inevitably fail. But there is a bright side to the fact that the cell is logically incomplete. Would any complete system—one able to detect any damage and repair itself perfectly—have the ability to evolve?

Have a comment? E-mail us at

F1000 Faculty Member Yves Barral is Full Professor at the ETH Zurich, Switzerland.

1. R.K. Mortimer, J.R. Johnston, “Life span of individual yeast cells,” Nature, 183:1751–52, 1959.
2. M. Ackermann et al., “Senescence in a bacterium with asymmetric division,” Science, 300:1920, 2003.
3. E.J. Stewart et al., “Aging and death in an organism that reproduces by morphologically symmetric division,” PLoS Biol 3:e45, 2005.
4. D. Hofstadter, Gödel, Escher, Bach, New York: Basic Books, 1979.
5. L.H. Hartwell, “Genetic control of the cell division cycle in yeast. IV. Genes controlling bud emergence and cytokinesis,” Exp Cell Res, 69:265–76, 1971.
6. Y. Barral et al., “Compartmentalization of the cell cortex by septins is required for maintenance of cell polarity in yeast,” Mol Cell, 5:841–51, 2000.
7. E. Caudron, Y. Barral, “Septins and the lateral compartmentalization of eukaryotic membranes,” Dev Cell, 16:493–506, 2009.
8. C. Luedeke et al., “Septin-dependent compartmentalization of the endoplasmic reticulum during yeast polarized growth,” J Cell Biol, 169:897–908, 2005.
9. Z. Shcheprova et al., “A mechanism for asymmetric segregation of age during yeast budding,” Nature, 454:728–34, 2008.
10. D.A. Sinclair, L. Guarente, “Extrachromosomal rDNA circles—a cause of aging in yeast,” Cell, 91:1033–42, 1997.

Interested in reading more?

Become a Member of

Receive full access to digital editions of The Scientist, as well as TS Digest, feature stories, more than 35 years of archives, and much more!
Already a member?