by Paul Nurse
As well as signalling through space, cells need ways to signal through time. To achieve this, biological systems must be able to store information. This means that cells can carry with them chemical imprints of their past experiences, which we can think of as working a bit like the memories we form in our brains. These cellular memories range widely, from transient impressions of what happened just a moment ago, to the extremely long-term and stable memories held by DNA. The cell uses short-term historical information during the cell cycle, when the status of events that occur early in the cycle are ‘remembered’ and signalled forward to later events in the cycle. For example, if the process of copying DNA has not yet been completed or has gone wrong, this fact needs to be registered and relayed to the mechanisms which bring about cell division. If not, the cell could attempt to divide before its entire genome has been properly copied, which could result in the loss of genetic information and the death of the cell.
The processes involved in gene regulation allow cells to store information over longer time scales. This was a particular interest of the British biologist Conrad Waddington during the mid-twentieth century. I met Waddington at Edinburgh University when I started my postdoctoral research there in 1974. He was a striking character, with wide interests in art, poetry and left-wing politics, but he is best known for coining the word epigenetics. He used it to describe the way cells gradually take on more specialized roles during the development of an embryo. Once the growing embryo instructs cells to commit to these roles, they remember that information and rarely change track. That way, once a cell has committed to forming part of the kidney, it will remain part of the kidney.
Today, the way most biologists use the word epigenetics is based on Waddington’s ideas. It describes the set of chemical reactions that cells use to turn genes either on or off in fairly enduring ways. These epigenetic processes do not change the DNA sequence of the genes themselves; instead, they often work by adding chemical ‘tags’ to the DNA, or to proteins that bind to that DNA. This creates patterns of gene activity that can persist through the lifespan of a cell and sometimes even longer, through many cell divisions. Occasionally, although far less commonly, they can persist from one generation to the next, potentially carrying information about an individual organism’s life history and experience directly, in chemical form, from parents to their offspring and on to subsequent generations. Some have argued that the cross-generational persistence of these patterns of gene expression poses a major challenge to the idea that inheritance is based only on the DNA sequences encoded in genes. However, present evidence indicates that cross-generational epigenetic inheritance only occurs in a few instances and seems to be very rare in humans and other mammals.
In addition to gene regulation, information processing is important for the ways living beings create ordered structures in space. Take my brimstone butterfly. It is an exquisitely complex construction: the wings are carefully shaped to allow it to fly; there are spots and veins placed on those wings with great precision. Moreover, all individual butterflies are built to the same plan: they all have a head, thorax and abdomen, six legs and two antennae, for example. These structures all form and grow in the same predictable proportion to the rest of their bodies. How is all this extraordinary spatial structure generated? How does it all emerge from a single uniform egg cell?
Even cells can take on a range of highly elaborate structures and shapes that are quite distinct from the regular, box-like cork cells that Robert Hooke described in the seventeenth century and that I observed in onion roots as a schoolboy – there are the comb-like hairs on lung cells, whose constant beating pushes mucus and infections out of your lungs; cube-shaped cells that live in and manufacture your bones; and neurons whose long, branching connections reach all parts of your body; among very many others. And within those cells, their organelles can be precisely located and grow and adjust their position as the cell changes.
How all of this spatial order develops is one of the more challenging questions in biology. Satisfactory answers will depend on understanding how information is signalled through both space and time. At present, we only really understand fully the structure of biological objects that are direct assemblies of molecules. The ribosome is a good example. The shapes of these relatively small objects are determined by the chemical bonds that form between their molecular components. You can think of these structures as if they are built up by adding pieces to a three-dimensional jigsaw, a bit like Lego. That means the information needed to assemble these structures is embodied in the shape of the ribosome components themselves – the proteins and RNAs. Those shapes, in turn, are ultimately specified very precisely by the information held in the genes.
Understanding how structures form at larger scales, in objects such as organelles, cells, organs and whole organisms is more difficult. Direct molecular interactions between components cannot explain how these structures form. That’s partly because they are larger, sometimes much larger, than objects like ribosomes. But it is also because they can produce and maintain perfect structures over a range of different sizes, even when cells or bodies grow or shrink. That is simply not possible with fixed, Lego-like molecular interactions. Take the division of a cell for example. A cell has a well-organized overall structure, and when the cell divides it generates two cells of approximately half the size and yet each of them has the same overall structure as the original ‘mother’ cell.
A similar phenomenon is seen with the development of an embryo, such as a sea urchin. A fertilized sea urchin egg undergoes repeated cell divisions and generates an elaborate and rather beautiful little organism. If the two cells formed after the very first division of the egg are split apart, then each cell will generate two perfectly formed sea urchins, but, amazingly, each one will be just half the size of a normal urchin of that age. This self-regulation of size and form is extraordinary and has puzzled biologists for more than a century.
However, by thinking about information, biologists are beginning to make sense of how these things take shape. One way that developing embryos generate the information they need to transform a uniform cell or group of cells into a highly patterned structure is by making chemical gradients. If you put a small drop of ink into a bowl of water, it will slowly diffuse away from the location of the original drop. The intensity of the ink colour gets lower further away from the drop, making a chemical gradient. That gradient can be used as a source of information: for example, if the concentration of ink molecules is high, we know we are close to the centre of the bowl, where the ink was dripped in.
Let’s now replace the bowl with a ball of identical cells and, instead of ink, we inject one side of the ball with a dose of a particular protein that can change the properties of cells. What this provides is a way to add spatial information to those cells so they can begin to build a pattern. The protein will diffuse through the cells, forming a gradient of high concentration at one side of the ball and low concentration on the other side. If cells react differently to high and low concentrations, the protein gradient can provide the information needed to start constructing a complex embryo. If, for example, a high protein concentration made head cells, a medium concentration made thorax cells, and a low concentration made abdomen cells, then one simple protein gradient could, in principle, lead to the beginnings of a new brimstone butterfly. In life, things are usually not quite as simple as that, but there is good evidence that gradients of signalling molecules across the bodies of developing organisms do indeed contribute to the appearance of sophisticated biological forms.
This was a set of problems that Alan Turing – he of Enigma code-cracking fame and one of the founders of modern computing – turned to during the early 1950s. He came up with an alternative, and imaginative, suggestion for how embryos generate spatial information from within. He devised a set of mathematical equations that predicted the behaviour of chemical substances interacting with each other, and so undergoing specific chemical reactions as they d
iffuse through a structure. Unexpectedly, his equations, which he called reaction-diffusion models, could arrange chemical substances into elaborate and often rather beautiful spatial patterns. By tweaking the parameters of his equations, the two substances could organize themselves into evenly spaced spots, stripes or blotches, for example. The attractive thing about Turing’s model is that the patterns emerge spontaneously, according to relatively simple chemical rules of interaction between the two substances. In other words, this provides a way for a developing cell or organism to generate the information it needs to take shape, entirely from within; it is self-organizing. Turing died before his theoretical ideas could be tested in real embryos, but developmental biologists now believe that this could be the mechanism that puts spots on cheetah’s backs and stripes on many fish; distributes the hair follicles on your head; and even divides each of a developing human baby’s hands into five distinct fingers.
When we look at life in terms of information, it is important to appreciate that biological systems have evolved gradually over many millions of years. As we have seen, life’s innovations arise as a consequence of random genetic mutations and variations. These are then sifted by natural selection, with those that work well being assimilated into the surviving, more successful, living organisms. This means that existing systems are changed progressively, by the gradual accretion of ‘add-ons’. This is in some ways analogous to your phone or computer, which frequently require the downloading and installation of new software updates. The devices gain new functions, but the software that drives them also becomes steadily more complicated. Similarly for life, all of these genetic ‘updates’ mean that the whole system of the cell will gradually tend to become more complex with time. This can lead to redundancy: some components will have overlapping functions; others will be the relics of superseded parts; and some will be wholly unnecessary for normal functioning but might be able to compensate if the primary component breaks.
This all means that living systems are often less efficient and rationally constructed than control circuits designed intelligently by human beings, another reason why analogies between biology and computing can only go so far. As Sydney Brenner observed, ‘Mathematics is the art of the perfect. Physics is the art of the optimal. Biology, because of evolution, is the art of the satisfactory.’ The life forms that survive natural selection persist because they work, not necessarily because they do things in the most efficient or straightforward way possible. All this complexity and redundancy makes the analysis of biological signalling networks and information flow challenging. Very often Occam’s razor – looking for the simplest adequate explanation to explain a phenomenon – simply does not apply. This can disturb some physicists who turn their attentions to biology. Physicists tend to be attracted to elegant, simple solutions, and can be less comfortable with the messy and less-than-perfect reality of living systems.
My lab has frequently wrestled with the redundancies and intricacies brought about by natural selection, because they can obscure the core principles of how biological processes work. To tackle this we genetically engineered yeast cells to generate a much simplified cell cycle control circuit. It was like stripping a car of all the components that are not essential for its critical functions, such as the bodywork, the lights and the seats, leaving only the essentials – the engine, transmission and wheels. This worked better than I had hoped. Our simplified cells could still carry out the major aspects of cell cycle control. Stripping a complex mechanism down to its basic elements made it easier for us to analyse information flow, and therefore gain new insights into the cell cycle control system.
Among the select group of indispensable cell cycle regulators highlighted by this experiment was the cdc2 gene. As a yeast cell moves through the cell cycle, the cell itself grows steadily and the amount of the Cdc2- and cyclin-containing CDK protein complex increases too. In terms of information, the cell uses the amount of active CDK complex present as both an input that reflects information about the size of the cell, and as the crucial signal that triggers the major events of the cell cycle. Proteins required early in the cell cycle are phosphorylated by the CDK complex early, leading to the copying of DNA during S-phase, and those required later are phosphorylated later, leading to mitosis and cell division at the end of the cell cycle. The ‘early’ proteins are more sensitive to the CDK enzyme activity than the ‘late’ ones, so they will be phosphorylated when there is less CDK activity in the cell.
This simple model of cell cycle control identified CDK activity as the crucial co-ordinating hub at the centre of cell cycle control. The explanation had just been obscured from our view by the superficial complexities of the network, the redundant functions of different components, the presence of less important control mechanisms, and perhaps also by the tendency of the human mind to embrace complexity, rather than seek out simplicity.
For much of this chapter I have focused on cells because they are the basic units of life, but the implications of thinking about life as information extend beyond the cell. There is real potential to gain powerful new insights into all parts of biology by looking for ways to understand how molecular interactions, enzyme activities and physical mechanisms produce, transmit, receive, store and process information. As this becomes a more prevalent approach, it is possible that biology will shift away from the rather common-sense and familiar world that it has generally occupied in the past, to one that is more abstract. In this, it might parallel the great shifts that took place in physics, from Isaac Newton’s essentially common-sense world to Albert Einstein’s universe, ruled by relativity, and on further to the quantum ‘weirdness’ revealed by Werner Heisenberg, as well as Erwin Schrödinger, in the first half of the twentieth century. It might be that the complexity of biology will lead to strange and non-intuitive explanations, and to work these out biologists will need ever more assistance from scientists in other disciplines, such as mathematicians, computer scientists and physicists – even philosophers, who are more used to thinking abstractly and are less focused on our everyday experiences of the world.
A view of life that is centred on information will also help us understand higher levels of biological organization. It can shed light on how cells interact with each other to generate tissues, how tissues make organs, and how organs work together to produce a fully operational living organism, such as a human being. The same is true at even bigger scales, when we look at how living organisms interact with each other, both within species and between species, and how ecosystems and the biosphere operate. The fact that information management occurs at all scales, from the molecular to the planetary biosphere, has important implications for how biologists try to make sense of life’s processes. Often, it is best to seek explanations close to the level of the phenomenon being studied. To be satisfactory, those explanations do not always need to be reduced down to the molecular-scale realm of genes and proteins.
However, it may well be that there are commonalities between the way information is managed at one scale that can illuminate how things work in a system that is either larger or smaller. For example, the logic underpinning feedback modules that control metabolic enzymes, regulate genes or maintain bodily homeostasis, will have similarities with the feedback modules that allow ecologists to make better predictions about how natural environments are likely to change when specific species go extinct or migrate out of their traditional ranges as a result of climate change or habitat destruction.
Given my interest in beetles and butterflies and insects in general, I am increasingly worried about the falling numbers and diversity of insects that are being observed in many parts of the world. What is particularly disturbing is that we do not know why this is happening. Is it habitat destruction, climate change, agricultural monocultures, light pollution, overuse of insecticides, or something else? There are many explanations proposed and some people feel very certain of their particular theories, but the truth is we do not really know. If we are to do somethin
g to help reverse declining insect populations, we need to understand the interactions between them and the rest of their worlds. This will be greatly informed by scientists who work in different ways, collaborating and thinking about these issues in terms of information.
Whichever level of biological organization we look at, attempts to deepen our comprehension will hinge on our ability to understand how information is managed within them. It is a way to move from describing complexity to understanding complexity. Once we can do this, we can start to see how flitting butterflies, sugar-consuming bacteria, developing embryos and all other life forms make the crucial leap of transforming information into meaningful knowledge that they can use to fulfil their purpose of surviving, growing, reproducing and evolving.
From our advancing understanding of the chemical and informational foundations of life springs the growing ability not only to comprehend life, but also to intervene in the workings of living things. So before I use the insights we have gained from climbing our five steps to define what life is, I want to consider how we can use knowledge of biology to change the world.
CHANGING THE WORLD
In 2012, I was due to travel to the Antarctic research station at Scott Base. I had always wanted to visit the vast frozen desert of the South Polar regions – literally the end of the Earth – and finally I had my chance. Before the trip I had to have a routine medical check-up, but the results turned out to be far from routine. For the first time in my life I had to directly confront my own mortality.
