Life's Greatest Secret (40 page)

Plants are much more likely than mammals to show epigenetic inheritance, partly because they do not have strictly separated germ and somatic cell lines. However, the division of germ and soma in animals is not quite as clear-cut as might be thought (indeed, not all animals have this division). Egg and sperm are cells that contain half your genomic DNA, but they also carry other material that may affect the fitness of your offspring. For example, there are ‘maternal effects’, whereby characters are differentially expressed depending on the mother, generally due to the presence of certain mitochondrial genes in the egg – we inherit our mitochondria from our mother, and this separate genome can exert specific effects before and during embryogenesis. Other molecules can have long-lasting effects down the generations, so for example if the nematode worm
C. elegans
becomes resistant to a virus, that resistance can be transmitted for several generations through the presence of small protective RNA molecules in the sperm cells; these molecules are directly involved in resistance to the virus, not in gene regulation.
³²

Even in plants there are limits to the significance of the inheritance of epigenetic factors. First, the environment can only alter gene regulation: there is no evidence that it can lead to any direct alteration of the genetic code. Second, for the moment, despite some tantalising hints that DNA methylation may be involved in a plant’s response to bacterial infection, there is not one clear example of an epigenetically based adaptation that is of a character that increases the fitness of the organism.
³³
However, epigenetic manipulation may have important consequences for plant breeding in the future. In the laboratory plant
Arabidopsis,
heritable epigenetic factors have been found to affect flowering time and root length, and can be subject to artificial selection. Heritable gene silencing in this species can occur through the activity of small RNA molecules, which may be transmitted in the gametes. For the moment these gene silencing effects have been found mainly in characters associated with the activity of transposons; one of the leaders in the field, David Baulcombe, has suggested that epigenetic effects in plants are probably associated with variation in particular traits, rather than in determining major characters.
³⁴

A recent review in the journal
Cell,
entitled ‘Transgenerational epigenetic inheritance: myths and mechanisms’, summarised our understanding of the contrasting situations in plants and animals in appropriately sober fashion:

although transmission of acquired states can occur in some animals (such as nematodes), proof that transgenerational inheritance has an epigenetic basis is generally lacking in mammals. Indeed, evolution appears to have gone to great lengths to ensure the efficient undoing of any potentially deleterious bookmarking that a parent’s lifetime experience may have undergone.

Nevertheless, some scientists have continued to suggest that evolutionary biology needs a major theoretical rethink; others have responded correctly that our current views already include the existence of epigenetic effects.
³⁶
Epigenetics remains fascinating, but it is an adjunct to our understanding of the complexities of gene regulation and the origin of plasticity, not a radical new model of inheritance and evolution. The fundamental claim of the central dogma is that once information has gone out of the DNA sequence into protein, there is no way for it to get back into the genome. Despite the existence of epigenetic inheritance in certain unusual circumstances, this statement remains true. We know of no way in which the information expressed in proteins can alter the DNA sequence.

The idea that an organism’s experience affects the characters shown by its offspring was an early contender for explaining evolution, and is associated with the name of the nineteenth century French naturalist Jean-Baptiste Lamarck. In fact, Lamarck never used the phrase by which he is now most widely known – ‘the inheritance of acquired characteristics’ – and according to his view only the youngest organisms could acquire new characters.
³⁷
The most widely-known example given by Lamarck was his suggestion that the giraffe acquired its long neck by gradual increases in length over many generations, caused by the act of stretching to reach higher branches to eat the leaves.

In its nineteenth-century context, Lamarck’s suggestion that acquired characteristics were the driver of evolutionary change provided a mechanism for a process that was still widely contested.
³⁸
The mechanism that Lamarck described was widely believed – Darwin also assumed that characters acquired during the organism’s lifetime affected the hereditary particles that he hypothesised existed in every tissue of the body and which were involved in reproducing those tissues in the next generation.
³⁹
In this sense, Darwin was also a Lamarckian. By the beginning of the twentieth century, with Weismann’s discovery of the germ and somatic cell lines in animals and the development of genetics, it soon became apparent that heredity was based on particles that were transmitted with no detectable effect of experience. Lamarck’s hypothesis was stripped from its context and became a caricature for the condescending amusement of students, while the fact that Darwin shared this concept was often passed over in silence.

The most notorious advocate of the inheritance of acquired characteristics was the Soviet agronomist Trofim Denisovich Lysenko. In the 1930s, Lysenko claimed that he was able to alter the make-up of wheat by environmental manipulation, and that genetics was a fantasy invented by the capitalist West. In the Cold War years, Lysenkoism led to the persecution of a whole generation of Soviet geneticists as their science was effectively outlawed. When Lysenko’s star waned in the late 1950s and 1960s, the Stalinist leaders of the USSR gradually allowed Soviet genetics to re-emerge, but even half a century on, the damage is still evident in terms of the relatively weak status of Russian genetics on a world scale as compared to other areas of science.

With the recent interest in transgenerational epigenetic effects, some science journalists have suggested that Lamarckian evolution is making a comeback (the name of Lysenko is rarely mentioned in such articles).
⁴⁰
Lamarckism is not on the rise, because there is no cellular pathway from protein to DNA. In the absence of such a route, this mode of evolution would require epigenetic effects to be reliably transmitted over many generations. There would then have to be some way of fixing the epigenetically determined character in the population in the absence of the environmental factor that initially induced it. This is a fundamental challenge for the idea that epigenetics plays a major role in evolution: Darwinian natural selection on protein-encoding genes leads to stable adaptations that persist even in the absence of the selection pressure that shaped them – a polar bear born in a southern European zoo still has thick white fur, even though it has never been near ice or cold. All its descendants will also show this character. This would not be the case with epigenetic inheritance, unless we were to discover new mechanisms for permanently fixing the epigenetic marks.

There is an even more fundamental problem: although Lamarckian inheritance apparently provides a common-sense explanation of adaptation – that is, of the appearance of characters such as a giraffe’s long neck – if there were a route from acquired characteristic to DNA, that route would carry information both for good things such as longer necks and for bad things such as high cholesterol or type 2 diabetes, which also occur during an individual’s lifetime. Unless the cell could know which changes would be advantageous in the future, both kinds of change would end up being translated back into DNA, or tagged epigenetically, and the next generation would show a confusing mixture of advantageous and disadvantageous characters.

*

Virtually all of the evolutionary adaptations we see around us are consequences of the sifting of DNA sequences by natural selection over many generations. However, other factors also affect the frequencies of sequences of DNA in a population. For example, shifts in apparently non-functional DNA sequences generally occur through completely random processes because they are not subject to natural selection. Even sequences that are subject to natural selection, which include both protein-coding genes and lots of sequences involved in gene regulation, can show changes in a population through random sampling, producing an effect called genetic drift. One of the easiest effects to understand occurs where a sudden reduction in the number of organisms in a population can affect gene frequencies and therefore the raw material that natural selection has to work on.

As well as being the primary motor that produces the fantastic adaptations we can see in the world around us, natural selection is also the only conceivable way in which it can be argued that information encoding adaptations can flow back from a character – be it a protein or whatever – into the gene sequence shared by a population. That is how natural selection works: the characters produced by different genotypes lead the organisms that carry those genes to leave different numbers of offspring in the next generation. The DNA sequences in the population therefore gradually change over time as random effects change the range of sequences available in a population, and the environment sifts the genetic composition of the population, with those individuals that can survive and reproduce in particular environmental conditions leaving more copies of their genes in the next generation. Natural selection is not a challenge to the central dogma, because the two phenomena function at different physical and temporal levels. Natural selection is a force that operates in populations over many generations, whereas the central dogma describes processes taking place in a single cell – or even in a single molecule – during the lifetime of an individual.
⁴¹

When Crick enunciated the central dogma, his aim was not to reframe Weismann’s division of cells into the somatic line and the germ line, or to defend the modern understanding of evolution by natural selection against the idea of the inheritance of acquired characteristics. The central dogma was based on known or assumed patterns of biochemical information transfer in the cell rather than any dogmatic position. As such it was vulnerable to being invalidated by future discoveries. Nevertheless, in its fundamentals it has been shown to be correct. Real or apparent exceptions to this rule, such as retrotranscription, prion disease or transgenerational epigenetic effects, have not undermined its basic truth.

Crick’s operational definition of genetic information was ‘the determination of a sequence of units’ either in a nucleic acid (DNA or RNA) or in the amino acid chain of a protein. A separate part of Crick’s argument was the sequence hypothesis – the idea that the three-dimensional structure of proteins is in some way inherent in the sequence and emerges as the amino acid chain is assembled. Crick admitted that it was ‘possible that there is a special mechanism for folding up the chain’ but felt that ‘the more likely hypothesis is that the
folding is simply a function of the order of the amino acids.’
⁴²
He therefore hypothesised that there is no separate genetic code for protein conformation. Nearly half a century later, this still seems to be correct. By comparing a novel DNA sequence with known amino acid sequences and three-dimensional protein structures, it is possible to make predictions about the three-dimensional conformation of the protein that could be produced by the DNA sequence. However, despite our deep understanding of the physicochemical rules underlying the shape that proteins take, there is as yet no way of predicting with absolute certainty the three-dimensional structure of a protein from a DNA sequence. Our predictions are becoming increasingly accurate because of novel computer algorithms and because the process is cumulative: as the number of correlations between sequence and structure increases, so, too, does accuracy.

There is something missing from this description of protein synthesis. In many cases – perhaps most, perhaps almost all – protein synthesis involves the presence of molecular chaperones, which alter the rate of folding.
⁴³
These chaperones can be molecules such as the heat shock proteins that prevent polypeptide chains from clumping together, or they may block competing biochemical reactions during the folding process. Chaperones can equally be hollow molecules that provide a favourable chemical micro-environment for the protein to be correctly folded, within a space known by the sci-fi-sounding term Anfinsen’s cage – named after Christian Anfinsen, who won the 1972 Nobel Prize in Chemistry for his work showing the link between the one-dimensional amino acid sequence and the three-dimensional protein structure.
⁴⁴

In a 1970 letter to Howard Temin, Crick, open-minded as ever, showed that he recognised the possibility that other factors apart from the amino acid sequence might determine three-dimensional protein structure: ‘I do not subscribe to the view that
all
“information” is
necessarily
located in nucleic acids. The central dogma only applies to residue-by-residue sequence information.’
⁴⁵

Crick would no doubt have been intrigued but unworried by some aspects of protein synthesis that might appear to contradict the central dogma. In a large number of bacteria and some eukaryotes such as fungi, some peptides are assembled without the direct involvement of DNA, mRNA or ribosomes. The existence of these oddities does not go against the key argument of the central dogma, because there is no route for information to flow from the protein into DNA. Instead, non-ribosomal peptides raise the question of how the information they contain is represented and transmitted genetically, if it is not contained in a nucleic acid sequence. There are two linked answers. Non-ribosomal peptides are synthesised using an assembly line of enzymes and the order of the amino acids is determined by the order in which the enzymes act on the growing chain – each functional part of each enzyme adds a particular amino acid. However, these enzymes are themselves encoded by genes, so ultimately the information that represents the peptide’s amino acid sequence is in fact contained in DNA sequences, albeit indirectly.
⁴⁶