Ch 12: extract 4

Why organisms are not machines: two more ways (at least three more to come)

4. Not one-way action – maybe not even interaction?

A further problem with a machine model is that it suggests a direction of action of one thing on another. I have just suggested that it is usually myriads of ‘things’ having an action on myriads of other things. But even that is wrong. In organisms there is never just action without both interaction and mutual construction. Cause and effect in organisms, if it can safely be applied, is never unidirectional, but reciprocal. When an organism interacts with its environment, both parties change; each influences the other. And not in some minor, incidental fashion, either. When DNA is ‘active’ (as Lewontin points out, it is largely inert), most of the work is done by the cell environment. This, crucially, includes the provision of information, which is conventionally thought of as the role of DNA. Both cell and genome depend on their mutual interaction. As Schrödinger put it: ‘physical action always is inter-action, it always is mutual.’^[1]

Research over the last 80 years or more demonstrates that it is not true that genetic change is simply due to accidents or damage to the DNA. Organisms actively reconstruct their genomes in response to their conditions.^[2] Shapiro points out that the genome is actively shaped by the cell, as conditions change, over three distinct time-scales: first during cell reproduction, involving the formation of nucleoprotein complexes; then during multicellular development, involving epigenetic formatting; and ultimately in evolutionary change, involving changes in DNA sequence structure:

one of the main adaptive features of DNA-based heredity is that DNA is a highly malleable storage medium, permitting rapid and major changes to complex organisms without disrupting their functional integrity.^[3]

This has interesting consequences that may surprise some readers. C.H. Waddington found that heat-shocking pupae of the fruit fly Drosophila in some cases led to abnormal patterns of veining on the wings of the adult fly. If one continued to breed such cases together, heat-shocking the offspring, a point was soon reached where not only virtually all flies were abnormal, but their offspring continued to be abnormal even without heat-shocking. The number of generations necessary for the abnormal patterning to become fixed in the population was only 14 – too few for this to be supposed a coincident chance novel mutation. The gene expression had, it seems, been altered directly, and quickly, by the environment. ^[4]

It is a standing joke among mechanists that cutting off the tail of a mouse does not produce tailless mice. However, scientists have developed microsurgery techniques that can alter the pattern of cilia on the surface of Paramecium aurelia (a unicellular organism). They found in a series of striking experiments that these changes were transmitted permanently to their offspring through the normal processes of cell division.^[5] Similar phenomena have been demonstrated in other ciliates.^[6]

‘Today’s biologists’, writes the physicist Evelyn Fox Keller,

recognise that however crucial the role of DNA in development and evolution is, it does not do anything by itself. It does not make a trait; it doesn’t even include a ‘program’ for development. Rather, it is more accurate to think of a cell’s DNA as a standing resource on which a cell can draw for survival and reproduction … it is always and necessarily embedded in an immensely complex and entangled system of interacting resources that collectively are what give rise to the development of traits.^[7]

Let us just pause there for a second. DNA is ‘a standing resource on which the cell can draw for survival and reproduction’. Hardly surprising, of course – unless you had been led to believe, somewhat improbably, that organisms were ‘robot vehicles blindly programmed to preserve the selfish molecules known as genes’.

Another instance of reciprocal rather than unidirectional action, writes Dupré, is that of an enzyme, often described as working like a lock and key, ‘fitting just-so into whatever it’s acting on. Yet increasingly it looks as if the two actually accommodate one another, less a key in a lock than a negotiation on the fly.’^[8] As one research group put it, ‘in reality, the activated receptor looks less like a machine and more like a … probability cloud of an almost infinite number of possible states, each of which may differ in its biological activity.’^[9]

On another level, all multicellular organisms are involved in mutual, symbiotic relationships of various kinds with a host of microorganisms.^[10] And the organisms, similarly, with the environment. Indeed, it is a good question whether the distinction between organism and environment can ever be other than approximate. In 1966, the experimental psychologist James Gibson coined the term ‘affordances’, referring to the opportunities afforded by an environment to an organism. The important point is that an affordance is not a thing, but resides in the relationship between an organism and its environment: the affordance requires both.^[11] What provides an affordance to one animal clearly may not to another, and may cease to be an affordance even for that animal at another moment in time. In this sense, affordances are emergent phenomena: they are opportunities for action that arise from a relationship and thereby define the organism as an agent.^[12]

Microbiologist Kriti Sharma, in her fascinating book, Interdependence, writes that ‘the cell is not exactly reacting to an environment, but is reacting with an environment, as oxygen reacts with iron and where both are transformed.’^[13] Niche construction is often thought of as what Dawkins refers to as the ‘extended phenotype’ of the organism, a one-way process, encoded in its genes and expressed through the animal creating external structures; yet, according to Dupré, this is to overlook that, equally, ‘the altered niche affects the behaviour and ultimately drives the evolution of the organism’. As he comments: ‘The difference in these perspectives nicely illustrates the difference between a thing- and a process-centred ontology.’^[14] But we might go further: Sharma’s formulation of interdependence is not just one of interaction, but of mutual constitution. Organisms and environments do not

co-construct one another sequentially, as when an organism digests food from the environment and excretes wastes into the environment, which in turn change the organism’s metabolism, and so forth. Such a view still separates organism from environment and makes one the causal agent, and then the other, and so forth, so that agency shuttles between the two. The ‘reactivity’ view suggests that organism and environment are transformed simultaneously, and at each moment … Rather, they arise new in each instant … Moreover, this ‘arising anew’ occurs dependently– that is, phenomena bring each other newly into being in each instant.^[15]

The orthodoxy is that DNA affects the fate of the cell, the cell affects the organism, and the organism the environment. This is the bottom-up view. At least as true is the top-down view: that the environment affects the organism, the organism accordingly restructures the cell, and the cell makes appropriate use of DNA in doing so. In a letter to Moritz Wagner, Darwin wrote, on 13 October 1876: ‘In my opinion the greatest error which I have committed, has been not allowing sufficient weight to the direct action of the environment, i.e., food, climate, &c, independently of natural selection … When I wrote the ‘Origin’, and for some years afterwards, I could find little good evidence of the direct action of the environment; now there is a large body of evidence’.^[16] As Fred Nijhout, Professor of Biology at Duke, makes clear: ‘When a gene product is needed, a signal from its environment, not an emergent property of the gene itself, activates expression of that gene.’^[17]

Sharma asks for a radical shift in the way we conceive of living systems. She suggests that not one, but two, steps are needed:

The first is a shift from considering things in isolation to considering things in interaction. This is an important and nontrivial move; it is also a relatively popular and intuitive concept … [But] to get to a thoroughgoing view of interdependence, I argue that a second shift is required: one from considering things in interaction to considering things as mutually constituted, that is, viewing things as existing at all only due to their dependence on other things. This second shift is potentially more subtle and difficult than the first, because though the first requires considering the mutual relations and influences between things, it does not actually require a change in the many habits and assumptions that usually commit us to viewing things as fundamentally independent.^[18]

One after another, causes that we conceived of as giving rise to an effect are shown to be at least to some extent, and in some cases to an even greater extent, themselves caused by the very effect to which they were supposed to give rise. We are used to thinking of metabolic processes as regulated by a cellular clock, though metabolic cycles can not only operate independently from such a clock, but themselves influence the circadian rhythms of that ‘clock’.^[19] In general, according to one researcher, ‘it seems that connections between the circadian clock and most (if not all) physiological processes are bidirectional.’^[20] It also seems that while some gene mutations clearly do cause cancer, the majority of mutations found in cancer are the consequences, rather than the causes, of disruption of intercellular communication.^[21] (The situation is made still more complex by the fact that, depending on context, the same gene can either promote or suppress the formation and growth of a tumour.)^[22]

The activity of individual genes reflects the choreography of chromosomes, which reflects the larger choreography of the nucleus, which reflects the choreography of the cell and organism as a whole with the environment. Who, then, is sculpting whom? How can we know the dancer from the dance?

5. The ‘parts’ are themselves changing

A machine is made of parts that do not typically alter with their context. A tappet, a widget or a gasket continues its existence effectively unaltered wherever it is put. In an organism, unlike a machine, the ‘parts’ are continually engaged in changing themselves, sometimes radically, depending on context. One of the most obvious examples is that each cell has precisely the same DNA, yet that same DNA results in dramatically different kinds of cells arising, and hence different kinds of tissue, depending on context. While we assume a mutated gene to explain a malignant trait, write Sui Huang and colleagues,

nobody will doubt that normal cells as distinct as a stem cell, a mature neuron, a blood cell or an epithelial cell all share the very same genome. No mutations are invoked to explain the vastly different phenotypes and their inheritance within a lineage. This opens the first fundamental question: how can the same set of genetic instructions produce a variety of discrete, persistent (non-genetically inherited) cell phenotypes?^[23]

The same ‘parts’ respond to quite different environments to produce utterly different effects.

An apparently ubiquitous phenomenon called ‘developmental system drift’ ensures flexibility in what a gene specifies.^[24] The result is that the same ends are repeatedly reached by quite distinct genetic means.^[25] In two species of roundworm, for example, ‘similar sexual structures arise from marked differences in signalling pathways, differences involving both novel wiring and novel protein modifications.’ ^[26] Different species of insects achieve segmentation using different mechanisms involving different genes, but with little overt difference in the outcomes.^[27] Detailed mechanisms can, of course, be identified for these developmental processes, but that’s not the point: ‘their details seem less important than the higher-level morphological “attractors” that exert a kind of downward causation … on their permissible variation, which is rather prolific.’^[28] Whole outcomes may determine how ‘parts’ behave in order to reach them, at least as much as any of the ‘parts’ determines an outcome. Thus many paths involving ‘wide-ranging differences in the number, order, and kind of molecular events’ may lead to the same reliable outcome, and considerable variations in genetic makeup can result in the expression of an identical phenotype, or form. Conversely, as we saw, identical genetic makeup does not necessarily result in identical outcomes in morphology; much depends on the context.^[29]

I have already referred to another instance, that there are far more proteins coded for by DNA than there are genes to make them: what those genes make depends on the context and what is required in it. We saw that epigenetic ‘dials’ can create 2,000 or more variations of proteins from the same gene blueprint.^[30] That the same genes can take on a number of roles provides a form of ‘belt and braces’ provision, ensuring that processes necessary to survival are preserved.^[31]

Similarly, some genes are widely conserved across species. Pax6, the gene playing the critical role in the development of an eye, occurs in almost identical form in a range of species, enabling the development of a fly’s eye, a frog’s eye or your eye: yet the types of eye, how they are structured, and how they function, are very widely different (see Plate 13[b]).

And it’s by no means just genes that are so malleable in their function. What any molecule does changes according to what is required by the context, and at many levels throughout the organism. The enzyme phosphoglucose isomerase is ‘best known for its role in the process that releases energy inside cells, for example – but when outside the cell, it can perform at least four distinct functions, such as promoting nerve growth.’^[32] Similarly with neurotransmitters: for example, serotonin (5-hydroxytryptamine, or 5-HT for short), like other neurotransmitters, has remarkably diverse effects. It is widely known for its effects on mood, which partly explain the action of many antidepressants. But it also has many other uses: in cell division, blood clotting, bone metabolism, breast-milk production, liver regeneration, bowel function, appetite regulation, and sexual function – depending on the context.

Enzymes are notorious for their shape-shifting qualities, but proteins more generally are not the static structures they are popularly conceived to be, but continually cycle through different conformations.^[33] Their structures and functions vary among cell types and change over the course of evolution. We now know that probably 30–50% of all proteins, the majority of which are not even enzymes, are so-called ‘intrinsically disordered’ proteins, lacking a fixed structure, and able to shift shape in a virtually unlimited field of possibilities (note the loaded language, which speaks of the left hemisphere preference for stasis and fixity: not ‘maximally flexible’, but ‘intrinsically disordered’).^[34]

Since structure and function are inseparable, it is not just the molecule’s function that changes, but its very structure. So it is with DNA. Genes are not things, not ‘batons’ to be handed on in the relay race of life. There is no such thing as Dawkins’s ruthlessly selfish gene, determined to pass on its lineage unscathed, in the process dominating and exploiting that poor, blind robotic vehicle, the organism, to which it belongs. ‘As genomes evolve, new genes are born and older genes may adopt novel functions, fuse, or disappear altogether,’ writes systems biologist Adrian Verster.^[35] Genes are malleable processes, subservient to the needs of the organism in which they happen for the time being to inhere.^[36]

This can be seen at two levels: changes in conformation, which affect gene expression of the same ‘line of code’; and changes in the ‘line of code’ itself.

An example of the former is that something as simple as the degree of licking and grooming of her pups by a mother rat leads to changes in gene conformation, and therefore expression, in her offspring.^[37] An example of the latter is the single-celled organism Oxytrichia trifallax, which gets rid of over 90% of its somatic genome, and re-organises the rest.^[38] This is certainly an extreme case, but, according to philosopher of science Denis Walsh,

not an especially exotic one. The engineering of the genome by the organism is commonplace. Cells actively cut, transpose, copy and fix their genomes. They do so in highly sensitive, adaptive ways.^[39]

What that means is that cells use DNA to adapt to new ends; it is not a matter of DNA using cells to further its own ‘selfish’ ones. The idea that genes are somehow (how?) ‘programmed’ to pass on their DNA does not sit well with the fact that cells are constantly acting on it to change it or to repair it; and such persistence as there is depends on ‘elaborate editing and correcting processes in the cell’.^[40] These include anything from point mutations to large-scale genome rearrangements and whole genome duplications.^[41] There are ‘massive levels of genomic rearrangement during development’.^[42]

It should not surprise us that the ‘parts’ of an organism change themselves, consume energy and turn over rapidly even though the whole structure persists, since life is at every level a dynamic process.^[43] ‘Logically, the elements of a process can be only elementary processes, and not elementary particles or any other static units’, wrote biologist Paul Weiss.^[44] One cannot explain a changing process by unchanging elements, though our extraordinary bias towards ‘thingness’, rather than forms or processes, encourages us to think that we could.

Sometimes biologists, says Dupré, talk as though they believe that organisms are put together

from bits of Lego … Rather, living things are processes that are capable of assuming many protean forms: dynamic, ever-changing, but balancing, for a time, on just the right side of chaos.

‘Just the right side of chaos’: they are, in other words, what Naseem Nicholas Taleb calls ‘antifragile’ systems.^[46] Taleb makes a distinction between robustness, which relies on resistance to change, but is susceptible to catastrophe when it finally can resist no more; and antifragility, which thrives on flexibility, makes small adjustments and thereby not only survives but evolves. All evolutionary processes are in this sense antifragile, constantly adapting, and because of their ability to permit constant small ‘errors’ (as they seem at the time) able to preserve themselves against final destruction. The expression ‘the right side of chaos’ reminds us that a system that is capable of flexible and rapid response to a multitude of possible occurrences must have inbuilt instability, as well as stability: the classic case of this is the Eurofighter Typhoon jet, a supremely agile fighter plane, which was built deliberately unstable so as to make it highly responsive. Because of this it can make swift and abrupt changes in direction that a more stable plane could accomplish only over many miles.

An organism must be capable of a vast repertoire of available responses, and some of them, unbelievably enough, are invented by organisms ‘on the hoof’, as we shall shortly see. Stuart Kaufmann writes that

genomic systems lie in the ordered regime near the phase transition to chaos. Were such systems deeply into the frozen ordered regime, they would be too rigid to coordinate the complex sequence of genetic activities necessary for development. Were they too far into the gaseous chaotic regime, they would not be orderly enough. Networks in the regime near the edge of chaos – this compromise between order and surprise – appear best able to coordinate complex activities and best able to evolve as well.^[47]

Ultimately, even what we conceive to be the ‘solid’ parts of cells are actually flows. The living cell is mainly fluid, principally water. Even surfaces, cell membranes, the cytoskeleton, and the various fibre systems, that look relatively solid, are ‘subject to more or less continuous dissolution and reconstitution.’^[48] We imagine the organism contains fixed structures because of the nature of left hemispheric attention, which replaces flow with frozen slices removed from time: our static diagrams, or photomicrographs.

---

^[1] Schrödinger 1951 (53: emphasis in original).

^[2] Walsh 2018 (179).

^[3] Shapiro 2013.

^[4] Waddington 1953.

^[5] Beisson & Sonneborn 1965.

^[6] Nelsen, Frankel & Jenkins 1989.

^[7] Keller 2014 (40–1).

^[8] Dupré 2017b.

^[9] Mayer, Blinov & Loew 2009.

^[10] Dupré & Nicholson 2018 (20–1); Leclerc 1972.

^[11] Gibson 1979b (285). See also Gibson 1979a (127).

^[12] Walsh 2018 (174).

^[13] Sharma 2015 (66: emphases in original). I am grateful to John Saunders for bringing this book to my attention.

^[14] Dupré 2017a.

^[15] Sharma 2015 (67–9: emphasis in original).

^[16] Darwin 1887, vol III (159: emphasis added).

^[17] Nijhout 1990.

^[18] Sharma 2015 (2: emphasis in original).

^[19] Kumar & Takahashi 2010.

^[20] Yang 2010.

^[21] Greenman, Stephens, Smith et al 2007. For a general overview of tissue organisation field theory, see Baker 2015.

^[22] Lobry, Oh, Mansour et al 2014; Liu, Radisky, Yang et al 2012.

^[23] Huang, Ernberg & Kauffman 2009 (871: emphasis in original).

^[24] True & Haag 2001.

^[25] Verster, Ramani, McKay et al 2014.

^[26] Wang & Sommer 2011; Robinson 2011.

^[27] Salazar-Ciudad, Solé & Newman 2001.

^[28] Newman 2014.

^[29] Webster & Goodwin 2011 (87); and Nicholson 2014 (167).

^[30] Bray 2003. See also Schmucker, Clemens, Shu et al 2000.

^[31] Anjum & Mumford 2018 (64).

^[32] Dupré 2017b.

^[33] Guttinger 2018 (315ff).

^[34] Dunker, Lawson, Brown et al 2001; Uversky 2010; Gsponer & Babu 2009.

^[35] Verster, Ramani, McKay et al 2014.

^[36] Dupré & Nicholson 2018 (42).

^[37] Champagne, Weaver, Diorio et al 2006.

^[38] Chen, Bracht, Goldman et al 2014.

^[39] Walsh 2018 (179: emphasis added).

^[40] Noble 2017. The reference is to Dupré 2017a.

^[41] Shapiro 2013 (287).

^[42] Chen, Bracht, Goldman et al 2014.

^[43] Mackintosh & Schmidt 2010.

^[44] Weiss 1962 (3).

^[45] Dupré 2017a.

^[46] Taleb 2013.

^[47] Kauffman 1995 (26).

^[48] Talbott 2011a (7).