Laws of Biology: A Complexity Law?
This post continues from the previous, which discussed the notion of laws in biology, and considered candidates for what might be the major contenders for universal ‘lawful’ status in this domain. Although at the end of this post, a series of possible universal biological dictates were briefly listed, the status of evolution as the primary biological law was highlighted. In turn, this places natural selection, the primer mover of evolution, in the spotlight.
But is this really the case? Is there a more fundamental law that underlies (or at least accompanies) all evolutionary processes? These issues are examined further in this post.
Complexity and Evolution
When discussing the priority of biological ‘laws’, semantics and definitions will inevitably enter the picture. Thus, while it might be acknowledged that evolution is ‘Law No. 1’, it might be proposed that more fundamental laws operate which enable evolution in itself. The ‘law’ of Darwinian natural selection immediately springs to mind, but other processes have been proposed as even more fundamentally significant.
In the previous post, the intriguing role of complexity as a putative ‘arrow of evolution’ was alluded to. Here it was also noted that this apparent ramping of complex functions could arise from ‘locking in’ of systems in increasingly intricate arrangements. In this view, where evolutionary success is associated with a structural or functional innovation which increases the overall complexity of an organism, reversing such a change in later generations may be unfeasible. This will occur when the selected evolutionary change has become enmeshed as a fundamental part of a complex network of interactions, where removing a key component cannot easily be compensated. Successive innovations which become ‘locked’ in a comparable manner thus lead to a steady increase in the net complexity of biological systems. Of course, this is not to say that all complex adaptations are irreversible, and a classic example of ‘complexity loss’ is the deletion of functional eyes from cave-dwelling animals living in complete darkness.
An interesting recent publication highlights a possible example of an evolutionary process that may lead to increased complexity. Gene duplication has long been acknowledged as an important driver of evolution, where formation of a duplicate gene copy allows continuation of the original gene function while allowing mutations in the second copy (or ‘paralog’), and accompanying exploration of new evolutionary space. An implicit assumption in such cases is that the expressed paralog does not interfere with the original gene function, but this may not apply where such functions depend on networks of co-operating protein-protein and protein-nucleic acid interactions. Johnson and colleagues have shown that this is indeed the case for certain duplicated transcription factor genes in yeast. As a result, a strong selective pressure exists for resolution of such paralog interference, one solution of which is mutational change which removes the interference effect, while simultaneously allowing the progressive emergence of novel functions. Such effects were noted by this group as being a potential source of increasing complexity, although this is not formally proven.
A complexity gradient is readily apparent between bacterial cells and single-celled eukaryotes, and in turn between the latter and the variety of multicellular organisms based on the eukaryotic cellular design. The key enabling event here was an ancient symbiosis between ancestral bacterial cells, resulting in the eventual evolution of mitochondria and chloroplasts as key eukaryotic organelles for energy production and photosynthesis respectively. The energetic endowment of mitochondria allowed the evolution of large genomes capable of directing the complex cell differentiation required for multicellular life. (And among the mechanisms for such genomic complexity acquisition, we can indeed note duplication events, as mentioned above).
And yet it is important to consider the putative evolutionary ‘drive’ towards complexity in terms of the biosphere as a whole. Whatever the evolutionary origin of biological systems of escalating intricacy, it is clearly not a global phenomenon. Only certain eukaryotic lineages have shown this apparent trend, while the prokaryotic biomass on Earth has existed in a similar state for billions of years, and will no doubt continue to exist as long as suitable habitats exist on this planet. (And here prokaryotes are clear masters at colonizing extreme environments).
Such observations are entirely consistent with the blind forces of natural selection. Change in a lineage will not occur unless variants emerge which are better-suited to existing or changing environments (including evasion or domination of natural predators or competitors). So, the question then becomes: is increased complexity simply a by-product or accompaniment to natural selection itself which may or may not occur, or is it an inevitability? Before continuing any further, it will be useful to briefly look at just how complexity might be assessed and measured.
Defining Biological Complexity
Perhaps consistent with its intuitive properties, a good definition of complexity in itself is not the essence of simplicity. One approach would seem logical: to perform a census of the range of different ‘parts’ that comprise an organism, where the total count provides a direct complexity index. (Obviously, by this rationale, the higher the ‘parts list’ number, the greater the perceived complexity). But a problem emerges in terms of the system level at which such a survey should be performed, since it can be in principal span hierarchies ranging from the molecular, organelle, and cell differentiation state, up to macroscopic organs in multicellular organisms. The physical size of organisms has also been noted as a correlate of complexity, but not a completely reliable one.
An additional and very important observations also suggests that a simple parts list of organismal components is at best a considerable under-rating of what may be the true underlying complexity. Biological systems are characteristically highly modular and parsimonious. This bland statement refers to the often incredible economy of informational packing in genomes, such that a basic human protein-encoding gene count of only approximately 20,000 can encode the incredible complexity of functioning human being. The baseline gene figure is greatly amplified by systems using separate gene parts (exons) in alternative ways, through RNA splicing and editing, and a gamut of post-translational modifications. But beyond this level of parsimonious modularity, the same gene products can perform quite distinguishable functions through differential associations with alternative expressed products of the same genome, corresponding to distinct cellular differentiation states. A far better account of complexity must therefore cover the entire interactome of an organism, but this is a far more onerous undertaking than a mere parts list.
And the levels of potentially encoded complexity don’t even stop there. Consider a protein A that interacts with proteins B and C in one cell type (α) within an organism, and D and E in another type of cell (β) within the same organism. The differential complexes ABC and ADE result from alternate programs of gene expression (cell type α having an expressed phenotype A+, B+, C+ D-, E-; while the β phenotype is A+, B-, C-, D+, E+), combined with the encoded structural features of each protein which enable their mutual interactions. The interaction of A with its respective partners is thus directly specified by the genome via regulatory control mechanisms. But indirect programming is also possible. There are numerous routes towards such a scenario, but in one such process a genomically-encoded gene A can be randomly assorted with other gene fragments prior to expression, such that a (potentially large) series of products (A*, A**, A***, and so on) is created. If a single cell making a specific randomly-created modification of the A gene (A*, for example) is functionally selected and amplified, then A* is clearly significant for the organism as a whole, yet is not directly specified by the genome. And the creation of A* thus entails a ramping-up of organismal complexity.
The ‘indirect complexity’ scenario is actively realized within the vertebrate adaptive immune system, where both antibody and T cell receptor genes are diversified by genomic rearrangements, random nucleotide additions (by the enzyme terminal transferase) and somatic hypermutation. And clearly the circuitry of the mammalian nervous system, with its huge number of synaptic linkages, cannot be directly specified by the genome (although here the details of how this wiring is accomplished remained to be sketched in).
These considerations make the point that defining and quantitating complexity in biological systems is not as straightforward as it might initially seem. In principle, a promising approach centers on treating the complexity of a system as a correlate of system information content. While this has been productive in many virtual models, it still remains an elusive goal to use informational measures in accounting for all of the above nuances of how biology has achieved such breath-taking levels of complexity.
A ‘Zeroth’ Law for Biology?
Where measures of complexity can be kept within a specified rein, multiple computer simulations and models have suggested that evolving systems do show a trend towards increasingly complex ‘design’. But in real biological systems, what is the source of burgeoning complexity? Is it somehow so inevitable that it needs the status of a ‘law’?
McShea and colleagues have proposed the ‘Zero-Force Evolutionary Law’ (ZFEL), which has been stated as: “In any evolutionary system in which there is variation and heredity, there is a tendency for diversity and complexity to increase, one that is always present but may be opposed or augmented by natural selection, other forces, or constraints acting on diversity or complexity.” This could be seen as a law where complexity is increasing ratcheted up over evolutionary time, through the acquisition of variations which may have positive, negative, or neutral selectable properties. If subject to negative selection, such variants are deleted, while positively-selected variants are amplified by differential reproductive success. Variants that are completely neutral, however, may be retained, and potentially serve as a future source of evolutionary diversity.
An interesting wrinkle on the notion of neutral mutations is the concept of conditional neutrality, where a mutation may be ‘neutral’ only under certain circumstances. For example, it is known that certain protein chaperones can mask the presence of mutations in their ‘client’ proteins which would be otherwise unveiled in the absence of the chaperone activity. (A chaperone may assist folding of an aberrant protein into a normal structural configuration, whereas with impaired chaperone assistance the protein may assume a partially altered and functionally distinct structural state). Such a masking / unmasking phenomenon has been termed evolutionary capacitance.
But is the ‘Zero-Force’ law truly that, or simply a by-product of the primary effect of Darwinian natural selection? (The latter was discussed in the last post as the real First Law of Biology). The above ZFEL definition itself would seem to embed the ‘Zero Force’ law as an off-shoot of evolution itself, by beginning with ‘In any evolutionary system……’. Certainly ZFEL may correctly embrace at least one means by which complexity is enhanced, but since the adoption or elimination of such candidate complexity is ultimately controlled by natural selection, it would seem (at least to biopolyverse) that it is a subsidiary rule to the overarching theme of evolution itself.
In any case, if a ‘zero-force’ law is operative, why has the huge biomass of prokaryotic organisms persisted within the biosphere for such immense periods of time? An interesting contribution to this question highlights the importance of an organism’s population size for the acquisition of complexity. In comparison with prokaryotes, eukaryotes (from single celled organisms to multicellular states) are typically larger in physical size but with smaller total population numbers. (Recall the above mention of the role of eukaryotic mitochondria in bioenergetically enabling larger genomes, and in turn larger cell sizes). In a large and rapidly replicating population, under specific circumstances a paralog gene copy arising from a duplication event (noted above as an important potential driver of complexity acquisition) has a significantly greater probability of being deleted and lost before it can spread and become fixed. Thus, from this viewpoint, a eukaryotic organism with a substantially reduced population base is more likely to accumulate genomic and ultimately phenotypic complexity than its prokaryotic counterparts. Once again, the origin of eukaryotes through the evolution of symbiotic organelles derived from free-living prokaryotes was an absolutely key event in biological evolution, without which complex multicellular life would never have been possible. And eons of prokaryotic existence on this planet preceded this development, suggesting that it was not a highly probable evolutionary step, perhaps dependent on specific environmental factors combined with elements of chance.
A complexity-enabling but highly contingent (and evidently rate-limiting) event such as eukaryogenesis does not create confidence in the operation of a regular biological law. And other ‘complexity breakthroughs’ are likely to exist. The ‘Cambrian Explosion’, where a variety of animal phyla with distinct body plans emerged during the beginning of the Cambrian era about 540 million years ago, may be a case in point. This ‘explosion’ of complexity in a relatively short period of geological time has long been pondered, although molecular phylogenetic data have suggested earlier origins of many phyla. Still, an intriguing suggestion has been that the first evolution of ‘good vision’ was an enabling factor for the rapid evolution (and thus complexification) of marine Cambrian fauna.
So increasing biological complexity seems to have more of a ‘punctuated’ evolutionary history than an inexorable upward trend. Fitting a ‘law’ into what is governed by environmental changes, contingency, and natural selection may be a tall order. But perhaps it is too early to say……
On that note, a non-complex biopoly-verse offering:
Within life, one often detects
A trend towards all things complex
Does biology have laws
That underlie such a cause?
Such questions can sorely perplex….
References & Details
(In order of citation, giving some key references where appropriate, but not an exhaustive coverage of the literature).
‘ Johnson and colleagues……’ See Baker et al. 2013.
‘….the eventual evolution of mitochondria and chloroplasts….’ See the excellent book by Nick Lane: Power, Sex, Suicide – Mitochondria and the Meaning of Life. Oxford University Press, 2005.
‘…..a census of the range of different ‘parts’….’ See McShea 2002.
‘…..protein-encoding gene count of only on the order of 20,000…..’ Note the ‘protein-encoding’ here; if non-coding RNA genes are added, the count is much higher. See the GENCODE data base.
‘ The physical size of organisms has also been noted as a correlate of complexity….’ See a relevant article by John Tyler Bonner (2004).
‘…..Biological systems are characteristically highly modular and parsimonious. / A far better account of complexity must therefore cover the entire interactome of an organism…..’ The need to address the modularity of living systems in order to fully apprehend them has been forcefully argued by Christof Koch (2012).
‘McShea and colleagues have proposed the ‘Zero-Force Evolutionary Law’……’ See Biology’s First Law . Daniel W. McShea and Robert N. Brandon , 2010 , Chicago University Press , Chicago, IL, USA; also Fleming & McShea 2013.
‘Such a masking / unmasking phenomenon has been termed evolutionary capacitance.’ See a recent interview (Masel 2013) for a background on such capacitance phenomena, and further references.
‘In comparison with prokaryotes, eukaryotes (from single celled organisms to multicellular states) are typically larger in physical size……’ Exceptions exist where very large prokaryotes overlap in size with single-celled eukaryotes. In such cases, the giant prokaryotes are powered by multiple genome copies. On that theme, see Lane and Martin 2010.
‘In a large and rapidly replicating population, under specific circumstances a paralog gene copy arising from a duplication event ……has a significantly greater probability of being deleted….’ This requires more detail regarding gene duplication outcomes: Following a gene duplication event, a resulting paralog copy can acquire deleterious mutations and be lost, or rarely acquire advantageous mutations providing a positive selection (neofunctionalization). But another possible outcome is where deleterious mutations occur in both gene copies, such that both are required for the continuing fitness of the host organism. In such circumstances of subfunctionalization, the original functions of the single encoded gene product are thus distributed between the two duplicate copies.
Another significant point with respect to the population size argument of M. Lynch and colleagues is that selectable fixation of mutational variants will always be longer in large replicating populations than in small ones. Where subfunctionalization occurs after gene duplication, additional mutational changes can occur which completely inactivate one copy, a terminal loss of fitness. In a large population base, such events will act against the species-wide retention of gene subfunctionalization much more so than in small populations. The latter, therefore, are subject to relatively increased complexification as a result of the preservation of this type of gene duplication.
‘ This ‘explosion’ of complexity in a relatively short period of geological time has long been pondered, although molecular phylogenetic data have suggested earlier origins of many phyla.’ See Jermiin et al. 2005 for some discussion of these themes.
‘….the first evolution of ‘good vision’ was an enabling factor for the rapid evolution…..’ See Zhao et al. 2013 for a recent study, and discussion of this notion.
‘Fitting a [complexity] ‘law’ into what is governed by environmental changes…..’ See Auerbach & Bongard 2014 for an in silico study of environmental effects on the evolution of complexity. They find environmental complexity and model organismal complexity are correlated, suggesting complexity may only be favored in certain biological niches.
Next Post: April.