Overproduction in Nature: a key principle?

One feature of biological systems that seems to be recurrent, from fertilization in reproductive systems, neuronal cell proliferation and connectivity in development, or the overproduction of gene products to confer pesticide resistance, is overproduction. The observation of overproduction in a range of biological contexts may reveal two distinct features of these systems. One is that overproduction results from stochastic processes. Because of this, there tends to be massive redundancies in overproduced systems. Structures are overproduced to ensure a functional architecture without locking the system in to a narrowly-defined operating range. The other feature is that overproducing systems are not tightly optimized energetically, which can lead to fluxes that may have productive and even adaptive consequences.

It could be suggested that overproduction is a key component of how natural selection and adaptive control operate on physiological systems. One aspect of this is the control of fluxes in biochemical production and cell proliferation. While overproduction appears at first glance to be a runaway process, it is in fact a self-regulating process with a high degree of gain. In this case, the gain is high because there is a default pathway for overproduction. However, there are energetic and physical upper-bounds that ultimately limit the rate of production. As we will see, natural systems have a tendency to overproduce. In some cases, overproduction serves as a means to build dissipative structures [1], or structures that result from the flow of energy, improving the efficacy of a particular biological process.

In the case of creating an animal embryo, millions of sperm are generally involved in the fertilization of a single egg. While only a single sperm actually donates its genetic material, many other sperm are involved in the overall process of fertilization. The fertilization of an egg is known as a receptor-ligand binding event, in which "specificity" (a highly selective and information-contingent process) is required. While it is debatable whether or not "collective" behavior exists among a population of ejaculated sperm, there is no doubt that many more sperm are produced than would be needed if fertilization were a deterministic process. Angiosperms use a similar strategy of overproduction to fertilize and reproduce. In this case, many spores are released, but only a fraction of these spores actually result in viable offspring. The reason for this overproduction in both animal and angiosperm reproduction is quite clear, as the dissemination of sperm is quite broad relative to the egg or stamen. Therefore, the delivery vehicle must be overproduced to hit its target.

Animal (above) and angiosperm (below) reproductive mechanisms

In the case of neuronal cell proliferation and connectivity in human brains, the number of neurons and synapses peaks early in development [2, 3]. Yet it is not the sheer number of these structures that leads to intelligence and adaptive brain function. This collection of neurons and synapses are both pruned during interaction with the environment, but not to the extent that they become too sparse to do their job. This property is also seen in the domain of learning and memory, as synapses are pruned during sleep in the course of consolidating memories acquired during the previous day [4]. Therefore, while overproduction can be governed by an upper-bound, the refinement that occurs in some systems as a counterbalance to overproduction seems to have a lower bound that corresponds with the maintenance of physiological function.

Active neurogenesis (above) and synaptogenesis (below)

Overproduction is driven by two requirements: the need to respond to stochastic, probabilistic events that characterize biological processes, and the need to recognize specific events and signals in the cell. Therefore, as seen in the sperm fertilization example, overproduction may also be a consequence of a system's specificity. In the case of enzyme specificity, this is defined by a lock and key model [5, 6]. In the lock and key model, a specific receptor binds to a specific ligand (a 1:1 relationship), and receptor-ligand pairings evolve towards this end. By contrast, overproductive regimes rely upon sheer numbers to confer recognition. In this case, there is a m:n relationship, with m being the number of signals and n being the overall population size. In this scenario, things become noisy (in the classic signal-to-noise ratio sense) only for very large numbers of n relative to m. Taking the two requirements of overproduction into account, it seems that natural systems prepare for the eventualities of the real world by producing a m (potential signal) embedded in a background of n (total number of products).

Overproduction is also a key component of adaptive responses to environmental stimuli. For example, the mosquito (Culex pipiens) overproduces a number of esterases which confers a resistance to certain insecticides [7]. This occurs according to two mechanisms: gene amplification and gene regulation. Specifically, overproduction of a specific esterase (A1) and gene amplification of a second esterase (B1) are required for this resistance mechanism. Both loci (mutants in this case) are expected be related in terms of a tight statistical association. Other alleles may play a role as well, which is variable by individual mosquito. This example does not correspond directly with the aforementioned m:n relationship, but the dynamics of gene amplification and gene regulation may operate in a similar way.

Picture of a mosquito (Culex pipiens)

The natural recycling processes within cells can also manage the rate of overproduction for certain products [8, 9]. While not supported by existing data, products with fast kinetics, a high rate of degradation, and variable survival times are expected to be overproduced. All of these properties both enable and drive the need for enhanced rates of product synthesis. In addition, the motility of these products can result in their own complexity. This can be demonstrated using a diffusion-limited aggregation (DLA) process, a physical model that simulates a large particle population aggregating at targets driven by purely diffusive processes. In a DLA model, a population of particles diffuses across a space at variable rates. This space also features a number of barriers to which particles can adhere to after a collision. What a DLA simulation will show after a period of time is that some particles (the n particles of our m:n relationship) will adhere to the barriers and create structure which other particles (the m particles of our m:n relationship) can use to guide themselves to a target or otherwise facilitate binding. This enhances the signal-to-noise ratio for products that carry a specific signal, arguing for overproduction as an adaptive response.

Example of a diffusion-limited aggregation process (above), and an example of "nanoflowers" [10], which uses a DLA process to construct therapeutic structures in vivo (below).

Finally, a collection of overproduced things can be thought of as a population of replicator vehicles [11]. Replicators are a ubiquitous feature of evolutionary systems, and are required for natural selection. Put simply (and recursively), replicators are objects with the ability to self-replicate. In a broader context, replicators constitute a population upon which natural selection and regulatory mechanisms can act. Thinking about overproduction in this way allows us to examine the role of variants in the population and better understand how they act in either a competitive or cooperative manner, which ultimately contributes to how overproduction is regulated over time.


REFERENCES:
[1] Bejan, A. (2000). Shape and Structure, from engineering to nature. Cambridge University Press, Cambridge, UK.

[2] Sanai, N. et.al (2011). Corridors of migrating neurons in the human brain and their decline during infancy. Nature, 478, 382-386.

[3] Chechik, G. and Meilijson, I. (1999). Neuronal Regulation: a mechanism for synaptic pruning during brain maturation. Neural Computation, 11(8), 2061-2080.

[4] Horn, D., Levy, N., and Ruppin, E. (1998). Memory maintenance via neuronal regulation. Neural Computation, 10(1), 1-18.

[5] Adami, C. (2006). Reducible Complexity. Science, 312, 61.

[6] Kondrashov, F.A. (2005). In search of the limits of evolution. Nature Genetics, 37(1), 9-10.

[7] Raymond, M. et.al (1998). An overview of the evolution of overproduced esterases in the mosquito Culex pipiens. Philosophical Transactions of the Royal Society of London B, 353, 1707-1711.

[8] Kundu, M. and Thompson, C.B. (2008). Autophagy: Basic Principles and Relevance to Disease. Annual Review of Pathology, 3, 427–455.

[9] Maxfield, F.R. and McGraw, T.E. (2004). Endocytic Recycling. Nature Reviews Molecular Cell Biology, 5, 121-132.

[10] http://uonews.uoregon.edu/archive/news-release/2011/5/forecast-calls-nanoflowers-help-return-eyesight

[11] Dawkins, R. (1996). Blind Watchmaker. W.W. Norton, New York.