15

Tradeoffs

In order to spend on one side, nature is forced to economise on the other side.

—Goethe, quoted by Darwin⁷⁴

Tradeoffsset the basis for design. To grow faster, it would seem that resources must be taken away from some other function. If there were no cost, why would the organism not already have achieved faster growth?

But focusing solely on tradeoffs will fail to reveal design. Suppose, for example, that growth rate, yield, and defense trade off against each other. Changed conditions may reduce allocation to defense, allowing both growth rate and yield to rise.³⁰⁰ If we do not measure defense, the rise in rate and yield would seem to contradict the rate-yield tradeoff.

There will always be another tradeoff not measured. One cannot win solely by collecting more data. There has to be a proper method behind the effort (Chapter 3).

Another problem is that organisms are not perfectly adapted. A cell might evolve a more efficient catabolic enzyme, capturing more free energy without costly reduction in another function. Some improvements can happen without tradeoff.

If the study of tradeoffs often fails to reveal design, then why focus on them? Because tradeoffs do play the central role in design. The problem is not the importance of tradeoffs in shaping design. The problem is our ability to infer the causes of design.

I have argued that comparative hypotheses provide the only approach to inferring the causes of design. Tradeoffs are the building blocks of comparative hypotheses. Before turning to comparative hypotheses in the following chapters, it is useful to consider in this chapter the potential tradeoffs that may be important in different situations.

We will never guess all of the important tradeoffs when developing theory. And we will never measure all of them in empirical study. But the more tradeoffs we know about and the more of those tradeoffs that we use to develop comparative predictions, the closer we will come to understanding the forces that shape organismal design.^{20,104,209,379}

Typically, one assumes that a tradeoff exists and then considers what predicted consequences follow from that assumption. How does one test for the tradeoff itself? As always, attempting direct measurement provides a limited signal about the forces of design.

Testing a comparative hypothesis provides the way forward. How do changing environmental conditions alter the likelihood that a particular tradeoff influences design? To develop such comparative hypotheses, one must have a good sense of the possible tradeoffs.

This chapter lists some tradeoffs for metabolic traits. These examples encourage thought about what might be important. The examples also provide the basis for developing broader lists, which would help to locate the boundaries for possible explanations of design. The actual is always a subset of the possible.

15.1 Biophysical Constraints and Cellular Allocation

How do physical limits, such as membrane surface area, impose tradeoffs between alternative functions? How do cells split investment between the structures needed for growth, such as ribosomes, and productive investment in actual growth, such as making new proteins?

LIMITED SPACE

Cells allocate their limited membrane surface area and their limited cellular volume to alternative functions. Those biophysical limits impose particular tradeoffs.

• Limited membrane surface area imposes tradeoffs. For example, food typically passes through membrane transporters. Oxidative phosphorylation uses the membrane to conduct electron transport. At high throughput, membrane space imposes a tradeoff between food uptake and electron transport.399
• Limited cellular volume constrains the total number of proteins, imposing tradeoffs. For example, the abundance of transcription factors that regulate cellular function trades off against the abundance of structural proteins that carry out cellular function.³²⁹
• Cell crowding slows diffusion and limits biochemical reactions.^16,359 To overcome slow diffusion, reaction cascades may colocalize into phase-separated compartments.³⁷ Tradeoffs likely arise in the maintenance of spatial partitioning within limited cellular volume. The implications for cellular design remain an open problem.^5,25,238

RESOURCE ACQUISITION

Cells must acquire various nutrients. Tradeoffs occur between finding, uptake, and efficiency.²⁴⁷ Tradeoffs also arise between the uptake of different nutrients.

• Exploration for new food resources by motility trades off against efficient exploitation of local food sources.^164,395,402 Tradeoff strength increases as the cost rises for motility and the benefit rises from the patchiness of food distribution.
• A related tradeoff occurs between costly motility for individual cells and the beneficial tendency to be at the edge of a growing colony. Motility in colonies segregates cells by their speed, potentially colocalizing cells with synergistic traits.⁴⁷⁴
• Patchy environments favor motility because of the high gains for finding new food patches. However, in resource-poor environments, rich patches tend to be rare, favoring the cost savings associated with relatively low motility.²⁹⁹
• Nutrient transporter affinity trades off against cost.¹⁰³ In rich environments, low affinity and low cost transporters maximize nutrient uptake rate per unit cost. In poor environments, higher affinity and higher cost transporters achieve greater nutrient uptake efficiency, as in the widely conserved ABC system.²⁹⁹
• Catabolic pathways for different food sources associate with matching uptake systems.⁸¹ Tradeoffs may occur between alternative uptake systems.

ALLOCATION TO ALTERNATIVE ASPECTS OF GROWTH

Tradeoffs occur between growth-related functions.^{156,240,440,464}

• The rate of protein production trades off against the yield efficiency of resource usage.³⁴³ Species with more ribosomal RNA operons make proteins faster, grow faster, and have lower biomass yield per unit carbon uptake.
• Increasing resources favor protein production rate over yield. In 1167 bacterial species, ribosomal RNA copy number increases with traits that are common in resource-rich environments, such as chemotaxis and larger genome size.³⁴³
• Allocation to alternative proteins varies with environmental challenge, suggesting broad tradeoffs.^188,240 Changes in the marginal costs and benefits of flux through different pathways may explain cellular shifts in proteome allocation to different cellular functions.¹⁸⁸
• In yeast, mitochondria mediate a growth rate versus yield tradeoff. The mitochondria change from primarily a biosynthetic anabolic hub during rapid fermentive growth to primarily a catabolic hub during slower and more efficient respiration.⁸³
• Limitation of essential elements induces tradeoffs between pathways. For example, in E. coli, phosphate limitation favored a mutant with lower TCA cycle flux. That mutant had reduced fitness under carbon limitation.³²
• Rapid growth requires more phosphorus for ribosomes, whereas other life histories demand relatively higher availability of carbon or nitrogen.^{95,147,211,245,281}

FREE ENERGY STORES FOR DELAYED FUNCTIONS

Molecular storage includes glycogen, trehalose, polyphosphates, polyhydroxyalkanoates, various lipids, and occasionally polypeptides.^208,378

When lack of essential nutrients limits current growth, microbes may gain by storing available nutrients and free energy for future benefits. Alternatively, tradeoffs between current and future benefits may favor storage. Possible future benefits include the following.

Figure 15.1   Biological phosphorus removal process. Redrawn from Fig. 17 of Curtin et al.72

• Many yeast and bacteria store glycogen to use during starvation conditions.⁴⁵⁰
• Storage may allow rapid transition between regulatory states. In E. coli, rapidly accessible glycogen stores reduced the lag time during shifts between alternative carbon sources.³⁷²
• Alternative glycogen molecular structures may trade off stability versus rapid accessibility.⁴³⁸ Stability favors starvation resistance. Accessibility favors rapid regulatory changes. Arguments continue about how glycogen’s structural branching influences long-term molecular stability versus short-term accessibility of free energy.^438,450

ALTERNATING ENVIRONMENTS FAVOR CYCLES OF STORAGE AND GROWTH

The microbiology of human-engineered water purification treatment provides a good example of such cycles.^72,88,297

Figure 15.1 shows a common process. Waste often contains common products of glycolytic fermentation, such as acetate. The first purification step begins with anaerobic bacterial activity. In the absence of oxygen or other strong electron acceptors, the bacteria cannot use acetate to generate a free energy gradient for growth.

Instead, cells draw on internal stores to generate usable free energy. They may derive glucose from stored glycogen or cleave phosphate bonds in stored polyphosphate. The free energy from those reactions drives uptake and transformation of acetate into polyhydroxyalkanoate (PHA) stores (Figs. 15.2a and 15.3b).

Figure 15.2   Metabolic pathways of storage and growth for common phosphate-accumulating bacteria in waste processing systems. (a) During the initial anaerobic phase, bacteria take up volatile fatty acids, such as acetate. Free energy from polyphosphate or glycogen stores drives transformation of the fatty acids into PHA storage. (b) During the subsequent aerobic phase, free energy obtained by oxidizing the stored PHA drives biomass production and the building of polyphosphate or glycogen stores. Metabolic variations occur among different species. For example, glycogen-accumulating bacteria use relatively little polyphosphate and so do not remove phosphates from the waste. When carbohydrates are present in the initial waste, anaerobic fermentation often transforms those carbohydrates into volatile fatty acids. From Fig. 2 of Dorofeev et al.88

After an anaerobic period, the treatment process aerates the waste. Cells derive usable free energy by oxidizing the PHA stored during the anaerobic phase. The free energy from PHA drives growth and biomass production. Cells also use the free energy from PHA to replenish their stores of polyphosphate or glycogen (Figs. 15.2b and 15.3b).

Figure 15.3   Bacterial uptake and storage of key nutrients over the cycle of anaerobic and aerobic environments in phosphate-removal processes. (a) Extracellular concentration levels of phosphate and volatile fatty acids (VFA), such as acetate. (b) Internal cellular stores for phosphate-accumulating species. In glycogen-accumulating species, the external phosphate changes relatively little and the internal polyphosphate remains low. Instead, glycogen storage falls more rapidly during the anaerobic phase and builds more rapidly during the aerobic phase. From Fig. 1 of Dorofeev et al.88

The bacterial biomass can be removed as waste sludge after one or more anaerobic-aerobic cycles. Sludge removal discards any stored carbon or phosphate, purifying the water.

One practical challenge concerns competition between bacteria that build polyphosphate stores and bacteria that build glycogen stores. The phosphate-accumulating bacteria provide efficient removal of phosphorus from waste. The glycogen-accumulating bacteria leave much of the phosphorus in the waste (Fig. 15.3). Tuning of the purification process requires finding the steps and environmental setpoints that favor phosphate-accumulating over glycogen-accumulating species.³⁵¹

This system provides an excellent model for the study of metabolic design.³⁸³ Although artificial, it captures the essence of alternating environments, the benefits of storage, the thermodynamic constraints imposed by varying electron acceptors, and the shaping of long-term fitness over a full demographic cycle.

Several tradeoffs likely influence the fitness of particular genotypes and the competition between genotypes and species.³⁸³

• During the aerobic phase, cells trade off biomass production against building polyphosphate or glycogen stores.
• Polyphosphate and glycogen production draw on the same ATP supply. A tradeoff may occur between building those alternative storage molecules.
• For example, polyphosphate and glycogen provide different precursors for other metabolic processes. They make different demands for uptake and membrane space. They likely have different redox characteristics, temperature sensitivities, pH sensitivities, and other chemical properties.
• During the anaerobic phase, uptake of volatile fatty acids may trade off against uptake of food sources that could drive glycolytic fermentation and growth. Limited membrane space or limited free energy to drive active transport may impose the tradeoff.

CHANGES IN LIMITING RESOURCES CAUSE CHANGES IN TRADEOFFS

Tradeoffs between alternative allocations often depend on context. Accounting for context improves comparative tests.

• When lack of particular nutrients limits growth, cells may store available nutrients for future benefits.^88,450 Building storage molecules may trade off against current maintenance, repair, and survival.
• By contrast, when available nutrients allow growth, the future benefits of storage may trade off against current growth.
• Restricted phosphorus limits ribosome and protein production.⁹⁵ Limited protein imposes a strong tradeoff between making proteins for new growth and making proteins for other functions, such as cellular regulation. Abundant phosphorus may relieve proteome limitation and reduce the importance of protein allocation tradeoffs.
• Tradeoffs that may dominate in one context may be weak or absent in another context. For example, limited free energy from available food sources means that each free energy consuming function trades off against all other free energy consuming functions.
• By contrast, free energy may be available in excess compared to an elemental limitation, such as nitrogen. Elemental limitation changes the particular tradeoffs that dominate design.
• In general, different elemental and resource limitations impose different costs and benefits for various cellular functions.^{66,67,101,104,249} Changes in costs and benefits alter the relative strength of tradeoffs between different functions.

15.2 Exploration versus Exploitation versus Regulation

Cells explore their environment to gain information and exploit their environment to gain resources. Cells use information obtained from exploration to regulate exploitation. Tradeoffs arise between exploration, exploitation, and regulation.

CELLULAR CONSTRAINTS

Genome size, cell size, and limited resources impose tradeoffs.

• More resources devoted to exploratory information-gathering imply fewer resources devoted to exploitation. For example, sensor abundance may trade off against enzyme abundance.
• More resources devoted to regulation imply fewer resources devoted to information and exploitation. For example, transcription factors may trade off against information and exploitation proteins.329
• Genome and cell size may trade off against exploration, exploitation, or regulation. In oceanic surface prokaryotes, small genomes correlate with nutrient-poor environments, slower growth, and fewer genes associated with adjusting to alternative conditions.⁴⁶³
• Small genomes may benefit in poor environments by more efficient maintenance and growth. Those benefits may trade off against a reduced proteome for information and regulation.

EXPLORATION VERSUS EXPLOITATION

Cells explore by moving through the environment and by using sensors of internal and external environmental states. Costly exploration reduces exploitation efficiency.

• Exploration by motility trades off against exploitation and growth efficiency (p. 229).^{164,299,395,402}
• In yeast, exploration by glucose sensors trades off against the exploitation efficiency of growth for already deployed glucose uptake receptors and catabolic pathways.⁴⁶⁵ Yeast have a broad range of sensors that regulate metabolism.⁶⁷

EXPLOITATION VERSUS REGULATION

Cells adjust by altering regulatory controls. Faster and more extensive regulation may reduce exploitation efficiency and growth rate.

• Fast mRNA decay speeds transitions to new cellular states by increasing message turnover. Fast mRNA decay also burns resources and degrades exploitation. Excess resources reduce the exploitation cost, favoring more rapid mRNA decay and regulatory adjustment.²⁸⁵
• Growth rate trades off against the rate of regulatory change when switching from one food source to another. Different regulatory mechanisms mediate this tradeoff (Section 14.5).^29,380
• Greater mRNA transcription rate imposes costs, degrading exploitation. Greater transcription rate also reduces stochastic fluctuations in protein production, improving regulatory precision. Observations across different genes and species follow this tradeoff between exploitation cost and regulatory precision.¹⁷⁴

DUAL FUNCTION

A membrane uptake receptor may also act as an information sensor.⁶⁷ In general, dual function reduces tradeoffs if a molecule can be tuned simultaneously for alternative actions.

Dual function does not contradict the central role of tradeoffs. Instead, the forces of design favor simultaneous improvement over tradeoffs, until further dual improvement cannot be achieved. At that point, improvement in one dimension often demands loss in another dimension. The tradeoff has returned (Fig. 3.1).

15.3 Thermodynamics and Biochemical Flux

Thermodynamic driving force and resistance to chemical transformation determine flux. The mechanisms that modulate driving force and resistance impose tradeoffs between components of success.

REGULATORY CONTROL TRADEOFFS

Various cellular mechanisms influence flux.100 Gain from modulating flux may trade off against loss in other components of success.

• Some glycolytic reactions have excess enzyme capacity and run near equilibrium.³¹⁰ Excess enzyme allows rapid increase in flux to meet the demand from increased resource flow. The benefit of flux adjustment trades off against the cost of excess enzyme.
• Raising enzyme concentration for a particular reaction and lowering concentration for another reaction may be a slow process. The efficiency gained by adjusting enzyme concentrations trades off against the slow response to changed conditions.
• Cells modify small parts of enzymes. Enzyme modification changes the resistance of reactions. Covalently modified enzymes provide cheap and fast flux control but may be less specific and efficient than modulating the concentrations of custom-designed enzymes.
• Thermodynamic driving force rises with greater reactant concentrations and smaller product concentrations. Stronger driving force increases flux but also dissipates free energy, lowering efficiency.

ADDITIONAL MECHANISMS THAT ALTER FORCE OR RESISTANCE

This subsection lists a few examples.

• Removing products to increase driving force may require costly processing or excretion of otherwise useful molecules.
• Separation of reactants modulates resistance. Biophysical changes to benefit one reaction may impose costly alteration of resistance for other processes.
• For example, lowering resistance by increasing membrane flux for a reactant may also increase the flux of toxins.
• Membranes impose resistance, maintaining chemical gradients and potential driving force.^35,86 Membranes also provide structural, sensor, and active transport functions. Altering a particular membrane property may influence the costs and benefits of several functions.

RATE VERSUS YIELD

Two commonly discussed tradeoffs arise.

• First, enhanced thermodynamic driving force increases the rate of a reaction. That increased rate associates with a reduced free energy yield that can do useful work.^47,317,444 This rate versus yield tradeoff arises from fundamental aspects of thermodynamics.
• Second, cellular growth rate often trades off against biomass yield. This rate versus yield tradeoff arises in observations.^257,326
• The two tradeoffs may be linked by assuming that cellular growth rate depends on biochemical reaction rates and that biomass yield depends on the usable free energy yield of reactions. Those assumptions often makes sense. But they are not guaranteed.

RATE VERSUS THERMODYNAMIC INHIBITION

When reaction products build up in concentration, thermodynamic driving force declines. In other words, reaction products thermodynamically inhibit flux.

• The benefits of increasing reaction rate may trade off against the costs of relieving product inhibition to maintain rapid flux. This tradeoff shapes overflow metabolism (Section 12.2).
• When driving force is low, small increases in product concentrations greatly reduce flux. Thus, low driving force’s gain in free energy efficiency trades off against greater sensitivity to product inhibition.
• Fast reactions incur costs to maintain high flux, whereas slow reactions incur costs to maintain positive flux.
• The total driving force over a metabolic cascade may be nearly constant. Greater driving force in one step trades off against lower driving force in another step. Reduced driving force of a step raises the risk that product inhibition blocks the entire cascade.
• In a cascade, a stronger final electron acceptor increases the total driving force. More steps reduce the force per step. Driving force and product inhibition impose tradeoffs when altering the final electron acceptor or the architecture of the cascade.

RATE VERSUS REDOX IMBALANCE

Cells maintain ratios of reducing and oxidizing compounds. Perturbed redox ratios interfere with regulation of biochemical processes.

• Rapid catabolic flux trades off against redox imbalance. In E. coli⁴²³ and yeast,^186,424 rapid flux creates an excess NADH–NAD +  redox imbalance. Lack of NAD +  limits the necessary electron acceptor to maintain flux through the TCA cycle.
• Electron transport takes electrons from NADH, yielding NAD + . Rapid upstream catabolic flux may produce NADH faster than electron transport can use it, creating an NADH–NAD +  imbalance (p. 167).
• Futile cycles dissipate the free energy in ATP–ADP and NADH–NAD +  disequilibria.³⁴⁸ Futile cycles suggest a tradeoff between some beneficial process that creates redox imbalance and the potentially costly dissipation of free energy to restore redox balance.

STRONG ELECTRON ACCEPTOR VERSUS OXIDATIVE DAMAGE

Oxygen is a strong electron acceptor that provides powerful thermodynamic driving force to catabolic cascades. Reactive oxygen species (ROS) also cause uncontrolled oxidative damage. Cells use various mechanisms to mitigate ROS damage.

• Strong electron acceptors trade off enhanced driving force against increased oxidative damage.⁴⁶⁷
• Growth rate trades off against ROS mitigation efficiency.¹²

15.4 Fitness Components and Life History

Life history analysis separates long-term success into fitness components. Components include reproduction, survival, and dispersal.393 In microbes, we typically equate reproduction and growth. By contrast, multicellular growth usually describes size increase from birth.

Life history emphasizes tradeoffs. Faster reproduction reduces survival. More dispersal lowers reproductive rate.

Fitness components may be divided into subcomponents by habitat, age, or other factors. Survival in food-limited habitats may trade off against reproduction in food-rich habitats. Dispersal out of old resource patches may trade off against reproduction in young resource patches.

Age-specific tradeoffs are important in life history. For example, reproduction early in life often trades off against reproduction late in life. In microbes, a lineage’s reproductive age is relatively younger at present and relatively older at a later time.

Lineages vary reproductively over time by accumulating damage or by intrinsic fluctuations, causing variation in age-specific fitness.³³⁰ Senescence occurs when age-specific fitness declines with age.^3,345

The following examples highlight tradeoffs between current reproductive rate and various fitness components. Typically, current reproductive rate associates with a lineage’s metabolic traits and growth rate.

GROWTH RATE VERSUS BIOMASS YIELD

A faster rate of cell division trades off against the biomass yield per unit of food intake. Prior sections described many examples of this tradeoff and possible underlying mechanisms.^{65,97,103,104,207,246,257,326,343,369,463}

GROWTH RATE VERSUS SURVIVAL

Greater current reproduction may trade off against lower survival. Reduced survival decreases future reproduction.

• Faster growing Schizosaccharomyces pombe yeast cells die at a faster rate.²⁹¹
• Among 16 bacteriophage species that attack E. coli, thinner capsid surfaces and more tightly packed genomes correlated with faster growth and lower survival.⁷⁹

GROWTH VERSUS MAINTENANCE AND STRESS RESISTANCE

Faster growth may trade off against maintenance and stress resistance. Such mechanistic links may explain the tradeoff between growth and survival.

• In E. coli, a few transcription factors act as master regulators of growth and maintenance. Stressed cells reduce growth and upregulate maintenance. Competition between these master regulators for access to RNA polymerase and transcription may mediate growth versus maintenance tradeoffs.^266,301
• RpoS is a master transcriptional regulator of E. coli’s stress response. Greater stress resistance associates with reduced catabolic processing of diverse carbon sources. Widespread polymorphism in the rpoS gene or its expression level occurs among strains, apparently tuning stress resistance versus growth to local conditions.¹⁰²
• Variable membrane permeability trades off nutrient uptake against susceptibility to oxidative stress.^102,177
• In S. cerevisiae, slower growth and higher biomass yield correlate with better stress resistance. That correlation occurs for different environments, nutrients, or mutations. Genomic knockout analysis reveals many mutations that mediate the tradeoff between growth and stress resistance.⁴⁶⁶
• Similarly, in Schizosaccharomyces pombe, a high-flux variant of the glycolytic enzyme Pyk1 associates with fermentation, fast growth, and decreased oxidative stress resistance. The low-flux variant causes a catabolic switch to respiration, slower growth, and greater stress resistance.²⁰²

GROWTH RATE VERSUS AGING RATE

Senescence occurs when survival or other fitness components decline with age. Lineage age can be measured as a particular amount of time or a particular number of cell divisions. The aging rate is the rate of decline in age-specific fitness.

• Aging in a lineage associates with accumulating cellular damage. In some species, damaged molecules segregate asymmetrically during cell division. One cell suffers a decline in age-specific fitness. The other cell maintains fitness or is rejuvenated.^{2,225,241,330}
• Greater stress-induced damage enhances asymmetry’s growth benefits.⁴²² If greater asymmetry imposes higher costs on other fitness components, then the increasing growth benefits of greater asymmetry trade off against those higher costs.
• In E. coli, stress resistance mediates a tradeoff between aging and growth. For example, increased expression of the general stress pathway reduces the aging rate and lowers the growth rate.⁴⁶¹

TRADEOFFS WITH DISPERSAL AND DORMANCY

Most resource patches eventually disappear. New ones open up. A lineage’s long-term success depends on colonizing new locations.

In a stable habitat, a lineage gains by dispersing a fraction of descendants to compete with others, reducing competition with itself.^111,169

Dispersal may occur over time rather than across space. For example, dormant spores reduce activity and protect themselves to survive the journey to a later time.²²⁹

The benefits of colonizing different locations or later times trade off against various costs.

• Dispersers typically pass through time or space nonreproductively.
• Dispersal often exposes microbes to mortality risk.
• Active movement requires resources.³⁴⁴ Dormant quiescence requires storage and maintenance. Those dispersal-related resources lower allocation to other functions.²²⁹
• Dispersal trades current reproduction for future reproduction in a different location. The reproductive value of growth in different patches depends on demographic processes. Reproductive value translates a unit of reproduction in different classes into a common valuation for long-term contribution to the population.^61,122,405
• Similarly, dormancy trades current reproductive opportunity for later opportunity. The reproductive value of current versus later reproduction depends on demography.
• Decaying conditions favor allocation to distant opportunities.

15.5 Warfare versus Productive Traits

Microbes attack competitors and defend against assault. Attack and defense trade off against productive growth, survival, and dispersal.

• Greater cell membrane permeability may enhance resource uptake and growth. That growth benefit trades off against greater susceptibility to antibiotics and other attack molecules that pass through the membrane.177,318
• Many resource uptake receptors provide the site of attack by bacteriophage. Increased resistance to attack trades off against reduced resource uptake.¹⁰⁴
• A significant fraction of microbial genomes may be devoted to attack and defense. Limitations on total coding, transcription, and proteome size impose tradeoffs between warfare and productive functions.^146,148,157

15.6 Cooperative Traits

Analyses of metabolic design typically focus on the success of a clonal lineage. How fast does the lineage grow? How efficiently does the lineage transform resources into biomass?

However, many studies identify microbial traits that reduce a lineage’s success while enhancing the success of other lineages.441 The tradeoff between growth rate and biomass yield provides a good example.³¹⁷

A lineage that reduces its growth rate leaves more resources for neighboring lineages, including itself. We must weigh a lineage’s cost for reduced growth against the yield benefit provided to neighboring lineages.

Ultimately, natural selection can favor traits that directly reduce the success of a lineage and indirectly increase the success of similar traits in other lineages. Hamilton’s¹⁶⁶ classic tradeoff compares a lineage’s direct cost, c, to the benefit for other lineages, b, weighted by the genetic similarity between lineages, r.

When a lineage is by itself, then all neighbors are the lineage itself, and r = 1. As neighbors become increasingly different from the focal lineage, r declines.

A cooperative trait is favored when the similarity-weighted benefit outweighs the direct cost, rb > c. For example, when considering slower growth, cooperation depends on how the similarity-weighted yield benefit trades off against the direct growth cost.

An alternative aspect of similarity can also be important.^122,405 Instead of measuring how a lineage’s cooperative trait benefits genetically similar neighbors, we can measure how a lineage’s cooperative neighbors provide benefit to the focal lineage (Section 5.2).

In this second case, a lineage that enhances group success has similarly cooperative neighbors in proportion to r. We once again get the condition rb > c for a cooperative trait to be favored. The distinction is that, in the first case, benefits flow from our focal cooperative lineage to genetically similar neighbors, whereas in the second case, benefits flow from phenotypically similar neighbors to our focal cooperative lineage.

The distinction is important because, in the second case, the cause of similarity between neighbors does not have to be genetic. For example, a lineage could have cooperative neighbors of a different species.¹¹³ Thus, similarity does not necessarily depend on kinship or pedigree, although common ancestry can be one cause of similarity.

Cooperative traits may also gain by payback through an ecological loop.⁷³ For example, a lineage may excrete a useful metabolic product, the loss of which directly reduces the lineage’s success.

That costly excretion may ultimately be favored by a synergistic feedback loop. Suppose an excreted product enhances the growth of another lineage that, in turn, excretes a product beneficial to the first lineage. In that case, the initial secreting lineage trades off an immediate cost for a later return benefit.

Cooperative tradeoffs have been widely discussed in the microbial literature, with many examples. The point here is that we must consider such tradeoffs when trying to understand the design of metabolic traits. In essence, sociality shapes biochemical and metabolic design.

SIMILARITY SELECTION

A lineage may trade the lost success from expressing cooperative traits for the benefits received from similar neighbors.

• Reduced growth rate imposes the cost of slower reproduction. Neighbors benefit from the resources left unused. In a yeast experiment, the relatively high similarity-weighted benefit from neighbors, rb, outweighed the cost of reduced growth, c, favoring a low growth and high yield strain.²⁵⁸
• Oil emulsions limit diffusion and keep beneficial unused resources close to a slow-growing cell. The smaller neighborhood also increases the similarity, r, between nearby strains. In laboratory evolution experiments, greater r favored slower growth and higher yield,²¹ with additional tradeoffs possibly playing a role.⁴⁰¹
• Microbes often secrete exoenzymes to break external food molecules into smaller pieces.^335,473 The secreting cell pays the cost of enzyme production. All nearby cells gain the benefit of extracellular digestion. A producing cell trades off the similarity-weighted digestion benefit, rb, against its production cost, c.
• In marine bacteria, extracellular enzymes break chitin into smaller pieces. Cooperative secretion of those enzymes decreases when diffusion of usable pieces exceeds uptake by nearby cells.⁹² Fast diffusion increases the spatial scale of benefits and reduces r, the similarity of neighbors that share benefits.

COOPERATIVE CROSS FEEDING

A lineage may trade its direct immediate success for a later return benefit through an ecological feedback loop.

• In an experiment, Salmonella enterica evolved costly secretion of methionine to enhance the growth of a methionine-dependent E. coli strain. The enhanced growth of E. coli increased its secretion of waste acetate, on which S. enterica depended for its growth.¹⁷⁰
• A follow-up experiment supplied lactose, which catabolically splits into glucose and galactose. Escherichia coli evolved to secrete galactose, boosting the growth of S. enterica and thus increasing the methionine supply on which that E. coli strain depends. This second step established mutually cooperative cross feeding.¹⁷¹

15.7 Timescale Tradeoffs

A beneficial trait on one evolutionary timescale may be a costly trait on another evolutionary timescale. Timescale tradeoffs pose one of the great challenges in the study of design.133

Consider a canonical problem of microbial life. A patch opens up with a fixed amount of resources. A colonizing microbe lands on the patch and begins to grow. A mutant arises that grows faster than its progenitor lineage but is less efficient at converting resources into biomass.

The mutant overgrows its competitors, dominating the population. Growth depletes local resources, ending the patch life cycle. Over the life cycle, some cells disperse and colonize new patches.

Over the short timescale of a cellular generation within the patch, direct competition favors rapid growth.²³⁶ Over the long timescale of the full life cycle, dispersal typically favors greater biomass production in a patch to increase the potential number of dispersing cells.

This tradeoff between growth rate and biomass yield expresses a design conflict between short and long timescales. Short-acting processes favor rapid growth to outcompete neighbors.²³⁶ Long-acting processes favor efficient yield to increase colonization of new patches. Different conditions cause different timescales to dominate (Section 5.9).^129,130

Various timescale tradeoffs arise in the study of design. A few examples follow.

SHORT-TERM VERSUS LONG-TERM GAINS

A rapidly growing mutant lineage that outcompetes neighbors is similar to cancerous overgrowth.¹³⁵ The short-term reproductive benefit favors highly competitive design.

Over longer timescales, a growth-enhanced cancerous design loses to yield-efficient designs tuned to increase dispersal success over the full demographic cycle of local growth and subsequent dispersal.

The particular design that is ultimately favored depends on the balance between short-term gains by local competitive growth versus long-term gains by dispersal.

The following list summarizes a few of the timescale tradeoffs.

• Short-term local gain trades off against long-term global gain.
• Rate versus yield compares gains on short versus long timescales.
• Short-term reproductive gain within a patch trades off against long-term dispersal gain between patches.
• Short-term competitive gain between genomic subsets within an individual trades off against the individual’s long-term success.

MULTILEVEL SELECTION

Natural selection can act at multiple levels. The different levels of selection associate with different timescales.^133,446

For example, growth rate affects relative success against neighbors within a locally competing group. Growth rate competition within groups happens over shorter timescales.

Biomass yield affects the productivity of a lineage or group when competing against other groups for dispersal and colonization of new resource patches. Biomass yield competition between groups happens over longer timescales.

Selection can also happen on a lower level, between genomic subsets within an individual. That low-level genomic competition happens on a shorter timescale than competition between individuals.

For example, plasmids may compete within cells over short timescales while lowering the success of host cells over relatively longer timescales.

At a higher level, selection can act more slowly between species or clades. A clade can outcompete other clades ecologically over broad spatial and temporal scales.

A clade can also reproduce in the sense of splitting into descendant lineages. On a relatively slow timescale, we can consider species births and deaths, determining the relative dominance of genera in terms of numbers of species.

The designs that we see in nature depend on the balance between selective success acting at these different timescales. In general, the lower levels and faster timescales dominate because selection almost always happens more quickly and intensely at lower levels.⁴⁴⁶

Selection at a lower, faster level may trade off against selection at a higher, slower level.¹⁶⁸

EVOLVABILITY: TRADEOFFS BETWEEN EXPLORATION AND EXPLOITATION

Organisms continually face new challenges. Long-term success demands the ability to adapt. Organisms gain by exploring alternative designs. Greater exploration speeds the rate of adaptation to new challenges and lowers the chance of lineage extinction.^311,436

Mechanistically, higher mutation rates and more mixing of genomes to create new gene combinations increase exploration of alternative designs.^267,308 Other variation-generating mechanisms may also speed adaptation.⁴⁴³

Mechanisms that generate variation and improve exploration often reduce efficiency in exploiting current resources. Long-term evolvability trades off against short-term fitness.

For example, most mutations are deleterious in the short term. Greater mutation rate reduces short-term exploitation success. However, as the environment changes, a higher mutation rate may allow faster adaptation to changing conditions by exploring a wider range of alternative designs.

15.8 Bet-Hedging Tradeoffs

Bet-hedging is another form of exploration versus exploitation. Bet-hedging describes alternative states for a trait.131,144,370

The different states may be alternatives for a single individual. For example, an individual may feed in different locations to hedge its bets against settling in a lower quality patch.

Or the states may be alternatives for different individuals from the same genetically identical clonal lineage. For example, some cells may stay and exploit a good patch, whereas other cells may disperse.

The clone’s overall success is the aggregate performance of its different bets on local versus distant gains. If the local patch disappears suddenly, the clone persists through its dispersers.

Bet-hedging may increase the success of a lineage by trading off the potential gains of exploration against the losses of reduced exploitation.

EXPLORATION ACROSS SPACE AND TIME

Exploration may occur spatially through motility and dispersal or temporally through persistence, dormancy, and sporulation.

• If an individual’s motility randomly samples alternative locations, then its bet-hedging movements trade off against motility costs.
• Cells in a clone may split into dispersers versus nondispersers, dormant versus active cells, and spores versus nonspores. These polymorphisms hedge a genotype’s bets across space and time.
• Cells within a clone may transiently enter and leave a persistent state that resists antibiotics.⁴²¹ For example, a temporary period of cellular reproductive quiescence protects against molecules that attack replicating DNA.
• Cells in a quiescent state trade a clone’s growth for a hedge against attack.
• During starvation, a clone may hedge between rapid resumption of growth and resistance to attack.²⁸² Cells in an active state resume growth rapidly. Cells in a quiescent state may survive antibiotic attack. A clone hedges by maintaining some cells in alternative states, achieving both rapid growth and resistance.
• Traits such as quiescence or dispersal may arise inevitably from biophysics rather than from the biological forces of design. Interpreting design depends on tests of comparative hypotheses.

EXPLORATION OF ALTERNATIVE FOOD SOURCES

In the classic diauxic shift, cells first feed on a preferred carbon source until it is depleted. A lag period follows during which cells shift gene expression to prepare the catabolic system for feeding on a second, less preferred carbon source (p. 214).

• After the preferred source is depleted, some cells may switch to the secondary source. Other cells may remain activated for the primary source. The primary-activated cells do not grow in the absence of the primary source but remain ready to resume growth rapidly if more of the primary source arrives.
• Similarly, some cells may switch to the secondary source before the first source is depleted. Having some cells pre-activated for the secondary source reduces the clone’s growth lag when the first source runs out.
• In each case, a clone adjusts more rapidly to changing availability of the two resources when it hedges between the alternative states. Hedging trades lower immediate growth for more rapid adjustment to changing conditions.

15.9 Control Tradeoffs

Control adjusts traits to internal and external conditions. Linking metabolic control to demography and life history poses a challenge for future work. Here, I briefly repeat a few key tradeoffs developed in Chapter 7.

• Error-correcting feedback trades off against the costs of control. Error is the difference between a trait’s value and the target value for the current environment. Costs arise from measuring the error and adjusting traits.
• Robust error correction at the system level trades off against error-prone system components.
• Fast adjustment trades off against stability. Fast adjustment requires a strong push, which risks overshooting the target. Overshoot destabilizes a system, potentially causing it to move dangerously far from its viable range.
• Responsiveness trades off against homeostasis. The more easily a system can respond to change, the more easily it can be perturbed from its homeostatic setpoint.

15.10 Summary

This chapter emphasized broad conceptual issues and common tradeoffs. That general scope provides a sense of the potential challenges of design faced by organisms.

In practice, tradeoffs can be very specific. Consider the title: “Evolutionary trade-off between vocal tract and testes dimensions in howler monkeys.”91

For microbial metabolism, nearly every biochemical and physiological detail could be associated with some tradeoff. In spite of that specificity, it is important to retain a broad sense of the main forces and tradeoffs that shape all of life.

Combining a wide conceptual approach with the specific understanding of natural history in each particular application emphasizes the many biological challenges and the many different tradeoffs faced by organisms.

The shifting dominance by different tradeoffs as conditions change raises a common difficulty in the study of design. As an example, I mentioned previously the complexity of analyzing tradeoffs between rate, yield, and antitoxin defense.

In a lab study that excludes interspecies toxin warfare, one may observe both rate and yield increasing. That joint increase may arise because rate trades off against antitoxin defense, and yield also trades off against defense. In the absence of attack, investment in defense declines, and both rate and yield increase.

One cannot test whether a tradeoff between two traits is important simply by measuring how the two traits change. Instead, one must develop comparative hypotheses about partial causation. For example, how does preventing attack alter the rate versus yield tradeoff in shaping design?

Comparative hypotheses do not solve all problems. But if observations tend to support a comparative prediction under a variety of conditions, then one may be on the right track. Further progress follows by adding additional partial causes to improve the success rate for predictions.

Finally, there remains a gap between the biochemical detail of prior chapters and the emphasis on fitness components in this chapter. Bridging that gap remains a central challenge in the study of design. The following chapters make a start.