A longer evolutionary history restricts novel adaptation in animals

Ariel E. Marcy Department of Biology, Stanford University

Five hundred and twenty million years ago, the Cambrian explosion produced the largest diversity

in body plans the world has ever known. About 100 different phyla emerged by the end of this era

including all 30 extant today (Rozhnov 2010). Since then, the number of phyla has dwindled despite

an overall increase in lower taxonomic ranks, such as families (Knoll 2006). Theories for the

explosion include greater ecological opportunity during the Cambrian, but this does not explain why

no new phyla appear after the Permian when other opportunities were granted.

The fossil record suggests the animal ability to generate novel adaptations, such as those that

distinguish higher taxonomic ranks, peaks early and decreases with time and subsequent

specialization (e.g. Erwin et al. 1987). Evolutionary theory predicts the pattern of biodiversity seen

in the fossil record by modeling how early organisms generate adaptations (Kauffman 1989). When

primitive, less fit animals readily mutate their early ontogeny, the resulting morphological

characters drive divergence of high level taxonomic groups (Kauffman 1989). A recent synthesis of

developmental genetics proposes gene regulatory networks as evolutionary significant generators of

morphological change (Davidson and Erwin 2010). Mirroring the hierarchical organization of taxa,

the hierarchical design of these DNA regions allows rapid evolution at first, but each viable

modification restricts the scope of subsequent developmental changes (Davidson and Erwin 2010).

Thus as animal life evolves over time, their ability to generate viable novel adaptations decreases,

producing higher taxonomic groups first and then filling in with families, genera, and species.

The Cambrian Explosion: interpretations of the fossil record

The rapid Cambrian proliferation of high taxonomic groups intertwines with early metazoan

animals at their most unconstrained. As such, the Cambrian explosion is central in understanding

why animal diversity fills in top-down. This section will discuss whether this event actually is an

“explosion,” the distinction between phyla and body plan, and the pivotal role of mutation in

radiations of high taxonomic groups.

An explosion characterized by rapid diversification of the highest taxonomic groups is

persuasive evidence for decreased evolutionary constraint earlier in multicellular animal evolution

because it requires higher order differences to appear first. Proponents of the long slow

accumulation origin of phyla, however, have suggested that the explosion is an artifact of geological

preservation (e.g. Cohen 2005). Most of our knowledge comes from a few well-preserved

Lagerstätten, which could give the impression of sudden phyla while small accumulations of

changes over millions of years went unpreserved. This taphonomic argument is rebuffed by the lack

of large metazoan animals in slightly earlier sites with comparable geological conditions (Knoll

2006). The diversity of well-preserved protist “acritarch” assemblages increased around the

Cambrian (Knoll 2006). As a reasonable marker for overall marine diversity, this supports a sudden

increase in animals (Knoll 2006).

The fossil record thus brackets the duration of the explosion – the divergence of major

taxonomic groups - to ~20 million years (Fig. 1) (Conway Morris 2000). While some suggest the

half a billion years granted by the Phanerozoic Eon would not be enough for the most primitive

phyla to gradually accumulate random changes to evolve into the most complex (Rozhnov 2010).

Furthermore, within extant phyla, defining body plan characters are remarkably stable rendering

them readily identifiable over 500 million years of fossil deposits (Rozhnov 2010). Thus the fossil

record demonstrates relative unconstraint before phyla diverge but not after.

Several hundred million years after the Cambrian explosion, the Permian extinction

eradicated 96% of species producing another large radiation in its wake (Kauffman 1989). In marine

organisms, the new genera and families proliferated at about the same rate as in the Cambrian, but

no phyla or classes emerged (Erwin et al. 1987, Rozhnov 2010). Kauffman summarized the

difference with, “The Cambrian filled in from the top down; the Permian witnessed filling in of

higher taxa from the bottom up” (Fig. 2). The entire Phanerozoic shows a decreasing trend in the

appearance of maximum taxonomic rank in radiations following major extinctions (Rozhnov 2010).

Since taxonomic rank roughly corresponds to degree of divergence, this result in comparison to the

Cambrian supports increasing constraints on animal morphology inherent in a longer evolutionary

history.

The most serious contradictory evidence comes from molecular data, which dates animal

divergences beginning 100 my earlier than the Cambrian explosion (Conway Morris 2000b). In an

effort to resolve the phylochronology of metazoans, Yang et al. analyzed molecular data in

arthropods, which they argue will be representative of other animal phyla (Fig. 1). Their molecular

evidence suggests the major metazoan phyla diverged in a blistering 15my time span in the great

Metazoan Radiation Event (MRE) (Yang et al. 2007). The MRE took place in the late Vendian about

100my earlier than the Cambrian explosion. This study pushes the appearance of arthropod phyla

away from the Cambrian explosion and back to the MRE, with sub-phyla divergences in this clade

occurring during the explosion (Yang et al. 2007). These data further corroborate the evolutionary

trend of highest taxonomic ranks radiating first followed by the next tier of taxonomic rank.

!Figure 1. The Precambrian context and other events leading up to the explosion: Metazoan chronology comparing early fossil record, molecular dates and palaeoenvironmental background. (A) Molecular dating results (B) Early metazoan fossils and the geological context in South China. (1) Strongly negative d13C anomalies (2) approximate position of arthropod-like fossils found in the Ediacara biota–Spriggina (3) interval of first fossil bilaterians found in South China (3’) molecular estimate for the origin of bilaterian crown groups (4) approximate position of the first large, spinose acritarchs with high diversity found in South China. Radiometric dates marked with stars. Grey boxes: time intervals of metazoan radiation recovered by molecular data (A) and the fossil record (B), representing the great Metazoan Radiation Event (MRE) and the Cambrian Explosion (CE), respectively (Yang et al. 2007).

But, molecular evidence cannot provide insights into the anatomical changes of organisms

(Conway Morris 2000b). Similarly, we must be careful not to equate the molecular divergence of

phyla with the instantaneous acquisition of all the body plan characters defining (Budd and Jensen

2000). Budd and Jenson’s phylogeny (Fig. 3) disentangles the concept of “phyla” from “body plan.”

A phyla represents the monophyletic clade arising after divergence from a sister phyla and a body

plan the set of characters defined by the crown group. As the cladogram shows, a body plan will be

a combination of primitive traits shared with the sister phylum and derived traits acquired after their

divergence.

!

Figure 3. The cladistics of phyla and body plans (Conway Morris 2000a)

This framework allows the unfamiliar Vendian “phyla” be interpreted as a stem-group that

possess some, but not all, of the traits that define Cambrian phyla (Conway Morris 2000b, Rozhnov

2010). Molecular evidence suggests the MRE witnessed rapid divergence of clades of phyla over a

span of 15 million years (Yang et al. 2007), however, recognizable body plans, resolve afterwards

with the input of derived traits (Conway Morris 2000b). Stem echinoderms, for example, share

bilateral symmetry with their sister phyla, the chordates (Rozhnov 2010). The characteristic 5-sided

radial symmetry of crown echinoderms associated with a reorganization event of Hox genes and

represents a derived trait acquired some time after the echinoderm-chordate divergence (Rozhnov

2010). As Conway Morris stated his synthesis, the cladistic understanding of phyla and body plan

will offer evolutionary insights into the roles of pre-adaptation, morphological constraint, and co-

options of gene function (Conway Morris 2000a). How these concepts relate to decreasing

evolutionary novelty will be discussed below and in the two remaining sections, respectively.

To explain why phyla diverge and body plans assemble during the Cambrian, some

hypotheses cite ecological opportunities as the main drivers and with good reason – important

events in animal evolution very often correspond to climatic signals (see Fig. 1). Animal

multicellularity occurred in a time of increased oxygen (Yang et al. 2007); the great Metazoan

Radiation Event and Cambrian explosion followed the extirpation of biomass by two ice ages and a

transient shallow water anoxia episode (Knoll 2006). Clearly, significant environmental changes

elicit correspondingly dramatic responses in organisms, but ecological upheavals have preceded

every radiation since the Cambrian without reproducing it (Kauffman 1989).

!

Figure 2. Taxa fills in top down: The evolution of biological diversity in the Phanerozoic: the upper portion shows the change in the number of genera of marine animals in the Phanerozoic (Sepkoski, 1995), the lower portion shows the maximum rank of new taxa, appearing at the major boundaries in the Phanerozoic (Rozhnov, 2001). Numbers show the geological age in Ma dating back from the present.

Evolution implies an accumulation of traits, which some have extended to the acquisition of

traits such as eyes or burrowing behavior to enter new ecological roles, like macroscopic predation

or filter feeding on the seafloor (e.g. Conway Morris 2000). Conway Morris posits that the

subsequent trophic interactions will cause cascading radiations, evidenced by sudden abundance of

defensive hard shells and bustling benthic communities (Conway Morris 2000). Cohen criticizes

this theory because it assumes these traits already exist in the populations of these organisms, pre-

primed for predatory selective pressures (Cohen 2005). Furthermore, the ancestors of the

mineralized metazoan lineages were all free-swimming pelagic organisms, if faced with sudden

predation, active evasion or camouflage would be more likely responses than mineralization (Cohen

2005).

Instead, Cohen posits mineralization as a pre-adaptation for defense, with an original

advantage for settling in the nutrient-rich, stable, yet depauperate benthic ecosystems following the

shallow water anoxic event (Cohen 2005, Knoll et al. 2006). Knoll predicted this event “would

create permissive ecologies in which novel (and not necessarily fitter) mutants generated by

surviving populations could persist and expand under low selective pressure, providing raw

materials for evolutionary innovation” (Knoll et al. 2006). Cohen’s data supports Knoll’s theory as

benthic mineralization was derived from shared metazoan biochemical pathways in at least seven

different and unrelated ways by as many pelagic lineages (Cohen 2005). All three scientific groups

posited subsequent trophic or community interactions with pre-adapted mineralized benthic clades

produced the sudden appearance of high benthic diversity.

While Cambrian researchers disagree on the details of what caused the explosion, as

demonstrated above, theories of ecological opportunity and trophic interaction complement each

other. I argue, however, that the importance of pre-adaptations in producing biodiversity is

understated. In the scenario above, biochemical pathway mutations allowed multiple lineages to

adapt in a low selective pressure environment, accommodating seven different colonizations and

subsequent niche broadening of each clade (Cohen 2005). The mutations then provide fodder for

pre-adaptations, which in this case fundamentally changed the way these organisms interacted and

co-evolved with the community assemblage (Conway Morris 2000). Thus mutations producing

niche-broadening pre-adaptations represent the rate-limiting step in a process determining which

morphological novelties persist. Therefore the amount of morphological constraint organisms

inherit from their evolutionary history has a direct, and I argue, the most salient effect on the rank of

taxonomic groups appearing after extinction events. This hypothesis can be tested from a theoretical

point of view.

Morphological constraint and theoretical evolutionary frameworks

Evolutionary theorists have proposed frameworks to model the evolution of organisms

through time, focusing in particular on the fate of mutants. Building on Wright’s adaptive landscape

framework (Fig. 4), they predict how early multicellular animals would explore the morphospace.

This section will outline the theoretical framework proposed to explain the pattern of diversification

around the Cambrian. As seen in the fossil record, the theories predict an increasing constraint with

time in animal abilities to produce novel adaptations and a decrease in the rate at which viable

mutations are introduced.

Wright’s adaptive landscape visualizes the relationship between phenotype and fitness (Fig.

4). The shape of the landscape is determined by the epistatic interactions among genes, which create

local optima for several phenotypes in the context of a niche (Kauffman 1989). A population of

organisms is represented as point on the landscape. Populations located in valleys have relatively

low fitness but also a greater number of viable mutations available and possible adaptive “walks”

across the landscape. When a population is not on a local optimum, directional selection acts on the

populations, such that it “hill-climbs” in small steps until it maximizes fitness by conquering a local

optima.

!

Kauffman extends this idea in his concept of “rugged” adaptive landscapes, which have a

multitude of peaks near and far away from any given point (Kauffman 1989). If the earliest

multicellular animals begin in a valley, he hypothesized two ways to navigate the landscape via

mutations (Kauffman 1989). The first is through the traditional gradual climbs, in which small point

mutations in protein-coding DNA create small substitutions in amino acid chains (Kauffman 1989).

This slow route of evolutionary change does not modify mutants much from original population –

i.e. move them far from their original adaptive valley (Kauffman 1989). Furthermore, mathematical

models predict a doubling in time to the next improvement uphill, further decreasing the rate of

evolution (Kauffman 1989). As the fossil record demonstrates, slow accumulations of change

cannot explain radiation events like the Cambrian explosion.

The second option, however, moves organisms across the landscape in long-jumps, via

mutations affecting early ontogeny (Kauffman 1989). The resulting novel mutants land in an

Figure 4. The adaptive landscape

unoccupied niche not adjacent to the original population in valley – called “hopeful monsters” their

likelihood of increasing fitness is at first, much higher than their hill-climbing relatives (Kauffman

1989). Within one generation, the hopeful monster may become reproductively isolated from its

parental population in this fast, yet risky venture of modifying early developmental features

(Kauffman 1989).

When Kauffman’s framework is applied in time, it predicts three rough time scales for a

theoretical metazoan radiation (Kauffman 1989), which I argue is supported by fossil and molecular

evidence (Fig. 5).

1) Fitter long jump mutants out-compete less fit hill-climbing relatives. In this stage, long-

jumping hopeful monsters find success much quicker than the ponderous hill-climbers.

Furthermore, Kauffman suggests the long-jump process can be iterative, allowing the organisms to

skip across the rugged landscape, populating a wide range of peaks (Kauffman 1989). These

conditions, where novel changes occur rapidly and phyla are founded (Kauffman 1989), correspond

to the emergence of phyla in MRE discussed in Yang et al. 2007. I suggest iterative long-jumping

underpins the successive input of derived characters accumulating into the cladistic definition of a

body plan. For example, the derived reorganization event of Hox genes conferring 5-sided radial

symmetry to echinoderms (Rozhnov 2010) fits the long-jump mutation requirement of affecting

early development.

2) The rate of finding fitter long-jump mutants decreases rapidly and eventually becomes

slower than hill-climbing mutants. Now nearby local optima are conquered and classes and orders

fill in around phyla (Kauffman 1989). These correspond to the taxonomic ranks that appeared in

arthropods in the Cambrian explosion (Yang et al 2007). The seven convergent mineralization

pathways provide examples of mutations effecting later stages of ontogeny (Cohen 2005) and thus

more local hill-climbing. Their most recent common ancestor would likely include the hopeful

monster that acquired the shared biochemical pathways (Cohen 2005).

3) Local optima are reached, and biodiversity reaches a stasis. Long-jumps are not favored,

so higher level taxa are not formed during these events (Kauffman 1989). Lower taxa can diverge as

slow environmental changes shift local optima, prompting small short bursts of mole-hill-climbing.

The signature of this in the fossil record is not as clear. The Paleozoic following the Ordovician

radiation is known for an unprecedented stasis in biodiversity but major and rapid climatic events

also occur during this time (L. Hadly, pers. comm.).

!

Overall the fossil record can be interpreted as following at least the first two stages of

Kauffman’s theory, which predicts the taxonomic rank of metazoans in the MRE and the Cambrian

explosion. The theory accounts for the high level of taxonomic rank diversification unique to the

Figure 5. Navigating the rugged adaptive landscape of the fossil record

Histogram displays of the distributions of times of appearance of animal phyla, classes, and orders in the marine fossil record. The width of each histogram bar is 10 MY. Solid bars indicate taxa with durably skeletonized representatives and open bars indicate taxa with exclusively soft-bodied or lightly sclerotized members. Note that all distributions are skewed with modes over the early Paleozoic (Erwin et al. 1987).

Numbers added by author to indicate Kauffman’s rough time scales with trends in taxonomic rank appearances/ stasis.

1) MRE

2) CE

3) stasis?

Precambrian/Cambrian radiation by reasonably assuming the low adaptive fitness of emerging

metazoans. In 1989, Kauffman positioned his theory as sufficient stand-alone alternative to

ecological opportunity and increasing robustness of developmental gene in explain the top-down

filling in pattern of diversity. I have already addressed how this theory may complement and inform

mechanisms invoking ecological opportunity. Recent work in evolutionary development suggests

the same genetic process underlies long-jumps and developmental robustness (Davidson and Erwin

2010).

A genetic mechanism

Canalization refers to the reduction in sensitivity of a phenotype to its underlying factors that

determine its expression, resulting in robustness of the phenotype despite mutation of those factors

(Loewe 2009). While Kauffman insists navigating the rugged adaptive terrain is sufficient to

explain taxonomic patterns in the Paleozoic, he also seems to imply the mutations responsible for

long-jumps also promote canalization. For example, “Early stages would lock in because they adapt

on more rugged landscapes than late stages” (Kauffman 1989).

A recent paper on gene regulatory networks suggests that this interpretation could be correct

for mutations affecting gene regulatory networks (GNR) (Davidson and Erwin 2010). They note

that gain-of-function mutations are overwhelmingly dominant and in GNRs would produce rapid

evolutionary change, even in large populations (Davidson and Erwin 2010). Furthermore, the

hierarchical design of gene regulatory networks conserves upstream feedback loops that force later

mutations downstream, such that these subsequent mutations have smaller effects on developmental

outcomes (Davidson and Erwin 2010). The diversity of these genes across animal lineages suggests

metazoans acquired these sophisticated patterning genes early in their history (Erwin 2009). At the

very least, this does not preclude mutations of GNRs as the genetic mechanism producing

morphological constraints and the hierarchical taxonomic patterns of the MRE, Cambrian

explosion, and beyond. The hierarchical gene regulatory networks certainly suggests an elegant

mechanism for the top-down appearance of hierarchical taxonomical rankings (Erwin 2009). More

work, however, is needed to determine the relative contribution of such genes to evolutionary trends

(Erwin 2009).

Conclusion and future directions

Fossil and molecular evidence, as well as adaptive landscape theories predict fast, high taxonomic

level changes happening earlier in animal evolution. This is a direct result of increasing

morphological constraint inherent in adaptation and at the genetic and developmental level. Fossil

evidence and evolutionary theory support the cladistic understanding of phyla as well as emphasize

the significance of pre-adaptation, morphological constraint, and co-options of gene function to the

metazoan trajectory. The intersection of morphology, community interactions and mutation is

another promising avenue of inquiry, and one with the potential to unite opposing views in

Cambrian literature.

A new field called Evolutionary Systems Biology (EvoSysBio) will investigate mechanisms

of evolution through modeling seven levels organismal biology of adaptive landscape – from

genetics to ecological interactions (Loewe 2009). As demonstrated in the preceding discussion

inputs on taxonomic diversity depends on a number of inputs that span these levels. Given

Kauffman’s elucidating 1989 model on the Precambrian/Cambrian radiation, I foresee sophisticated

adaptive landscape models resolving more finely the dynamics of evolution and diversity during

such events in deep time. The relative stasis of biodiversity in the Paleozoic despite environmental

upheaval comes to mind. Surprisingly, this review had a relatively short-sighted view of

evolutionary history. If I may a suggestion to the direction of this budding field, train these

theoretical models on unusual events in the fossil record, but when comparing across time consider

evolutionary history a salient character of the organisms.

Bibliography

Budd, G. E., and S. Jensen. 2000. A critical reappraisal of the fossil record of the bilaterian phyla.

Biological Reviews 75:253-295.

Cohen, B. L. 2005. Not armour, but biomechanics, ecological opportunity and increased fecundity

as keys of the origin and expansion of the mineralized benthic metazoan fauna. Biological

Journal of the Linnean Society 85:483-490.

Conway Morris, S. 2000a. Nipping the Cambrian "explosion" in the bud? Bioessays 22:1053-1056.

Conway Morris, S. 2000b. The Cambrian "explosion": slow-fuse or megatonnage? Proceedings of

the National Academy of Sciences of the United States of America 97:4426-4429.

Davidson, E. H., and D. H. Erwin. 2010. Evolutionary Innovation and Stability in Animal Gene

Networks. Journal of Experimental Zoology Part B-Molecular and Developmental Evolution

314B:182-186.

Erwin, D. H. 2009. Early origin of the bilaterian developmental toolkit. Philosophical Transactions

of the Royal Society B-Biological Sciences 364:2253-2261.

Erwin, D. H., J. W. Valentine, and J. J. Sepkoski. 1987. A comparative-study of diversification

events – the early Paleozoic versus the Mesozoic. Evolution 41:1177-1186.

Kauffman, S. A. 1989. Cambrian explosion and Permian quiescence – implications of rugged fitness

landscapes. Evolutionary Ecology 3:274-281.

Knoll, A. H., E. J. Javaux, D. Hewitt, and P. Cohen. 2006. Eukaryotic organisms in Proterozoic

oceans. Philosophical Transactions of the Royal Society B-Biological Sciences

361:1023-1038.

Loewe, L. 2009. A framework for evolutionary systems biology. Bmc Systems Biology 3:34.

Rozhnov, S. V. 2010. From Vendian to Cambrian: the beginning of morphological disparity of

modern metazoan phyla. Russian Journal of Developmental Biology 41:357-368.

Yang, Q., J. Y. Ma, X. Y. Sun, and P. Y. Cong. 2007. Phylochronology of early metazoans: combined

evidence from molecular and fossil data. Geological Journal 42:281-295.