Once we were worms

New Scientist vol 179 issue 2406 - 02 August 2003, page 34

What could we humans possibly have in common with sea squirts, bag-like creatures that attach themselves to the seafloor and feed by filtering seawater? Simon Conway Morris goes in search of missing links

"WE are taking you to a seafood restaurant," said my friend. "It is one of the best in Marseille." Splendid, I thought, anticipating the feast. As the table filled with dishes I realised this wasn't just going to be a treat for the gastronome in me but also for the zoologist. Alongside the familiar oysters, snails, clams, crabs and fish there were some rather more exotic delicacies. "Try these," gestured my companion. Large spiny balls had been cracked open to reveal thin orange strips plastered to the inside. They might not be to everyone's taste, but in Marseille the bright orange gonads of sea urchins are very popular. "Terrific!" I murmured. "Better than the ones I had in Japan." My friend was obviously impressed. "What about these? We call them violets-de-mer - bit of an acquired taste." I forked what looked like yellow-grey sludge out of a bag of leather and swallowed. "First rate," I gasped, reaching quickly for the wine. I had eaten my first sea squirt.

A gastronomic excursion, no doubt, but our meal that evening also posed something of a biological conundrum. On the face of it, sea squirts and sea urchins have little in common with the fish on the table - let alone with the large-brained primates who were eating them. Yet, despite their remarkable variety of forms, all the creatures on and around the table belong to a group called the deuterostomes. So what did our common ancestor look like? And how on earth did its descendants come to take the series of dramatic evolutionary leaps that have allowed them to explore some of the most peculiar recesses in biology, from animals that have visited the moon to others that have thrown away their brains?

Until recently these questions remained largely unanswered, but the evolutionary history of deuterostomes is now one of the hottest topics in biology. Working at the amazing Chengjiang fossil deposits in south-western China, my Chinese colleagues and I are starting to crack the code of deuterostome evolution. We have found what we believe are some of the earliest members of the group. These bizarre creatures, dating from the evolutionary big bang known as the Cambrian explosion more than half a billion years ago, could help us understand our real origins and trace the evolutionary path from humans back to a time when we were little more than worms.

Along with the deuterostomes the other most complex animals are the protostomes: literally "first mouths". This super-assemblage includes groups as varied as the insects (arthropods), earthworms (annelids) and lamp-shells (brachiopods). It is unclear whether protostome preceded deuterostome in evolution but the crucial distinction between them arises during early embryology. In both, the fertilised egg divides, forming a tiny ball of cells. Then one end of the embryo folds inwards in a process known as gastrulation, to give a hollow sphere with an opening called a blastopore. In protostomes the blastopore usually forms a mouth. In deuterostomes it forms an anus and another opening must be created for the mouth - hence deuterostome, or "second mouth".

It is more than a century since biologists first recognised that all higher animals could be divided into two groups in this way. Some of the evolutionary connections are obvious. For example, our own fishy ancestry is clearly imprinted on our bodies. Not only do we have a backbone made of vertebrae, we also have the makings of gills that appear early in our embryology as shallow depressions in the neck, only to disappear shortly afterwards. Less obvious is the connection between vertebrates and sea squirts - one of a group of animals known as tunicates. At first glance we have nothing in common with these bag-like animals that attach themselves to the seafloor and feed by filtering seawater. But take a closer look, and you see large numbers of gill slits, which work with two funnels to direct the flow of water and filter out plankton. What's more, before the adult tunicate settles on the seafloor, its larvae take the form of tiny tadpoles, which are more interesting from an evolutionary point of view. Although they are simple animals their distinct head and tail make them obviously fish-like. Hardly surprising then that the tunicate larva has long been among the candidates for explaining the origin of the vertebrates.

Gill slits, fish-like larvae and other less obvious but nonetheless telltale signs leave no doubt that tunicates are closely related to vertebrates. But they also have some fascinating peculiarities. Perhaps the best known is that the outer wall, or tunic, is rich in the carbohydrate cellulose. This is the principal structural biopolymer of plants and, via wood pulp, of the New Scientist you are holding. Strangely, despite its versatility, cellulose has never been exploited by animals - with the conspicuous exception of the tunicates. Did they make this innovation independently or, as now seems more likely, did tunicates pick up a cellulose gene from a more primitive organism sometime in their evolutionary past? Another oddity, revealed by a recently published genome study of the tunicate Ciona, is that it uses the copper-based haemocyanin as a respiratory protein (Science, vol 298, p 2157). Protostomes have independently evolved the use of haemocyanin at least twice, but it has never been found in a deuterostome before. Again, did tunicates evolve the protein independently or did they smuggle it across the barrier with protostomes by gene transfer?

Tunicates are odd members of the deuterostome clan, but top marks for strangeness must surely go to the echinoderms, which include sea urchins and starfish. The most glaring peculiarity is the fivefold (or pentaradial) body plan. Why they adopted this is a complete mystery. What we do know is that the ancestral echinoderm had a body with bilateral symmetry - a plane of mirror symmetry running from front to back and defining left and right-hand sides - the pattern retained by virtually all other animals. Also, bizarrely, although this ancestor almost certainly had an advanced nervous system with a brain and nerve cord, echinoderms threw this away and replaced it with a diffuse net of nervous tissue. And echinoderms have evolved a porous calcite skeleton - yet another unique construction. Molecular biology confirms the strange nature of these beasts, revealing that there has been a radical redeployment of the developmental genes. Echinoderms could almost come from another planet. But the fossil record does contain one intriguing piece of evidence linking them with other deuterostomes: at least one group of primitive echinoderms has gill slits.

Meet the ancestors

Echinoderms, tunicates, fish, humans - it is hard to spot any sort of evolutionary coherence. And that's even before you add in the other member of this unlikely roll-call, a group known as the hemichordates, especially the so-called acorn worms. Their membership of the gang is secured by gill slits and larvae similar to those of some echinoderms. But in almost every other way acorn worms are different again from their fellow deuterostomes, comprising three sections that include a swollen acorn-like proboscis, a central part with gill slits and an elongated posterior. How do we solve this evolutionary puzzle? Molecular biology can offer profound insights, both in terms of the development of the embryo and discerning deep evolutionary relationships. But this approach reveals nothing about the appearance of the ancestral forms, how they functioned, or the sequence in which evolutionary innovations appeared. The only way to get this sort of information is from the fossil record. It is notoriously imperfect, riddled with holes and open to interpretation, but if we want to find out what primitive deuterostomes really looked like and how body plans were remoulded again and again to give the diverse forms we see today, then we must blow the dust off the book of the fossil record.

That is exactly what we have been doing in China for the past decade. I have been working with a group based at Northwest University in Xian, led by my friend Degan Shu. The evidence comes from deposits quite similar to the famous Burgess Shale in British Columbia, Canada, but somewhat older. These are the amazing Chengjiang deposits, dated at about 530 million years old, and appearing in outcrops across a 100-kilometre-wide area around the city of Kunming. Today it is one of the most attractive parts of China, fertile with lakes and hills that roll towards the border with Vietnam. But 530 million years ago it was a shallow sea, dotted with islands, and close to the equator. Tropical storms were frequent, and when the seabed was stirred up millions of animals were entombed and transformed into exquisitely preserved fossils that provide a snapshot of an ancient world. We collect the fossils with the help of teams of people from nearby villages who, despite their lack of academic training, have a keen eye for rarities. The scientific work, of careful preparation and detailed study has to wait until the fossils are unpacked in Xian. Every year's collecting produces surprises, and every time I arrive in Xian I'm not disappointed.

In May 2001 the fossil record really came up trumps. We seemed to have found the key to the ground floor of deuterostome evolution in the form of a group of peculiar-looking fossils known as vetulicolians. These are abundant in the Chengjiang deposit, but also seem to have been widespread in the Cambrian world. When first discovered in 1987 by Xianguang Hou, who is now at the University of Kunming, vetulicolians were classified as arthropods, the phylum that includes crabs and flies. Like some of the more primitive arthropods, their body consists of a large anterior with a shield-like carapace and a segmented tail. But, as we found more fossils, it became apparent that the arthropod resemblance was superficial. In fact, the vetulicolians found at Chengjiang take several forms but all share two features which turn out to be central to the deuterostome story. First, the body is composed of two distinct sections and secondly they have a row of perforations that look like rudimentary gill slits.

The vetulicolian bipartite body consists of a posterior that acted as a propulsive tail and an anterior with a gaping mouth and voluminous interior cavity with the line of perforations on either side. In the advanced species these perforations have a complex structure consisting of large pouches with access to both the animal's interior and the outside world, via another opening. This suggests the vetulicolians were filter-feeders swallowing large quantities of seawater to capture tiny food particles, and expelling excess water through the perforations. Only the deuterostomes have similar perforations in the anterior body. So, are we seeing evidence of one of the first stages in the evolution of the gill slits? It is possible, but even if the gill slits evolved independently, there remains the question of the distinctive bipartite body. Many years ago, the famous American palaeontologist Alfred Sherwood Romer argued that the ancestor of the fish must be a composite animal, constructed of two sections: a large anterior with gill slits, and an elongated and segmented tail. Suspiciously like a vetulicolian, in fact.

From here deuterostome genealogy takes several different pathways, including those towards the echinoderms and the ancestors of vertebrates. There are still plenty of gaps and not every step is known but the clues are beginning to point to a consistent story. Consider first the echinoderms. Among a medley of early forms is a very odd group, sometimes referred to as the calcichordates. The calcareous skeleton is exactly like that found in all other echinoderms, but the overall arrangement of the body is bipartite and curiously reminiscent of the vetulicolians. Perhaps the first and most primitive echinoderms were relatives of the vetulicolians, acquiring the calcareous skeleton as the first step to a radical upheaval of their body plan that eventually led to the pentaradial form and diffuse arrangement of nervous tissue that we see today.

The fossils at Chengjiang also hold clues to the earliest steps towards the vertebrates. The deposit has yielded numerous fossils of primitive jawless fish as well as rarer amphioxus-like animals, which are generally considered to be the precursors of fish. Although the stages connecting these vertebrates to the first deuterostomes are unclear, some answers may come from another Chengjiang group called the yunnanozoans which, like the vetulicolians are forcing us to rethink deuterostome evolution. Some experts interpret the yunnanozoans as primitive fish, but comparisons with the undisputed Chengjiang fish do not bear this out. In fish, and the more primitive amphioxus animal, the muscle blocks are arranged in a unique cone-in-cone arrangement. These so-called myomeres appear as zigzags on the animal's flank - the same characteristic pattern you find on a side of smoked salmon. This arrangement cannot be seen in the yunnanozoans. These animals may be on the road to becoming fish but, if so, they are far more primitive.

Yunnanozoans do, however, have the bipartite body plan. At first sight the arrangement seems rather different from the vetulicolians because there is a slender anterior bearing a series of gills and a posterior with prominent segments on the upper side. Imagine, however, that the front and back sections of a vetulicolian were to grow past each other: it could be transformed into a yunnanozoan. Also, intriguingly, a discovery published by our team earlier this year (Science, vol 299, p 1380) suggests that the prominent feathery gills of the yunnanozoans were attached to the creature's outside, whereas in amphioxus and vertebrates they are internal. This raises the possibility that the modern vertebrate gill slit has two origins. Perhaps the openings in the vetulicolian pharynx served to expel excess water, while the more active yunnanozoans evolved external gills to meet the demands of respiration, and these two functions then became integrated.

Vetulicolians and yunnanozoans have only just appeared on the palaeontological map, but they hold great promise for those of us trying to understand the roots of deuterostome evolution. Yet this is just the start of the story. Chengjiang is clearly a treasure trove and so far we have only picked up a few of the brightest evolutionary jewels. If we are lucky, we may find other fossils that represent intermediate forms between calcichordates and less primitive echinoderms, perhaps giving us an insight into the evolution of pentaradiality and the radical reconfiguration of the body plan in echinoderms. Then there is the still-incomplete story of vertebrate ancestry. Here the amphioxus animal is particularly fascinating because it is poised on the edge of vertebrate complexity, with all the genes in position but crucially lacking the evolution of key developmental networks. How and why did it make that evolutionary leap? Chengjiang may one day reveal the answer. So the evolutionary story continues to unfold, offering the tantalising prospect that we are on the way to tracking every major step of our origins, from vetulicolians in the Cambrian oceans to the bipedal hominid reading this copy of New Scientist.