Serhiy Yastrebov. I. Origin of chordates: a modern perspective on the problem

Living room "Batrachos". This publication is the first of its kind on this site. Sergey Alexandrovich Yastrebov, a lecturer at MSU and author of the wonderful blog Caenogenesis (primarily devoted to developmental genetics) suggested posting three of his popular articles (and perhaps more) on Batrachos. They were published in the journal Potential. The second article is here, the third – here. S.A. Yastrebov Origin of chordates: a modern view of the problem From whom did chordates originate? The reasons for interest in this question are clear. First, chordates are one of the largest and most diverse animal types (by number of species they are probably fourth – after arthropods, molluscs and roundworms1). Second, chordates attract zoologists’ attention with their large size, complex brain and complex behaviour. Third, we belong to this type ourselves. It is not surprising that the question “from whom did chordates originate?” began to interest zoologists already in the 19th century. However, this problem turned out to be surprisingly tangled. Of course, questions of the kind “who originated from whom?” are never simple in biology, but the question of chordate origins stands out even against that background. By the beginning of the 20th century about a dozen hypotheses on chordate origins had accumulated. As possible ancestors, a wide range of animal groups were considered, from arachnids to Cnidaria inclusive. In 1910 a zoological conference devoted specifically to the question of chordate origins was held in London. History preserved the words of one participant when he returned home. When friends asked about the discussion results he answered: “Everyone agreed on one statement: that their opponents are wrong!” Over the past century the number of chordate origin theories, of course, has grown even more. To review them all would require a whole book (and such books have been written – for example by Olha Mykhailivna Ivanova‑Kazas2). If the history of this question is presented in an extremely brief form, it looks roughly as follows. At the end of the 19th century the most popular hypothesis was that chordates originated from annelids, developed by the German zoologist Anton Dorn. Chordates and annelids indeed resemble each other. Both have a coelomic (so‑called “secondary”) body cavity and well‑developed segmentation. Annelids possess a ventral nerve cord, chordates a spinal cord, which is also segmented (segmental nerves branch from the vertebrae). The circulatory system in both includes dorsal and ventral vessels (in chordates they are called dorsal and ventral aortae), connected by lateral vessels (in chordates called arterial arches). A fairly important similarity lies in the excretory system. In the school textbook example of a chordate – the lancelet – the excretory system is represented by segmental tubules‑nephridia, similar to the nephridia of many annelids. The lancelet is a lower chordate, not a vertebrate. Yet the kidney of vertebrates developed from a segmental series of nephridia (this follows clearly from its embryonic development). Principal differences between the body plans of annelids and chordates are as follows: 1. In chordates segmentation does not cover the skin (more precisely, the epidermis). In annelids, as is known, segmentation does cover the coverings and is visible externally, which is why they are called “ringed”. 2. The central nervous system is located on the ventral side in annelids and on the dorsal side in chordates. 3. In annelids blood flows forward in the dorsal vessel and backward in the ventral one; in chordates the opposite. 4. Annelids have a perioral neural ring, meaning their oesophagus literally passes through the brain. Chordates lack this. Combining points 2 and 3 leads easily to the following supposition: the ancestor of chordates was an annelid whose dorsal and ventral body sides were swapped. This is exactly the conclusion drawn by Dorn (Fig. 1). [IMG_1] Fig. 1. Hypothesis of dorsal‑ventral side reversal in chordate ancestors (after Gerhart). In the scheme the animal is shown with body sides as in annelids and arthropods (dorsal up, ventral down). If it is turned over, it “becomes” a chordate. Arrows indicate blood flow direction. Point 1 received little serious attention in the 19th century – who knows what may happen to the coverings. However, point 4 created a major problem. Interestingly, Dorn considered the absence of the perioral neural ring to be the main distinction of chordates from annelids. He was convinced that former elements of this ring are retained in chordates and form the brain. Supposing a gradual “migration” of the mouth to the opposite side of the body was very difficult: the mouth in annelids is inside the neural ring and would have to tear it during movement, for which there is no evidence. Thus the chordate mouth must have arisen anew. But where did the “old” mouth go, and what occupies its place in modern vertebrates? Dorn never managed to give a convincing answer. Similar difficulties were faced by authors who hypothesised chordate origins not from annelids but from arthropods. It is worth recalling that in the 19th century most zoologists were convinced that annelids and arthropods were closest relatives. Indeed, they share very similar segmentation and nervous system organization. It was commonly thought that arthropods are direct descendants of annelids. From which of them chordates originated was then considered of minor importance. The situation changed dramatically at the beginning of the 20th century. In 1908 the Austrian zoologist Karl Grobben proposed dividing all bilaterally symmetric animals into two large groups: primary deuterostomes and secondary deuterostomes (Fig. 2). This division is embryological. Multicellular animals have a developmental stage called the gastrula, in which the embryo is a bilayered sac. The entrance to the internal cavity of this sac is called the blastopore. In primary deuterostomes the blastopore partially or completely becomes the mouth opening. In secondary deuterostomes the mouth is not connected to the blastopore and breaks through anew, often at the opposite end of the body. This division is still used in zoology. Primary and secondary deuterostomes are indeed two large, long‑separated evolutionary branches. Their embryonic differences are profound (they affect not only the fate of the blastopore but also, for example, the mode of egg cleavage). Primary deuterostomes are far more numerous. Already a hundred years ago it was clear that annelids and arthropods belong to them, while chordates are secondary deuterostomes. Thus they are not close relatives of annelids or arthropods. [IMG_2] Fig. 2. Primary and secondary deuterostomes (after Malakhov, with modifications). The annelid shown here is marine and sedentary, so it looks somewhat unexpected to those familiar with the group mainly from the earthworm. In the modern system, only three animal types are placed among secondary deuterostomes: chordates, echinoderms and hemichordates. Echinoderms include such relatively common organisms as sea urchins, starfish and sea lilies. Hemichordates are far less widely known. They are marine animals that feed by filtering water. The hemichordate group is divided into two classes: enteropneusts (worm‑like organisms living in burrows in the substrate) and pterobranchs (animals with tentacles living in tubes, mostly sessile and colonial). Could chordates have originated from one of them? Such hypotheses appeared already in the 19th century, but after Grobben’s work they gained many more supporters. Of course, these are not all the variants. For example, the Englishman Adam Sedgwick and the Belgian Auguste Lamère thought that chordates originated directly from Cnidaria. There was also a hypothesis of chordate origin from ctenophores – transparent marine animals remotely resembling jellyfish, which some zoologists considered ancestors of the first worms. Finally, there were quite exotic versions. For instance, a hypothesis of chordate origin from cephalopod molluscs, based on the fact that both have large brains and complex eyes. Or a hypothesis of chordate origin from nemertines – a special type of marine worms with very long bodies and an eversible proboscis (it was assumed that the chord derived from this proboscis). Or from flatworms – such an idea was also voiced. In total, several dozen different chordate origin hypotheses have been counted. For convenience they can be grouped into four categories according to “from whom chordates are said to have originated”: 1. From animals with pronounced segmentation (such as annelids or arthropods). 2. From other secondary deuterostomes (echinoderms, hemichordates or some extinct group). 3. Directly from the common ancestors of bilaterally symmetric animals (most often Cnidaria or ctenophores, though other versions existed). 4. The last group – a mixed bag including “exotic” ideas that lead nowhere. What can modern biology say about which of these hypotheses is correct? Recall the school zoology textbook. It first describes Cnidaria, then flatworms, then roundworms, then annelids, then molluscs, then arthropods, and finally – chordates. Usually it is explicitly or implicitly assumed that this order of presentation reflects evolutionary sequence (“the ladder of nature”, as it was said in the 18th century). Flatworms are the simplest bilaterians, they appeared first. Chordates are the most complex, they appeared last. The whole history of animals is a history of gradual multistage complexity increase, and chordate evolution is the final stage of this process. Such views were indeed widespread for more than a century. At least this picture usually resulted from zoology courses for non‑specialists (specialists always had far more diverse opinions). But now the situation has changed. According to some scholars, at the turn of the 20th and 21st centuries zoology underwent a genuine scientific revolution3. If this is true, the starting point of the revolution is undoubtedly 1997, when a group of American researchers constructed one of the first phylogenetic trees of bilaterians based on ribosomal RNA. A brief explanation is needed. As is known, ribosomes are very small (invisible in light microscopy) cellular organelles whose function is protein synthesis. Each ribosome consists of several ribonucleic acid molecules – RNA. An RNA molecule is a polymer composed of many monomeric units – nucleotides. The closer the animals are related, the more similar their RNA nucleotide sequences are. Ribosomes synthesize protein, and only that. This function exists in every cell. Therefore the ribosomal RNA sequence is likely independent of animal anatomy or physiology. Differences in it can only be random. The more random differences have accumulated between the rRNA of two given animals, the longer ago their common ancestor lived. The number of differences in rRNA, in the first approximation, is proportional to the time of evolutionary divergence and depends on nothing else – whereas if we studied, for example, respiratory or muscle proteins, the differences would strongly depend on physiological factors and lifestyle. That is why analysis of rRNA (abbreviated rRNA) is so convenient for establishing relationships. However, the results turned out rather unexpected. The phylogenetic tree constructed from rRNA looked like this (Fig. 3): [IMG_3] Fig. 3. Molecular phylogenetic tree of bilaterian animals (compiled from the works of Aginaldo et al.) This tree differs from the classical one in one, but very important, point: arthropods here are placed not near annelids but near roundworms. We have already mentioned that arthropods are organized very similarly to annelids. At the beginning of the 19th century the great French zoologist Georges Cuvier even grouped them into a single type called “Articulata”. In the 20th century there were biologists who considered such a grouping correct. In any case, few doubted that these two groups were close relatives. The idea that arthropods derived from annelids entered textbooks and – without exaggeration – for more than a hundred years was regarded as one of the most established facts in zoology. Roundworms, at first glance, seem completely unlike arthropods. Already because they lack a coelom (arthropods have a typical coelom, at least in embryos) and, as a rule, lack segmentation. The first reaction of most zoologists to this discovery was very skeptical, in the spirit of “What will these molecular biologists come up with?” Yet after a few years it became clear that the new phylogenetic tree cannot be dismissed, because accumulating new data (primarily molecular, but not only) continuously confirm it. It is now part of university curricula. The group that includes roundworms and arthropods together is called “molting animals” (Ecdysozoa). Annelids have no relation to this group, except that they are also primary deuterostomes. And now back to the chordate origin problem. We have already seen how the overall phylogenetic tree of bilaterians looks after the discovery of the molting group. Chordates belong to the branch of this tree called “secondary deuterostomes”. The phylogenetic tree of the secondary deuterostomes themselves, according to modern data, is as follows (Fig. 4): [IMG_4] Fig. 4. Modern phylogenetic tree of secondary deuterostomes (after Malakhov, with modifications) Recall that enteropneusts and pterobranchs together are called hemichordates. It turns out that secondary deuterostomes at the very beginning of their evolution split into two branches: one leading to chordates, the other to the common ancestors of hemichordates and echinoderms. Chordates are the group that diverged first from the common stem of other secondary deuterostomes. How do secondary deuterostomes fare with the characters we began the discussion with – segmentation and a coelom? All types of secondary deuterostomes have a coelom, and it is well developed4. Segmentation is more complicated. Echinoderms and hemichordates have body regions that can be considered segments, but there are only three of them. This segmentation is called oligomeric. Chordates, however, usually have several dozen segments – like annelids and arthropods (recall how many vertebrae we have). This segmentation is called polymeric. What about these characters in the other bilaterians? Consider the overall phylogenetic tree (highly simplified, without many types), on which they are indicated (Fig. 5): [IMG_5] Fig. 5. Modern phylogenetic tree of bilaterian animals with indications of some characters (original). Explanation in the text A solid line on this scheme highlights the names of animal types that possess a coelom (or at least its remnants), a dashed line – those that have polymeric segmentation. Note that flatworms are absent from the scheme. The position of flatworms on the evolutionary tree is currently actively discussed (there is a well‑grounded hypothesis that most of their groups are simplified descendants of annelids), but we will not consider it here. We see that a coelom is present in representatives of three different large evolutionary branches: 1) secondary deuterostomes, 2) annelids with molluscs, and 3) arthropods. The question arises: did the coelom appear independently in these three branches, or was it already present in the common ancestor of primary and secondary deuterostomes, indicated on the scheme by a black circle? In the latter case one could suppose that roundworms lost the coelom secondarily. Polymeric segmentation is present in the same three branches, although not in all types. Is it necessary to assume that it arose three times independently? Or perhaps the common ancestor of bilaterians was already coelomic and segmented – like an annelid, an arthropod or a chordate? This hypothesis was already proposed in 1884 by the aforementioned Englishman Adam Sedgwick. He never had many supporters. But at the beginning of the 21st century Sedgwick’s hypothesis received a second life. Many new data – genetic, embryological and even paleontological – have appeared that can be easily interpreted in its favour5. For example, French biologists Guillaume Balavuand and André Adutt note that Sedgwick’s hypothesis fits very well with our knowledge of genetic regulation of segment formation in embryos of various animal types. Again, an explanation is needed. Each gene has an expression domain – the part of the organism in which its product is present in cells. For instance, there is a group of genes called “segment polarity genes”. The expression domain of such a gene is either in the anterior or posterior part of each segment, determining where its front and rear ends will be. These genes, like other segmentation genes, are often the same in arthropods, annelids and chordates. Their action is strikingly similar: for example, the expression domains of the gene engrailed in lancelet and Drosophila embryos are located strictly in the anterior part of each prospective segment. Such similarity cannot be accidental. Either it is a deep parallelism (independent development of similar adaptations on a common basis), or evidence that primary and secondary deuterostomes derived from a common segmented ancestor. Thus, perhaps two of the four main chordate‑origin hypotheses we listed are simultaneously correct: chordates originated directly from the common ancestor of bilaterians (hypothesis 3), which possessed polymeric segmentation (hypothesis 1). Both Adam Sedgwick and Anton Dorn were right – each in his own way.However, it must be emphasized that this conclusion is merely a hypothesis that is currently under discussion, has its supporters and opponents, and may well be refuted by new facts. Let us look once more at the evolutionary tree. If Sedgwick’s hypothesis is indeed correct, then it is completely pointless to search for the ancestor of chordates among any other types of bilaterally symmetrical animals—at least among modern ones. Paradoxical as it may seem, chordates occupy a position closer to the base of the bilaterian evolutionary tree than to its crown. Paleontology can reveal much more about the beginning of chordate evolution, and to some extent already does. However, that is a topic for a separate discussion. Brief list of references Malakhov V.V. Proiskhozhdenie khordovykh zhivotnykh. // Sorosovskii obrazovatel’nyi zhurnal, no. 7, 1996. — A popular article by one of the leading contemporary zoologists. Some parts have become outdated, but it is very well written and covers aspects of the problem not discussed in detail in our article. Online: here. Aguinaldo A., Turbeville J., Linford L., Rivera M., Garey J., Raff, Lake J. Evidence for a clade of nematodes, arthropods and other moulting animals. // Nature, 1997, vol. 387, pp. 489–493. — That very historical article on rRNA that initiated a revolution in zoology. Online: here. Balavoine G., Adoutte A. The Segmented Urbilateria: A Testable Scenario. // Integr. Comp. Biol. (2003) 43 (1): 137–147. — A very interesting article elucidating the essence of new views on the evolutionary tree of bilaterians. Online: here. 1 In fact, “roundworms” are usually not a single type in modern systems, but several types (different authors recognize different numbers). The classification of these animals remains rather confusing. For convenience, we will use the traditional term “roundworms,” but it should be borne in mind that the reality is far more complex. Back to article text. 2 Ivanova-Kazas O.M. Ocherki po filogenii nizshikh khordovykh. Leningrad, 1995. Available online in this library. Back to article text. 3 Malakhov V.V. Revoliutsiia v zoologii: novaia sistema bilatertii. // “Priroda,” 2009, no. 3. The journal can be downloaded here. Back to article text. 4 There is a caveat. At the end of the twentieth century, a strange marine worm lacking a coelom, protonephridia, and even a through gut was discovered—Xenoturbella. It proved very difficult to classify, but molecular data apparently supported the hypothesis that Xenoturbella represents a separate type, Xenoturbellida, belonging to the Deuterostomia. If this is true, then there are not three but four types of deuterostomes, and one of them lacks a coelom. Back to article text. 5 This issue is discussed in detail in V.V. Malakhov’s article “Novyi vzgliad na proiskhozhdenie bilatertii” (“Priroda,” 2004, no. 6). Online: here. Back to article text.