The neutral theory of molecular evolution, the basic theory for explaining evolutionary changes at the molecular level, claims that most evolutionary changes at the molecular level are caused by random drift of neutral mutants. This is sharply contrasted to the evolution at the organismal level, where it is generally thought that natural selection is the major driving force by which evolutionary changes occur. How can we explain the evolutionary changes at the two levels in a unified way. This is the major problem that remains unsolved for molecular evolutionists. As a first step for understanding the final problem, we are investigating relationships of evolutionary diversifications between molecular and organismal levels, based on molecular phylogenetic approach.

I. Divergence pattern of animal gene families and relationship with the Cambrian explosion

In multicellular animals, a variety of gene families involved in cell-cell communication and developmental control have evolved through gene duplication and gene shuffling, basic mechanisms for generating diverse genes with novel functions. Each of these animal gene families is thought to have originated either from a few ancestral genes which are shared with plants and fungi or from an ancestral gene created uniquely in animal lineage. The major groups of bilateral animals are thought to have diverged explo-sively at or prior to the Vendian - Cambrian boundary. No direct molecular evidence has been provided to date as to whether the Cambrian explosion was triggered by a dramatical increase in the number of genes involved in cell-cell commu-nication and developmental control either immediately prior to or in concert with the Cambrian explosion.

A molecular phylogeny-based analysis of several animal-specific gene families has revealed that the gene diversification by gene duplication occurred during two active periods interrupted by a long intervening quiescent period. Intriguingly, the Cambrian explosion is situated in the silent period, indicating that there is no direct link between the first burst of gene diversification and the Cambrian explosion itself. The above result also suggests the importance of gene recruitment as a possible molecular mechanism for morphological diversity.

To understand a possible origin of animal-specific gene families, we have carried out cloning and sequencing of genes related to the animal-specific protein tyrosine kinase (PTK) genes from choanoflagellates, a group of unicellular plotists known to be the closest relatives of multicellular animals. Many PTK related genes including both the receptor and non-receptor type genes have been identified, some of which are likely to be orthologous to animal PTKs. A molecular phylogenetic tree of the PTK family members including those from animals and choanoflagellates showed an unexpected pattern of the PTK gene family: Most gene duplications that gave rise to different PTK subfamilies recognized in all animal groups occurred at very ancient times before the divergence of animals and choanoflagellates.

II. The origin of land vertebrates and evolutionary relationships among the coelacanth, lungfishes, and tetrapods based on nuclear DNA-coded genes

The phylogenetic relationships among tetrapods and two living groups of lobe-finned fishes, the coelacanth and the lungfishes are still unsolved and debated, despite many studies based on morphological and molecular data. Recent analyses based on complete mitochondrial sequences resulted in confusing phylogenetic trees, which are obviously inconsistent with generally accepted trees. To resolve this difficult phylogenetic question at statistically solid bases, we have cloned and sequenced ten nuclear DNA-coded genes from fourteen major groups of vertebrates, including the coelacanth and the three lungfish species. These sequences, together with sequences available from databases, have been subjected to phylogenetic analyses based on the maximum likelihood (ML) method, using a cyclostome and a lancelet as an outgroup. The obtained ML tree supports the close association of the coelacanth and the lungfish, suggesting that the coelacanth and the lungfish equally closely related as sister groups of tetrapods.

Publication List:

Hashimoto-Gotoh, T., Tsujimura, A., Watanabe, Y., Iwabe, N., Miyata, T. and Tabira, T. (2003) Differential involvements of presenilins-1 and –2 in the genesis of Alzheimer’s desease. Gene in press.

Terakita, A., Koyanagi, M., Tsukamoto, H., Yamashita, T., Miyata, T. and Shichida, Y. (2003) Counterion displacement in the molecular evolution of rhodopsin family. Nature Structure & Molecular Biology in press.

Kikugawa, K., Katoh, K., Kuraku, S., Sakurai, H., Ishida, O., Iwabe, N., and Miyata, T. (2003) Basal jawed vertebrate phylogeny inferred from multiple nuclear DNA-coded genes.BMC evolutionary biology in press.

I. Mathematical method for biological phenomena

We are studying biological phenomena of higher orders by using mathematical models. Mathematical models give us integrative understanding for complex behavior of biological systems including a lot of factors.
Study of the mechanisms responsible for morphological difference between species is an important research focus of the current developmental biology. One way to answer this question is to model the pattern formation process and to show how the species-specific pattern can be formed by the same model with different parameter values. Such a theoretical study would be useful in identifying candidates of cell-cell interaction that are likely to be responsible, to which future experimental study can be focused.

II. Pattern formation of the cone mosaic in the zebrafish retina

In teleost fish, there are several subtypes of cone cells, which are sensitive to different wave-lengths of light. In retinas of some species of teleost fish, regular arrangements of cone cells are observed, where each subtype of cone cells appears periodically in the two-dimensional retinal sheet. These patterns are called "cone mosaics." The biological mechanism of the pattern formation is still under examination. Some species show quite different mosaic patterns between peripheral region and more central region in the retina, which suggests mobility of cone cells in retinal space.

Different patterns are observed in different species. For example, in the zebrafish retina, there are four subtypes of cones, maximally sensitive to blue, red, green and ultra-violet. A green-sensitive cone cell and a red-sensitive cone cell are in tight contact and form a double cone. The cells of the other two subtypes are called single cones. The blue-, ultra violet-, red-, and green-sensitive cones are also called long single, short single, long double, and short double cones, respectively. In an adult zebrafish retina, rows of single cones and those of double cones appear alternatively. This pattern is called a “row mosaic”.

Figure 1. Cone mosaic of fish retina, (Left) in zebrafish, and (Right) in medaka. In fish retina, there are four kinds of differentiated cells. A pair of red-sensitive and green-sensitive cones is in close contact, constituting a double cone. Here four kinds of cones are indicated by capitals (R, G, B, U). In the regular mosaic pattern of zebrafish (Left), the rows of double cones and those of single cones (blue- and UV-sensitive cones) are in parallel and alternately. Two small cones of a red-light sensitive and a green-light sensitive cones form a “double cone”.

In contrast, in the medaka retina, four types of cones are arranged in a different manner as illustrated in Fig. 1. In the medaka pattern, each blue-sensitive cone is surrounded by four double cones, and the red-sensitive part of a double cone is close to the green-sensitive part of another. This is called a “square mosaic”.

We have studied the mechanism of the pattern formation by using mathematical models. In this study, we examined a process of cell-cell interaction to generate the regular mosaic pattern -- namely cell rearrangement. In the model, cells have already been determined as to their final subtypes, but they change their locations in the pattern formation process and the cell movement is affected by their neighbours.

Figure 2. The examples of obtained patterns by cell rearrangement model. The obtained patterns were completely the same as the actual zebrafish retina, when the used adhesion was appropriate.

We show that the same model can produce both row and square mosaic pattern. If the cell-cell interaction is restricted to nearest neighbors only, the square mosaic pattern cannot be generated. We studied the model considering the adhesion working between nearest neighbors and next nearest neighbors, with different weighting between them. Two “shape factors” specifying how to combine adhesion in different geometric are very important in determining whether row mosaic (zebrafish pattern) or square mosaic (medaka pattern) is to be formed.

Figure 3. The region in which zebrafish pattern is formed. Dots indicate the parameters that generated the zebrafish pattern obtained by computer simulations. Planes specified the borders separating the parameter region by comparing the zebrafish pattern with four other patterns. A region determined by theoretical analysis was consistent with the numerical results.

Figure 4. Alternative irregular patterns obtained the parameter regions around the area shown in (A).

We study the conditions for generating mosaic pattern. The condition for generating zebrafish mosaic are shown in Fig. 3, 4. They are the necessary condition and that should be satisfied even in actual organisms if they make the arrangement based on cell adhesion and cell movement.

III. Directionality of Stripes Formed by Anisotropic Reaction-Diffusion Models.

The pattern formation of animal coating has been studied mathematically by a pair of partial differential equations, named a reaction-diffusion (RD) model. By the model, starting from an initial distribution very close to uniformity, a spatial heterogeneity emerges and a stable periodic pattern is formed spontaneously. This simple mechanism suggests that the reaction of a small number of chemicals and their diffusion might create stable non-uniform patterns. When we analyze the model in a two dimensional plane, striped patterns in addition to spotted patterns often emerge. This was considered as the basic mechanism explaining the stripe patterns observed among animal coating.

We focused on the directionality of the stripes. Most of the stripes observed in the fish skins are either parallel or perpendicular to their anterior-posterior (AP) axis. The direction of stripes is considered of importance in the behavioral and ecological viewpoints. However, very little is known about the mechanisms that makes the strong directionality either in the actual fish skin or in the theoretical models. The standard RD model doesn't determine the direction of stripes. To explain the directionality of stripes on fish skin in closely related species, we have studied the effect of anisotropic diffusion of the two substances on the direction of stripes, in the cases in which both substances have the high diffusivity in the same direction.

We also studied the direction of stripes in more general situations in which the diffusive direction may differ between the two substances. We derived a formula for the direction of stripes, based on a heuristic argument of unstable modes of deviation from the uniform steady state. We confirm the accuracy of the formula by computer simulations. When the diffusive direction is different between two substances, the directions of stripes in the spatial pattern change smoothly with the magnitude of anisotropy of two substances. When the diffusive direction of the two substances is the same, the stripes are formed either parallel to or perpendicular to the common diffusive direction, depending on the relative magnitude of the anisotropy. The transition between these two phases occurs sharply.

Figure 5. The stripe formation in the skin of Genicanthus. These two species are closely related.

Figure 6. Summary of the direction of stripe patterns obtained by the anisotropic diffusion model. Horizontal and vertical axes indicate anisotropy of activator and that of inhibitor, respectively. The left-upper corner indicates the distortion of diffusion range. Each point indicates the direction of the observed stripe: horizontal; vertical; or not-determined. The direction is determined only by the difference between anisotropies. A small difference in diffusion anisotropy can alter the final pattern to one with opposite directionality. Note the transition of the stripes by changing the anisotropies from O to A (or I).