Molecular phylogeny of Enchytraeidae (Oligochaeta) indicates separate invasions of the terrestrial environment

Introduction

Phylogeny of the family Enchytraeidae has been given little attention compared with other oligochaete families. The most recent discussion including all known genera date as far back as Černosvitov (1937) where the then accepted genera were assembled in six subfamilies. The suggested subdivision has had limited impact because classification of the subfamilies was not convincing and the discussion included genera with doubtful connection to the family Enchytraeidae. In the 1980th Coates published a series of papers dealing with various aspects of enchytraeid phylogeny based upon morphological observations. Coates (1986, 1987) are focussed upon the genus Propappus and conclude that it cannot be retained within the family and a new family, Propappidae, is erected with a suggested sisterelationship to Enchytraeidae. In Coates (1989) the phylogeny of Enchytraeidae is discussed based upon the assumption that a bisetate condition, two setae in each of the four bundles per segment, is the ancestral state in Enchytraeidae, and the discussion mostly includes bisetate species and species without any setae (achaetinous).

Enchytraeid species are also included in more recent studies using DNA data to evaluate the phylogeny of clitellate annelids. Siddall et al. (2001) and Erséus and Källersjö (2004) using nuclear 18S data place the Enchytraidae in a sister group relation to the earthworms. In a quite recent paper, Marotta et al. (2008), by using a combination of morphological and 18S data add Lumbriculidae and leeches to this sister clade thus supporting the use of the two outgroup species chosen in this study. Erséus (2005) review the then known molecular studies and Rousset et al (2008) provide strong support for an ‘aquatic’ ancestral state of clitellate annelids.

The aim of the present study was to evaluate the intra-familiar phylogeny of Enchytraeidae by using molecular data involving both nuclear and mitochondrial genes covering a total of 4977 bp. The species studied are representatives of well defined species-rich genera from the Northern Hemisphere. Lumbriculus variegatus and Lumbricus terrestris are used as outgroup species.

Materials and methods

Results

The result of the Bayesian analysis is shown in Fig. 1. Enchytraeus sp. appears as sister group to the other in-group species studied. Among the latter again a single species, Lumbricillus lineatus, turn up as sister group to a strongly supported monophyletic group comprising the remaining taxa. Thus the molecular data indicate the segregation of the two genera Lumbricillus and Enchytraeus from a major later diversifying branch including the remaining enchytraeid genera of this study.

Within this major grouping a strong bipartition is revealed with one branch comprising members of the genera Mesenchytraeus and Cernosvitoviella and another more diverse lineage containing representatives of genera that dominate most terrestrial habitats on the northern Hemisphere. Once again these are distributed among two branches, one which embraces Henlea nasuta and Cognettia sp. and another consisting of Achaeta affinis, Fridericia sp. and Buchholzia sp. with the Achaeta species appearing as sister-related to the two latter.

Discussion

In the older literature (Michaelsen 1916; Stephenson 1930; Ĉernosvitov 1937; Brinkhurst and Jamieson 1971) it is generally agreed that the Enchytraeidae are classified with some mostly aquatic families such as Tubificidae, Naididae and Phreodrilidae in a subgroup of Oligochaeta called Microdrili referring to their moderate size and contrasting another subgroup of oligochaetes, Megadrili, comprising the generally larger terrestrial earthworm families. Consequently, much discussion on the intra-familiar phylogeny has been based on the assumption that Enchytraeidae descended directly from aquatic ancestors as an invasion of terrestrial habitats independent from that which gave rise to the megadrile earthworms. However, very few morphological features support the above sketched relationships and even if terrestrial oligochaetes may ultimately stem from an aquatic ancestor of a microdrile appearance, i.e. Rousset et al (2008), nothing indicates that this is specifically the case for Enchytraeidae, and the view expressed by Černosvitov (1937) that enchytraeid genera with morphological features resembling similar organs in aquatic Microdrili hold an ancestral position within the family is doubtful. Here exemplified by the genera Mesenchytraeus and Cernosvitoviella, which through the possession of sigmoid setae with nodulus and nephridia with little interstitial tissue from a morphological point of view are supposed to hold an ancestral position, but do in fact join the derived branch of the family comprising most of the genera included in this study. Thus if the molecular phylogeny indicates the true relationship, it remains doubtful whether the above mentioned features represents the ancestral states in Enchytraeidae or are secondary adaptations to the relatively moist environments inhabited by Mesenchytraeus. and Cernosvitoviella species.

The isolated position of Enchytraeus and Lumbricillus in the molecular phylogeny may at first sight appear surprising. However, both genera have morphological characteristics in which they differ not only from other enchytraeid genera included in this study but also from Oligochaeta in general, features that may have gained too little attention in phylogenetic discussions. As pointed out by Stephenson (1930) some species of Enchytraeus and Pachydrilus (now Lumbricillus) are unique among enchytraeids in having their testis enclosed in a dense sac (testis sac) within which the male cells mature. As now defined this would seem to be the case in all members of the genera Enchytraeus and Lumbricillus while in the other enchytraeid genera studied immature male cells are released from the testes and mature freely in the cavity of segment XI (see Fig. 1). Thus in Enchytraeidae the term seminal vesicle covers morphologically quite different structures. In say Fridericia galba the testes release immature sperm cells directly into the lumen of segment XI where they mature and the seminal vesicle is bounded anteriorly by septum X/XI and posteriorly by septum XI/XII while in say Enchytraeus albidus immature sperm cells are released into a closed sac surrounding the testes and the growing seminal vesicle is a separate structure within the lumen of segment XI. This strong difference is directly seen in the positioning of the so-called sperm funnel in relation to the seminal vesicle. The sperm funnel is the first part of the vas deferens and in most species a conspicuous cylindrical structure located in segment XI and in Fridericia galba and other species and genera with the same type of seminal vesicle the sperm funnel is found within the same cavity as the developing sperm cells, i.e. within the seminal vesicle while in Enchytraeus albidus and other members of the genera Enchytraeus and Lumbricillus sperm funnels are found outside this structure. The so-called sperm sac characteristic of the genus Mesenchytraeus is a backward directed pouching of septum XI/XII in which sperm cells mature, but since this lumen is a mere extension of segment XI the situation in Mesenchytraeus is in principle identical to that described for Fridericia galba. According to Stephenson (1930) the latter situation is with minor modifications universal in Oligochaeta indicating that it might be the ancestral condition while Lumbricillus and Enchytraeus and perhaps the megascolecid genus Ocnerodrilus represent a unique specialization.

Another partitioning which organizes enchytraeids in the same two groups as described above and which may have important ecological implications is a strong difference in the number and size of eggs. This is most obvious when medium sized to large species are compared and in both Enchytraeus and Lumbricillus species the eggs are relatively small compared to the size of the worm and usually 10–20 and occasionally more than 30 eggs mature simultaneously (Christensen 1956). They are laid in the same cocoon and the vast majority develop and hatch out. This is in sharp contrast to species of the other genera such as Fridericia, Henlea and Mesenchytraeus which produce relatively much larger eggs and normally only one egg is laid in each cocoon (see Fig. 1).

Under optimal conditions sexually mature worms usually deposit one cocoon per day which means that Enchytraeus and Lumbricillus have a much higher reproductive rate measured in number of offspring than members of the other genera mentioned above (B. Christensen, pers. obs.). On a scale covering the reproductive output in enchytraeids, members of Enchytraeus and Lumbricillus may be called r-strategists while members of Fridericia etc relative to that may be called K-strategists and this may be an adaptation to the preferred habitat types of the species in question.

As stated above species of Enchytraeus and Lumbricillus are dominant members of the enchytraeid fauna in the littoral zone along the sea, a habitat type which according to conventional ecological thinking favour r-strategists because it is so unpredictable, one day heaven on earth for enchytraeids with an ideal water regime and an abundance of food in the decaying seaweed washed ashore, the next day rough weather may have swept the wrack bed away and more or less destroyed the habitat for a while. In contrast, inland terrestrial habitats are relatively more predictable both with respect to water regime and food supply and such environments are supposed on the enchytraeid reproductive scale to favour K-strategists. Thus the molecular data presented may indicate an early split into two main strategies: one adapting to exploit decaying plant material washed ashore along the sea and another main line primarily adapting to exploit decaying plant material of inland soils.

However, the above scenario is painted with broad strokes and extant members of the genus Enchytraeus, in particular, are in fact more diverse in their ecologies. A group of medium sized to large species including the nominal species of the family Enchytraeus albidus mainly inhabit the littoral zone, they have the same chromosome number (n = 21) and they all reproduce sexually (Christensen 1980). This is in sharp contrast to another group, the Enchytraeus buchholzi group, which differ in all these features. The species in question inhabit a wide variety of inland soils, the basic chromosome number is different (n = 18) and several cases of polyploidy have been observed even as high as decaploidy (Christensen op.cit.), and in addition to normal sexual reproduction a case of so-called subamphimixis is recorded in which some members of the chromosome complement behave as in a normal sexual species while others follow a course as in mitotic parthenogenesis (Christensen and Jensen 1964). Finally multiplication through transverse fission of the adult body and subsequent regeneration of new anterior and posterior ends is found in other species (Christensen 1964, 1973). This high degree of intra-generic heterogeneity may be in line with the long prehistory of the genus Enchytraeus indicated in the molecular phylogeny.

Compared with their supposed aquatic ancestors terrestrial oligochaetes must have required several adaptations among which regulation of the water balance in these occasionally dry environments represents an important challenge to the functioning of the nephridium. This is clearly seen in the Enchytraeidae where the nephridia in most genera are elaborate organs of a characteristic shape and with an extremely narrow nephridial canal embedded in interstitial tissue. The only exceptions are species of the genera Mesenchytraeus and Cernosvitoviella where the nephridial canal is loosely wound, relatively short and wide and the interstitial tissue is poorly developed and as such resembling the situation in aquatic oligochaete families. In members of Enchytraeus and Lumbricillus the ciliated nephrostome or funnel is the only anteseptal part of the nephridial canal while the tightly coiled postseptal part embedded in interstitial tissue form a solid ovoid structure and the fairly wide efferent duct has a postero-ventral origin. In species of Henlea and Cognettia coils of the canal may or may not extend forward into the anteseptal part but in either case the nephrostome is a distinct forward directed structure. The post-septal part is elongate, occasionally somewhat pear-shaped and the long and slender efferent duct usually with an antero-ventral origin. In Fridericia, Buchholzia and Achaeta the ante-septal part is usually large with dense coils of the canal and in contrast to the former group the nephrostome is completely embedded in the tissue; the postseptale elongate and may gradually extend into a fairly wide efferent duct or the duct may arise as a more distinct structure in a postero-ventral position. Thus, a sub-division into four groups of genera indicated by the molecular data is reflected also in the evolution of the enchytraeid nephridium and this organ thus appears as a fairly conservative structure illustrating the major subdivisions of the family (see Fig. 1).

This contrasts a more diversified pattern seen in other much used taxonomical characters such as shape of setae, origin of dorsal vessel and various specializations of the intestinal canal. If the molecular tree reflects the true relationships functionally identical specializations appear to have arisen independently within different lineages and a closer inspection usually shows that the morphological similarities between these are only superficial.

Irrespective of the ancestral state of the setae (straight or sigmoid) the same type has developed independently and repeatedly within all the major branches: Lumbricillus has sigmoid setae while they are straight in Enchytraeus; in Cognettia they are sigmoid and in Henlea straight and finally in Buchholzia they are sigmoid but in Fridericia straight. A further specialization where the innermost setae in a bundle is shorter than the outer ones is seen in two different lineages namely in members of the genera Henlea and Fridericia with four or more setae per bundle and again this situation has undoubtedly arisen independently in the two genera and the similarity is only superficial, in Fridericia we see a perfect pair wise symmetric arrangement around the mid-axis of the setal bundle while in Henlea the setae are not arranged in distinct pairs.

A somewhat similar situation is seen in species of Henlea and Buchholzia in the combination of a heart-like ante-clitellar origin of the dorsal vessel and the presence of various appendages in the very same region of the intestine. The ancestral situation in the family is undoubtedly a smooth origin of the dorsal vessel immediately behind the clitellum and absence of intestinal specializations. The presence of these may perhaps improve digestion and the large pulsating heart-like expansion of the dorsal vessel in the nearby segments may secure an efficient blood supply to this physiologically active region of the intestine. The molecular data indicate an independent origin in Henlea and Buchholzia and a detailed morphological analysis support this. In Buchholzia the intestinal appendages are dorsal while in Henlea they are mostly of a lateral position. In Buchholzia the dorsal vessel arises from the summit of the intestinal appendage while in Henlea it arises independently of the appendages from the blood sinus of the intestine itself which usually widens considerably in this region.