Abstract 1 Genetical mtDNA relationships of 41 taxa of weevils were examined using cladistics. Ingroup taxa belong to Otiorhynchus scaber and O. nodosus and outgroup comparison was made with O. singularis. All three species are minor forest pests.
2 Otiorhynchus scaber specimens are either diploid sexuals or diploid, triploid and tetraploid clones, from two different populations (Slovenia and Austria) that belong to two different evolutionary lineages. Otiorhynchus nodosus specimens are tetraploid clones. Both species show geographical parthenogenesis, as do many other Otiorhynchus species.
3 Mitochondrial data indicate that O. nodosus clones are more closely related to Slovenian sexuals of O. scaber than these are to sexuals from Austria. It also shows that almost all clones of O. scaber collected in one of the two regions where sexuals are found are more closely related to sexuals from the other region.
4 Three different hypotheses that may explain the distribution of O. scaber, mechanisms important for the evolution of the clones and implications of the presence of Wolbachia are discussed.
5 We conclude that parthenogenesis is likely to be linked to hybridization in O. scaber and that hybridization events between ancestors of O. nodosus and O. scaber are the probable cause of the presence of O. nodosus in the ingroup. We also find that polyploid clones are superior colonizers compared to sexuals and diploid clones, in O. scaber.
6 The results suggest that systems where both sexuals and clones exist are more complex than previously suggested. The mapping of genetic variation in clonal complexes and the tracing of clonal origins may be useful in pest management.
Asexual reproduction in animals was shown for the first time by Bonnet (1745), who observed that aphid females produce offspring without males. His results were met with scepticism and it was not until the end of the 19th century before scientists became convinced that aphids are not hermaphroditic and that they can reproduce asexually. Many asexual animals have been discovered since then. Eighty vertebrate taxa from 14 families of fish, amphibians and reptiles have been found to reproduce asexually (Alves et al., 2001). Among invertebrates, asexuality is much more common than among vertebrates, and new asexual forms are still being described (e.g. Scholtz et al., 2003). White (1973) even suggested that most well studied animals (including humans) have some potential for asexual reproduction. However, this phenomenon is best studied in insects. So far over 900 species of insects are known to be asexual or have asexual forms (Normark, 2003). This is likely to be an underestimate, as many asexual forms are probably hidden in sexual populations (Lynch & Gabriel, 1983; Stenberg et al., 2003). Asexual reproduction is found all across the ‘tree of life’ and is present in most animal groups (excluding mammals). It has originated independently many times (as reviewed by Normark, 2003) and takes a variety of forms. Although many animals reproduce asexually, all are assumed to have originated from sexually reproducing ancestors. Asexual animals are of interest not only because they hold clues to the still unanswered question: why do we have sex? They are also fairly common, especially among insects and in particular among beetles, and therefore are likely to play an important role in evolution and the biological diversity that has developed.
In sexual reproduction there are two major ways of maintaining the parental chromosome number when reproducing without males. One way is automixis, where the eggs are still produced through meiosis (Suomalainen et al., 1987), and the diploid chromosome number is commonly restored by fusing two of the haploid meiotic products. The other way is to abandon meiosis altogether, the eggs then being produced by mitosis (apomixis). Apomictic animals lack meiotic recombination and without recombination all offspring will be identical clones except for random mutations. This is thought to make clonal lineages more vulnerable to parasites in comparison to sexuals (The Red Queen hypothesis: van Valen, 1973) and, in the long run, accumulation of mutations should gradually decrease the fitness of the clones (Muller's Ratchet: Muller, 1964). The two main advantages of clonality are a theoretical twofold increase in reproductive output and the ability of colonization by a single individual.
Even though thousands of clonal species are known, very little is known about the origin of apomictic reproduction (Normark, 2003) and only a few reports explain causes behind the phenomenon. In these cases the event that caused the transition from sexuality to clonality was a hybridization between two different sexual species (e.g. Dufresne & Hebert, 1997; Normark & Lanteri, 1998). The underlying molecular mechanisms that caused the transition are not known.
Even if most animals reproduce sexually, and most asexual forms are found to have a recent origin (Avise et al., 1992), there are several cases where asexual organisms have more or less outcompeted their sexual ancestors (e.g. Lynch, 1984; Suomalainen et al., 1987). There are a number of clonal species where close sexual relatives have not been observed. This could mean that the sexuals have been outcompeted, either by their own clonal relatives, or by other species. It could also mean that the sexual ancestor and the clonal form have diverged over time and no longer may be recognized as close relatives. The third reason could be that the clonal form is a hybrid between two ancestral sexual species, making identification difficult.
Many asexual forms share a common distributional pattern known as geographical parthenogenesis (Vandel, 1928, 1940; Glesener & Tilman, 1978; Bell, 1982; Suomalainen et al., 1987). Sexuals are found within a small central area, outside of which only clones exist. The clonal distribution is often shifted to higher altitudes or harsher climates. This pattern is found in both plants and animals all over the world. Examples include dandelions (Asker & Jerling, 1992) and many weevil species in Europe (Suomalainen et al., 1987) and some automictic lizards in Australia (Moritz, 1983). If different ploidy levels of clones are present, the extent of distribution may be positively correlated with ploidy level. The correlation between ploidy level and extent of distribution observed in, e.g. weevils (Suomalainen et al., 1987; Stenberg et al., 2000), is not included in Vandel's definition of geographical parthenogenesis and we have chosen to call this special form of distribution geographical polyploidy (Stenberg et al., 2003).
Of the 900 insect species known to have asexual forms, almost 200 are weevil species (Normark, 2003). Apomictic parthenogenesis (true clonality) is the most common form of asexual reproduction in insects and has been found in most major insect orders. All asexual weevils studied so far are apomictic and lack meiosis and recombination (Suomalainen et al., 1987). One particularly well-studied genus of weevils that has more than 60 known parthenogenetic forms is Otiorhynchus (e.g. Mikulska, 1960; Suomalainen et al., 1987; Normark, 2003). Most of these parthenogenetic forms are major or minor agricultural and forest pests, e.g. the vine weevil, O. sulcatus (Fabricius, 1775), the clay coloured weevil, O. singularis (Linnaeus, 1767) and the strawberry root weevil, O. ovatus (Linnaeus, 1758) (Palm, 1996). The genus has been studied since the beginning of the last century (e.g. Penecke, 1922; Székessy, 1937; Suomalainen et al., 1987). The species with clonal forms typically show geographical parthenogenesis and if several ploidy levels are present, geographical polyploidy. Most ancestral sexual forms are assumed to be known and their distribution is limited to the eastern Alps (Jahn, 1941), while the surrounding areas are populated by all female populations. An interesting feature of the areas where many of the sexuals reside is that all are forest refugia. These are small moist areas that were not glaciated during the last Ice Age, and within these areas different plants and animals are believed to have overwintered. Most forest refugia are valleys characterized by a distinct flora that remained free of ice even if the surroundings were glaciated (Ellenberg, 1986). Thorough allozyme investigation of many species in the genus Otiorhynchus has been performed (e.g. Suomalainen & Saura, 1973; Saura et al., 1976a, 1976b).
The species that received most attention in the population genetic surveys was Otiorhynchus scaber (Linnaeus) 1758. Otiorhynchus scaber lives on spruce roots and has four different forms that show geographical polyploidy, namely diploid sexuals and diploid, triploid and tetraploid clones (Stenberg et al., 2003). The diploid forms have a very limited distribution in the Austrian and Slovenian Alps. The triploids are found in the central European mountain range and the tetraploids have conquered almost all European spruce forests (Fig. 1). Clonal diversity is high in the central area and decreases towards the margins (Stenberg et al., 2000). Both parthenogenesis and polyploidy have originated at least three times and the Austrian and Slovenian sexuals are separated with about 5% sequence divergence into two distinct mitochondrial lineages. Otiorhynchus scaber is sluggish, flightless and considered a minor forest pest.
Based on the results of extensive studies on O. scaber (Stenberg et al., 1997, 2000, 2003) we will in this paper discuss: (1) the possible mechanisms involved in the transition to clonality; (2) three different hypotheses that may explain the distribution of clones; (3) that when clonal forms originate in this species several ploidy levels can be created almost simultaneously; and (4) the influence of clonal reproduction on sexual evolution.
Materials and methods
From the data presented in Stenberg et al. (2003) the two populations of O. scaber that were most extensively sampled and contained both sexuals and diploid clones (Plesh and Mozirje) were selected and reanalysed. Using nona vs. 2.0 (Goloboff, 1998) in the Winclada vs. 0.99m24 (Nixon, 1999) software, we examined the mtDNA relationships between 41 taxa. Seventeen were sexually reproducing O. scaber specimens and 21 are clonal. Two taxa were tetraploid clones of O. nodosus (Müller 1764) from Sweden, and the last was a sexual diploid of O. singularis collected in middle Europe. The three partial mtDNA markers, COI, COIII and CytB (see Stenberg et al., 2003) contained 113 informative characters resolving the ingroup when all sequence sites with four or more ambiguities had been removed (total sequence was 1132 bases). ‘Most parsimonious cladograms’ (MPR) were generated with a heuristic search approach (100 000 maximum trees, 30 replications with 20 starting trees per replication and multiple tree bisection and reconnection), all unsupported branches were collapsed and suboptimal trees discarded. Characters were optimized using acctran optimization and evaluated on a subset of the MPRs. A majority rule consensus, with 80% cut-off level (only branches that are present in more than 80% of the MPRs are represented in the consensus), was also created to illustrate the consistent information from the MPRs.
PCR using Wolbachia-specific 16S rRNA primers similar to those used by O'Neill et al. (1992) was performed on 14 diploid specimens from Austria (population Plesch) and 14 diploids from Slovenia (population Mozirje). Five males and nine females were selected from both populations (identical to the individuals used in Stenberg et al., 2003). The two primers used were: W16S-F (5′-TTGTAGCYTGCTATGGTATAACT-3′) and W16S-R (5′-GAATAGGTATGATTTTCATGT-3′). PCR was performed using Ready-to-go PCR beads (Amersham Pharmacia Biotech Inc., Denver, USA) under the following conditions: one cycle of 95 °C for 1 min, 35 cycles of 95 °C for 30 s, 52 °C for 30 s and 74 °C for 1 min and one last cycle of 74 °C for 5 min. Samples were then run on a 2% agarose (Shelton Scientific Inc., Shelton, USA) gel to detect presence of PCR products.
Two bands were cut out from the agarose gel and the DNA purified using JETsorb (Genomed GmbH, Löhne, Germany). The products were then sequenced using DYEnamic ET terminator kit (Amersham Pharmacia Biotech Inc.) and an ABI Prism 377 apparatus (Perkin Elmer, Wellesley, USA). Sequences were then used as query in a BLAST search on NCBI.
After hard collapsing of unsupported nodes and deletion of suboptimal trees the heuristic search resulted in 288 equally parsimonious cladograms with a length of 188, an ensemble Consistency index (Ci) of 74 and an ensemble Retention index (Ri) of 95 (data not shown). The resulting majority rule consensus is illustrated in Fig. 2.
Wolbachia sp. was detected in four out of five males and two out of two sexual females from Plesch but only one out of five males and none out of five sexual females from Mozirje. None of the diploid clones harboured Wolbachia. The 16S rRNA primers amplified a region in these individuals that after sequencing was confirmed to be from Wolbachia. A comparison of the sequences to the public database allowed us to further classify them as representatives of a Wolbachia type B strain.
Mechanisms involved in the origin of clonality
The transition from sexual reproduction with meiosis to apomictic reproduction with mitotic production of eggs is in general seen as a complex process (White, 1973), and has only been observed cytologically. We believe that what appears to be mitosis actually could be meiosis II. It is only the first meiotic division that can easily be distinguished. The second meiotic division is, in essence, a mitotic division. If the first meiotic division is inhibited and the egg starts to divide without fertilization, apomixis would then be the result. Interestingly, Seiler (1947) observed what he interpreted to be vestiges of the first meiotic division in Otiorhynchus oocytes. In animals, including humans, it is also fairly common to find that unfertilized eggs start to divide (e.g. Kaufman, 1983). To avoid changes in ploidy level in the offspring, the inhibition of meiosis I and the ability to initialize egg division without fertilization must occur more or less simultaneously (Van Dijk & van Damme, 2000), despite the fact that evidence found in apomictic plants that the two processes are not genetically linked (Roche et al., 2001). The frequency of transitions from sexuality to clonality (e.g. Normark, 2003) and the fact that most cases of cyclical parthenogenesis (e.g. monogonont rotifers, aphids and cladocerans) involve an alternation between sex and apomictic parthenogenesis (Suomalainen et al., 1987) suggests that the process need not be so complex.
Wolbachia has been shown to have a number of effects on insect reproductive systems (Werren, 1997). These intracellular bacterial parasites are known to cause asexuality in hymenopterans. However, the hymenopteran wasps normally produce haploid males parthenogenetically and the asexual forms caused by Wolbachia are automictic (reviewed in Stouthamer, 1997). Wolbachia are known to infect many different species of insects (Werren et al., 1995) but have never been shown to induce apomictic parthenogenesis (Normark, 2003). Therefore, the importance of our finding that O. scaber carries these parasites is unknown. Wolbachia was almost exclusively found in sexual specimens from Plesch (only one incidence in Mozirje). It is unlikely that Wolbachia or a similar endoparasite (Zchori-Fein et al., 2001) plays a role in the origin of clones in O. scaber, because it is not more common in Mozirje, where most of the recent transitions to clonality seem to have occurred (Stenberg et al., 2003).
Distribution of clones
In our study we did not observe clones clustering close to the sexuals from the same site. There is, for example, almost 6% sequence divergence between the clones and the sexuals collected in Mozirje (Stenberg et al., 2003). Based on a divergence time of 2% per million years (Brower, 1994; Ribera et al., 2001), the last common ancestor of these forms existed approximately 3 million years ago. It is also striking that all clones from Mozirje cluster closer to the Plesch sexuals, and that most Plesch clones cluster closer to the Mozirje sexuals, than to the sexuals from the same collecting site. There are three different hypotheses that may explain this pattern.
The distribution of O. scaber must have been oscillating with the expansion and retraction of the Pleistocene glaciations. The sexuals from Austria and Slovenia have probably been isolated from each other for a long time. They must have overwintered, at least the latest Ice Age, in different locations. The Austrian refugia have evidently been very small and the sexuals must have shared them with a number of clones (Stenberg et al., 2000). In Slovenia the situation was probably very different, because the southern parts of Slovenia were not glaciated. The Slovenian habitat is relatively open and the sexual populations seem to be more extensive than in Austria. We believe that the populations in Slovenia migrated south during glaciations and then back north, up the mountains, in the warmer periods. In the interglacials the two major mitochondrial lineages may have had overlapping distributions, creating mixed populations. Subsequently, in the cold periods there must have been strong lineage sorting. If the two major mitochondrial lineages independently have given rise to clones, lineage sorting could create the cluster pattern we observe. However, if lineage sorting were the main force behind the extant distribution, we should expect to find clones in Mozirje that cluster in the same lineage as the sexuals of the same population, as lineage sorting is expected to have been weaker in the large Slovenian refugia.
Another explanation may be that the small morphological differences observed between Austrian and Slovenian sexual populations (Braun, 1992), and the mtDNA divergence, could indicate the presence of two distinct species. If both species produce clones, the distributional pattern could be explained by both species showing classical geographical parthenogenesis. If the clonal forms of these two species were forced to marginal areas of the distribution of the sexuals (as discussed by Vandel, 1928, 1940) it could fit our data where clones found in, e.g. Mozirje are actually more related to the Plesch population. However, both clones collected in Plesch that cluster closer to Plesch sexuals than Mozirje sexuals and the O. nodosus specimen nested deep within the ingroup argue against this hypothesis.
The third explanation, which we find more plausible, is that the distribution of clones in the cladogram in relation to their geographical origin and the presence of the two very distinct mitochondrial lineages suggests a hybrid origin. The two mitochondrial lineages may represent two isolated conspecific taxa that occasionally have overlapping distributions between the Ice Ages. Genetical differences created during the glaciations may not be so large that complete species barriers have emerged; sporadic matings could then result in the creation of hybrid clones. This view agrees with Hewitt's (1999) study of the European biota, where he found the Alps to be a major hybridization zone in the periods between Ice Ages. The similar distribution of O. nodosus and O. scaber makes hybridization events possible, and may explain the presence of O. nodosus in the ingroup.
Suomalainen (1940) suggested that diploid parthenogenesis originates first and that polyploidy results from accidental fertilizations of diploid asexual eggs by sperms of males belonging to related species. One of Suomalainen's (1940) cytological observations strongly supports the hybridization theory as well as the ability of males to occasionally fertilize clonal eggs. Suomalainen found polyploids with more than one metaphase plate in a single egg in the oogenesis. It is unlikely that discrete metaphase plates would be present in a polyploid that is not of hybrid origin and has not acquired its higher ploidy level with the help of unrelated males. The presence of more than one metaphase plate may be predicted in the cell division, if a male fertilizes a parthenogenetic egg. The main reason for this prediction is that if weevil sperm contributes the centrosome, as in most known animal cases (Sluder et al., 1993), additional fertilizations by males will introduce additional centrosomes and possibly also metaphase plates. Even though these phenomena are poorly understood in insects, the presence of two spindle systems has also been observed in human (Simerly et al., 1999) and sea urchin (Holy & Schatten, 1991) eggs fertilized by two sperms at the same time. In conclusion we believe that some, if not all, clones are the result of hybridization between the two different lineages observed in our study (that may or may not deserve species status).
The large and unresolved clade containing most of the clones collected in Plesch contains all three ploidy levels. Most individuals differ by only one or a few nucleotide substitutions and some are identical. All these clones must have a common sexual ancestor. The fact that these clones have not diverged significantly since their origin indicates that the different ploidy levels originated more or less simultaneously. This pattern is also found in other clades when additional taxa are included (Stenberg et al., 2003). Although our mitochondrial data are not sufficient to understand the cause of this phenomenon, this may indicate that the ploidy level is initially unstable in a clonal lineage.
It is still possible that some changes in ploidy level are the result of non-disjunctions of metaphase plates, but if so, most non-disjunctions should lead to a higher ploidy level. If the ploidy level became lower through non-disjunction we would expect to find occasional triploid individuals or populations within tetraploid areas. One interpretation of this is that changes in ploidy level are caused by chance fertilizations by males. This interpretation is supported by the pattern of geographical polyploidy, where all forms of clones seem to be produced in the sexual areas. It can further be assumed that if non-disjunctions are fairly rare, most of them should lead to higher ploidy levels, as clones will be accumulating recessive deleterious mutations that will have a high risk of being unmasked if ploidy level is lowered.
Causes for clonal success and evolutionary implications
All the results indicate that polyploid clones are superior colonisers when compared with diploid clones and sexuals. All forms of O. scaber that are now found in northern Europe have over-wintered the latest Ice Age in refugia in the Alps or similar areas. When the ice retracted, a uniform environment surrounding the refugia was most likely created. Clonal forms of weevils must have had a huge advantage in colonizing such areas and forming stable populations compared to the sexuals, both because of the higher reproductive output and because it must have been difficult for males to find sexual female colonisers in an area with high clonal density. It has been argued that in species with poor dispersal abilities, asexuality is favoured (Lynch, 1984) and virtually all asexual insects, including O. scaber, are flightless (Suomalainen et al., 1987). In addition, there seems to be a trade-off between wings and female fecundity, with females without wings producing more offspring (Wagner & Liebherr, 1992). Comparing clones with different levels of ploidy, there are several aspects connected to higher ploidy level that may have played a role in allowing the tetraploids to expand further than the triploids, e.g. larger size and more favourable general purpose genotypes (Lynch, 1984).
Clonal populations are usually dense because of their faster reproduction (compared to sexually reproducing forms). Therefore, competition is probably fierce within the clonal community. If a clone is slightly better adapted to its environment it will, most likely, swiftly become dominant in the population. That is probably one reason why clonal diversity declines towards the margins of geographical parthenogenesis, even if dispersal would be common. The constant production of clonal lineages, and possible fertilization of clonal eggs, keeps diversity high in the central areas. As all clones originate from the sexuals, most will have the ability to survive in the central area, whereas only some have the ability to colonize the marginal areas. New generations of clones from sexuals are not the only source of variation in a clonal population. There is evidence of adaptation (Stenberg et al., 1997) and observed cases of ‘niche shifts’ in clonal pests of Otiorhyncus, e.g. O. sulcatus (Palm, 1996). Hybridization combined with adaptation could be the reason behind evident cases of weevil pests with clonal forms that have mastered a variety of novel environments.
Clones, even if they only persist only for a limited time, may also be an important factor in the evolution of related sexual species. When new clones appear they may be equally well adapted to their habitat as the sexual ancestors and share the same or overlapping niches. This might force the sexual populations to evolve and expand their niches. If abundant clones then die out because of mutational breakdown, parasite pressure or inability to cope with a changing environment there will most likely be a relaxed selective pressure on the sexual population. These shifts in clonal abundance could be an important evolutionary force on sexual species where clonal forms are common. It is possible that the constant pressure of clones, together with hybridization events, in part, drives the high diversity seen, for example, in weevils.
The differences between sexual and asexual weevils, clones with different levels of ploidy, and other insects with clonal forms are of interest not only from an evolutionary view, but also for pest management. Our studies show that the complexity in clonal evolution and advantages of polyploidy have been underestimated. The rapid infestation of new areas and obvious adaptation of clones to new habitats (e.g. new host plant species) increases the demand for knowledge of clonal pests. Understanding that not all ‘Vine Weevils’ are the same might be increasingly important in identifying infection routes and engineering targeted pest control, when more and more pesticides are being banned.
This study was supported by the Nilsson–Ehle Foundation, Philip Sörensen Foundation, Lawski foundation and the J. C. Kempe Foundation. We would like to thank Prof. Anssi Saura for help and inspiration with the manuscript and Kerstin Kristiansson for laboratory assistance.