Lineage Specificity of the Evolutionary Dynamics of the mtDNA D-Loop Region in Rodents

This paper reports an intraorder study on the D-loop-containing region of the mitochondrial DNA in rodents. A complete multialignment of this region is not feasible with the exception of some conserved regions. The comparative analysis of 25 complete rodent sequences from 23 species plus one lagomorph has revealed that only the central domain (CD), a conserved region of about 80 bp in the extended termination-associated sequences (ETAS) domain, adjacent to the CD, the ETAS1, and conserved sequence block (CSB) 1 blocks are present in all rodent species, whereas the presence of CSB2 and CSB3 is erratic within the order. We have also found a conserved region of 90 bp located between tRNAPro and ETAS1 present in fat dormouse, squirrel, guinea pig, and rabbit. Repeated sequences are present in both the ETAS and the CSB domain, but the repeats differ in length, copy number, and base composition in different species. The potential use of the D-loop for evolutionary studies has been investigated; the presence/absence of conserved blocks and/or repeated sequences cannot be used as a reliable phylogenetic marker, since in some cases they may be shared by distantly related organisms but not by close ones, while in other ones a relationship between tree topology and presence/absence of such motifs is observed. Better results can be obtained by the use of the CD, which, however, due to its reduced size, when used for tracing a phylogenetic tree, shows some nodes with low statistical support. Received: 26 February 2001 / Accepted: 6 June 2001     

A Survey of the Molecular Evolutionary Dynamics of Twenty-Five Multigene Families from Four Grass Taxa

We surveyed the molecular evolutionary characteristics of 25 plant gene families, with the goal of better understanding general processes in plant gene family evolution. The survey was based on 247 GenBank sequences representing four grass species (maize, rice, wheat, and barley). For each gene family, orthology and paralogy relationships were uncertain. Recognizing this uncertainty, we characterized the molecular evolution of each gene family in four ways. First, we calculated the ratio of nonsynonymous to synonymous substitutions (d _N/d _S) both on branches of gene phylogenies and across codons. Our results indicated that the d _N/d _S ratio was statistically heterogeneous across branches in 17 of 25 (68%) gene families. The vast majority of d _N/d _S estimates were <<1.0, suggestive of selective constraint on amino acid replacements, and no estimates were >1.0, either across phylogenetic lineages or across codons. Second, we tested separately for nonsynonymous and synonymous molecular clocks. Sixty-eight percent of gene families rejected a nonsynonymous molecular clock, and 52% of gene families rejected a synonymous molecular clock. Thus, most gene families in this study deviated from clock-like evolution at either synonymous or nonsynonymous sites. Third, we calculated the effective number of codons and the proportion of G+C synonymous sites for each sequence in each gene family. One or both quantities vary significantly within 18 of 25 gene families. Finally, we tested for gene conversion, and only six gene families provided evidence of gene conversion events. Altogether, evolution for these 25 gene families is marked by selective constraint that varies among gene family members, a lack of molecular clock at both synonymous and nonsynonymous sites, and substantial variation in codon usage. Received: 25 May 2000 / Accepted: 16 October 2000     

Evolutionary Analysis of the ErbB Receptor and Ligand Families

We have compared all available deduced protein sequences of the ErbB family of receptors and their ligands. Analysis of the aligned sequences of the receptors indicates that there are some differences in the receptors that are specific to invertebrates. In addition, comparison of the vertebrate ErbB receptors suggest that a gene duplication event generated two ancestral receptors, the ErbB3/ErbB4 precursor and the ErbB1/ErbB2 precursor. Subsequent gene duplications of these precursors generated the four receptors present in mammals. Analysis of the sequences for the known ligands of the ErbB receptors suggests that the vertebrate ligands segregate into the ErbB1 ligands and the ErbB3/ErbB4 ligands, paralleling the evolution of the receptors; however, it is difficult to ascertain any correlation between the invertebrate and the vertebrate ligands. Even though ErbB3 is kinase-impaired, there is significant conservation of the kinase domain within the vertebrate lineage (human, rat, and F. rubripes), suggesting some function for this domain other than kinase activity, such as mediating protein–protein interactions that are involved in receptor dimerization and/or activation of the kinase domain of the heterodimerization partner. To date, no ligand for ErbB2 has been identified, and comparison of the extracellular domains of ErbB2 reveals two regions that are not conserved across the mammalian species. These two regions of divergence align with sequences in ErbB1 that have been shown to be proximal to the amino-terminus and to the carboxyl-terminal region, respectively, of bound EGF. Further, one of these regions contains an insertion, relative to the other members of the mammalian ErbB family, which might affect the ligand binding site and provide a structural basis for this receptor's apparent inability to bind ligand independently. Received: 8 September 1999 / Accepted: 17 January 2000     

Evolutionary Studies on Uricases of Fungal Endosymbionts of Aphids and Planthoppers

Aphids belonging to the three genera Tuberaphis, Glyphinaphis, and Cerataphis contain extracellular fungal symbionts that resemble endocellular yeast-like symbionts of planthoppers. Whereas the symbiont of planthoppers has a uricase (urate oxidase; EC 1.7.3.3) and recycles uric acid that the host stores, no uric acid was found in Tuberaphis styraci, and its fungal symbiont did not exhibit the uricase activity. However, the fungal symbionts of these aphids, including that of T. styraci, were shown to have putative uricase genes, or pseudogenes, for the uricase. Sequence analysis of these genes revealed that deleterious mutations occurred independently on each lineage of Glyphinaphis and Tuberaphis, while no such mutation was found in the lineage of Cerataphis. These genes were almost identical to those cloned from the symbionts of planthoppers, though the host aphids and planthoppers are phylogenetically distant. To estimate the phylogenetic relationship in detail between the fungal symbionts of aphids and those of planthoppers, a gene tree was constructed based on the sequences of the uricase genes including their flanking regions. As a result, the symbionts of planthoppers and Tuberaphis aphids formed a sister group against those of Glyphinaphis and Cerataphis aphids with high bootstrap confidence levels, which strongly suggests that symbionts have been horizontally transferred from the aphids' lineage to the planthoppers'. Received: 29 March 2000 / Accepted: 31 May 2000     

High Evolutionary Rates in Nuclear Genes of Squamates

We compared nonsynonymous substitution rates (Ka) of nuclear coding genes between four major groups of living sauropsids (reptiles): birds, squamates, crocodiles, and turtles. Since only 9 orthologous genes are known in all the four taxonomic groups, we searched for orthologous genes known in chicken and at least one of any representative of poikilotherm sauropsids. Thus, we analyzed three additional data sets: 28 genes identified in chicken and various squamates, 24 genes identified in chicken and crocodilians, and 20 genes identified in chicken and turtles. To compare nonsynonymous substitution rates between all lineages of sauropsids, we used the relative-rate test with human genes as the outgroup. We show that 22/28 nuclear coding genes of squamates, especially snakes (15/16), have an higher evolutionary rate than those in chicken (in mean, 30–40% faster). However, no such difference is detected between crocodiles, turtles and chicken. Higher substitution rate in squamates nuclear coding genes than in chicken, and probably than in other sauropsids, could explain some of the difficulties in resolving the molecular phylogeny of reptiles. Received: 5 July 2000 / Accepted: 13 February 2001     

Evaluating Evolutionary Divergence with Microsatellites

We report the use of microsatellites (MS) to track the recent evolution of swine. Allelic frequencies for nine MS loci linked on swine chromosome 6 (SSC6) representing four western and one Chinese swine breeds were used to estimate genetic distances and times of breed divergence. A phylogenetic tree was constructed which partitioned into western and Meishan breed branches. Yorkshire and Hampshire breeds exhibited the most recent divergence with a calculated distance of 391 years. The oldest divergence, of 2,227 years, was between Meishan and Hampshire swine. Estimates of breed divergence are consistent with historical records. Additional analysis suggests that polymorphic MS linked on a single chromosome are sufficient to determine evolutionary relationships within a single species. Received: 23 September 1996 / Accepted: 26 April 1997     

Evolutionary dynamics of the human endogenous retrovirus family HERV-K inferred from full-length proviral genomes

Several distinct families of endogenous retroviruses exist in the genomes of primates. Most of them are remnants of ancient germ-line infections. The human endogenous retrovirus family HERV-K represents the unique known case of endogenous retrovirus that amplified in the human genome after the divergence of human and chimpanzee lineages. There are two types of HERV-K proviral genomes differing by the presence or absence of 292 bp in the pol-env boundary. Human-specific insertions exist for both types. The analyses shown in the present work reveal that several lineages of type 1 and type 2 HERV-K proviruses remained transpositionally active after the human/chimpanzee split. The data also reflect the important role of mosaic evolution (either by recombination or gene conversion) during the evolutionary history of HERV-K. Received: 5 February 2001 / Accepted: 22 March 2001     

Molecular Evolutionary Analysis of the Thiamine-Diphosphate-Dependent Enzyme,Transketolase

Members of the transketolase group of thiamine-diphosphate-dependent enzymes from 17 different organisms including mammals, yeast, bacteria, and plants have been used for phylogenetic reconstruction. Alignment of the amino acid and DNA sequences for 21 transketolase enzymes and one putative transketolase reveals a number of highly conserved regions and invariant residues that are of predicted importance for enzyme activity, based on the crystal structure of yeast transketolase. One particular sequence of 36 residues has some similarities to the nucleotide-binding motif and we designate it as the transketolase motif. We report further evidence that the recP protein from Streptococcus pneumoniae might be a transketolase and we list a number of invariant residues which might be involved in substrate binding. Phylogenies derived from the nucleotide and the amino acid sequences by various methods show a conventional clustering for mammalian, plant, and gram-negative bacterial transketolases. The branching order of the gram-positive bacteria could not be inferred reliably. The formaldehyde transketolase (sometimes known as dihydroxyacetone synthase) of the yeast Hansenula polymorpha appears to be orthologous to the mammalian enzymes but paralogous to the other yeast transketolases. The occurrence of more than one transketolase gene in some organisms is consistent with several gene duplications. The high degree of similarity in functionally important residues and the fact that the same kinetic mechanism is applicable to all characterized transketolase enzymes is consistent with the proposition that they are all derived from one common ancestral gene. Transketolase appears to be an ancient enzyme that has evolved slowly and might serve as a model for a molecular clock, at least within the mammalian clade. Received: 13 September 1995 / Accepted: 14 November 1996     

Evolutionary Lability of Context-Dependent Codon Bias in Bacteria

In bacteria, synonymous codon usage can be considerably affected by base composition at neighboring sites. Such context-dependent biases may be caused by either selection against specific nucleotide motifs or context-dependent mutation biases. Here we consider the evolutionary conservation of context-dependent codon bias across 11 completely sequenced bacterial genomes. In particular, we focus on two contextual biases previously identified in Escherichia coli; the avoidance of out-of-frame stop codons and AGG motifs. By identifying homologues of E. coli genes, we also investigate the effect of gene expression level in Haemophilus influenzae and Mycoplasma genitalium. We find that while context-dependent codon biases are widespread in bacteria, few are conserved across all species considered. Avoidance of out-of-frame stop codons does not apply to all stop codons or amino acids in E. coli, does not hold for different species, does not increase with gene expression level, and is not relaxed in Mycoplasma spp., in which the canonical stop codon, TGA, is recognized as tryptophan. Avoidance of AGG motifs shows some evolutionary conservation and increases with gene expression level in E. coli, suggestive of the action of selection, but the cause of the bias differs between species. These results demonstrate that strong context-dependent forces, both selective and mutational, operate on synonymous codon usage but that these differ considerably between genomes. Received: 6 May 1999 / Accepted: 29 October 1999     

Taking Variation of Evolutionary Rates Between Sites into Account in Inferring Phylogenies

As methods of molecular phylogeny have become more explicit and more biologically realistic following the pioneering work of Thomas Jukes, they have had to relax their initial assumption that rates of evolution were equal at all sites. Distance matrix and likelihood methods of inferring phylogenies make this assumption; parsimony, when valid, is less limited by it. Nucleotide sequences, including RNA sequences, can show substantial rate variation; protein sequences show rates that vary much more widely. Assuming a prior distribution of rates such as a gamma distribution or lognormal distribution has deservedly been popular, but for likelihood methods it leads to computational difficulties. These can be resolved using hidden Markov model (HMM) methods which approximate the distribution by one with a modest number of discrete rates. Generalized Laguerre quadrature can be used to improve the selection of rates and their probabilities so as to more nearly approach the desired gamma distribution. A model based on population genetics is presented predicting how the rates of evolution might vary from locus to locus. Challenges for the future include allowing rates at a given site to vary along the tree, as in the ``covarion' model, and allowing them to have correlations that reflect three-dimensional structure, rather than position in the coding sequence. Markov chain Monte Carlo likelihood methods may be the only practical way to carry out computations for these models. Received: 8 February 2001 / Accepted: 20 May 2001     

Reconstructing the Complex Evolutionary History of Hepatitis B Virus

A detailed analysis of the evolutionary history of hepatitis B virus (HBV) was undertaken using 39 mammalian hepadnaviruses for which complete genome sequences were available, including representatives of all six human genotypes, as well as a large sample of small S gene sequences. Phylogenetic trees of these data were ambiguous, supporting no single place of origin for HBV, and depended heavily on the underlying model of DNA substitution. In some instances genotype F, predominant in the Americas, was the first to diverge, suggesting that the virus arose in the New World. In other trees, however, sequences from genotype B, prevalent in East Asia, were the most divergent. An attempt was also made to determine the rate of nucleotide substitution in the C open reading frame and then to date the origin of HBV. However, no relationship between time and number of substitutions was found in two independent data sets, indicating that a reliable molecular clock does not exist for these data. Both the pattern and the rate of nucleotide substitution are therefore complex phenomena in HBV and hinder any attempt to reconstruct the past spread of this virus. Received: 5 December 1998 / Accepted: 23 February 1999     

Evolutionary History of MHC Class I Genes in the Mammalian Order Perissodactyla

We carried out an analysis of partial sequences from expressed major histocompatibility complex (MHC) class I genes isolated from a range of equid species and more distantly related members of the mammalian order Perissodactyla. Phylogenetic analysis revealed a minimum of six groups, five of which contained genes and alleles that are found in equid species and one group specific to the rhinoceros. Four of the groups contained only one, or very few sequences, indicating the presence of relatively nonpolymorphic loci, while another group contained the majority of the equid sequences identified. These data suggest that a diversification of MHC genes took place after the split between the Equidae and the Rhinocerotidae yet before the speciation events within the genus Equus. Received: 17 November 1998 / Accepted: 7 April 1999     

Evolution in a chronic RNA virus infection: selection on HTLV-I tax protein differs between healthy carriers and patients with tropical spastic paraparesis

HTLV-I causes T-cell leukemia and tropical spastic paraparesis (TSP) in a minority of infected people, whereas the majority remain healthy. The virus differs little in sequence between isolates but has been shown to have a quasispecies structure. Using the Nei and Gojobori algorithm, we have shown that the proportion of nonsynonymous to synonymous changes in HTLV-I proviral tax gene sequences from healthy seropositive subjects (Dn/Ds= 0.9 to 1.3) is significantly higher than those from TSP patients (Dn/Ds= 0.3 to 0.6). Here we show that the distinction between healthy seropositives and TSP patients can only be seen with proviral tax sequences, but not with cDNA, the amino-terminal or carboxy-terminal half of tax, or the rex gene. The Dn/Ds ratio of proviral tax sequences was used to analyze two TSP patients with atypical features and to investigate the influence of cytotoxic T cells (CTL) on the viral quasispecies. Received: 18 January 1995 / Accepted: 7 November 1995     

Evolutionary Discontinuity of the Carabine Ground Beetles

The Carabine ground beetles are mostly hind wing-less and cannot fly, so that there is more chance of diversification by geographic isolation compared with winged insects. The relationships between morphological diversification and phylogeny of the ground beetles of the world have been inferred mainly by comparisons of mitochondrial ND5 gene sequences. Based on dating by a mitochondrial DNA ``clock,' it has been deduced that an explosive radiation of the major carabine groups took place 50–40 MYA. This was followed by occasional radiations on various scales, sometimes accompanied by parallel morphological changes. There are also a good number of examples showing that the fundamental morphology has remained unchanged for a long time among geographically isolated populations within the same species. Thus, carabid evolution would have proceeded discontinuously, with phases of rapid morphological change alternating with silent phases. Received: 26 October 2000 / Accepted: 16 February 2001     

Archaeabacterial Seryl-tRNA Synthetases: Adaptation to Extreme Environments and Evolutionary Analysis

The aminoacyl-tRNA synthetases are ubiquitous enzymes which catalyze a crucial step of the cell life, the specific attachment of amino acids to their cognate tRNA. The amino acid sequences of three archaeal seryl-tRNA synthetases (SerRS) from Haloarcula marismortui and Methanococcus jannaschii, both belonging to the group of Euryarchaeota, and from Sulfolobus solfataricus, of the group of Crenarchaeota, were aligned with other eubacterial and eukaryal available SerRS sequences. In an attempt to identify some features of adaptation to extreme environments of these organisms, amino acid composition and amino acid substitutions between mesophilic and thermophilic SerRS were analyzed. In addition, universal phylogenetic trees of SerRS including the three known archaeal sequences, rooted by the threonyl-tRNA synthetases were inferred. Amino acid analyses of the SerRS revealed two ways of adaptation to thermophilic environments between the Eubacteria and the Archaea; most of the usually described amino acid substitutions were nonsignificant in the case of archaeal thermophilic SerRS and most amino acid composition biases seemed to be linked to the genome G+C content pressure. The phylogenetic analysis of the SerRS showed the Archaea to be paraphyletic, H. marismortui emerging with the Gram-positive Bacteria, M. jannaschii being near the root of the tree, and S. solfataricus branching with Eucarya. Received: 30 March 1998 / Accepted: 14 July 1998     

Evolutionary History of 4.5SI RNA and Indication That It Is Functional

To date, the small nuclear 4.5S_I RNA has only been studied in the rat (Rattus norvegicus). Combining PCR and hybridization analyses, we have revealed 4.5S_I RNA homologues sequences in the genomes of four myomorph rodent families (Muridae, Cricetidae, Spalicidae, and Rhizomyidae), and not in other myomorph families (Dipodidae, Zapodidae, Geomyidae, and Heteromyidae) or sciuromorph and caviomorph rodents. By Northern-hybridization, 4.5S_I RNA has been detected in the common rat (R. norvegicus, Muridae), golden hamster (Mesocricetus auratus, Cricetidae), and Russian mole rat (Spalax microphthalmus, Spalacidae), but not in the related great jerboa (Allactaga jaculus, Dipodidae) or in four non-myomorph rodent species tested. cDNA derived from 4.5S_I RNA of M. auratus and S. microphthalmus has been cloned and sequenced. The hamster RNA is found to differ from rat 4.5S_I RNA by only one nucleotide substitution. For the mole rat, two variants of 4.5S_I RNA are detected: short (S) and long (L) with length 101 and 108 nt, respectively. The L variant differs from the S variant as well as from murid and cricetid 4.5S_I RNAs by both a 7 nt insertion and a varying number of nucleotide substitutions. The sequence similarity between the spalacid S-variant and murid/crecitid variants of 4.5S_I RNA is 90%. Judging from species distribution, 4.5S_I RNA genes emerged during the same period of time as the related short interspersed element B2 arose. This occurred after the divergence of Dipodidae lineage but before the branching of Spalicidae/Rhizomyidae lineage from a common myomorph rodent stem. S variant genes seemed to emerge in a common ancestor of spalacids and rhizomyds whereas L variant genes formed in spalacids following the divergence of these two families. The low rate of evolutionary changes of 4.5S_I RNA, at least, in murids and cricetids (6 × 10⁻⁴ substitutions per site per million years), suggests that this RNA is under selection constraint and have a function. This is a remarkable fact if the recent origin and narrow species distribution range of 4.5S_I RNA genes is taken into account. Genes with narrow species distribution are proposed to be referred to as stenogenes. Received: 11 December 2000 / Accepted: 27 August 2001     

Evolutionary History of the 11p15 Human Mucin Gene Family

The four human mucin genes MUC6, MUC2, MUC5AC, and MUC5B are located at chromosome 11p15.5. It has been demonstrated that the three mucins MUC2, MUC5AC, and MUC5B contain several Cys-subdomains of 108 amino acid residues. In contrast, little information is available concerning MUC6. These Cys-subdomains contain 10 cysteine residues that have a highly conserved position. We present here a coherent probable evolutionary history of this human gene family after comparison of the nucleotide sequences of these Cys-subdomains. The three MUC loci MUC2, MUC5AC, and MUC5B may have evolved from a common ancestral gene by two successive duplications. Moreover, we can postulate that MUC5AC and MUC5B have evolved in a concerted manner, while MUC2 has evolved separately. Received: 30 January 1997 / Accepted: 17 April 1997     

Exploring Nonnatural Evolutionary Pathways by Saturation Mutagenesis: Rapid Improvement of Protein Function

Random point mutagenesis does not access a large fraction of protein sequence space corresponding to primarily nonconservative amino acid substitutions. The cost of this limitation during directed evolution is unknown. Random point mutagenesis over the entire gene encoding the psychrophilic protease subtilisin S41 identified a pair of residues (Lys211 and Arg212) where mutations provided significant increases in thermostability. These were subjected to saturation mutagenesis to test whether the amino acids not easily accessible by point mutagenesis provide even better ``solutions' to the thermostabilization challenge. A significant fraction of these variants surpassed the stability of the variants with point mutations. DNA sequencing revealed highly hydrophobic residues in the four most stable variants (Pro/Ala, Pro/Val, Leu/Val, and Trp/Ser). These nonconservative replacements, accessible only by multiple (two to three) base substitutions in a single codon, would be extremely rare in a point mutation library. Such replacements are also extremely rare in natural evolution. Saturation mutagenesis may be used advantageously during directed evolution to explore nonnatural evolution pathways and enable rapid improvement in protein traits. Received: 15 March 1999 / Accepted: 28 June 1999     

Evolutionary histories of highly repeated DNA families among the artiodactyla (mammalia)

Six highly repeated DNA families were analyzed using Southern blotting and fluorescence in situ hybridization in a comparative study of 46 species of artiodactyls belonging to seven of the eight extant taxonomic families. Two of the repeats, the dispersed bovine-Pst family and the localized 1.715 component, were found to have the broadest taxonomic distributions, being present in all pecoran ruminants (Giraffidae, Cervidae, Antilocapridae, and Bovidae), indicating that these repeats may be 25–40 million years old. Different 1.715 restriction patterns were observed in different taxonomic families, indicating that independent concerted evolution events have homogenized different motifs in different lineages. The other four satellite arrays were restricted to the Bovini and sometimes to the related Boselaphini and Tragelaphini. Results reveal that among the two compound satellites studied, the two components of the 1.711a originated simultaneously, whereas the two components of the 1.711b originated at two different historical times, perhaps as many as 15 million years apart. Systematic conclusions support the monophyly of the infraorder Pecora, the monophyly of the subfamily Bovinae (containing the Boselaphini, Bovini, and Tragelaphini), an inability to resolve any interrelationships among the other tribes of bovids, paraphyly of the genus Bos with respect to Bison, and a lack of molecular variation among two morphologically and ecologically distinct subspecies of African buffaloes (Syncerus caffer cafer and S. c. nanus). Cytogenetically, a reduction in diploid chromosome numbers through centric fusion in derived karyotypes is accompanied by a loss of centromeric satellite DNA. The nilgai karyotype contains an apparent dicentric chromosome as evidenced by the sites of 1.715 hybridization. Telomeric sequences have been translocated to the centromeres without concomitant chromosomal rearrangement in Thompson's gazelle. Received: 18 June 1995 / Accepted: 1 September 1995     

Evolutionary History of B1 Retroposons in the Genus Mus

Short interspersed DNA elements (SINEs) amplify by retroposition either by (i) successive waves of amplification from one or a few evolving master genes or by (ii) the generation of new master genes that coexist with their progenitors. Individual, highly conserved, elements of the B1 SINE family were identified from the GenBank nucleotide database using various B1 subfamily consensus query sequences to determine their integration times into the mouse genome. A comparison of orthologous loci in various species of the genus Mus demonstrated that four subfamilies of B1 elements have been amplifying within the last 1–3 million years. Therefore, B1 sequences are generated by coexisting source genes. Additionally, three B1 subfamilies have been concurrently propagated during subspecies divergence and strain formation in Mus, indicating very recent activity of this retroposon family. The patterns of intra- and interspecies variations of orthologous loci demonstrate the usefulness of B1 integrations as a phylogenetic tool. A single inconsistency in the phylogenetic trends was depicted by the presence of a B1 insert in an orthologous locus exclusively in M. musculus and M. pahari. However, DNA sequence analysis revealed that these were independent integrations at the same genomic site. One highly conserved B1 element that integrated at least 4–6 million years ago suggests the possibility of occasional function for B1 integrations. Received: 25 February 2000 / Accepted: 5 June 2000