| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Systematics |
2Department of Biology, University of Joensuu, P.O. Box 111, FIN-80101 Joensuu, Finland; 3Department of Conservation Biology and Genetics, Evolutionary Biology Centre, Uppsala University, Norbyvägen 18 D, 752 36 Uppsala, Sweden
Received for publication September 11, 2003. Accepted for publication June 3, 2004.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Key Words: ADH • Betula • Betulaceae • birch • matK • phylogeny
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
). The genus Betula belongs to the birch family, Betulaceae, which is placed in the order Fagales. The genus is included in one of the two clades or subfamilies of Betulaceae, Betuloidae, which includes the genera Alnus and Betula (Furlow, 1990
; Bousquet et al., 1992
). The genus is taxonomically problematic (Schilling, 1989
), and it has been considered to be the most complex of all circumpolar genera (Hultén, 1971
). The number of accepted species ranges from about 30 to over 60 species (Furlow, 1990
; Atkinson, 1992
; de Jong, 1993
). The family Betulaceae (Bousquet et al., 1992
; Kato et al., 1998
; Chen et al., 1999
) and the tribe Betuleae (Savard et al., 1993
) have been the subject of molecular studies, but the phylogeny of the genus Betula has received less attention.
Regel (1865)
, the original monographer of Betula, divided birches into four main series, namely the Albae, Costatae, Acuminatae, and Nanae series. Ever since this division, different attempts to identify sections or subgenera and the relationships among the Betula species have been made on the basis of morphology, biochemical characters, and/or chromosomal numbers (Winkler, 1904
; Komarov, 1936
; Pawlowska, 1983
; de Jong, 1993
; Julkunen-Tiitto et al., 1996
; Keinänen et al., 1999
). De Jong (1993)
suggests that species of the subgenus Betulenta (e.g., B. alleghaniensis Britt. and B. lenta L.) are most closely allied to birches from the Eocene. The subgenera Betulaster (e.g., B. maximowicziana Reg.) and Neurobetula (e.g., B. ermanii Cham. and B. schmidtii Reg.), which is a rather heterogeneous and partly artificial group, are regarded as closely related to the subgenus Betulenta. The homogenous subgenera Betula (e.g., B. pendula Roth., B. papyrifera Marsh., B. platyphylla Suk. var. japonica (Miq.) Hara, B. populifolia Marsh., B. pubescens Ehrh., and B. resinifera Britt.) and Chamaebetula (e.g., B. fruticosa Pall., B. nana L., and B. humilis Schrenk) are seen as derived from ancestors related to the subgenus Neurobetula. The results of Keinänen et al. (1999)
, based on secondary metabolites, supported the classification presented by de Jong, except for B. schmidtii, which exhibited a flavonoid composition similar to white birches (subgenus Betula) and B. pubescens, which contained several flavonoids not detected in other white birches.
The difficulties associated with Betula phylogeny and taxonomy are, in part, due to extensive hybridization and introgression. Hybridization is common among birches (Tutin et al., 1964
; Flora of North America Editorial Committee, 1997
; Jonsell, 2000
; Thórsson et al., 2001
), and it is therefore expected that introgression may have played an important part in the evolution of the genus (Alam and Grant, 1972
; Furlow, 1990
; Atkinson, 1992
). The basic chromosome number of Betula is n = 14. Species of Betula form a polyploid series, with chromosome numbers of 2n = 28, 56, 70, 84, and 112 (Furlow, 1990
). Polyploidy is found in species of various subgenera, which indicates that several independent polyploidizations have occurred within the genus. A combination of both chloroplast and nuclear DNA sequences can provide complementary information on the evolution of the genus. Analysis of the chloroplast genome will provide information about the maternal evolutionary relationships between the species, and the biparentally inherited nuclear DNA should provide independent data from which to infer evolutionary. relationships
In the present study both coding and noncoding nuclear DNA and chloroplast DNA sequences were used to investigate the phylogeny of the genus Betula. Alcohol dehydrogenase genes (ADH) encoding enzymes related to anoxia and glycolysis are among the best-characterized nuclear genes in plants. ADH is an essential enzyme in anaerobic metabolism and its expression increases under oxygen stress as well as in response to cold stress in Zea mays and Arabidopsis thaliana and to dehydration in A. thaliana (Freeling and Bennett, 1985
; Dolferus et al., 1994
). In the majority of flowering plants that have been studied, both monocots and dicots, two or three ADH loci have been identified, each containing 10 exons and nine introns (Dennis et al., 1984
; Gaut and Clegg, 1991
, 1993
; Morton et al., 1996
; Sang et al., 1997
; Gaut et al., 1999
). In A. thaliana and Arabis gemmifera, however, a single ADH gene is present, which consists of seven exons and six introns (Chang and Meyerowitz, 1986
; Miyashita et al., 1996
). Phylogenetic studies of ADH genes have been made at both high and low taxonomic levels (see Cohn and Moore, 1988
; Yokoyama and Harry, 1993
; Sullivan et al., 1994
; Russo et al., 1995
; Betrán and Ashburner, 2000
). In plants, phylogenetic relationships between grasses and palms (Morton et al., 1996
) and also within many plant families and genera (Gaut and Clegg, 1991
, 1993
; Sang et al., 1997
; Charlesworth et al., 1998
; Miyashita et al., 1998
; Gaut et al., 1999
) have been analyzed using ADH sequences.
The matK gene is located within the intron of the trnK gene, which in tobacco is located to the large single-copy region of the chloroplast DNA molecule (Shinozaki et al., 1986
). It is a maturase-coding gene, which seems to splice not only the intron of trnK but also other introns (Vogel et al., 1999
). DNA sequences from matK have been used for estimating phylogenies of vascular plant families, such as Pinaceae (Wang et al., 2000
) and Taxaceae (Cheng et al., 2000
), and it has also been successfully used within genera like Juglans (Stanford et al., 2000
), Chrysosplenium (Soltis et al., 2001
), and Lycium (Fukuda et al., 2001
). In many studies matK has proven to be more variable and contain more phylogenetic information than other chloroplast genes (Olmstead and Palmer, 1994
; Liang and Hilu, 1996
; Manos and Steele, 1997
; Hu et al., 2000
).
The species for this study were chosen from all the major subgenera or sections (Regel, 1865
; Winkler, 1904
; de Jong, 1993
) of the genus and from all three major parts of the Betula range: Europe, Asia, and North America (Table 1). Specifically we wanted to address the following questions: (1) Are the chloroplast- and nuclear-based sequences congruent? (2) Are the DNA-based phylogenies congruent with phylogenies based on morphological or biochemical properties? (3) Are the polyploid species in the subgenus Betula evolved from one origin or has the polyploidization occurred several times independently in the evolution of the subgenus?
|
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
) were collected either from individuals growing in botanical gardens, from natural populations of Betula pendula, or obtained from experiments located at the research station of the Finnish Forest Research Institute at Punkaharju, Finland (Table 1).
The seven individuals used to investigate the within-species variation in B. pendula have been analyzed with chloroplast polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP) in a phylogeographic study of this species (Palmé et al., 2003
). The individuals included here were chosen because they have different chloroplast haplotypes (identified with PCR-RFLP), and come from different geographic regions. The samples were frozen in liquid nitrogen and stored at –80°C.
The Corylus avellana individual used in the ADH analysis was sampled in Joensuu Botanical garden, Finland. The two C. avellana individuals used as outgroup for the matK analysis were sampled in Halltorps Hage, Sweden, and Montejo de la Sierra, Spain, and used in a phylogeographic study of C. avellana (Palmé and Vendramin, 2002
). These particular individuals were chosen since they belong to the two major chloroplast haplotypes found in C. avellana. Ideally we would have preferred to use a member of the genus Alnus as an outgroup since it is, according to earlier studies (Bousquet et al., 1992
; Chen et al., 1999
), more closely related to Betula than Corylus, but we were unable to amplify the desired sequences.
Isolation of the genomic clone of BpADH and Southern hybridization
A genomic clone of BpADH (Betula pendula ADH) was isolated from the lambda FixII genomic library (M. Kiiskinen and J. Kangasjärvi, University of Helsinki, Helsinki, Finland, personal communication) using the partial, 32P-labeled cDNA of BpADH (nucleotides 55–417, accession number AJ279698, received from M. Korhonen, University of Helsinki) as a probe in screening. The filters were hybridized at 42°C using a low temperature hybridization solution and washed to a final stringency of 0.1 x saline-sodium citrate (SSC)/ 0.5% sodium dodecyl sulphate (SDS) at 37°C. The genomic clone of BpADH was isolated, subcloned using pGEM-T EasyVector System II (Promega, Madison, Wisconsin, USA) and sequenced. The genomic sequence of BpADH is available from EMBL Nucleotide Sequence Database (accession number AJ549107).
Southern blots were prepared by standard methods (Sambrook et al., 1989
). DNA was isolated from leaves of B. pendula by using the Dneasy Plant Mini kit (QIAGEN, Valencia, California, USA), digested with EcoRI, BamHI, HindIII, AccI, and EcoRV restriction enzymes and 10 µg of total DNA was loaded per lane. The same probe was used as for the screening of genomic library. The filter was hybridized at 65°C (Church buffer, Church and Gilbert, 1984
) for 16 h and washed to a final stringency of 0.1 x SSC/1% SDS at 65°C.
Isolation and sequencing of ADH alleles
Total DNA for both ADH and matK analyses was extracted from young leaves using Dneasy Plant Mini kit (QIAGEN) or with a protocol from Doyle and Doyle (1990)
modified according to Palmé et al. (2003)
. The primers for the ADH alleles were designed based on the genomic clone of BpADH, and were as follows: 5'-GCACCACCACAAGTAGGTGAAG-3' (forward) and 5'-AATCTTGAAGCCCCAGCAATCC-3' (reverse). The PCR was performed with 30 cycles of 30 s at 95°C, 45 s at 59°C (53°C for C. avellana), and 90 s at 68°C, followed by 5 min at 68°C. The PCR reactions were conducted using Expand enzyme mix (Boehringer Mannheim, Mannheim, Germany), composed of Taq and Pwo (proofreading activity) DNA polymerases. The Pwo polymerase can reduce error frequency by a factor of 10 compared to Taq polymerase. Our own data shows no PCR mistakes made by Expand in another birch gene, BpMADS2 (about 3000 base pairs [bp]), and this suggests (based on Poisson distribution of errors) that the 95% upper bound for the error rate is 1 x 10–3 (Järvinen et al., 2003
).
After amplification, PCR-products were cloned using pGEM-T EasyVector System II (Promega, Madison, Wisconsin, USA), and the positive clones were selected using Easypreps (Berghammer and Auer, 1993
) and digestion (EcoRI and HindIII). Both strands were sequenced using primers 300 bp apart on average. Sequencing was conducted on an ABI Prism 310 sequencer (Applied Biosystems, Foster City, California, USA). All sequence polymorphisms were visually rechecked from the chromatograms. The reliability of the sequences was also rechecked with direct sequencing. The sequencing reaction follows the dideoxy chain termination method directly using PCR products as template. For direct sequencing, the PCR products were purified using QIAquick PCR purification kit (QIAGEN). On the average, about 500 bp were rechecked from every species with this method. The sequences are available from the EMBL Nucleotide Sequence Database (accession numbers from AJ535640 to AJ535656, ALIGN_000515).
DNA isolation and sequencing of matK alleles
Four pairs of primers were used to amplify four overlapping DNA segments comprising the matK gene as well as part of both upstream and downstream flanking regions. First, DNA from the seven B. pendula individuals was amplified using primers from Demesure et al. (1995)
, Grivet and Petit (2002)
, and Hu et al. (2000)
: K1-matK1, matK2-matK4L, matK4-trnK5, and matK6-K2. The PCR reactions were conducted using Taq DNA polymerase (Fermentas, Hanover, Maryland, USA). A PCR program with either 48°C or 50°C as annealing temperature was used: 94°C 4 min, 35 times of 94°C 45 s, 48/50°C 1 min, 72°C 1 min 30 s, and the final step 72°C 10 min. Double strand sequencing of the PCR products was performed by Genome Express (Meylan, France).
The overlap between the last two fragments was small and the ends of the sequences were of insufficient quality to make a reliable alignment between the matK4-trnK5 and matK6-K2 fragments. To increase the general overlap between the separate fragments and to improve the quality of the PCR product the following primers were used for the other Betula species: K1-matK7, matK2-matK4L, matK4-trnK685F, and matK6-K2 (Steele and Vigialys, 1994
; Demesure et al., 1995
; Hu et al., 2000
; Grivet and Petit, 2002
). The PCR protocol described above was used also here except that all amplifications were done with annealing temperature of 48°C. The resulting PCR product was either cleaned with a QIAquick PCR Purification Kit (QIAGEN) (the matK2-matK4L fragment) or with a QIAquick Gel Extraction Kit (QIAGEN) (the K1-matK7, matK4-trnK685F, matK6-K2 fragments). The sequencing reactions of the PCR products were performed on a PTC-100 Programmable Thermal Controller (MJ Research, Waltham, Massachusetts, USA) with a heated lid. The analysis of the sequencing fragments was done in the BM-unit (Lund, Sweden) at Lund University. The sequences are available from EMBL Nucleotide Sequence Database (accession numbers from AY372008 to AY372027, AY373441, AY373442).
Analysis of the ADH sequences
Nucleotide and amino acid sequences of ADH were analyzed using GCG program package release 10.0, program PileUp (Genetic Computer Group, Madison, Wisconsin, USA). Both nucleotide and amino acid alignments were constructed with Genedoc (Nicholas and Nicholas, 1997
) and ClustalX (Thompson et al., 1997
) programs and refined visually. Amino acid sequences in data banks (Table 2) were used in alignments together with the sequence of BpADH. The parsimony trees (Eck and Dayhoff, 1966
; Fitch, 1971
) of amino acid sequences were constructed using the programs available in Phylip 3.5 (Felsenstein, 1993
).
|
). The presence of recombination and/or gene conversion among ADH sequences was tested with the program Geneconv version 1.81 (Sawyer, 1989
, 1999
). Given aligned DNA sequences, Geneconv looks for aligned segments for pairs of sequences that are sufficiently similar to suggest past gene conversion or recombination.
The phylogenetic trees of the nuclear DNA sequences were inferred using two methods: maximum parsimony and maximum likelihood as implemented in the program Phylip 3.5 (Felsenstein, 1993
). In the maximum parsimony program all characters were equally weighted (a character-based algorithm). In the maximum likelihood program the transition/transversion ratio was set to 2.0 and the base frequencies were estimated from the real nucleotide frequencies in the data set. The reliability of the trees was tested using bootstrapping with 100 replicates for maximum likelihood and 1000 for maximum parsimony.
Phylogenetic analysis of B. papyrifera and two other species, B. nana and B. pubescens, was somewhat problematic because of the presence of long indels (from position 66 to 525) in three alleles. Maximum likelihood ignores indels, whereas parsimony method can accommodate them. While it is unsatisfying to ignore some variation, it can also be problematic to include variation that is difficult to interpret. If indels are viewed as character states originating from particular and well-defined biological events, then they can be assumed to contain historical information suitable for phylogenetic analysis and ignoring them may yield misleading topologies (Giribet and Wheeler, 1999
). For example, analysis of three different nuclear genes in bumble bees (66 species, 23 subgenera) shows that indels can be a reliable source of phylogenetic information (Kawakita et al., 2003
). However, because the mechanism leading to indels is often unknown, it is not yet obvious how those should be taken into account. In the present study, three different parsimony analyses were carried out. In the first one, each base pair within an indel was considered as a character, in the second one indels were ignored, and in the third one each indel was considered as a single character.
Analysis of chloroplast DNA sequences
The chloroplast DNA sequences were edited and assembled into one continuous sequence per individual with BioEdit 5.0.9 (Hall, 1999
). The resulting 20 sequences were aligned using ClustalX (Thompson et al., 1997
) and all the polymorphic sites were rechecked from the chromatograms. The computer package DnaSP 3.51 (Rozas and Rozas, 1999
) was used to calculate nucleotide diversity and haplotype diversity. Three data sets were used in the analysis: the first one includes all the sequenced individuals, the second one includes only one individual per species to make the comparison with the ADH data set more appropriate, and the third data set includes only the B. pendula individuals. The B. pendula sequence included in the second data set was of haplotype 1 (see Supplementary Data accompanying the online version of this article).
The phylogenetic trees of the chloroplast DNA sequences were inferred using same two methods as for the nuclear DNA sequences. The reliability of the trees was tested using bootstrapping with 100 replicates for maximum likelihood and 500 for maximum parsimony.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
) and Z. mays ADH1 (Dennis et al., 1984
). The structure of the isolated gene resembled a typical ADH gene structure (Fig. 1) with 10 exons and nine introns (Dennis et al., 1984
; Gaut and Clegg, 1991
, 1993
; Morton et al., 1996
; Sang et al., 1997
; Gaut et al., 1999
). Comparison of the coding sequence of BpADH at amino acid level with other sequences revealed rather high similarity to AtADH with 80% identity and 86% similarity, to PhADH1 (Petunia hybrida) with 81% identity and 86% similarity and to MdADH (Malus domestica) with 85% identity and 90% similarity.
|
|
Nucleotide variation in the ADH gene
The region of the ADH gene that was sequenced covers portions of five exons (exons 2–6) and four introns (introns 2–5). With the primers used, an 
1070-bp fragment was amplified from 11 birch species. From B. pubescens and B. papyrifera two fragments were amplified, one being 1070 bp, the other about 1500 bp. From B. nana an 1500-bp fragment was amplified. The length variation was due to one long indel (from positions 66–525 bp) in the first intron region used. The length of the fragment amplified from C. avellana was 973 bp.
Eighty-two sites, out of 1037, were variable (see Supplementary Data accompanying the online version of this article). Of these, 44 sites were phylogenetically informative. In addition to this, 25 sites contained length variation. The length variation as well as 33 parsimony-informative polymorphic sites and 29 singletons were located to the introns. Twenty polymorphic sites, 11 out of which were parsimony informative, six nonsynonymous, and 14 synonymous, were located in the coding region. The nucleotide diversity, 
, varied from 0.0211 (SD 0.00193, data set including long alleles of B. papyrifera and B. pubescens) to 0.0218 (SD 0.00170, data set including short alleles of B. papyrifera and B. pubescens). All the 16 genes sequenced were different, and thus the haplotype diversity was 1.00 (SD 0.022).
For the long alleles of B. nana, B. papyrifera and B. pubescens, the nucleotide diversity, , was 0.0190 (SD 0.00754). In addition to 30 singletons, the long indel contained two indels (from 88 to 111 and from 121 to 135) separating B. papyrifera from B. nana and B. pubescens. Outside the long indel, the sequence identity between the short and the long alleles within the same species, including both coding and noncoding regions, was high. The coding regions of the two different B. papyrifera alleles were, for instance, completely identical.
Significant evidence (P = 0.01) for gene conversion/recombination was only found between the short alleles of B. pubescens and B. pendula and between the short alleles of B. pubescens and B. schmidtii.
Nucleotide variation in matK
In the chloroplast DNA fragment nucleotide variation was limited. A total of 10 sites were polymorphic for point mutations (of which four were polymorphic only within B. pendula) and three additional sites contained length variation in the analyzed fragment that varied from a total of 2431 bp to 2445 bp (see Supplementary Data accompanying the online version of this article). The length variation, as well as four of the point mutations, was located to the noncoding flanking regions while six mutations were found in the part coding for the matK protein. Of these six mutations, three were synonymous (966, 1497, 1699) and three nonsynonymous (890, 1015, 1701). The nucleotide diversity, , was 0.00076 (SD 0.00023) for the whole data set and 0.00080 (SD 0.00026) if only one individual per species is included. Within B. pendula was 0.00047 (SD 0.00017).
Seven haplotypes were identified in the 20 investigated individuals when length variation was considered (see Supplementary Data accompanying the online version of this article). Four of these haplotypes were found within B. pendula and the most common haplotype in B. pendula (haplotype 1) was also found in most of the other species. Betula maximowicziana has haplotype 5 that differs from haplotype 1 by only one point mutation. Haplotypes 6 and 7 were very similar to each other but differ from haplotype 1 by five point mutations. Two of these differences, at site 1015 and 2397, were shared with the Corylus outgroup. These haplotypes were found in three American species: B. alleghaniensis, B. papyrifera, and B. lenta. The haplotype diversity, if the size variation is excluded, is 0.574 (SD 0.122) for the whole data set or 0.473 (SD 0.136) if only one individual per species is included. Within B. pendula the haplotype diversity was 0.714 (SD 0.181).
Phylogenetic trees
Phylogenetic analyses of nuclear and chloroplast DNA sequences of the 14 Betula species were performed and strict consensus trees were constructed. The nuclear ADH variation with 44 parsimony informative characters divided the genus Betula into three main groups: B. fruticosa, B. humilis, B. ermanii, and the short allele of B. pubescens formed the first group; B. maximowicziana, B. lenta, and B. alleghaniensis the second one; and the remaining species the third group (Figs. 3–5). Both maximum parsimony (MP) and maximum likelihood (ML) methods recovered the same three main groups when indels were ignored or considered as a single character, although the support for different groups differed among methods. When each base pair within an indel was considered as character the only notable difference between the maximum likelihood tree (ML tree) and maximum parsimony trees (MP tree) was the placement of the long allele of B. papyrifera, which in the MP tree clustered with the other long alleles (B. nana and B. pubescens; Fig. 3) and in the ML tree with the short allele of B. papyrifera (Fig. 4).
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
ADH = 0.0211–0.0218 compared to cp = 0.00080). Although the matK gene has proven to be variable in many other genera (Stanford et al., 2000
; Fukuda et al., 2001
; Soltis et al., 2001
), variation among the birch species studied here was limited. More variation might have been expected, since B. pendula showed within species variation in chloroplast PCR-RFLP analysis (Palmé et al., 2003
). Also the seven B. pendula individuals, which were chosen for this study based on different chloroplast haplotypes, showed some variation at nucleotide level ( = 0.00047).
There were only five parsimony-informative sites in the matK fragment in the 14 Betula species studied here, and the phylogenetic tree of matK (Fig. 6) clearly shows that the haplotypes can only be divided into two well-supported groups: one including the three American species and the other containing the remaining species. The reason for this low level of variation in the matK region might be that the chloroplast haplotypes of Betula have a recent common ancestor. It could also be due to directional selection on the chloroplast DNA, but our data are obviously too limited to say anything meaningful about the presence of selection on cpDNA. Finally, low variation could be due to a particularly low mutation rate. Interestingly, low levels of variation were also found in matK within Corylus (seven informative characters in seven sequences; Palmé and Vendramin, 2002
). Kato et al. (1998)
found 25 phylogenetically informative nucleotide sites in the matK fragment in all the six genera of Betulaceae (six species). Also the results with the rbcL gene indicate that the chloroplast genome evolves slowly in Betulaceae (Bousquet et al., 1992
). The chloroplast genome is inherited clonally and generally maternally in angiosperms (Radetzky, 1990
; Rajora and Dancik, 1992
; Dumolin et al., 1995
) and is therefore expected to have a smaller effective population size (Ne) than nuclear genes. In monoecious species, such as Betula, Ne for chloroplast genes is expected to be half of that for nuclear genes and the level of genetic variation is therefore expected to be smaller. However, the difference here far exceeds the expected difference. The fact that the substitution rates in chloroplast genes are generally much lower than in nuclear genes (Gaut et al., 1996
; Li, 1997
; Small et al., 1998
) could at least explain part of the difference.
Phylogeny of genus Betula
The results obtained from ADH sequences were not in complete agreement with the subgeneric classifications based on morphological characters (Regel, 1865
; Winkler, 1904
; de Jong, 1993
; Figs. 3– 5). Betula schmidtii (subgenus Neurobetula) and B. nana (subgenus Chamaebetula) grouped with the species in subgenus Betula, and B. ermanii (subgenus Neurobetula), grouped with species in subgenus Chamaebetula, B. humilis and B. fruticosa.
Betula schmidtii is considered to be a rather peculiar species among birches because of the blackness of its bark, hard and heavy wood, and slow growth (Ashburner, 1980
). The ADH sequence of B. schmidtii resembled closely that of B. pendula and other white birches (subgenus Betula), and preliminary results with two other genes support the placement of B. schmidtii into subgenus Betula (P. Järvinen and T. Sopanen, University of Joensuu, Joensuu, Finland). Also the flavonoid profiles of B. schmidtii support the placement of B. schmidtii into subgenus Betula (Wollenweber, 1975
; Keinänen et al., 1999
). The flavonoid composition of the leaves of B. schmidtii resembled closely those of white birches (Keinänen et al., 1999
).
Based on its shrubby dwarf phenotype, B. nana is usually placed in the subgenus Chamaebetula (or in the series Humiles), along with other dwarf birches, B. humilis and B. fruticosa. In this study however, B. nana was clustered with B. pubescens, within the subgenus Betula. A direct connection between B. nana and B. pubescens was already proposed in 1920s (Woodworth, 1929
). Betula nana and B. pubescens are known to hybridize quite extensively (Ashburner, 1980
; Anamthawat-Jónsson and Tomasson, 1990
; Jonsell, 2000
; Thórsson et al., 2001
; Palmé et al., 2004
), and their flavonoid profiles are very similar, whereas flavonoid profiles of B. nana and B. humilis buds diverge considerably (Wollenweber, 1975
).
According to de Jong (1993)
, the subgenera Betulenta and Betulaster are closely related to each other and contain old species, which are closely allied to the birches from Eocene. Our study with ADH gene supports the assumption of close relatedness of these two subgenera: B. maximowicziana, B. lenta, and B. alleghaniensis group together (Figs. 3–5). This grouping, however, might partly be artificial, because only one species of the subgenus Betulaster was used in this analysis.
Origin of the polyploid birches
From two of the polyploid species, B. pubescens and B. papyrifera, two versions of the ADH gene were amplified, and these different alleles can potentially be useful in deducing their origin. There are several hypotheses about the origin of B. pubescens, generally involving B. pendula, but it was also suggested that one of its parents might be B. humilis (Walters, 1968
; Howland et al., 1995
). The clustering of the short allele of B. pubescens with B. humilis supports this hypothesis. The long allele of B. pubescens clusters with B. nana suggesting a common ancestor to these two species. However, an alternative explanation could be introgression between B. nana and B. pubescens, two species that are known to hybridize extensively (Tutin et al., 1964
; Jonsell, 2000
; Thórsson et al., 2001
).
Betula papyrifera has been considered to be the result of synthesis of a pendula-like species and an autotetraploid of other origin (Johnsson, 1949
). In the MP-tree (Fig. 3), when gaps within indels were considered as characters, the long allele of B. papyrifera clusters with other long alleles, while the short allele of B. papyrifera clusters with B. pendula and the other diploid "wart-birches" suggesting allopolyploid origin of this species. Preliminary results with another birch gene, BpMADS2, support this assumption (P. Järvinen and T. Sopanen, unpublished data). On the other hand, the fact that in the ML tree (Fig. 4) the two B. papyrifera ADH alleles cluster together would indicate that B. papyrifera is an autopolyploid. However, in MP and ML trees the bootstrap values for clustering of the short allele of B. papyrifera with "wart-birches" are low. If all branches with <50% bootstrap support are collapsed, this short allele cannot be clustered with confidence to either of these two groups. The chloroplast haplotype of B. papyrifera clusters with B. lenta and B. alleghaniensis, and although it seems likely that this haplotype has an introgression origin, it could also mirror one of the parent species of the polyploid B. papyrifera.
The tetraploid B. ermanii is phenotypically extremely variable (Ashburner, 1980
). At the nucleotide level, its ADH allele closely resembled the short allele of B. pubescens and of the members of subgenus Chamaebetula. Also the flavonoid profiles of the buds of B. ermanii and B. pubescens were almost identical (Wollenweber, 1975
) supporting the idea of at least one common ancestor of these two species, potentially B. humilis (Walters, 1968
).
There is no correspondence between the ADH phylogeny and polyploidy levels. Both diploids and polyploids are present in all major groups. If the ADH gene tree mirrors the species tree, this should indicate that several independent polyploidizations have occurred within the genus. Similar results were obtained with the genus Fraxinus (Oleaceae), where polyploidy appeared to have evolved several times independently (Jeandroz et al., 1997
).
Reconciling gene trees to a species tree
Phylogenetic trees of the ADH and matK genes used in this study do not give fully congruent results. Incongruence between nuclear and cytoplasmic markers has often been reported (e.g., Soltis et al., 1996
; Erdogan and Mehlenbacher, 2000
; Semerikov et al., 2003
), and there can be many reasons for this incongruence, for example, lineage sorting, introgression, horizontal transfer, paralogous genes, and interlocus interactions (Wendel and Doyle, 1998
). Recombination can be a severe problem when reconstructing phylogenetic trees, as it can create "mixed" sequences. Basing a phylogeny on such sequences can result in an incorrect tree topology (Wendel and Doyle, 1998
). Evidence for recombination in ADH has been found in a selfer, A. thaliana (Innan et al., 1996
). Also gene conversion can result in phylogenetic incongruence, and some indications of gene conversion have been found in ADH among some grass species (Gaut et al., 1999
) but no evidence of it was found in peonies (Sang et al., 1997
). In this study, Sawyer's statistical test only detected evidence of gene conversion between B. pendula, B. pubescens, and B. schmidtii. It should, however, be noted that methods to detect gene conversion or recombination are not very powerful (Posada and Crandall, 2001
; Wiuf et al., 2001
). Finally, an additional problem when studying duplicated genes is the danger of inadvertently including paralogous genes resulting in a gene tree that reflects the duplication of the gene rather than a possible species tree. The fact that the coding regions of the short and long alleles of B. papyrifera are identical and noncoding sequences almost identical (see Supplementary Data accompanying the online version of this article) strongly suggest that these two sequences represent two different alleles of the same ADH gene, not two different ADH genes.
In the present case cytoplasmic introgression seems to be the most likely explanation for the differences between phylogenetic trees of ADH and matK genes, although the extensive hybridization in the genus Betula makes it difficult to separate introgression from polyploidization, as examples such as B. papyrifera and B. pubescens clearly demonstrate. Cytoplasmic introgression has proven to be common among other birch species (Palmé et al., 2004
). The three species in group B in the phylogenetic tree of matK (Fig. 6) are all American species with overlapping geographical distributions, so there has been plenty of opportunity for hybridization. Also, both morphological data and the phylogenetic tree of ADH indicate that B. papyrifera belongs to the white birches (subgenus Betula). Although hybrids between B. alleghaniensis and B. papyrifera have been identified (Flora of North America Editorial Committee, 1997
), the present data do not permit an inference regarding the origin of the introgression of B. papyrifera.
When studying molecular phylogenies it is important to keep in mind that an individual gene tree does not necessarily mirror the species tree (Pamilo and Nei, 1988
). Processes such as those mentioned above can frequently cause incongruences among different gene trees and the species tree. In a genus that is known for its high levels of hybridization and introgression, such as Betula, transfer of genes across the species boundaries is undoubtedly most extensive. If such an introgressed allele is sequenced instead of one of the "original" alleles this will of course affect the structure of the gene tree and normally it will not mirror the majority of the other genes in the species. If introgression is widespread then bifurcating trees may simply no longer reflect the evolutionary process, which becomes graph-like. This could potentially be the case for several Betula species.
Conclusions
Here we have made an effort to investigate the phylogeny within the genus Betula. As many other studies based on morphology, biochemical characters, and chromosome numbers have shown, the relationships among the Betula species are complex. This is probably at least partly due to hybrid introgression resulting in the transfer of genes and therefore also other characteristics between species. Our study is a first step towards understanding the relationship among genes in different Betula species. To further understand the relationship among the species, a larger number of unlinked genes has to be investigated. Indeed, if introgression is as extensive as some evidence suggests (Thórsson et al., 2001
), it might not be possible to represent the true species phylogeny as a tree at all but as a graph representing multiple transfers of genes.
|
FOOTNOTES |
|---|
4 E-mail: Pia.Jarvinen{at}joensuu.fi
; FAX: +358 13 251 3590
|
LITERATURE CITED |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Anamthawat-Jónsson K. T. Tomasson 1990 Cytogenetics of hybrid introgression in Icelandic birch. Hereditas 112: 65-70[CrossRef][ISI]
Ashburner K. 1980 Betula—a survey. Plantsman 2: 31-54
Atkinson M. D. 1992 Betula pendula Roth (B. verrucosa Ehrh.) and B. pubescens Ehrh. Journal of Ecology 80: 837-870[CrossRef]
Berghammer H. B. Auer 1993 "Easypreps": fast and easy plasmid minipreparation for analysis of recombinant clones in E. coli. Biotechniques 14: 524, 528
Betrán E. M. Ashburner 2000 Duplication, dicistronic transcription and subsequent evolution of the alcohol dehydrogenase and alcohol dehydrogenase-related genes in Drosophila. Molecular Biology and Evolution 17: 1344-1352
Bousquet J. S. H. Strauss P. Li 1992 Complete congruence between morphological and rbcL-based molecular phylogenies in birches and related species (Betulaceae). Molecular Biology and Evolution 9: 1076-1088[Abstract]
Bucher M. K. A. Brander S. Sbicego T. Mandel C. Kuhlmeier 1995 Aerobic fermentation in tobacco pollen. Plant Molecular Biology 28: 739-750[CrossRef][ISI][Medline]
Chang C. E. M. Meyerowitz 1986 Molecular cloning and DNA sequence of the Arabidopsis thaliana alcohol dehydrogenase gene. Proceedings of the National Academy of Sciences, USA 83: 1408-1412
Charlesworth D. F. L. Liu L. Zhang 1998 The evolution of the alcohol dehydrogenase gene family by loss of introns in plants of the genus Leavenworthia (Brassicaceae). Molecular Biology and Evolution 15: 552-559[Abstract]
Chen Z. D. S. R. Manchester H. Y. Sun 1999 Phylogeny and evolution of the Betulaceae as infered from DNA sequence, morphology, and paleobotany. American Journal of Botany 86: 1168-1181
Cheng Y. R. G. Nicolson K. Tripp S. M. Chaw 2000 Phylogeny of Taxaceae and Cephalotaxaceae genera inferred from chloroplast matK gene and nuclear rDNA ITS region. Molecular Phylogenetics and Evolution 14: 353-365[CrossRef][ISI][Medline]
Church G. M. W. Gilbert 1984 Genomic sequencing. Proceedings of the National Academy of Science, USA 81: 1991-1995
Cohn V. H. G. P. Moore 1988 Organization and evolution of the alcohol dehydrogenase gene in Drosophila. Molecular Biology and Evolution 5: 154-166[Abstract]
de Jong P. C. 1993 An introduction to Betula: its morphology, evolution, classification and distribution, with a survey of recent work. In D. Hunt [ed.], Proceedings of the IDS Betula symposium, 2–4 October 1992. International Dendrology Society, Richmond, UK
Demesure B. N. Sodzi R. J. Petit 1995 A set of universal primers for amplification of non-coding regions of mitochondrial and chloroplast DNA in plants. Molecular Ecology 4: 129-131[Medline]
Dennis E. S. W. L. Gerlach A. J. Pryor J. L. Bennetzen A. Inglis D. Llewellyn M. M. Sachs R. J. Ferl W. J. Peacock 1984 Molecular analysis of the alcohol dehydrogenase (Adh1) gene of maize. Nucleic Acids Research 12: 3983-4000
Dolferus R. M. Jacobs W. J. Peacock E. S. Dennis 1994 Differential interactions of promoter elements in stress response of the Arabidopsis Adh gene. Plant Physiology 105: 1075-1087[Abstract]
Doyle J. J. J. L. Doyle 1990 Isolation of plant DNA from plant tissue. Focus 12: 13-15
Dumolin S. B. Demesure R. J. Petit 1995 Inheritance of chloroplast and mitochondrial genomes in pedunculate oak investigated with an efficient PCR method. Theoretical Applied Genetics 91: 1253-1256
Eck R. V. M. O. Dayhoff 1966 Atlas of protein sequence and structure. National Biomedical Research Foundation, Silver Spring, Maryland, USA
Erdogan V. S. A. Mehlenbacher 2000 Phylogenetic relationships of Corylus species (Betulaceae) based on nuclear ribosomal DNA ITS region and chloroplast matK gene sequences. Systematic Botany 25: 727-727[CrossRef][ISI]
Felsenstein J. 1993 Phylip (Phylogeny Inference Package) version 3.5. Computer program and documentation distributed by the author. Website: http://evolution.genetics.washington.edu/phylip/software.pars.html#PHYLIP
Fitch W. M. 1971 Toward defining the course of evolution: minimum change for a specified tree topology. Systematic Zoology 20: 406-416[CrossRef][ISI]
Flora of North America Editorial Committee [eds.] 1997 Flora of North America north of Mexico, vol. 3. Oxford University Press, New York, New York, USA
Freeling M. D. C. Bennett 1985 Maize Adh1. Annual Review of Genetics 19: 297-323[ISI][Medline]
Fukuda T. J. Yokoyama H. Ohashi 2001 Phylogeny and biogeography of the genus Lycium (Solanaceae): inferences from chloroplast DNA sequences. Molecular Phylogenetics Evolution 19: 246-258[CrossRef][ISI][Medline]
Furlow J. 1990 The genera of Betulaceae in the southeastern United States. Journal of the Arnold Arboretum 71: 1-67[ISI]
Gaut B. S. M. T. Clegg 1991 Molecular evolution of alcohol dehydrogenase 1 in members of the grass family. Proceedings of the National Academy of Science, USA 88: 2060-2064
Gaut B. S. M. T. Clegg 1993 Molecular evolution of the Adh1 locus in the genus Zea. Proceedings of the National Academy of Science, USA 90: 5095-5099
Gaut B. B. Morton B. C. McCaig M. T. Clegg 1996 Substitution rate comparisons between grasses and palms: synonymous rate differences at the nuclear gene Adh parallel rate differences at the plastid gene rbcL. Proceedings of the National Academy of Science, USA 93: 10274-10279
Gaut B. S. A. S. Peek B. R. Morton M. T. Clegg 1999 Patterns of genetic diversification within the Adh gene family in the grasses (Poaceae). Molecular Biology and Evolution 16: 1086-1097[Abstract]
Gregerson R. M. Mclean M. Beld A. G. Gerats J. Strommer 1991 Structure, expression, chromosomal location and product of the gene encoding ADH1 in Petunia. Plant Molecular Biology 17: 37-48[CrossRef][ISI][Medline]
Giribet G. W. C. Wheeler 1999 On gaps. Molecular Phylogenetics and Evolution 13: 132-143[CrossRef][ISI][Medline]
Good A. G. L. E. Pelcher W. L. Crosby 1988 Nucleotide sequence of a complete barley alcohol dehydrogenase 1 cDNA. Nucleic Acids Research 16, 7182
Grivet D. R. J. Petit 2002 Phylogeography of the common ivy (Hedera sp.) in Europe: genetic differentiation through space and time. Molecular Ecology 11: 1351-1362[CrossRef][Medline]
Hall T. A. 1999 BioEdit: a user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucleic Acids Symposium Serries 41: 95-98
Howland D. R. Oliver A. Davy 1995 Morphological and molecular variation in natural populations of Betula. New Phytologist 130: 117-124[CrossRef][ISI]
Hu J. M. M. Lavin M. Wojciechowski M. Sanderson 2000 Phylogenetic systematics of the tribe Millettieae (Leguminosae) based on chloroplast trnK/matK sequences and its implications for evolutionary patterns in Papilionideae. American Journal of Botany 87: 418-430
Hultén E. 1971 The circumpolar plants. II. Dicotyledons. Kungl. Svenska Vetenskapsakademiens Handlingar, Fjärde Serien, Band 13. Almquist and Wiksell, Stockholm, Sweden
Innan H. F. Tajima R. Terauchi N. Miyashita 1996 Intragenic recombination in the Adh locus of the wild plant Arabidopsis thaliana. Genetics 143: 1761-1770[Abstract]
Järvinen P. J. Lemmetyinen O. Savolainen T. Sopanen 2003 DNA sequence variation in BpMADS2 gene in two populations of Betula pendula. Molecular Ecology 12: 369-384[CrossRef][Medline]
Jeandroz S. A. Roy J. Bousquet 1997 Phylogeny and phylogeography of the circumpolar genus Fraxinus (Oleaceae) based on internal transcribed spacer sequences of nuclear ribosomal DNA. Molecular Phylogenetics and Evolution 7: 241-251[CrossRef][ISI][Medline]
Johnsson H. 1949 Studies on birch species hybrids I. Betula verrucosa x B. japonica, B. verrucosa x B. papyrifera and B. pubescens x B. papyrifera. Hereditas 35: 115-135[ISI]
Jonsell B. [ed.] 2000 Flora Nordica, vol. 1. The Bergius Foundation, Royal Swedish Academy of Sciences, Stockholm, Sweden
Julkunen-Tiitto R. M. Rousi J. Bryant S. Sorsa M. Keinänen H. Sikanen 1996 Chemical diversity of Betulaceae species: comparison of phenolics and terpenoids in northern birch stems. Trees 11: 16-22
Kato H. K. Oginuma Z. Gu B. Hammel H. Tobe 1998 Phylogenetic relationships of Betulaceae based on matK sequences with particular reference to the position of Ostryopsis. Acta Phytotaxonomica et Geobotanica 49: 89-97
Kawakita A. T. Sota J. S. Ascher M. Ito H. Tanaka M. Kato 2003 Evolution and phylogenetic utility of alignment gaps within intron sequences of three nuclear genes in bumble bees (Bombus). Molecular Biology and Evolution 20: 87-92
Keinänen M. R. Julkunen-Tiitto M. Rousi J. Tahvanainen 1999 Taxonomic implications of phenolic variation in leaves of birch (Betula L.) species. Biochemical Systematics and Ecology 27: 243-254[CrossRef]
Komarov V. 1936 Flora of the U.S.S.R., vol. 5. Moskva-Leningrad, Izdatelstvo Akademii Nauk SSSr, Russia
Li W. H. 1997 Molecular evolution. Sinauer, Sunderland, Massachusetts, USA
Liang H. K. W. Hilu 1996 Application of the matK gene sequence to grass systematics. Canadian Journal of Botany 74: 125-134
Llewellyn D. J. E. J. Finnegan J. G. Ellis E. S. Dennis W. J. Peacock 1987 Structure and expression of an alcohol dehydrogenase 1 gene from Pisum sativum. Journal of Molecular Biology 195: 115-123[CrossRef][ISI][Medline]
Manos P. K. Steele 1997 Phylogenetic analyses of "higher" Hamamelididae based on p
50% are indicated next to the relevant nodes (1000 replicates)
