Primer design and introns in the barcode region
Four sets of barcode primers were designed and successfully tested during our experiments. The primer pair AHyFu-F and AHyFu-R was successful for amplifying and sequencing COX1 for 24 of 29 Fusarium strains tested. These primers were designed to amplify a 567-bp fragment of COX1 but yielded PCR fragments of three sizes: ~570 bp (no introns), ~2000 bp and ~3000 bp (introns present). Sequencing of Fusarium strains revealed introns at three of the known intron positions (3, 4, 11, see Fig. 1); at two positions two different sequences types were present (4a, 4b and 11a, 11b). In our initial survey, the COX1 of most Fusarium strains had either no introns present, or introns 3 and 4b (Figs 2–4, Table 2).
Figure 2. Gene tree of Fusarium COX1 from intron 4b to the end of the barcode only with Clonostachys rosea as outgroup. The neighbour-joining tree is on the left and the consensus maximum parsimony tree on the right (CI = 0.60; RI = 0.84; 138 most parsimonious trees length = 203). Roman numerals on the consensus tree indicate the main clades I–IV, each of which may represent a distinct COX1 copy. Names in bold text indicate species that occur in multiple clades. Side panels show heterozygous bases detected in barcode sequences of two species. An asterisk indicates heterozygous bases in the sequence, suggesting paralogous copies. Superscript codes on each strain refer to barcode type: 1 = 567 bp code obtained with the barcode primers AHyFu-F and AHyFu-R; a and b denote different sequences from different amplification replicates; 2 = intron 3 and 4b present but removed for this analysis, sequences obtained with barcode primers; 3 = Intron 4b present, obtained with primers Fus-I4b-F and AHyFu-R; 4: Intron 11a present, sequence obtained with barcode primers; 5 = Intron 11b present, sequence obtained barcode primers. The final gapless alignment was 411 bp, with 63 parsimony informative characters.
Download figure to PowerPoint
Figure 3. Gene tree of intron 3 recovered from Fusarium species using primers Fus-I3-F and Fus-I3-R, rooted with Fusarium delphinoides. The neighbour-joining tree is on the left and the consensus maximum parsimony tree on the right (CI = 0.86; RI = 0.96; 208 most parsimonious trees length = 34). The side panel shows a heterozygous base pair detected in the intron of one strain. Although this intron was also recovered in some sequences using the barcode primers, it is unproven that all sequences in this tree actually occur in the COX1 gene; some could represent mobile mitochondrial introns. The alignment had 290-bp positions, of which 12 were parsimony informative.
Download figure to PowerPoint
Figure 4. Gene tree of intron 4b recovered from Fusarium species using primers Fus-I4b-F and AHyFu-R, intron segment only, rooted with Fusarium delphinoides. The neighbour-joining tree is on the left and the consensus maximum parsimony tree on the right (CI = 0.84; RI = 0.93; 28 most parsimonious trees length = 27). Because the reverse primer is located in the COX1 barcode, these introns definitively occur in the COX1 gene. The alignment had 198-bp positions, of which 15 were parsimony informative.
Download figure to PowerPoint
Table 2. Summary of the number of different COX1 copies detected in individual strains of Fusarium sequenced in this study. Asterisks indicate the sequence had heterozygous bases, indicating the existence of additional multiple copies. Sequences which originally produced a barcode with no introns and later amplified intron 3 are labeled with a ^; or later amplified intron 3 and 4b are labeled with a °
| ||Barcode copies||No introns||Intron|
|Fusarium acuminatum BBA 62149||1|| ||1||1|| || || |
|Fusarium avenaceum BBA 71782||1|| ||1||1|| || || |
|Fusarium cf. avenaceum BBA 63784||2||1°||1||1|| || || |
|Fusarium babinda DAOM 235678||3|| ||2*|| || ||1|| |
|Fusarium boothii DAOM 235624||2||1°||1||1|| || || |
|Fusarium circinatum DAOM 235752||2||2|| || || || || |
|F. circinatum DAOM 235753||2||2|| || || || || |
|F. circinatum DAOM 235758||4||2*^||2*|| || || || |
|Fusarium coeruleum BBA 64400||1||1|| || || || || |
|Fusarium concolor DAOM 235731||1|| ||1||1|| || || |
|Fusarium delphinoides DAOM 235647||2||1°||1||1|| || || |
|Fusarium diversisporum BBA 63606||1|| ||?||1|| || || |
|Fusarium equiseti BBA 67786||1|| ||1||1|| || || |
|Fusarium flocciferum BBA 64535||1|| ||1||1|| || || |
|Fusarium cf. heterosporum DAOM 235643||1|| ||1||?|| || || |
|Fusarium incarnatum BBA 68459||1|| ||1||1|| || || |
|Fusarium lateritium BBA 62244||2|| ||1|| ||1|| ||1|
|F. lateritium var. buxi BBA 64798||1|| ||1||1|| || || |
|Fusarium pallidoroseum BBA 69792||1|| ||1||1|| || || |
|Fusarium sacchari DAOM 235795||3||2*^||1|| || || || |
|Fusarium solani DAOM 235651||2|| ||1|| || || ||1|
|Fusarium torulosum BBA 64465||2||1^||1|| || || || |
|Fusarium cortaderiae DAOM 235621||3||2°||1||1|| || || |
|Fusarium graminearum DAOM 235800||2||1^||1|| || || || |
Fusarium circinatum and F. sacchari had heterozygous bases in the barcode sequences, indicating more than one COX1 copy in these putatively haploid strains (Fig. 2, Table 2). This, along with the presence of additional PCR products of sizes that would indicate the presence of introns (i.e. ~2000 or 3000 bp), led us to design intron-specific primers for the two most commonly found introns (3 and 4b). We used these to test for introns in Fusarium strains where they were not seen in the initial barcode amplifications. The intron primers targeted a 337-bp fragment of intron 3 (Fus-I3-F and Fus-I3-R) and a second primer (Fus-I4b-F) paired with the barcode reverse primer (AHyFu-R) to yield a 724-bp fragment starting in intron 4b. When these primers were tested on Fusarium genomic DNA that had previously yielded no introns, eight of the nine samples yielded intron-containing amplicons, suggesting the presence of multiple copies of COX1 (Table 2). It is possible that the intron 3 primers amplified copies of the intron outside of the COX1 gene. However, the primer designed for intron 4b, used in conjunction with the reverse barcode primer, would amplify only copies of this intron associated with COX1; four of nine samples yielded this product (Table 2).
We observed that when intron 3 was present, intron 4b was usually also present (Table 2). Although data were missing on intron 4b for some strains, none were confirmed as missing this intron when intron 3 was present. Some strains, such as Fusarium circinatum (DAOM 236752, 235753, 235758) may have additional copies, because multiple bands of variable length were seen in PCR amplifications. Two of these three strains (DAOM 235752, 235753) yielded different barcodes during replicate amplifications, while the final strain (DAOM 236758) had multiple heterozygous bases in the one isolated barcode (Table 2). Some strains had other COX1 copies with less common introns, Fusarium babinda (DAOM 235678) was the only strain to contain intron 11a; F. solani (DAOM 235651) and the F. verticilloides genome contained intron 11b. Fusarium lateritium (BBA 62244) had intron 11b as well as intron 4a, an intron otherwise only known in Podospora anserina (Tables 1 and 2). Some strains, such as F. babinda (DAOM 235678) and F. circinatum (DAOM 235758), had heterozygous bases in intron 3, suggesting multiple copies of that intron (Table 2, Fig. 3).
Outside the genus Fusarium, the forward primers AHyFu-F, AHyMe-F and AHyLe-Fa paired with AHyFu-R, were able to amplify Acremonium cf. chrysogenum (DAOM 226667), Metarhizium anisopliae (DAOM 237735), Monocillium mucidum (DAOM 226847), Beauveria bassiana (DAOM 233520, 210087), Lanatonectria flocculenta (DAOM 229273) Emericellopsis minima DAOM (226707), Hypocrea jecorina (DAOM 232048), and Neonectria ditissima (KAS 2832). All yielded a barcode with no introns, with the exception of L. flocculenta (DAOM 229273), which had introns 3 and 4b present (Table 1). The barcode sequence lacking introns for Hypocrea jecorina (DAOM 232048) confirmed that this species has two potential barcodes, because a second fragment with two introns was reported previously (Table 1; Chambergo et al. 2002).
The primer pair Pez-F and Pez-R was tested on a relatively small number of members of the fungal subdivision Pezizomycotina, but was successful at amplifying the barcode region of COX1 from the following member of the class Sordariomycetes: Acremonium murorum var. felina (DAOM 226750), Clonostachys rosea (DAOM 226795), C. compactiuscula (DAOM 226738; all Hypocreales, Bionectriaceae), Gliocladium viride (DAOM 226717; Hypocreales, Hypocreaceae), Stilbella aciculosa (DAOM 165520; Hypocreales, Nectriaceae), Arthro-botryum hyalospora (JCM 3809; Chaetosphaeriales), and Monilochaetes infuscans (CBS 379.77; Glomerellales). All yielded PCR products and barcode sequences of the expected length of c. 650 bp, none of which had introns.
Neighbour-joining and maximum parsimony analyses of the barcode sequences obtained for the Fusarium species are shown in Fig. 2. This tree cannot be interpreted as a species tree, because individual strains occur repeatedly in different clades. There appear to be at least four major copies of COX1 in this tree (indicated by the Roman numerals I–IV), and possibly more. Our interpretation is that each major clade represents a distinct, homologous COX1 copy. This is supported by the fact that barcodes without introns (superscript 1 in Fig. 2) group together and barcodes with intron 4 (superscripts 2 and 3) group together on the tree in clades I and II. Oddly enough, clade III showed identical barcodes in both Fusarium cf. avenaceum (BBA 63784) and F. cortaderiae (DAOM 235621), one of which has no introns and the other has at least intron 4 present. It is difficult to interpret the significance of the smaller clades, whether they represent sporadically sampled copies, or phylogenetic signal from one copy overlaid on the multicopy tree.
The low barcoding utility of this data was evident by the relative lack of phylogenetic structure within each of the major clades. For example, clade I included four species with identical COX1 barcodes; they are classified in three different taxonomic sections (Gerlach & Nirenberg 1982), viz. Fusarium boothii and Fusarium graminearum in section Discolor (= F. graminearum species complex), Fusarium circinatum in section Liseola (= Gibberella fujikuroi species complex) and Fusarium oxysporum in section Elegans (= F. oxysporum species complex, O'Donnell et al. 2007). Clades II, III and IV included representatives of two different sections each.
The species resolution observed with intron 3 (Fig. 3, 304 bp) and intron 4b (Fig. 4, 198 bp) was similar to that seen with the exonic barcoding region (Fig. 2). With intron 3, the occurrence of two species from section Liseola (= Gibberella fujikuroi species complex), F. circinatum and F. sacchari, in two clades, and the relatively distant placement between the closely related species F. babinda and F. concolor, suggest the possible existence of paralogues of this intron. The distantly related species Fusarium equiseti and F. torulosum shared a common intron 3 sequence. Relatively few copies of intron 4b were recovered, but the inferred phylogeny was relatively congruent with the RPB2 phylogeny of O'Donnell et al. (2007), except that F. cortaderiae (DAOM 235621) was out of place. The phylogeny of intron 3 (Fig. 3) matched that of the partial barcode sequences with superscripts 2 and 3 (Fig. 2) for many strains but included additional strains for which the equivalent barcode was missing in Fig. 2. For example, the three members of the F. graminearum species complex, F. graminearum, F. cortaderiae and F. boothii formed a well-supported clade in Fig. 3. But F. cortaderiae occurred only in clade III in Fig. 2, suggesting that it may have barcodes yet to be isolated that would group in clades I and II.