Fungi belonging to the genus Botryosphaeria Ces. and De Not (Pleosporales, Loculoascomycetes) are plurivorous phytopathogens colonizing a wide range of host plants of agricultural, forestry, ecological and economic importance causing diseases (cankers, gummosis, sudden wilting, root rot and branch and whole tree dieback), as well as rots in pre-harvested fruits. These fungi are distributed worldwide in temperate and tropical climate regions (Jacobs and Rehner, 1998; Alves et al., 2005).
A species of Botryosphaeria described as ligninolytic and characterized
to the species level as B. rhodina has been recognized as having commercial
applications for the bioremediation of polyaromatic hydrocarbons and xenobiotic
compounds (Barbosa et al., 1996; Dekker at al., 2002). This species
also produced an exopolysaccharide (botryosphaeran) when grown on basal medium
(Barbosa et al., 2003) that has potential applications as an immuno-potentiator.
Several aspects complicate species identification in the Botryosphaeria genus. Morphological diversity among the teleomorphs is often insufficient to allow a clear identification of the species. Thus, species identification is based mainly on morphological characteristics of the anamorphs; the most common form of Botryosphaeria found in nature (Jacobs and Rehner, 1998; Denman et al., 2000). These characteristics include shape, pigmentation and size of pycnidia. In addition, colony morphology, conditions for pycnidia formation, chromogenicity and temperature effects on mycelial growth rate have also seen used for species recognition (Jacobs and Rehner, 1998).
The Polymerase Chain Reaction (PCR), have been used extensively in inter-, as well as intra-, specific comparisons of several fungi (Guarro et al., 1999; Crous et al., 2001; Rodrigues et al., 2004; Magnani et al., 2005). Ribosomal DNA (rDNA) sequences are universal and contain conserved and variable regions allowing discrimination of fungi of different taxonomic levels (Guarro et al., 1999). Non-coding regions, such as the Internal Transcribed Spacers (ITS), evolve more rapidly and are consequently more variable than coding regions. These regions are easy to access using universal PCR primers ITS1 and ITS4 (White et al., 1990).
Sequence analysis of the ITS region alone (Jacobs and Rehner, 1998; Denman et al., 2000; Zhou and Stanosz, 2001a; Alves et al., 2004), or in combination with sequences from others regions (Slippers et al., 2004a; Phillips et al., 2005), have made a significant contribution to resolving taxonomic problems in the genus Botryosphaeria and have been used to study relationships among species and to distinguish closely-related Botryosphaeria sp. (Smith et al., 2001; Denman et al., 2003; Alves et al., 2004).
No data has yet been reported for the least-length minimum size of the ITS sequence capable of identifying fungal species with confidence. The uncertainty in the minimum size often necessitates having to sequence the complete ITS region. This exercise is labor intensive and costly involving the utilization of bioinformatic tools for processing and contig assembly before performing comparisons with sequences deposited in public databases.
The objective of this study was to determine a minimal DNA sequence length of the ITS region for the identification of Botryosphaeria species by a method that is rapid and economical, making the technology already available more accessible.
MATERIALS AND METHODS
Download and Selection of ITS Sequences from Genbank Database
The NCBI (National Center for Biotechnology and Information) GenBank database
via the Entrez nucleotide portal was accessed to search for entries under the
terms Botryosphaeria and Internal Transcribed Spacer. The entries found
were downloaded in FASTA format and the DNA sequences processed to remove the
genes surrounding the ITS fragment (18S rDNA and 28S rDNA). Thus, the following
analyses were done with the remaining fragments, which included only the sequences
of the first Internal Transcribed Spacer (ITS1), the complete 5.8S rRNA gene
and the second Internal Transcribed Spacer (ITS2).
Sequences selected were subdivided into separate files for each species belonging to the genus Botryosphaeria and the incomplete sequences ITS1-5.8S-ITS2 discarded. Sequence alignments were performed using BioEdit software (Hall, 1999) and manual adjustments were made where necessary. Alignments were converted into genetic distances by the Kimura two-parameter model and phylogenetic trees constructed using the Neighbor-Joining (NJ) method (Saitou and Nei, 1987) within the Phylogenic Inference Package (PHYLIP version 3.57; Felsenstein, 1993). Two sequences of each
Botryosphaeria species were selected for the analysis, with the exception of B. corticis and B. eucalyptorum. The selection was based upon the distribution of sequences at the tree examining, in particular, those sequences that were more divergent.
For each species the selected sequences were processed to obtain fragments
of different sizes to allow the construction of databases with 50, 100, 150,
200, 250, 300, 350, 400 and 450 bp as well as the complete ITS region (~500
bp). This procedure was undertaken from both sequence extremities, forwards
from the 5 to 3
end of the ITS1 region, increasing the size of the fragment up to 500 bp. The
same procedure was taken in the reverse direction from the 3 to 5 end of the
ITS2 region increasing by 50 bp fragments in direction of ITS1(Fig.
For each species, both selected sequences as well as the DNA fragments with
different sizes were compared with those already deposited in the GenBank database
using BLASTn search tool in default conditions (Altschul et al., 1990).
Species Recognition was Based on Two Criteria (I) Comparison between score average
obtained with a match of the same species (identification zone) with the best
score obtained with the second species matched; (ii) Recognition of the greater
number of GenBank deposited sequences of the querys same species inside
the identification zone. Thus, species identification was considered positive
when the average of the score at the same species was higher than the best score
showed by the second species. Sequences deposited without identification were
Phylogenetic Relationships among the Species by ITS Fragments
Sequence alignments of the complete ITS1-5.8S-ITS2 region, as well as the
fragments with different sizes were used for genetic distance analyses and grouping
using PHYLIP software package, version 3.5c (Felsenstein, 1993) as described
above. Tree robustness was estimated by 1000 bootstrap re-sampling using the
program SEQBOOT (Felsenstein, 1985).
||Schematic representation of the model developed for the attainment
of the ITS fragments used in Botryoshaeria species discrimination
From the GenBank database and following the criteria as described in the materials and methods 23 entries representing 13 Botryosphaeria species were selected (Table 1). Sequence alignments with these entries were performed using the BioEdit software. Length of the complete ITS1-5.8S rRNA gene-ITS2 sequences ranged from 454 bp (B. rhodina) (AY343483) to 504 bp (B. stevensii) (AY236955). The sequences were processed to obtain fragments of different sizes and compared with those in the GenBank database by BLASTn tool. The results obtained with the comparison of the different size fragments set from the 23 selected entries with those already deposited at the GenBank are shown in Table 1. The data revealed that the use of fragments from the 50 first nucleotides of the ITS1 region (50F) of B. corticis, B. corticola, B. rhodina, B. stevensii and B. tsugae, made possible the identification of these species, discriminating them from other Botryosphaeria sp.
An increase in the size of the fragments to 100 bp (100F) made an increment in the number of recognized species, allowing the identification of eight species (Table 1). An increment in the number of recognized species (11 species) could be observed with fragments of 150 and 200 bp (150F and 200F), but based on the score values, only fragments from 200 bp or more demonstrated enough sequence complexity for a meaningful identification.
The use of fragments corresponding to the last 50 nucleotides of the ITS2 region (50R) of B. australis, B. corticis and B. corticola allowed the identification of these species separating them from the other Botryosphaeria (Table 1). An increase in the fragment size to 100 bp (100R) made an increment in the number of recognized species, allowing the identification of eight species (Table 1). A total of 10 species could be identified with fragments of 150 bp (150R) and an increase of the size to 200 bp allowed the discrimination between B. obtusa and Sphaeropsis sapinea; other species belonging to the Botryosphaeriaceae family.
Species identification from the DNA fragments with different sizes was performed
considering the data obtained from the best score, in spite of the sequence
sense (starting from the beginning of ITS1 or from the end of ITS2).
||Discrimination of Botryosphaeria species from different
sized ITS region sequence fragments based on comparison with sequences deposited
in the GenBank database using BLASTn search tool
|(+) Positive species discrimination; (-) Negative species
discrimination, (a) Positive identification related to Botryosphaeriaceae
family, with the exception of Sphaeropsis sapinea, (b-g) Positive
identification related to the Botryosphaeria genus, with exception
of; (b) B. lutea; © B. mamane; (d) B. dothidea;
(e) B. australis; (f) B. ribis; (g) B. parva
Thus, the ability of identification of one fragment only was considered when
the best scores were obtained with the same species studied and when the average
of these score values were higher than that observed for other species, recognizing
the greater number of GenBank deposited sequences.
Thus, in taking the data obtained with the use of fragments with 200 bp (200F and 200R) of both B. rhodina entries (AY343483 and AY612337), the best score obtained was 396 (Table 2). The average of the scores obtained with the B. rhodina entries (27 to 200F and 28 to 200R) was 381 and 375 for the 200F fragments and 391 for 200R. At the same time, those scores were higher than those observed with the second species (200 and 357 for 200F and 200R, respectively). Thus, we consider positively the capacity of both 200 bp (200F and 200R) ITS fragments to recognize and discriminate B. rhodina from the other species deposited at the public nucleotide database. Similar results can be observed with the other species with the exception of B. parva and B. ribis, that could be separated from the other species, but without separation between themselves.
Considering these results, both ITS fragments with 200 bp (200F and 200R) were capable of identifying Botryosphaeria species in comparison with the data deposited at the GenBank. The results with the 200 bp fragments were the same as those observed with the complete ITS region (Table 1).
The 23 entries were aligned and used to construct a phylogenetic tree based on neighbor-joining method considering the complete ITS region. The analyses of datasets show two groups with a strongly supported (bootstrap 1000) separation. The first group was divided into two subgroups, one (bootstrap 999) containing B. parva and B. ribis (which were indistinguishable); B. australis, B. lutea and B. eucalyptorum and another with B. corticis, B. dothidea and B. mamane (bootstrap 1000). The second group (bootstrap 985) was composed of B. corticola, B. obtusa, B. stevensii, B. tsugae and B. rhodina with all species distinguishable.
Neighbor-joining trees were also constructed with 200 bp fragments (200F and
200R) (Fig. 2a-c). The trees topologies
were very similar to that obtained with the complete ITS region. The same groups
can be identified with some changes at the positioning of B. obtusa,
B. tsugae and B. stevensii at the tree constructed with 200F and
B. eucalyptorum and B. tsugae at the tree constructed with the
||Results (best score, average score and next species score)
obtained with the comparison of 200 bp fragments (F and R) from 13 Botryosphaeria
species from the GenBank database using BLASTn search tool
|Last verification at GenBank in 07/27/05, ND: Not Determined
||Phylogenetic relationships among Botryosphaeria species
based on neighbor-joining analyses of the complete ITS region DNA sequence
data (a), 200 bp F fragments (b) and 200 bp R fragments (c). Bootstrap values
(1000 replicates) are indicated at (below) the nodes. The bars represent
the nucleotide changes
The correct identification of Botryosphaeria species is of practical importance in plant pathology, biotechnology and environmental studies. Formerly, the classification systems adopted were based on the phenotypic approach (Guarro et al., 1999) that, in Botryosphaeria species, were based mainly on morphological criteria. However, the phenotypic approach has been largely criticized for its lack of consistency, standardization, stable terminology and for its high subjectivity, still using laborious and time-consuming methods. In addition to the classical methods, the DNA approach has been extensively used in the identification of fungal species, including the Botryosphaeria genus (Smith et al., 2001; Slippers et al., 2004 a, b; van Niekerk et al., 2004).
The group of genes usually used in identification of Botryosphaeria species were those that codified rDNA, mainly the ITS region. This region has been used alone (Jacobs and Rehner, 1998; Denman et al., 2000; Zhou and Stanosz, 2001a; Alves et al., 2004), or in combination with other regions, such as β-tubulin and the EF1-α genes (Slippers et al., 2004a, b; van Niekerk et al., 2004; Phillips et al., 2005).
In this study, we tested the potential of different sized ITS rDNA sequences to recognize and discriminate the many species from the Botryosphaeria genus, aiming to determine the minimal DNA sequence of the ITS region useful for species identification.
Table 1 shows that ITS sequences sized with at least 200 bp (F and R) made possible the identification and the recognition of the majority of the sequences from Botryosphaeria already deposited in the database search. These results were similar when the complete ITS region was used with the same objective. The identification was supported by the BLASTn alignment score values between the fragments and the sequences already deposited at the GenBank. Thus, with the exception at B. ribis and B. parva, all the other fragments with at least 200 bp always showed preferential alignments with entries that corresponded to the same species of the query. These alignments showed that intra-specific score values were always higher than the inter-specific ones. The differences between score values show that the sequence complexity in this region was enough to discriminate the Botryosphaeria species (Table 2).
Based on the second criteria stated for each query, the greater number of GenBank available sequences corresponding to the same species were recognized. Fragments with 200 bp (F and R) were able to recognize almost the same number of entries than the complete sequence (Table 2). Entries were considered recognized when they were inside the identification zone. An exception was observed only for B. stevensii and B. dothidea sequences that recognize almost 50% of the total number of GenBank deposited sequences. However, the complete ITS region sequence recognizes the same number. The large number of partial ITS sequences deposited in the database can explain this phenomenon. Once the score was calculated not only by the quality of alignment, but also by the size of sequence aligned, partial sequences showed lower scores, which excluded them from the identification zone. This does not mean that those entries were not valid as the same species, but they did not satisfy the established criteria. Classification errors before sequence submission to the database should also be considered to explain these exceptions.
The robustness of the recognition capacity was tested by phylogenetic tree
analyses. The phylogenetic relationships between the 13 species tested with
the complete sequence of the ITS region were similar to those previously published
with the same dataset (Slippers et al., 2004a; van Niekerk et al.,
2004). Almost the same tree topologies were observed when the 200 bp (F and
R) fragment dataset was used to construct the trees. Minor changes were observed
at the positioning of some species, but the major original groups were maintained.
Based on the ITS rDNA analyses, B. parva and B. ribis could not be distinguished (Table 1 and 2). This result was in agreement with that reported for the ITS region, EF1-α or the mt SSU rDNA sequences (Alves et al., 2004; van Niekerk et al., 2004; Slippers et al., 2004a; Zhou and Stanosz, 2001b). However, those species can be securely separated from other Botryosphaeria species using fragments from 100 bp (Table 1). Considering that B. ribis and B. parva are distinct species, but closely related and probably recently derived on an evolutionary scale, the discrimination of those only should be performed by the use of the multiple-gene genealogy approach, combining the sequence datasets of the ITS rDNA region, β-tubulin, EF1-α genes (Slippers et al., 2004a).
In almost all of the studied species, fragments with 200 bp contained the ITS1 (ranging from 134-185 bp) or the ITS2 (ranging from 154-172 bp) complete region. This observation leads to the conclusion that both ITS regions are sufficient for conclusive identification of Botryosphaeria species. In this way, we can recommend for sequencing the choice of one of the universal ITS1 or ITS4 primers (White et al., 1990), as these primers are anchored at the 18S rDNA and 28S rDNA genes, respectively, both flanking ITS1 and ITS2 regions.
The score differences observed between the identification zone and the second best match with the ITS1 region corresponding fragments (200F) were higher than those observed with the corresponding ITS2 region fragments (200R). Thus, we can recommend the preferential use of the ITS1 region for Botryosphaeria species identification due to its higher sequence complexity when compared with the ITS2 region.
The knowledge that for identification of Botryosphaeria species, a 200 bp simple sequence read is sufficient, which makes the identification faster and easier and greatly reduces costs of DNA sequencing and contig assembly programs. The major advantages of this procedure are its simplicity, the universal availability of the PCR primers, its reproducibility and its amenability to computer database analysis.
Despite all of the advantages cited above, this procedure should be treated with caution when phylogenetic closely-related species are being examined. These problems are often detected when the identification zone is not clearly demarcated showing different species with the same score. In these cases, other genomic regions, classical characters (such as morphology and biochemical), relative pathogenicity, distribution and ecology should be accessed.
Fungi identification and detection in environmental samples are stressed in the literature (González-Lamothe et al., 2002; Magnani et al., 2005). The results presented here suggest that the application of this procedure for the identification of other genera should be useful for identifying fungi species from different sources as the metagenomic approach. Thus, a knowledge of the minimal DNA sequence from the ITS region will make the in-silico identification routine an easier, simpler and dynamic procedure.
This research was supported by grants from CNPq (Conselho Nacional de Desenvolvimento Científico e Tecnológico), CAPES (Coordenação de Aperfeiçoamento de Pessoal de Nível Superior) and Fundação Araucária, Brazil. G.T.V.B. and R.F.H.D. were supported by fellowships from CNPq (pos-doc and senior visiting research professor, respectively) and J.E.G. was supported by a fellowship from CAPES/ProDoc program.