Mycopathologia

, 168:11

Characterization of Penicillium Species by Ribosomal DNA Sequencing and BOX, ERIC and REP-PCR Analysis

Authors

  • Cristina Redondo
    • Laboratorio de Patología Vegetal, Departamento de Protección VegetalInstituto Nacional de Investigación Agraria y Alimentaria (INIA)
  • Jaime Cubero
    • Laboratorio de Patología Vegetal, Departamento de Protección VegetalInstituto Nacional de Investigación Agraria y Alimentaria (INIA)
    • Laboratorio de Patología Vegetal, Departamento de Protección VegetalInstituto Nacional de Investigación Agraria y Alimentaria (INIA)
Article

DOI: 10.1007/s11046-009-9191-y

Cite this article as:
Redondo, C., Cubero, J. & Melgarejo, P. Mycopathologia (2009) 168: 11. doi:10.1007/s11046-009-9191-y

Abstract

The genus Penicillium is one of the largest and widely distributed fungal genera described to date. As a result, its taxonomic classification and species discrimination within this genus has become complicated. In this study, 52 isolates that belonged to the Penicillum genus and other related genera were characterized using two DNA-based methods: (i) analysis of the nucleotide sequences of internal transcribed spacers in ribosomal DNA and (ii) analysis of DNA fingerprints that were generated by polymerase chain reactions with specific primers for enterobacterial repetitive intergenic consensus (ERIC) and repetitive extragenic palindromic (REP) sequences, and BOX elements. Using both methods, Penicillium species were discriminated from other fungal genera. Furthermore, Penicillium species that include strains which are used as biocontrol agents, such as P. glabrum, P. purpurogenum, and P. oxalicum, could be distinguished from other Penicillium species using these techniques. Based on our findings, we propose that a polyphasic approach that includes analysis of the nucleotide sequences of ribosomal DNA and detecting the presence of highly conserved, repeated nucleotide sequences can be used to determine the genetic relationships between different Penicillium species. Furthermore, we propose that our results can be used as a start point to develop a strategy to monitor the environmental presence of particular strains of Penicillium species when they are used as biocontrol agents.

Keywords

BiocontrolDNA fingerprintingFungal monitoringInternal transcribed spacersrep-PCR

Introduction

The genus Penicillium is one of the largest and widely distributed of all fungal genera described to date. In the period from the publication of the first taxonomic study on Penicillium in 1930 to the latest classification in 2004, 225 new Penicillium species have been described [13]. The results of previously published taxonomic studies, in which only morphological characteristics were used to identify fungal species, yielded different classification proposals because of strain variation [1, 4, 5]. Strain variation in a fungal species is common due to differences in the environmental conditions of their habitat [6]. Since the species concept is based primarily on morphology, the exact identification of some fungal species is not always possible, and further investigation is then necessary in order to identify the fungus [7]. For this reason, a polyphasic approach that incorporates a battery of morphological, physiological, biochemical, and chemotaxonomic characterizations is recommended in order to develop an accurate and consensual classification for fungi. However, the results of taxonomic studies on the Penicillium subgenera can be ambiguous even using this polyphasic approach because not all the strains from the same species display identical characteristics or profiles [810]. In addition, mycotoxin detection does not always provide information about those fungi that are currently present in the environment because the mycotoxins may be contaminants and/or residues from other micro-organisms that are no longer present in the environment [11]. Moreover, identification methods that rely on recognizing a limited number of enzymes and/or other proteins are not sufficiently discriminating to ensure accurate identification of fungal species [11, 12].

In order to overcome the limitations of the classical microbiological techniques, advanced molecular methods with high resolution and accuracy are now being used for taxonomic classification. For example, nucleotide sequencing of biosynthetic genes [1315] and ribosomal and mitochondrial DNA [10, 14, 16, 17] has been used to identify new Penicillium species [18, 19], to characterize different Penicillium strains, and to distinguish among different species within the Penicillium genus [2024].

Identifying families of repetitive DNA sequences has been shown also to be a useful and reliable strategy to determine the genetic relationships within groups of microorganisms. Although repetitive sequence-based polymerase chain reactions (rep-PCRs) are used to identify highly conserved, repeated nucleotide sequences in bacterial DNA [25], Gillings and Holley reported that rep-PCR fingerprinting can also amplify other regions and even non-bacterial sequences of the bacterial genome [26]. Despite this limitation, rep-PCR fingerprinting has been used to analyze the DNA of various fungal species [2729], and to characterize the Penicillium genus and other closely related fungal genera [10, 30]. Although the reliability of rep-PCR fingerprinting for fungal characterization has been validated by other DNA fingerprinting techniques, such as random amplification of polymorphic DNA (RAPDs) [10] and amplified fragment length polymorphism (AFLPs) [31, 32], incidental similarities among the DNA fingerprints of distantly related organisms can occur when using these techniques [32].

The development and use of fungi as biocontrol agents require a risk assessment of their potential hazards, and compliance with regulatory and registration requirements is essential in order to ensure the efficacy and safety of a biocontrol agent [33]. As part of the registration requirements [34], the risk assessment must include monitoring of the biocontrol agent in the field in the presence of the indigenous fungal population.

Some Penicillium species are well-known biocontrol agents for controlling several important plant diseases. For example, Penicillium glabrum Wehmer (sin Penicillium frequentans Westling) [35] and P. purpurogenum [36] have been used successfully to control brown rot and twig blight in stone fruit, both of which are caused by Monilinia laxa. Another example is the use of P. oxalicum to control tomato wilt, which is caused by Fusarium oxysporum f.sp. lycopersici [37]. Therefore, molecular markers may be useful for fungal screening in the field in order to satisfy this requirement [38].

The aims of this investigation were (a) to characterize Penicillium species using two DNA-based techniques in order to establish an accurate method for taxonomic purposes and (b) to evaluate their utility as a first approach for monitoring Penicillium species in the environment when a member of these species is used as a biocontrol agent. To this end, some representative species of the genus Penicillium were characterized using two DNA-based methods: (i) analysis of the nucleotide sequences of internal transcribed spacers (ITS) in ribosomal DNA (rDNA) and (ii) analysis of the rDNA fingerprints that were generated after PCR using primers for enterobacterial repetitive intergenic consensus (ERIC) and repetitive extragenic palindromic (REP) sequences, and BOX elements.

Materials and Methods

Fungal Isolates

Table 1 lists the isolates that were used in this study. These isolates included 26 P. glabrum isolates, 45 Penicillum species and seven isolates from species of other closely-related fungal genera and commonly found as saprophytes in fruit trees. All the isolates were characterized phenotypically on their colony morphology in different culture media and other morphological characteristics [1, 5]. The isolates of the biocontrol agents, P. glabrum PF-909 (ATCC number 66108) and P. purpurogenum iPP828 (ATCC number 66107), were obtained from the mycobiota of peach tree twigs in Spain [39]. The isolate of the biocontrol agent, P. oxalicum (ATCC number 201888), was identified originally from soil samples, and was kindly provided by Dr. F. Reyes [40]. All the isolates were maintained in 20% glycerol at −80°C in the collection of the Plant Protection Department of INIA (Madrid, Spain), and were grown in 20 ml of potato-dextrose agar (PDA) (Difco, Detroit, USA) in the dark at 25°C for seven days in 90-mm-diameter plastic Petri plates for the production of mycelial and conidial inoculum.
Table 1

The number, collection date, host and origin of the fungal isolates used in this study

Fungal Isolate

Isolate number

Year of isolation

Sample of origin

Location

Alternaria spp.

Alt sp

2004

Peach tree twigs

Madrid (Spain)

Aspergillus niger

Asp

1987

Peach tree twigs

Zaragoza (Spain)

Cladosporium cladosporioides

908,87,9

1987

Peach tree twigs

Zaragoza (Spain)

Cladosporium cucumerinum

Cc

1987

Peach tree twigs

Zaragoza (Spain)

Epicoccum nigrum

282- 96794a

19

Peach tree twigs

Zaragoza (Spain)

Fusarium oxysporum fsp lycopersicy

FUS 3- 201829a

1986

Peach tree twigs

Zaragoza (Spain)

Paecilomyces farinosus

408,87,5

1987

Peach tree twigs

Zaragoza (Spain)

Penicillium adametzoides

432,87,5

1987

Peach tree twigs

Zaragoza (Spain)

Penicillium camemberti

313,86,5

1986

Peach tree twigs

Zaragoza (Spain)

Penicillium camemberti

944,86,9

1986

Peach tree twigs

Zaragoza (Spain)

Penicillium chrysogenum

753,87,8

1987

Peach tree twigs

Zaragoza (Spain)

Penicillium chrysogenum

722,87,8

1987

Peach tree twigs

Zaragoza (Spain)

Penicillium citrinum

608,86,6

1986

Peach tree twigs

Zaragoza (Spain)

Penicillium citrinum

918,86,9

1986

Peach tree twigs

Zaragoza (Spain)

Penicillium concentricum

613,87,7

1987

Peach tree twigs

Zaragoza (Spain)

Penicillium expansum

PE

1986

Peach tree twigs

Zaragoza (Spain)

Penicillium glabrum

252-87

1987

Peach tree twigs

Zaragoza (Spain)

Penicillium glabrum

516-87

1987

Peach tree twigs

Zaragoza (Spain)

Penicillium glabrum

PF 1

2004

Peach tree twigs

Madrid (Spain)

Penicillium glabrum

PF 2

2004

Peach tree twigs

Madrid (Spain)

Penicillium glabrum

PF 3

2004

Peach tree twigs

Madrid (Spain)

Penicillium glabrum

PF 4

2004

Peach tree twigs

Madrid (Spain)

Penicillium glabrum

PF 5

2004

Peach tree twigs

Madrid (Spain)

Penicillium glabrum

PF 6

2004

Peach tree twigs

Madrid (Spain)

Penicillium glabrum

PF 7

2004

Peach tree twigs

Madrid (Spain)

Penicillium glabrum

PF 8

2004

Peach tree twigs

Madrid (Spain)

Penicillium glabrum

PF 10

2004

Peach tree twigs

Madrid (Spain)

Penicillium glabrum

PF 11

2004

Peach tree twigs

Madrid (Spain)

Penicillium glabrum

PF 12

2004

Peach tree twigs

Madrid (Spain)

Penicillium glabrum

PF 13

2004

Peach tree twigs

Madrid (Spain)

Penicillium glabrum

PF 14

2004

Peach tree twigs

Madrid (Spain)

Penicillium glabrum

PF909a

1986

Peach tree twigs

Zaragoza (Spain)

Penicillium glabrum

7-86

1986

Peach tree twigs

Zaragoza (Spain)

Penicillium glabrum

9-86

1986

Peach tree twigs

Zaragoza (Spain)

Penicillium glabrum

46-87

1987

Peach tree twigs

Zaragoza (Spain)

Penicillium glabrum

118-87

1987

Peach tree twigs

Zaragoza (Spain)

Penicillium glabrum

212-87

1987

Peach tree twigs

Zaragoza (Spain)

Penicillium glabrum

229-87

1987

Peach tree twigs

Zaragoza (Spain)

Penicillium glandicola

732,87,8

1987

Peach tree twigs

Zaragoza (Spain)

Penicillium lanosum

929,87,9

1987

Peach tree twigs

Zaragoza (Spain)

Penicillium oxalicum

PO 201888a

1989

Peach tree twigs

Zaragoza (Spain)

Penicillium pulvillorum

129,87,2

1987

Peach tree twigs

Zaragoza (Spain)

Penicillium purpurogenum

PP828a66107a

1986

Peach tree twigs

Zaragoza (Spain)

Penicillium purpurogenum

119.87.2

1987

Peach tree twigs

Zaragoza (Spain)

Penicillium raistrickii

104,87,2

1987

Peach tree twigs

Zaragoza (Spain)

Penicillium raistrickii

134,87,2

1987

Peach tree twigs

Zaragoza (Spain)

Penicillium raistrickii

150,87,2

1987

Peach tree twigs

Zaragoza (Spain)

Penicillium sp.

451,87,5

1987

Peach tree twigs

Zaragoza (Spain)

Penicillium sp.

718,87,8

1987

Peach tree twigs

Zaragoza (Spain)

Penicillium sp.

921,87,9

1987

Peach tree twigs

Zaragoza (Spain)

Penicillium thomii

650,86,6

1986

Peach tree twigs

Zaragoza (Spain)

Penicillium verrucosum

226,87,3

1987

Peach tree twigs

Zaragoza (Spain)

aIsolates that were deposited in the American Type Culture Collection (ATCC) and their ATCC numbers

DNA Extraction From Pure Fungal Cultures

Flasks that contained a broth medium (5 g malt extract (Difco), 5 g yeast extract, and 5 g glucose L−1 of sterile distilled water) were inoculated with an aliquot of each actively growing fungal culture, and then incubated with shaking at room temperature for one week. After one week, mycelia were recovered by filtration, and then lyophilized. DNA extraction from the mycelia was performed according to the method described by Keijer et al. [41] with minor modifications. Briefly, 5–10 mg aliquots of the lyophilized biomass were added to tubes that contained 500 μl of extraction buffer (20 mM Tris HCl, 5 M NaCl, 0.5 M EDTA, 4% SDS), and the tubes were then incubated for 15 min at 65°C. Upon completion of the incubation, the tubes were centrifuged at 12500 rpm, and the DNA was extracted by subjecting the supernatant to two phenol:chloroform and chloroform:isoamyl alcohol extractions [42]. Finally, the DNA was precipitated by two consecutive isopropanol and ethanol precipitations, resuspended in 50 μl of sterile water, and then stored at −20°C until use.

PCR of Ribosomal DNA

The various fungal rDNAs were amplified by PCR with primers ITS4 (5′-TCC TCC GCT TAT TGA TAT-3′) and ITS5 (5′-GGA AGT AAA AGT CGT AAC AAG G-3′). [43]. The PCRs were carried out in a 25-μl reaction volume that contained the DNA templates (80 ng of purified DNA), 1× PCR buffer (Invitrogen, Carlsbad, California, USA), 1.5 mmol l−1 MgCl2, 0.1 μM of primers ITS4 and ITS5, 0.1 mmol l−1 of each dNTP and 1 unit of Taq polymerase (Invitrogen). The PCR conditions were: 95°C for 1.5 min, 50°C for 1 min, and 72°C for 2 min for 40 cycles, plus an initial step of 95°C for 3 min and a final step of 72°C for 10 min. The PCR products were purified using the Wizard SV Gel and PCR Clean-up system (Promega), and then sequenced at the CISA Sequencing Service of INIA.

Analysis of the Nucleotide Sequences of the Fungal Ribosomal DNAs

The nucleotide sequences of fungal rDNAs were aligned and compared with those in the GenBank database using the BioEdit Sequence Alignment Editor 5.0.6 [44] and the CLUSTAL W 1.82 programmes [45]. In addition, nucleotide sequences obtained in this study were deposited in the GenBank database under the accession numbers DQ681321 to DQ681353.

The dendrograms were constructed by an analytical programme (MEGA version 3.1) [46] using the neighbor-joining method [47]. The Jukes–Cantor model [48] and the “p” distance were used to calculate genetic pair-wise distances among the sequences. The reliability of the clusters was evaluated by bootstrapping with 1,000 replicates.

PCR for BOX Elements, and REP and ERIC Sequences (rep-PCRs) for Fungal rDNAs

The BOX-PCRs were carried out in a 25-μl reaction volume that contained the rDNA templates (80 ng of purified DNA), 1× PCR buffer, 6 mmol l−1 MgCl2, 2.4 μmol l−1 of primer BOXA1R (5′-CTA CGG CAA GGC GAC GCT GAC G-3′) [49], 0.2 mmol l−1 of each dNTP, and 2 units of Taq polymerase. The PCR conditions were an initial denaturation step of 95°C for 5 min, followed by 10 cycles of 94°C for 1 min, 40°C for 30s, and 72°C for 1 min, and 30 cycles of 90°C for 1 min, 48°C for 30s, and 72°C for 1 min with a final extension of 72°C for 10 min.

The REP-PCRs were carried out in a 25-μl reaction volume that contained the rDNA templates (80 ng of purified DNA), 1× PCR buffer, 4 mmol l−1 MgCl2, 1.2 μM of primers REP 1R (5′-III ICG ICG ICA TCI GGC-3′) and REP 2R (5′- ICG ICT TAT CIG GCC TAC -3′) [49], 0.2 mmol l−1 of each dNTP, and 2 units of Taq polymerase. The PCR conditions were an initial denaturation step of 94°C for 5 min, followed by 40 cycles of 94°C for 1 min, 40°C for 1 min and 72°C for 4 min with a final extension of 72°C for 10 min.

The ERIC-PCRs were performed in a 25-μl reaction volume that contained the rDNA templates (80 ng of purified DNA), 1× PCR buffer, 2 mmol l−1 MgCl2, 1.2 μM of primers ERIC 1R (5′-ATG TAA GCT CCT GGG GAT TCA C-3′) and ERIC 2R (5′-AAG TAA GTG ACT GGG GTG AGC G-3′) [25], 0.2 mmol l−1 of each dNTP, and 2 units of Taq polymerase. The PCR conditions were an initial denaturation step of 94°C for 5 min, followed by 40 cycles of 94°C for 1 min, 52°C for 1 min and 72°C for 8 min with a final extension of 72°C for 10 min.

BOX, REP, and ERIC DNA fingerprints were visualized in 3% agarose gels under ultraviolet light after their staining with ethidium bromide. Each rep-PCR was performed at least twice to verify the reproducibility and reliability of the reaction.

Analysis of rep-PCR Fingerprints

DNA fingerprints were first binarized (0 = absent, 1 = present), and Dice similarity coefficients for pairs of isolates were calculated using the NTSYS programme (version 2.1, Exeter Software, Setauket, New York, USA) [50]. Dendrograms were then constructed by the unweighted pair-group (UPGMA) clustering method using the MEGA programme (version 3.0) [43] (Fig. 2).

Results

Classification of Penicillium spp. Based on the Results of the Analysis of the Nucleotide Sequences of the rDNAs

The nucleotide sequences of ITS1, ITS2, and the 5.8S ribosomal subunit were first analysed independently. Dendrograms that were based on pair-wise comparison of the nucleotide sequences of the three rDNAs always showed that the Penicillium species grouped together. Results of the analysis of the nucleotide sequences of ITS1 showed that the mean level of similarity between species of the Aspergilloides subgenus and other Penicillium spp. was 99.0 ± 1.0% with an average of 2.7 ± 1.3 nucleotide differences and a maximum of three nucleotide discrepancies over a sequence of 157 nucleotide positions. The mean level of similarity for the nucleotide sequences of the 5.8S ribosomal subunit and ITS2 between species of the Aspergilloides subgenus and other Penicillium spp. was 100.0 ± 0.1% in both instances (data not shown).

The most consistent and robust clustering was found when the nucleotide sequences of ITS1, ITS2, and the 5.8S ribosomal subunit were analysed together. The results of this combined analysis yielded a mean level of similarity between Penicillium spp. and other fungal isolates of 71.0 ± 0.1% with an average of 149.3 ± 6.0 nucleotide differences and a maximum discrepancy of 187 nucleotides over a sequence of 629 nucleotide positions. The results of this combined analysis revealed also that the mean level of similarity within Penicillium spp. was 94.0 ± 0.1%, with an average of 33.9 ± 2.5 nucleotide discrepancies and a maximum of 108 nucleotide differences. The dendrogram that was constructed using pair-wise comparison of the three rDNAs revealed that the variability in nucleotide sequences of the three rDNAs among the different fungal isolates was sufficient to discriminate Penicillium spp. from the other related fungal genera (Fig. 1). This dendrogram revealed also that all isolates of Penicillium spp. were grouped together with Cladosporium spp., Paecilomyces spp., Fusarium spp., Alternaria spp., and Epicoccum spp. as outgroups.
https://static-content.springer.com/image/art%3A10.1007%2Fs11046-009-9191-y/MediaObjects/11046_2009_9191_Fig1_HTML.gif
Fig. 1

The dendrogram was constructed using the neighbor-joining method and based on phylogenetic analysis of the nucleotide sequences of ITS1, 5.8S and ITS2 rDNA. It shows the relationships between Penicillium species and other fungal genera in which the Jukes–Cantor model and the ‘p’ distances were used to calculate genetic pairwise distances among the sequences. The distance between isolates was obtained by adding the lengths of the connecting horizontal branches. Bootstraps values (1,000 replicates) are indicated at the nodes

https://static-content.springer.com/image/art%3A10.1007%2Fs11046-009-9191-y/MediaObjects/11046_2009_9191_Fig2_HTML.gif
Fig. 2

The dendrogram that was based on the combined BOX-REP similarity matrix. Dice similarity coefficients for pairs of isolates were calculated using the NTSYS (verion 2.1) programme, and the dendrogram was constructed by the UPMGMA clustering method using the MEGA programme (version 3.0)

Three groups within the Penicillium cluster could be identified from the results of the analyses of the combined three rDNA regions. One group (Group 1), which included Penicillium species that belonged to the Penicillium and Furcatum subgenera, had a mean level of similarity of 97.4 ± 0.3% and a maximum of 35 nucleotide differences. A second group (Group 2) that consisted of P. adametziodes, P. glabrum, and P. thomii had a mean level of similarity of 99.2 ± 0.3% and a maximum of 6 nucleotide differences over a sequence of 642 nucleotides. The third group (Group 3) included all isolates of the Biverticillium subgenus, and this group had a mean level of similarity of 88.1 ± 0.3% and a maximum of 61 nucleotide differences (Fig. 1). Similar clustering results were obtained when nucleotide sequences of isolates from the GenBank database were included in the analyses (data not shown).

Classification of Penicillium spp. Using BOX, ERIC, and REP DNA Fingerprints

The DNA products that were generated by PCR with the BOX, REP, and ERIC primers yielded bands (BOX, ERIC, and REP DNA fingerprints) whose sizes ranged from 100 to 1200 bp. The Penicillium classification that was obtained from the results of the cluster analysis of the BOX, ERIC, and REP DNA fingerprints was similar to that obtained after analysis of the nucleotide sequences of the three rDNAs. However, the Penicillium spp. could be distinguished clearly from the other fungal genera after analyzing the ERIC, BOX, and REP DNA fingerprints. In addition, P. glabrum, P. adametzioides, and P. thomii grouped together, and were distinct from other Penicillium species. Although the resolution of the BOX, ERIC and REP DNA fingerprints was lower than that of nucleotide sequencing of rDNA, it was possible to distinguish between those species with strains that are used as biocontrol agents, such as P. glabrum, P. purpurogenum, and P. oxalicum using this method, and that this distinction was unequivocal.

After the BOX-PCR analysis, we found that the mean level of similarity between Penicillium spp. and other fungal genera was approximately 28%, while the mean level of similarity among species within the Penicillium cluster was 69%. The mean level of similarity of those Penicillium isolates that belonged to the Aspergilloides subgenus was 98% within this group, and was 60% between this group of isolates and other Penicillium spp. Finally, there was 100% similarity for all P. glabrum isolates, and when these DNA fingerprints were compared to those of other two members of Group 2, namely, P. adametziodes and P. thomii, the mean level of similarity was 91%.

After the REP-PCR analysis, we found that the mean level of similarity between Penicillium spp. and other fungal genera was about 38, and 59% within the Penicillium cluster. The mean level of similarity of P. adametziodes, P. glabrum and P. thomii isolates (Group 2) was 98% within the group, and 46% when the DNA fingerprints of this group were compared with other Penicillium species. There was 100% similarity for all P. glabrum isolates, and when these DNA fingerprints were compared to those of the other two members of group 2, the mean level of similarity was 85%.

After the ERIC-PCR analysis, we found that the mean level of similarity between Penicillium spp. and other fungal genera was approximately 40, and 56% within the Penicillium cluster. The mean level of similarity of group 2 isolates was 98% within the group and 39% when the DNA fingerprints of this group were compared with those of other Penicillium species. There was 100% similarity for all P. glabrum isolates, and when the DNA fingerprints were compared to those of other two members of group 2, the mean level of similarity was 88%.

In order to improve the clarity of the dendrogram, different similarity matrices were generated by combining the banding patterns that were generated from the BOX, REP, and ERIC PCRs. Although some discrepancies were found in the lowest branches of the dendrogram, we selected the combination of the BOX and REP-PCR banding patterns for the cluster analysis because this combination gave the most information, and was the most robust of all the combined analyses according to the biological characteristics of the fungi. The incorporation of the ERIC PCR banding pattern into the analysis did not contribute additional information on the relationship among Penicillium species. In fact, we found that the addition of the ERIC PCR banding patterns sometimes disturbed the cluster analysis because biologically different isolates that were separated previously in the BOX- or REP-PCR banding patterns became grouped.

The dendrogram that was based on the combined BOX-REP PCR similarity matrix showed two clusters: one cluster contained isolates of all Penicillium spp. and the other cluster contained isolates of Cladosporium sp., Alternaria sp., and Paecilomyces sp. The mean level of similarity between the two clusters was 36, and 65% within the Penicillium cluster. In the Penicillium cluster, those species that belonged to the subgenus Aspergilloides could be distinguished from other Penicillium species: the mean level of similarity was 54%. Within the subgenus Aspergilloides, P. glabrum could be discriminated from P. adametziodes and P. thomii, despite the mean level of similarity being 92% in both instances. All P. glabrum isolates showed 100% similarity.

Discussion

One of the aims of this research was to develop an accurate and rapid method to distinguish among Penicillium species, and especially the strains of those species that are used as biocontrol agents. Relying on morphology and ability to grow on selective culture media as the method to identify Penicillium species is time consuming and laborious. In addition, considerable expertise is needed to clearly differentiate between closely related species [8, 22, 51].

Direct analysis of DNA polymorphisms is a useful approach to establish genetic variation in micro-organisms. These techniques are especially valuable to enhance epidemiological studies because typing is a necessary first step in epidemiological studies [52]. The study of DNA polymorphisms involves the selection of a target sequence, and several approaches have been used to achieve this objective. One approach involves the exploitation of ubiquitously conserved known genes that display sequence variation. In particular, comparative nucleotide sequencing of rDNA subunits, such as ITSs, has been used widely for distinguishing between fungal species, and to develop specific protocols for identifying fungal species [20, 22]. Another approach involves the screening of random parts of the genome to identify distinctive nucleotide sequences by techniques, such as RAPD and rep-PCR.

When used for taxonomic purposes, the results of all these methods yielded intra-specific polymorphisms. Therefore, their precision is not consistent. Analysis of the nucleotide sequences in rDNA is useful because related fungal species have variations in the nucleotide sequences of ITS1 and ITS2 [53]. Therefore, the results of an analysis of the nucleotide sequences of ITS1 and ITS2 can be exploited for the specific detection and classification of fungi [54, 55]. Although the results of an analysis of the nucleotide sequences of ITSs have been used previously for phylogeny studies of the Penicillium genus [20], Skouboe and others [22] claim that such analyses are not sufficiently precise and accurate for identifying species. Here, we confirmed the previous reports on the limitations of analyzing nucleotide sequences of rDNA in fungal species, and in particular within the Penicillium genus [3, 14, 20, 22, 56]. We found also that the results of a combined analysis of the nucleotide sequences of ITS1, 5.8S and ITS2 rDNA enabled discrimination among Penicillium species. Although the nucleotide sequences of the ITSs may display sufficient variability to enable discrimination within a subgenus, we showed that the value of the discrimination method was increased when we included the nucleotide sequence of the 5.8S, ribosomal subunit into our analysis. This improvement was seen also when the nucleotide sequences from the GenBank database were combined with those that were determined in this investigation, and then analyzed. The results of our combined analysis of the nucleotide sequences of these three rDNAs are consistent with that obtained following the analysis of the CO1 DNA barcodes in Penicillium species [17]. Although analysis of the nucleotides sequences of the three rDNAs is a powerful clustering method in a polyphasic approach for discriminating between fungal species, we still felt that other techniques were needed for this purpose.

Analysis of DNA fingerprints using techniques such as RAPDs [10] and AFLPs [32] is valuable when the techniques are used to determine the genetic variation and to discriminate between Penicillium species. Although the complex AFLP patterns allow a distinction of very closely related strains, incidental similarities among the fingerprints of distantly related organisms may still occur [32].

In this study, we found that rep-PCR fingerprinting, which is a widely used method for typing of the bacterial genome, can be used also to distinguish among Penicillium species. We found that this method is sufficiently sensitive to enable discrimination of those species that could not be discriminated using phenotypical characteristics. We found also that this method enabled a discrimination between closely related species that could not be distinguished using the results of an analysis of the nucleotide sequences of ITSs. The analysis of rep-PCR fingerprints has three major advantages: it is simple to do, the PCR primers are universal, and it is tolerant for a wide range of DNAs. Our results showed also that rep-PCR is a practical, rapid, and accurate method for characterizing members of the Penicillium genus. We believe also that rep-PCR fingerprinting can be used to distinguish between those fungal species with strains that are used as biocontrol agents, such as P. glabrum, P. purpurogenum, and P. oxalicum, and pathogenic fungi such as P. expansum [57].

We found that the precision of the three rep-PCR fingerprinting techniques, in which highly conserved, repeated DNA sequences are amplified, is not identical. For this reason, we decided to combine the banding patterns that were generated by ERIC-, REP-, and BOX-PCRs in order to distinguish better between the various Penicillium species. When the results of the BOX- and REP-PCRs were combined, we found that the results of this cluster analysis was the most informative and the most robust of all the analyses according to biological characteristics of the fungi. Moreover, the dendrogram that was constructed from the results of this combined analysis was similar to that found after the analysis of the nucleotide sequences of the rDNAs. We found also that the precision of the DNA fingerprints that were generated by ERIC-PCR was less than that of those that were generated by the other two rep-PCR fingerprinting techniques. In fact, its inclusion in the clustering analysis disturbed the resultant fungal classification. According to Gillings and Holley, the products of ERIC-PCR are not always products from genuine ERIC sequences because complex ERIC-PCR banding patterns can be readily produced from many targets at low annealing temperature [26]. Furthermore, the specificity of the primers that are used in ERIC-PCR is limited and consequently, similar to that found when a non-specific PCR, such as RAPD, is used.

We found that the DNA fingerprints from each fungal species were different, even though some species were closely related. We found also that the results of the rep-PCR analysis agreed with those of the analysis of the nucleotide sequences of the three rDNAs: closely related species that displayed high levels of similarity when results of the analysis of the nucleotide sequences of the rDNAs were used also had high levels of similarity when the products of rep-PCR were used to determine similarity. Therefore, the molecular techniques that were used in this study allow us to simultaneously monitor several fungal species that are present in a common environment.

We found that accurate discrimination among isolates that belonged to the same species of Penicillium could not be achieved using rep-PCR. Beng and others [30] reported that considerable intraspecific variability existed within Penicillium strains when BOX DNA fingerprints were used to distinguish between the strains. They reported also that this variability correlated with the geographical origin and plant host species, as has been described for bacteria [49, 58]. Nevertheless, any method of DNA fingerprinting must be validated for its effectiveness for discriminating among strains, and must consider also the possibility that the method could isolate the same strain several times in a particular geographical area [59]. Although the detection of different DNA fingerprints should indicate the existence of different strains, the detection of an identical fingerprint in two isolates is not a guarantee that a unique strain has been found because of possible limitations of the technique. The techniques described herein do not allow fungal discrimination at the isolate level, and therefore they are not sufficiently sensitive in most instances to track a biocontrol agent after its release into the environment. This inability hampers their utility to monitor the micro-organism in the environment when it is used as a biocontrol agent, despite its high population concentration and increased likelihood of its detection. Although these methods may be useful to detect an introduced micro-organism at the site of its environmental release, the inability of the methods to distinguish between individual strains hampers also their utility to detect the biocontrol agent at a site that is different from the site of its introduction. Detection of the biocontrol agent at a second distant site should not be interpreted to mean that its presence at this site was definitely due to its spread from the site of its introduction. Furthermore, any subsequent detection of the biocontrol agent should not be interpreted to mean its persistence or replacement of the original population by the biocontrol agent at the two sites. Due to the fact that these options need to be considered for the environmental monitoring of a biocontrol agent, we believe that these methods cannot be used for this purpose at this point in time.

Finally, we found that the results of the analysis of the nucleotide sequences of rDNAs and rep-PCR fingerprints are able to provide information on the genetic relationships among Penicillium species. Therefore, we propose also that further studies should be done in order to validate the ability of these techniques to distinguish between isolates that belong to the same species, and especially those that are used as biocontrol agents.

Copyright information

© Springer Science+Business Media B.V. 2009