Author for correspondence: Tina C. Summerfield Tel: +1 765 494 0560 Fax: +1 765 496 1496 Email: email@example.com
• Pseudocyphellaria crocata, P. neglecta and P. perpetua specimens were examined to investigate links between genetic variation and morphology, geographical distribution and cyanobiont specificity.
• Fungal internal transcribed spacer (ITS), β-tubulin and cyanobacterial tRNALeu (UAA) intron sequences were used to investigate symbiont diversity in these lichens.
• Specimens were morphologically distinct but could not be distinguished by ITS sequences. Phylogenetic analyses split the P. crocata specimens into two clades, the larger of which contained P. neglecta and P. perpetua. Five cyanobionts were identified; two of these were in a number of specimens, while three were each restricted to a single lichen thallus.
• Fungus-specific molecular markers indicated that all specimens belonged to a single phylogenetic species. However, this may contain a cryptic species. Geography was linked to genetic diversity with Canadian specimens forming a monophyletic group, and most Southern Hemisphere specimens grouping together, although Chile represented a hot spot of genetic diversity. There was no connection between fungal genetic diversity and cyanobiont choice, consistent with the presence of a common pool of cyanobionts.
A lichen is the symbiotic association of a fungus with a green alga and/or a cyanobacterium. Traditionally, taxonomic assignment has been based on the dominant fungal partner, using characters such as chemistry, morphology, reproductive mode and habitat preference, but increasingly DNA markers have been used to examine existing taxonomic assignments, and modifications to previous classifications have often been suggested by these data (Grube & Kroken, 2000). Molecular techniques have been utilized in fungi at a range of taxonomic levels to reveal deep relationships and to identify cryptic species within morphologically homogeneous groups (Lutzoni et al., 2004; Myburg et al., 2004; O'Donnell et al., 2004).
Genetic variation of the photosynthetic symbiont of lichens has also been examined. Studies have reported algal photobionts to be shared among fungal species, although some fungi have shown a selective preference for particular photobiont genotypes (Beck et al., 1998; Helms et al., 2001; DePriest, 2004; Yahr et al., 2004). The lichen species examined in this study are bipartite with cyanobacterial photobionts belonging to the genus Nostoc (Galloway, 1988). In other cyanolichens, it has been shown that the identity of the Nostoc strain is influenced more by fungal identity than geographical distribution, although the degree of specificity of the cyanobacterial associations varied between lichen species (Paulsrud & Lindblad, 1998; Paulsrud et al., 1998, 2000). Such studies on Nostoc-containing lichen species led Rikkinen et al. (2002) to suggest that cyanobiont specificity occurred at a community level, with lichens from a particular habitat type exhibiting specificity for a group of cyanobacterial strains. However, low cyanobiont selectivity was reported in lichen specimens from Antarctica, although this may be explained by selective pressure in this harsh environment (Wirtz et al., 2003).
The objective of the present study was to investigate genetic variation within and between fungal species and, in particular, whether or not this is linked to lichen morphology, geographical distribution or the cyanobacterial strain present. The lichen species examined were Pseudocyphellaria crocata (L.) Vainio, which has a cosmopolitan distribution; Pseudocyphellaria neglecta (Müll. Arg.) Magnusson, with a more restricted distribution in Australasia and possibly the palaeotropics; and Pseudocyphellaria perpetua McCune & Miadlikowska, a new species from north-west America. The species P. crocata and P. neglecta exhibit the same chemistry and similar habitat ranges, but distinct morphologies (Galloway, 1988). Specimens of these two species from New Zealand were found to have identical fungal ITS sequences, but their photobionts were distinct Nostoc strains (Summerfield et al., 2002). Pseudocyphellaria perpetua is morphologically similar to some P. crocata forms but can be distinguished by its yellow medulla and predominately laminal soralia, and may have a distinct terpenoid composition (Miadlikowska et al., 2002). In North American specimens, the ITS sequences of P. perpetua were reported to be distinct from those of P. crocata; however, no information is available on cyanobiont specificity (Miadlikowska et al., 2002).
Identification of the primary photobiont can be established from cyanobacterial tRNALeu (UAA) intron sequences. The sequence of this intron has been shown to be highly variable in the Nostocaceae (Xu et al., 1990) and has previously been used to examine cyanobacterial diversity in lichen specimens (Paulsrud & Lindblad, 1998; Paulsrud et al., 1998, 2000). Two variable regions within the intron have been identified: region 1 is comprised of degenerate heptanucleotide repeats and nonrepetitive elements, with sequence variation arising from different types and numbers of repeats and the presence and length of nonrepetitive elements (Paulsrud & Lindblad, 1998; Costa et al., 2002). As a result, Nostoc strains have been identified using this region, with each type of variable region 1 representing a different strain (Paulsrud et al., 2000).
Here we present data on specimens from Australia, Chile and Canada and compare these with specimens from North America and New Zealand, in order to examine these species over a broad geographical range. Fungus-specific primers were used to amplify either the ITS of the nuclear-encoded ribosomal repeat unit, or part of the gene encoding β-tubulin, in order to examine the variation in the fungal component of the lichen samples between and within species. Additionally, the specificity of the cyanobionts in these bipartite cyanolichens was investigated by comparing tRNALeu (UAA) intron sequences.
Materials and Methods
Pseudocyphellaria crocata and P. neglecta thalli were identified according to Galloway (1988) and the P. perpetua thallus was identified according to Miadlikowska et al. (2002). The P. crocata thalli were designated numbers PC7 to PC14; P. neglecta thalli PN5 and PN6; and the P. perpetua specimen was named PP1. Previously collected P. crocata and P. neglecta thalli from South Island, New Zealand have been designated PC1–PC6 and PN1–PN4, respectively (Summerfield et al., 2002).
DNA was extracted from lichen thalli as described by Cubero et al. (1999). Fungus-specific primers were used to amplify the ITS region of nuclear-encoded rRNA genes from the DNA isolated from lichen thalli. Nested PCR was carried out using external primers ITS1-f and NL6Amun designed by Gardes & Bruns (1993) and Egger (1995), respectively, and internal primers ITS5 and ITS4 designed by White et al. (1990). The PCR was carried out as described by Summerfield et al. (2002). Part of the gene encoding β-tubulin was amplified using the fungus-specific primers 5′-GTACTGGAGCTGGTATGG-3′ and 5′-CCAACACGCTTGAAGAGCTC-3′. The reaction conditions were 94°C for 1 min, followed by 30 cycles of 94°C for 30 s, 56°C for 30 s, 72°C for 1 min. The PCR was concluded with an extension at 72°C for 2 min. The cyanobacterial tRNALeu (UAA) intron was amplified from lichen DNA using nested primers designed to the tRNALeu (UAA) exons by Paulsrud & Lindblad (1998). All PCR products were gel purified using the QIAquick gel extraction kit (Qiagen, Valencia, CA, USA). The PCR products were sequenced with the primers used for PCR, using an ABI PRISM 377 DNA Sequencer and BigDye Terminators (Applied Biosystems, Foster City, CA, USA).
The alignment of DNA sequences was carried out using ClustalX with default parameters, checked by eye, and ambiguous regions excluded from further analysis (Thompson et al., 1997). Where ITS sequences from lichen specimens collected from the same region were identical, only one sequence was included in the phylogenetic analyses. This was the case for the New Zealand P. crocata specimens PC2, PC4 and PC6; the four P. crocata specimens from Oregon, USA; all five New Zealand P. neglecta sequences, and the four P. perpetua sequences from Oregon. In addition, the ITS sequences from specimens PC10 and PC12 were not included in analyses, as these showed no variation at parsimony informative sites compared with those from specimens PC9 and PC11, respectively. The ITS sequence of the P. perpetua specimen McCune et al. 24838 was identical to that of Reeb7-V99/1, and was also not included (Table 1). Additionally, New Zealand P. crocata specimens PC1, PC3 and Thomas 900 were excluded as they differed at only one or two nucleotide positions from New Zealand specimen PC2.
Table 1. Lichen specimens, collection sites and accession numbers of fungal internal transcribed spacer (ITS), β-tubulin and cyanobacterial tRNALeu (UAA) intron sequences
Aligned ITS sequences were used to construct maximum likelihood (ML), maximum parsimony (MP) and neighbour-joining (NJ) trees using paup* 4.0b10 (Swofford, 2002). To construct the ML tree, a heuristic search was performed with stepwise addition followed by tree bisection–reconnection (TBR) branch swapping (100 random-order replicates). The Akaike information criterion in the program modeltest ver. 3.06 was used to select the model of sequence evolution for ML analyses (Akaike, 1974; Posada & Crandall, 1998). The HKY + G model was selected (Hasegawa et al., 1985). Parameters were as follows: transition : transversion ratio = 1.8982; base frequencies: A = 0.2092, C = 0.2922, G = 0.2752, T = 0.2234; gamma distribution shape parameter = 0.2353. Bootstrap support was assessed (100 replicates) by a heuristic search using stepwise addition (10 random-order replicates) followed by TBR branch swapping. Maximum parsimony trees were produced using a heuristic search with 100 replicates of random-order sequence addition followed by TBR branch swapping. Support for the groups was estimated by 1000 bootstrap replicates, each with 10 random-order sequence addition replicates followed by TBR. Neighbour-joining trees were created with 1000 bootstrap replicates using the proportion of invariable sites and the rate variation parameters estimated by ML.
The β-tubulin sequences of representative P. crocata specimens from New Zealand, Australia, Chile and Canada, and P. neglecta specimens from New Zealand and Australia, were aligned and used in phylogenetic analyses. The SYM + I model of sequence evolution was used for ML and distance analyses; this model was selected using the Akaike information criterion in the programme modeltest ver. 3.06 (Zharkikh, 1994). The parameters used were: R(a) [A–C] = 2.4149; R(b) [A–G] = 6.2480; R(c) [A–T] = 0.0216; R(d) [C–G] = 2.9522; R(e) [C–T] = 15.2508; R(f) [G–T] = 1.0000, and the proportion of invariable sites = 0.6417. The MP analyses and assessment of bootstrap support for ML, MP and NJ trees were carried out as for the analyses of the ITS sequences.
Lichen specimens of three bipartite species of Pseudocyphellaria were obtained from Australia, Chile and Canada (Table 1). Two P. crocata specimens (PC7 and PC8) and two P. neglecta specimens (PN5 and PN6) were from Australia. Four P. crocata specimens (PC9–12) were collected from three locations in Chile, and two P. crocata samples (PC13 and PC14) and one P. perpetua sample (PP1) were from Canada. Representative thalli of each species were photographed (Fig. 1). All P. crocata specimens had white medulla, marginal and laminal yellow soralia, and yellow pseudocyphellae; in addition specimen PC10 had sparse apothecia. Both P. neglecta thalli had a white medulla, some phyllidiate margins and yellow pseudocyphellae, and specimen PN7 had apothecia. The P. perpetua specimen had a yellow medulla, and the marginal and occasional laminal soralia were yellow, as were the pseudocyphellae.
Fungal ITS sequences were compared to investigate mycobiont diversity, and the GenBank accession numbers of these sequences are listed in Table 1. In addition to the ITS regions, the PCR products included part of the genes encoding the SSU and LSU rRNA as well as the gene encoding the 5.8S rRNA. The SSU and LSU partial sequences were identical for all specimens. Sequence variation occurred at only one position within the 5.8S rRNA gene: here a nucleotide of the Chilean specimens PC11 and PC12 differed from those of the other specimens (data not shown).
A similarity matrix was constructed to examine variation between the ITS sequences of different specimens (data not shown). Most of the specimens from the same country had high percentage sequence identity. The ITS sequences of Australian P. crocata specimens, PC7 and PC8, had 99.6% identity; identity between the two P. neglecta specimens was 99.8%. Two of the Chilean specimens, PC9 and PC10, had 100% sequence identity and the other two specimens, PC11 and PC12, had 99.3% identity. However, overall there was considerable sequence diversity among Chilean samples, as sequence identity between the two groups, PC9/PC10 vs PC11/PC12, was much lower, ranging from 88.3 to 89.0%. Interestingly, the two Chilean P. crocata specimens collected from the same location (PC10 and PC11) did not have similar ITS sequences (89.0% sequence identity). The Canadian P. crocata specimens had 99.8% sequence identity and the Canadian P. crocata and P. perpetua ITS sequences had at least 99.8% identity: this was more than observed between P. crocata sequences from different countries which had between 86.7 and 99.3% identity. This analysis also showed that the Australian P. crocata ITS sequences had higher sequence identity to the P. neglecta sequences than to P. crocata sequences from Canada and Chile.
A comparison was made between the ITS sequences of PC7–14 and PN5,6 and the New Zealand P. crocata and P. neglecta specimens PC1–6 and PN1–4 reported by Summerfield et al. (2002). The ITS sequences of Chilean specimens PC11 and PC12 were least similar to those of the P. crocata and P. neglecta specimens from New Zealand with identity ranging from 88.1 to 89.0%. However, ITS sequences of Chilean specimens PC9 and PC10 had 98.9–100% identity to those of P. crocata and P. neglecta specimens PC1–6 and PN1–4. Additionally, the ITS sequences from P. crocata specimens from Australia and Canada had a high percentage identity to New Zealand P. crocata and P. neglecta specimens, ranging from 95.1 to 99.6%.
The ITS sequence data from this study and ITS sequences from GenBank were used to create ML, MP and NJ trees, all of which exhibited similar topology. Two ML trees were obtained, one of which is shown in Fig. 2. The topology of the second tree differed only in that the two Australian P. neglecta specimens were placed with the Australian P. crocata PC7 specimen. The P. crocata, P. perpetua and P. neglecta samples formed a moderately supported monophyletic group with bootstrap support of 71, 80 and 69% for ML, MP and NJ analyses, respectively. Within this clade, the three species did not form separate monophyletic groups; however, there was strong support for two clades (bootstrap values for one clade were 100, 99 and 100%; for the second clade, 87, 99 and 97% for ML, MP and NJ, respectively). Both clades contained P. crocata specimens: the smaller clade was comprised of P. crocata from the USA and one of the Chilean P. crocata specimens, while all remaining P. crocata, P. perpetua and P. neglecta specimens were in the larger sister clade. In this larger clade, the P. perpetua specimens and Canadian P. crocata specimens formed a monophyletic group with good bootstrap support from ML and MP analyses (90 and 89%, respectively) and moderate support from NJ (72%). The relationships among the P. crocata and P. neglecta specimens from Chile, New Zealand and Australia were poorly resolved in all analyses.
As observed by Miadlikowska et al. (2002), there was good bootstrap support for the two more distant monophyletic groups containing Pseudocyphellaria anomala and P. anthraspis, and P. rainierensis and P. aurata, respectively. The P. murrayi and P. dissimilis specimens included in our analysis formed a sister clade to the P. crocata/P. neglecta/P. perpetua cluster, although bootstrap support for this was moderate (66, 76 and 52% for ML, MP and NJ, respectively). Therefore our data were similar to those of Miadlikowska et al. (2002); however, the additional P. crocata and P. neglecta specimens included in Fig. 2 were resolved in a monophyletic group with P. perpetua specimens.
Sequences were also analysed for part of the β-tubulin gene from representative P. crocata and P. neglecta specimens from different geographical locations, and P. dissimilis and P. murrayi specimens were used as outgroups. The sequences contained 60 parsimony informative sites. The ML tree (Fig. 3) had near-identical topology to the MP and NJ trees (data not shown). The P. crocata and P. neglecta specimens formed a strongly supported monophyletic group with bootstrap support of 100% for the ML, MP and NJ analyses. This group contained two clades, the smaller of which contained two P. crocata specimens, one from the USA and one from Chile. Bootstrap support for this group was 100% for all three analyses. The second, larger clade contained P. crocata specimens from New Zealand, Australia, Chile and Canada as well as P. neglecta specimens from New Zealand and Australia. This clade had moderate bootstrap support of 67, 94 and 81% for ML, MP and NJ, respectively. Within this larger group, the P. crocata and P. neglecta specimens did not form distinct monophyletic groups, nor did the specimens cluster according to the geographical location of collection sites. This was in contrast to the clustering of Canadian specimens observed in Fig. 2; however, only one Canadian specimen was included in the analysis for Fig. 3. New Zealand, Australian and Chilean P. crocata specimens had identical sequences, and a second specimen from New Zealand (PC5) differed at three nucleotide positions in an ambiguous region that was not used in the phylogenetic analyses. However, outside this group P. crocata and P. neglecta specimens were intermixed.
Phylogenetic analyses of ITS sequence data for this subset of specimens resolved two monophyletic groups that were identical in species composition to the two equivalent clades of the β-tubulin trees (data not shown). The ITS trees showed excellent bootstrap support for the clade containing two P. crocata specimens from the USA and Chile (98, 100 and 99% for ML, MP and NJ analyses, respectively) and for the larger clade containing the P. crocata and P. neglecta (100% for all three analyses, data not shown). As observed for the β-tubulin data the specimens in this larger clade were not grouped by species or geographical origin, although when additional specimens were included there was some clustering according to geographical location (Fig. 2).
As a number of studies have employed the use of multiple DNA markers, we investigated the potential of other markers for distinguishing Pseudocyphellaria specimens. A fingerprinting method using a primer specific to fungal tRNA genes has been applied to the investigation of lichen diversity (Schmitt et al., 2002). However, for P. crocata this tRNA gene primer was not fungus-specific, as PCR products were obtained from the amplification of cyanobiont DNA (data not shown). Two studies, on different fungal species, found sequences of the gene encoding glyceraldehyde-3-phosphate dehydrogenase more informative than ITS (Camara et al., 2002; Myllys et al., 2002). Investigation of Pseudocyphellaria specimens found that the primers of Myllys et al. (2002) did not result in a single, specific PCR product, although a small region of this gene (< 200 bp) including one intron was amplified from Pseudocyphellaria specimens using redesigned primers (Summerfield, 2003). However, the partial glyceraldehyde-3-phosphate dehydrogenase gene sequences from P. crocata, P. neglecta and P. perpetua contained only seven parsimony-informative sites, and therefore were not used for further analyses (data not shown).
It was previously reported that P. crocata and P. neglecta from New Zealand had indistinguishable ITS sequences, but the two species were found to associate with different cyanobionts (Summerfield et al., 2002). To investigate whether this cyanobiont specificity was observed in specimens from a broader geographical range, comparison of cyanobacterial tRNALeu (UAA) intron sequences from the lichen specimens was undertaken. All the specimens examined in this study had variable region 1 sequences that contained repeats corresponding to class 1 as defined by Costa et al. (2002). Variation in the number of repeats and the presence and length of nonrepetitive insertions in these repeats accounted for the major differences between the introns (Fig. 4). Based on these differences, five intron types were identified in the Pseudocyphellaria specimens examined (Table 2).
Table 2. Cyanobacterial tRNALeu (UAA) intron associated with each lichen specimen (Pseudocyphellaria spp.)
Cyanobiont intron type
The two Australian P. crocata samples had different intron types with a 24-bp nonrepetitive insertion in the intron of specimen PC8 compared with that of PC7. These intron types were different from that of the Australian P. neglecta specimens, which were identical to each other and contained two fewer AATYHAA repeats than the Australian P. crocata sequences, and did not contain the insertion seen in the intron of PC8. The type 1 intron observed in specimen PC8 was the same as that identified in the six P. crocata specimens from New Zealand, and the type 3 intron of the Australian P. neglecta specimens was the same as reported in the four P. neglecta specimens from New Zealand (Summerfield et al., 2002).
Chilean specimens had intron types that were different from each other. The type 3 intron found in P. neglecta specimens was also present in P. crocata specimen PC9 and a different intron type (type 4), containing an additional 21 bp nonrepetitive insertion, was identified in PC10. Specimen PC11 had the same intron type as the Australian specimen PC8 (type 1), and PC12 had a type 5 intron which had two fewer TDNGATT repeats than the other intron types.
The cyanobacterial tRNALeu (UAA) introns from the Canadian P. crocata and P. perpetua thalluses were of type 1. This type 1 intron was the most frequently identified, being found in thalli from New Zealand, Australia, Chile and Canada.
New Zealand P. crocata and P. neglecta specimens have previously been shown to have indistinguishable ITS sequences, suggesting that they represent morphotypes of the same species (Summerfield et al., 2002). In the current study, 36 specimens were used to investigate the genetic variation of P. crocata, P. neglecta and P. perpetua over a broader geographical range using ITS sequences obtained experimentally, as well as data available in GenBank. Specimens of these three species were morphologically distinct, but phylogenetic analyses of their ITS sequences indicated they did not form separate monophyletic groups. Additionally, phylogenetic analyses of the β-tubulin sequences found that P. crocata and P. neglecta specimens did not form distinct monophyletic groups. The β-tubulin data provide support for the robustness of the ITS data in Fig. 2 as trees with similar topologies were obtained with both data sets. These data therefore suggest that P. crocata, P. neglecta and P. perpetua are morphotypes of the same phylogenetic species. Previously, morphological characters have been observed to be incongruent with phylogenetic data (Sochting & Lutzoni, 2003; for review see DePriest, 2004). At the species level, studies have shown both agreement and disagreement between morphology and phylogenetics. This includes the investigation of the Peltigera canina species complex that found concordance between molecular data and morphology in 15 of the 17 species examined (Miadlikowska et al., 2003). However, in the same study, the species Peltigera degenii was divided into two putative cryptic species that were not morphologically distinct. In contrast, the cryptic species Parmelia serrana, identified using molecular techniques, was subsequently found to be morphologically distinct from the species P. saxatilis (Crespo et al., 2002; Molina et al., 2004).
Morphological differences between the species P. crocata and P. neglecta are seen in their reproductive structures, P. crocata having soralia while P. neglecta has phyllidia and often has more apothecia (Galloway, 1988). Previous studies have indicated that the presence of soralia is not a reliable marker for delimiting species. Species pairs including Usnea florida/U. subfloridana and Umbilicaria antarctica/U. kapenii, which differ only in reproductive strategy, and putative species pairs from Roccellina capensis and Dendrographa leucophaea, both of which contain an asexual and sexual morphotype, have been examined using molecular techniques. In all four cases, specimens were found to form monophyletic groups in which the two putative species are intermixed (Lohtander et al., 1998a, 1998b; Articus et al., 2002; Ott et al., 2004). In these cases the phylogenetic groupings have led to the suggestion that these species be reduced to a single phylogenetic species exhibiting variety in reproductive mode.
The species P. perpetua is distinguished from P. crocata and P. neglecta by its yellow medulla and predominantly marginal soralia. Previous phylogenetic analyses suggested that P. crocata and P. perpetua form distinct monophyletic groups and this, coupled with morphological differences, led to the suggestion that they were distinct species (Miadlikowska et al., 2002). The P. crocata specimens used in the study of Miadlikowska et al. (2002) were from Oregon, USA. In our analyses (Fig. 2) these Oregon P. crocata specimens are located in a clade separate from the P. perpetua specimens (P. crocata US AF401979, P. perpetua US AF401977, P. perpetua Can AF401973). However, P. crocata specimens from different locations were grouped with P. perpetua specimens, therefore the two species are no longer resolved as two distinct monophyletic groups.
In some cases the ITS sequences appeared to vary more with geographical distribution than morphology. For example, the Canadian P. crocata and Canadian and US P. perpetua formed a single monophyletic group with good bootstrap support (Fig. 2). In contrast, the Chilean specimens did not group together: P. crocata specimens from Chile were placed in clades with P. crocata from the USA or P. crocata specimens from New Zealand. These Chilean specimens could not be distinguished by morphology (Figs 1b,c). These two clades may represent cryptic species with overlapping geographical ranges: one identified only in the USA and Chile; the other with a broader habitat range in both Northern and Southern Hemispheres. Similarly, Miadlikowska et al. (2003) suggested that P. degenii may be comprised of two cryptic species, one found in North America and the other in Europe. Additionally, in Nephroma helveticum some agreement was found between genetic variation and geographical origin, leading to the suggestion that some populations of this lichen could be recognized as distinct species (Lohtander et al., 2002).
The specific association of one cyanobiont with P. crocata and a different cyanobiont with P. neglecta specimens from New Zealand, observed by Summerfield et al. (2002), was not found in thalli collected from other countries. In total, five different cyanobionts were identified associated with P. crocata specimens. We found no relationship between the genetic variation of the fungus and that of the cyanobiont. Nevertheless, the Chilean examples had the most genetically diverse fungi and also had the most cyanobiont variation, the four specimens each having a different cyanobiont. Three cyanobionts were found in one lichen thallus each: these were a P. crocata specimen from Australia (PC7), and two specimens from Chile (PC10 and PC12). However, the cyanobiont with the type 1 intron was found in lichen thalli from all three countries and in both P. crocata and P. perpetua specimens (Table 2). Furthermore, the cyanobiont with the type 3 intron, previously found in P. neglecta thalli from New Zealand, was identified in thalli from Australia and Chile that were P. neglecta and P. crocata specimens, respectively.
Comparison with sequences in GenBank indicated that intron types 1, 2 and 4 had been identified in other lichen species from Finland and western North America. The lichen species containing these cyanobionts were classified into the Nephroma guild by Rikkinen et al. (2002). Fungi belonging to this lichen guild are hypothesized to share a common pool of compatible cyanobacteria. The Nephroma guild contains epiphytic lichens, and the Pseudocyphellaria species examined in this study can grow epiphytically but are also found in grassland and heaths. The remaining two cyanobionts (type 3 associated with thalli PC9, PN5 and PN6; type 5 associated with thallus PC12, Fig. 4) had introns that have not been identified in other lichen species. The closest match to the cyanobiont with the type 3 intron was with Nostoc commune strain MOA (data not shown). Free-living cyanobacteria were not included in the group of cyanobacteria associating with the Nephroma guild; however, the cyanobiont found in PC9, PN5 and PN6 has an intron with the same class of repeats as those cyanobionts associated with this guild. Interestingly, the cyanobiont with intron type 5 from PC12 did not match any existing intron type in GenBank. These data are largely consistent with the hypothesis that, rather than highly specific mycobiont/photobiont associations, there are pools of symbiotically competent cyanobacteria and green algae (Rikkinen et al., 2002; Yahr et al., 2004).
We thank Drs D. J. Galloway, J. Elix, G. Kantvilas and T. Goward for supplying specimens. We also thank Dr Galloway for verifying the identity of the new specimens collected for this study. We are grateful to Dr Judy Broom for advice on phylogenetic analyses and critical reading of the manuscript. This research was supported by a New Zealand Marsden grant (UOO805).