Introduction

The lichen association is a mutualistic interaction between a mycobiont (fungal partner) and a photobiont (green algal or cyanobacterial partner). The name of a lichen is applied to the fungal partner only; the photobiont has no nomenclatural status for classification of lichen associations. The association between the mycobiont and the photobiont produces a lichen with species-specific phenotypic characters. Even though phenotypic characters of lichens are the result of the interaction between two unrelated organisms, the evolutionary history of lichen associations is generally inferred from fungal phenotypic characters and incorporated into phylogenetic studies with nonlichenized fungi (Lutzoni 1997; Lutzoni et al. 2001; Gargas et al. 1995). Evolutionary studies of the photobiont have relied on phenotypic features of the photobiont maintained artificially in culture media (Friedl 1989; Hildreth and Ahmadjian 1981) or by DNA sequences using algal specific primers (Beck et al. 1998; Friedl and Zeltner 1994; Friedl and Rokitta 1997; Helms et al. 2001).

Phenotypic characters such as morphology and secondary metabolites and genotypic characters such as nucleotide sequences are commonly used to determine similarity and infer evolutionary relationships among species. Similar to other phenotypic characters, secondary structures of ribosomal RNA (rRNA) have also been used in evolutionary comparisons (Coleman et al. 1998; Lott et al. 1998; Piercey-Normore et al. 2006; Hausner and Wang 2005). Secondary structures of the spacer regions in rRNA have been reported to be involved in the rRNA processing for maturation of the ribosome (Good et al. 1997; Lalev and Nazar 1998; Lalev et al. 2000). The secondary structures retain a highly conserved configuration, especially where proteins interact and cleavage occurs during rRNA maturation, reflecting the critical nature of this process to survival of living organisms. The structure of the spacer regions thus provides an additional character to be used to examine evolutionary relationships (Coleman 2003).

Both the internal transcribed spacer 1 (ITS1) and the ITS2 rDNA regions contain distinct core structural features described in eukaryotes including fungi and algae. The fungal ITS1 and ITS2 secondary structures, based on Saccharomyces pombe and S. cerevisiae, contain smaller hairpins immediately adjacent to a centrally extended hairpin (Lalev and Nazar 1998; Côté et al. 2002). Differences between ITS1 and ITS2 occur in the primary DNA sequence and the size and location of small hairpins and bulges. Two working models have been used to describe ITS rRNA secondary structures; the hairpin model and the ring model (Joseph et al. 1999; Côté et al. 2002). A dynamic conformational model has been proposed for the ITS2 structure providing distinct roles for the ring model early in rRNA processing and the hairpin model later in rRNA biogenesis (Côté et al. 2002). Although most published reports on fungal rRNA secondary structures use the hairpin model as the working model (Lalev and Nazar 1998; Côté et al. 2002), algal ITS1 and ITS2 structures are based on a ring model in the Volvocales (Coleman et al. 1998; Mai and Coleman 1997). The algal ITS structures contain four hairpin loops around an open ring. The conserved core structures of both ITS regions contain specific subsequences and structures thought to be involved in enzymatic interactions during rRNA processing.

Phenotypic characters of lichens are influenced by the interaction between mycobiont and photobiont and by environmental conditions (Purvis 1997; Hawksworth 1973). Even though most lichen symbionts can be cultured separately, many are ecologically obligate (Honegger 1996), where only the symbiotic phenotype is found in nature, forming a thallus unlike either of the cultured partners. Variation in the symbiotic phenotype with the same fungal species and a different photobiont has been illustrated in photomorphs, species pairs, and sibling species (Purvis 1997; Hawksworth 1973). The Peltigerales contain several photomorphs (Laundon 1995) where a single fungal species will associate with either a green algal partner or a cyanobacterial partner at different times or in different habitats. Photomorphs may show differences in morphology, chemistry, or habitat preferences (Goffinet and Bayer 1997; Stenroos et al. 2003 and references therein). A species pair refers to one member of a fungal species that is fertile with apothecia (primary species) and another member that produces vegetative propagules (secondary species), which appear to be genetically isolated and have been given different names (Purvis 1997). Sibling species are species that show negligible morphological differences but produce different chemical substances (chemospecies) or are physiologically or ecologically different (W. L. Culberson 1986).

Because of the obligate and intimate relationship between the photobiont and the mycobiont, it has been hypothesized that lichen symbionts undergo long-term coevolution (Ahmadjian 1987). Coevolution could be examined using a diversity of approaches including congruent patterns in phylogenetic trees (cospeciation [Paterson and Gray 1997]), but also the evolution of polymorphisms, directional selection, host specificity, speciation, and adaptation (Thompson 1994; but see Janzen 1980). The lichen forming fungal family Cladoniaceae has revealed no overall cospeciation between algal and fungal partners in the large and widely distributed genus Cladonia (Piercey-Normore and DePriest 2001). Since DNA sequences encode conserved secondary RNA structures, we examined both phylogenetic comparisons and the evolution of polymorphisms in the secondary structures within one of the seven taxonomic sections of the genus Cladonia collected only within North America. We report phylogenetic histories of both symbionts within Cladonia using primary ITS rDNA sequences and secondary rRNA ITS structures. We also provide evolutionary comparisons to invoke coevolutionary hypotheses and revisit species concepts in lichens.

Materials and Methods

Taxon Sampling and Identification

Specimens from the genus Cladonia section Cladonia were collected from Manitoba and Nova Scotia (Table 1). Species identifications were determined by morphological (Ahti 2000; Brodo et al. 2001; Thomson 1984) and chemical analysis using standardized thin-layer chromatography (C. F. Culberson 1972). Identifications were confirmed by T. Ahti (Helsinki University). Collected materials are deposited in the University of Manitoba Herbarium (WIN). Sectional taxonomy follows the classification of Ahti (2000) with the exception that Cladina is subsumed within Cladonia (Ahti and DePriest 2001); the C. chlorophaea group follows Thomson (1967); the C. gracilis group follows Ahti (1980); the C. pyxidata group follows Aptroot et al. (2001); and the C. verticillata group as “the group with centrally proliferating scyphi” follows (Ahti 2000, p. 83), currently listed as C. cervicornis ssp. verticillata (Esslinger 2006). Fungal and algal ITS nucleotide sequences representing species within Cladonia sect. Cladonia were retrieved from NCBI GenBank and included in the analyses (Table 1).

Table 1. Collection information and accession numbers for symbionts in the Cladoniaceae and Trebouxia
Full size table

DNA Isolation and Amplification

Single podetia that were in good condition were used for isolation of genomic DNA using a CTAB (cetyltrimethylammonium bromide) extraction protocol modified from Grube et al. (1995). Polymerase chain reaction (PCR) was performed on isolated genomic DNA to selectively amplify the fungal and algal ITS1 and ITS2 and the 5.8S from nuclear ribosomal DNA (rDNA) using taxon-specific primers 1780F-5′ (fungal specific), 1780A-5′ (algal specific) (Piercey-Normore and DePriest 2001), and ITS4-3′ (White et al. 1990). Amplification reactions were performed in a total reaction volume of 300 μl (six 50-μl reaction volumes combined) containing 10–50 ng DNA per reaction, PCR buffer (50 mM KCl, 100 mM Tris-HCl [pH 8.3]), 2.0 mM MgCl2, a 200 μM concentration of each dNTPs, a 0.5 μM concentration of each primer, and 2 units of Taq DNA polymerase (GibcoBRL, Burlington, ON, Canada). For samples that were difficult to amplify 5% DMSO or 1× PCRx Enhancer (GibcoBRL, Burlington, ON, Canada) was added to reactions. DNA amplifications were performed on a Techne Genius thermal cycler (Fisher Sci, Nepean, ON, Canada). Both algal and fungal ITS rDNA regions were amplified with an initial denaturing temperature of 95°C for 5 min, followed by 33 cycles at a denaturing temperature of 94°C for 45 s, an annealing temperature of 56°C for 45 s, and an extension temperature of 72°C for 1.5 min.

DNA Sequencing and Alignment

PCR products were agarose gel purified by crushing frozen blocks of agarose, pipeting the buffer containing DNA into a tube, and subsequently precipitating it with 0.2 vol 5 M NaCl and 2.5 vol 100% ethanol. DNA was quantified on 1% agarose gel and stained with ethidium bromide. DNA was sequenced using BigDye terminators on a 377 ABI DNA sequencer (University Core DNA and Protein Services, University of Calgary, Calgary, Alberta). Sequences were edited using Sequencher 4.2.2 (GenCodes Corp., USA), aligned visually using Se-Al v2.0a11 (Rambaut 2001), and imported into PAUP 4.0 (Swofford 2003).

ITS Secondary Structure

The ITS regions and the 5.8S rDNA were defined based on the conserved sequence at the 3′ end of the 18S gene, the 5′ and 3′ ends of the 5.8S gene, and the 5′ end of the 26S gene (Hausner and Wang 2005). Since these regions were not present in all sequences, they were compared with DQ219313, and similarity of the ends of the ITS1 and ITS2 delimited the regions in all sequences. The coding regions, required for the basal stems in ITS secondary structures, were omitted from the models presented in Figs 1–3. Similarly, algal sequences were delimited by comparison between MN1172 (this study) and known sequences of Trebouxia s.s. (e.g., AF242470). Algal ITS2 sequences are missing seven bases on the 3′ end because of ambiguous base calling. The stem loop structures were folded using the mfold web server (http://www.bioinfo.rpi.edu/applications/mfold/old/rna/form1.cgi [Zuker 2003]). In addition to folding each ITS region using the entire ITS sequence, each hairpin was also folded separately to test the reliability of the folding. If more than one fold was produced, the final fold was based on comparisons with previously published ITS1 and ITS2 folds, maximization of the hydrogen bonding forming solid stems, and the largest negative Δg value (free energy). Free energy values were reported for structures excluding the stems formed from the SSU and LSU coding regions.

Fig. 1.
figure 1

Secondary structures of ITS1 rRNA for eight mycobiont species in the Cladoniaceae. The two major hairpins are labeled I and II. Lateral hairpins (and bulges) are labeled with lowercase letters. Labels on the first structure apply to all structures. The location and substitution for members within a species are indicated in a box adjacent to the base that was substituted. The inset near C. pyxidata shows an alternative fold to create compensatory base pairings in hairpins I and Ia. The free energy value (dG), obtained from mFold (Zuker 2003), number of samples compared (n), and length of the ITS region (L) are shown for each species.

Full size image
Fig. 2.
figure 2

Secondary structures of ITS2 rRNA for seven mycobiont species in the Cladoniaceae. Cladia was not included because the ITS2 sequence was incomplete. Four major hairpins are labeled I–III in C. merochlorophaea and IV in C. pocillum, representing two models. Labels on the first structure apply to all structures. The location and substitution for members within a species are indicated in a box adjacent to the base that was substituted. The free energy value (dG), obtained from mFold (Zuker 2003), and length of the ITS region (L) are shown for each species.

Full size image
Fig. 3.
figure 3

Secondary structures of ITS1 and ITS2 rRNA for the photobiont including changes for all taxa in this study. The four major hairpins are labeled I–IV. The location and substitution for 59 taxa are indicated in a box adjacent to the base that was substituted. The universally alignable “GGU” is indicated by a solid vertical line on the ITS2 structure. The free energy value (dG) was obtained from mFold (Zuker 2003). The inset shows an alternative fold for ITS1.

Full size image

Phylogenetic Analyses

Algal and fungal datasets were analyzed by maximum parsimony methods implemented in PAUP 4.0 (Swofford 2003). Maximum parsimony was performed using the tree bisection and reconnection (TBR) branch swapping option. Heuristic searches were conducted using 100 random addition replicates with a limit of 10 trees per search and 500 bootstrap replicates (Felsenstein 1985). Cladia aggregata was assigned as the outgroup taxon in the fungal phylogeny. Alternative analyses were performed to test whether polyphyly was a result of ambiguous DNA sequence alignment, large distance with the outgroup taxa, or variability in RNA secondary structure. These analyses were as follows. (1) Positions 463–472 were omitted from the analysis. (2) The two Cladonia and the non-Cladonia outgroup taxa were omitted from the analysis and the tree was midpoint rooted. (3) A neighbor joining (NJ) analysis was performed on the entire dataset using the default settings of uncorrected “p” distances and systematically broken ties. (4) A parsimony analysis was performed on hairpins I and Ia in ITS1. (5) Alternatively, a maximum parsimony analysis was performed on hairpins II, IIa, and IIb in ITS1. (6) Maximum parsimony analyses were performed on the ITS1 and ITS2 regions separately.

Trebouxia erici was the outgroup taxon in the algal phylogeny. Trees produced from the data subset were midpoint rooted. An incongruence test was performed to compare subsets of the data that included the nine symbiont partners. The Kishino-Hasagawa (KH; Kishino and Hasagawa 1989) test was implemented in PAUP 4.0. The null hypothesis was that the algal ITS tree produced the same phylogenetic history as the fungal ITS tree. If p< 0.05, the null hypothesis was rejected and the symbionts produced the same phylogenetic histories, suggesting that coevolution has occurred (Paterson and Gray 1997).

Results

Sequence Variation

The combined ITS alignment was 586 bp long in the fungus and 570 bp long in the alga. The range of the fungal ITS actual lengths was 558–566 bp for complete sequences and 528 bp (including 62 missing bases) for Cladia aggregata. The range of the algal ITS actual lengths was 552–559 bp. Pairwise similarities within Cladonia sect. Cladonia were 88%–100% and included two species outside sect. Cladonia similarities that were >77%. The fungal ITS1 had 144 polymorphic sites, the 5.8S had 5 polymorphic sites, and the ITS2 had 69 polymorphic sites. Pairwise ITS similarities within the natural lichen algae and known species of Trebouxia were 95%–100%. The algal ITS1 had 31 polymorphic sites, the 5.8S had none, and the ITS2 had 28 polymorphic sites in addition to seven 1- to 5-bp insertions.

The mycobionts had a shorter ITS2 (164–170 bp) than ITS1 (238–248 bp). The photobiont showed the opposite trend. It had a shorter ITS1 (198–200 bp) than ITS2 (206–214 bp). The range of GC content in each of the algal ITS1 (53%–63%) and ITS2 (43%–55%) was greater than that in the mycobiont ITS1 (49%–57 %) and ITS2 (52%–57%). A larger proportion of compensatory base changes (CBC) was present in the algal ITS2 than the ITS1 stem regions (Fig. 3). The alga also had a larger number of loop substitutions and indels in the ITS2 than the ITS1. The fungus showed the opposite trend. A larger proportion of CBC was present in the fungal ITS1 than the ITS2, but fewer total loop substitutions (Figs. 1 and 2).

Variation in rRNA Secondary Structure

The mycobiont ITS1 rRNA secondary structure showed similar morphology in all species examined (Fig. 1). The model consisted of a hairpin (I) at the 5′ base of a single central hairpin (II). Hairpin I contains a short lateral hairpin (Ia) located near a highly conserved hairpin base containing GAG paired with CUC, which has a single CBC in C. cervicornis ssp. verticillata (Fig. 1). The central hairpin is highly conserved and has a short hairpin (IIa) before the 5′ end of the apex of the central hairpin. Another short hairpin (IIb) with a basal bulge is present on the opposite side of the central hairpin near the 3′ end of the ITS1. Bulges of variable sizes are present on the lower 5′ end of hairpin I where the outgroup, Cladia, has a short lateral hairpin. Even though substitutions are present in hairpin II, the structure and the terminal loop are preserved among all species examined. Structural variation in the ITS1 was evident in the size of the 5′ bulge, the size of the hairpins, and the occurrence of bulges adjacent to hairpins IIa and IIb. All hairpins in ITS1 were folded separately to test for alternative folds.

The mycobiont ITS2 rRNA structure was more conserved among species than the ITS1 structure but showed two possible models. One of the ITS2 models contained three hairpins (I, II, and III) including the central hairpin (III) and the other model contained four hairpins (IV in C. pocillum; Fig. 2). Hairpin I is identical in all specimens examined. Hairpin II shows more variation than hairpin III in five species examined (Fig. 2). The fourth hairpin is shown in C. pocillum, which has the lowest free energy value, at –65.3 kcal/mol. A similar ITS2 structure for C. pocillum was folded that had no 3′ hairpin and it was less stable, at –63.5 kcal/mol energy. Alternative folds for the other species produced a fourth hairpin similar to that presented in C. pocillum. Since free energy values for both models are similar, they may reflect two possible morphologies for the ITS2 structure. The single-stranded region that connects the four hairpins is purine rich. Since the 3′ end of the ITS2 in the outgroup taxon was absent from the GenBank accession, the ITS2 structure of Cladia could not be compared with the ITS2 structures in species of Cladonia. Structural variation was present in small bulges on the hairpins and the occurrence of a hairpin on the 3′ side of the central hairpin in C. pocillum. All hairpins in ITS2 were folded separately to test for alternative folds.

The photobiont ITS1 secondary structure contains four main hairpins (I–IV) (Fig. 3). A short hairpin before hairpin I, and two stems in place of the single-stranded ring, is present in some models but they have no CBC support. A second model was produced from a shift in the base pairing of helix II (Fig. 3). The photobiont ITS2 structure was more similar to the mycobiont ITS2 than the photobiont ITS1 was to the mycobiont ITS1 structure. The photobiont ITS2 structure also contains four hairpins (I–IV) around an open ring of single-stranded bases. The fourth hairpin is very short but conserved among all algal sequences examined in this study. All secondary structures are putative models and biochemical experimentation is necessary to support their configuration.

Fungal Evolution

All species complexes and many species were polyphyletic, with clades separating the members supported by high bootstrap values and large numbers of CT transitions (Fig. 4). The Cladonia chlorophaea species complex, represented by C. chlorophaea s.s., C. merochlorophaea, and C. grayi, is polyphyletic, forming three groups. Within the C. chlorophaea complex C. merochlorophaea was monophyletic with 87% bootstrap support, C. grayi was polyphyletic, and the only sample of C. chlorophaea included in the analysis clustered with C. fimbriata with 87% bootstrap support. A larger number of substitutions were present in a single chemospecies, C. grayi, in both ITS1 and ITS2 structures than between two well-defined species, C. chlorophaea s.s. and C. fimbriata (Figs. 1 and 2).

Fig. 4.
figure 4

Evolutionary history of mycobiont species in Cladonia sect. Cladonia represented by 1 of 71 most parsimonious trees of 421 steps based on the combined ITS1, 5.8S, and ITS2 nucleotide sequences. Bootstrap values >50% are shown above the branches. Number of CT transitions is shown below the branches. Cladia aggregata was assigned as the outgroup. The phylogenetic analysis was based on 129 informative characters of 586 total characters, CI=0.6532, RI=0.7033.

Full size image

Cladonia pyxidata and C. pocillum were also polyphyletic and their members were separated into three clades with bootstrap support and an unsupported branch that collapses in the strict consensus tree. Cladonia pocillum formed a sister clade to four different species and also formed a strongly supported clade with C. macrophyllodes with 100% bootstrap. Two individuals of Cladonia pyxidata were unresolved but outside a highly supported clade. One member of C. pyxidata clustered with C. cornuta with 91% bootstrap support and five CT transitions.

The Cladonia gracilis species complex, including C. cornuta, C. ecmocyna, C. gracilis, and C. maxima, was polyphyletic and the lineages contained only one to three CT transitions. One member of C. ecmocyna was basal to all other members of sect. Cladonia in a clade supported by 98% bootstrap and nine CT transitions. The largest cluster with eight members collapsed in the strict consensus tree (data not shown). Two of the three members of C. cornuta clustered together with 98% bootstrap support. The other C. cornuta formed a strongly supported dichotomy with C. pyxidata. C. coniocraea formed a strongly supported clade with C. rei, C. fimbriata, and C. chlorophaea.

The C. cervicornis species complex, consisting of C. cervicornis ssp. verticillata and C. macrophyllodes, fell into two separate but strongly supported clades. The three members of C. cervicornis ssp. verticillata formed a monophyletic clade and C. macrophyllodes clustered with C. pocillum at 100% bootstrap support.

Most of the alternative analyses produced similar results to that presented in Fig 4. The differences were in the positions of the unresolved taxa such as DQ530203 and AF455192 C. ochrochlora, DQ219313 C. gracilis, and DQ530196 C. ecmocyna. Analyses of the separated ITS regions produced slightly different trees. Two of the Cladonia outgroup taxa (C. subcervicornis and C. rangiferina) clustered with the ingroup. The unresolved taxa were also placed in different clades. However, the highly supported clades and dichotomies were retained in both ITS analyses. The pattern of polyphyly was also consistent between ITS1 and ITS2.

Algal Evolution

Five of six known species of Trebouxia placed in the analysis for reference showed close relationships with amplified regions in natural lichen algae, not only those associated with Cladonia but also two other genera, Cladia and Pycnothelia (Fig. 5). Photobionts of eight species of Cladonia and one genus (Pycnothelia) clustered with T. glomerata, T. pyriformis, and T. irregularis with 100% bootstrap support. However, photobionts associated with other species of lichen-forming fungi fell outside this clade. In some cases collection location seems to have some influence on photobiont association such as with C. gracilis and C. rangiferina from Manitoba and C. perforata from Florida. In other cases collection location has little influence on the photobiont since photobionts from C. pocillum collected in Manitoba fell into three different clades in the tree.

Fig. 5.
figure 5

Evolutionary history of natural lichen algae amplified from Cladonia sect. Cladonia and six known species of Trebouxia represented by one of nine most parsimonious trees of 113 steps based on the combined ITS1, 5.8S, and ITS2 nucleotide sequences. Bootstrap values >50% are shown above the branches. Number of CT transitions is shown below the branches. Trebouxia erici was assigned the outgroup. General habitat is in parentheses following the accession number: TN, temperate boreal and noncalcareous; TC, temperate boreal and calcareous; BN, boreal and noncalcareous; BC, boreal and calcareous; SC, subarctic and calcareous. The phylogenetic analysis was based on 39 informative characters of 570 total characters, CI=0.8584, RI=0.9512.

Full size image

The most common type of substitution was CT transitions, which were mapped onto the algal phylogeny (Fig. 5). The two most distant species groups, T. erici and T. magna, contained long branches and had the highest number of CT transitions and the highest total number of substitutions. The clade containing T. irregularis, T. pyriformis, and T. glomerata was supported by 100% bootstrap and two CT transitions. A clade containing four species from South America and one from North America was supported by 93% bootstrap and three CT transitions. A lineage of photobionts from C. perforata collected in Florida was supported by 87% bootstrap and two CT transitions. Other photobionts also from C. perforata and collected in Florida were not included in this clade.

Symbiont Evolutionary Comparisons

Three observations can be made from phylogenies of symbiont partners in this study (Fig. 6). (1) Natural lichen algae clustered together corresponding to geographic location. One strongly supported cluster represented algae that occur on highly calcareous gravel. This clade was split into two highly supported clades representing algae in a temperate climate and algae in an open subarctic climate. The third clade represented algae in a less calcareous sandy area with a temperate climate. This result was not as strongly supported in Fig. 5. (2) Terminal branches of the fungi are long and the internal branches are short. The opposite trend is present in the algal partner; the terminal branches are short and the internal branches are long. (3) The phylogenetic trees of symbiont partners are not congruent (Fig. 6; KH p=0.0002). Two fungal lineages corresponded with two algal lineages collected at the same location in Manitoba.

Fig. 6.
figure 6

Evolutionary histories including only those lichen associations from which the ITS1, 5.8S, and ITS2 rDNA regions of both symbionts were sequenced. Bootstrap values >50% are shown above the branches. Both trees were midpoint rooted. A The single most parsimonious tree for the natural lichen algae; length = 20 steps, CI=1.0000, RI=1.0000. B The single most parsimonious tree for the lichen-forming fungal partner; length = 140 steps, CI=0.7857, RI=0.5385.

Full size image

Discussion

Variation in Primary DNA Sequences

The length of the ITS1 in the genus Cladonia (238–248 bp) is longer than that reported for outgroup Cladia (210 bp), but it is within the range reported for other ascomycetes (Hausner and Wang 2005) and plants, and less than that reported for vertebrates (Torres et al. 1990). The length of the fungal ITS2 region ranges from 164 to 170bp, also within the range reported for ascomycetes, but shorter than that reported for plants and vertebrates (Torres et al. 1990). The algal ITS1 and ITS2 regions are within the length reported for Chlamydomonas species (Coleman and Mai 1997) and natural lichen algae that associate with Cladonia (Piercey-Normore and DePriest 2001; Yahr et al. 2004; Piercey-Normore 2004) but shorter than that reported for the Volvocales (Coleman et al. 1998), the genus Cladophora (Bakker et al. 1995 [except the ITS2 length was within range]), and natural lichen algae that associate with Evernia (Piercey-Normore 2006) and Letharia (Kroken and Taylor 2000).

CT transitions provide a mechanism to maintain secondary structure by providing CBC and are thought to accumulate with evolutionary distance (Coleman et al. 1998). This is supported by the algal phylogeny (Fig. 5). The numbers of CT transitions were mapped onto the algal tree to investigate an additional form of support for taxonomic relationships. Some long lineages and those with high bootstrap support also had a larger number of CT transitions but the pattern was not consistent throughout the tree. For example, the lineage between T. magna/T. erici and the remainder of the taxa has bootstrap support less than 50% and three CT transitions. On the other hand, the lineage to the algae from C. rangiformis/C. fimbriata has 93% bootstrap support but no CT transitions. A similar trend is present in the fungal tree where the higher bootstrap support and longer branches, indicating a larger number of total substitutions, have higher numbers of CT transitions. However, the trend is not consistent throughout the tree (Fig. 4). Examination of CT transitions, other types of substitutions, and CBC where both nucleotide positions changed may yield patterns in a higher-level taxonomic study.

RNA Secondary Structures Are Conserved

The ITS2 secondary structure is more conserved than that of the ITS1 in both symbionts. The ITS1 rRNA secondary structure in the mycobiont is similar to that proposed by Hausner and Wang (2005) and Lalev and Nazar (1998), with variable secondary hairpins and bulges extending from a single central hairpin (Fig. 2). Hairpins I and Ia, in the ITS1 model (Fig. 1), are structurally similar to those proposed for the ITS2 of Ceratocystis by Hausner and Wang (2005). The short hairpin and bulge (hairpins IIa and IIb) are variable in size in different species and form a longer hairpin in the outgroup, Cladia (Fig. 1), but even the highly compact ITS folds of the Microascales formed a hairpin on the right side of the central hairpin (Hausner and Wang 2005). The ITS2 rRNA secondary structures in the mycobiont are consistent with the structures reported by Hausner and Wang (2005). Highly divergent eukaryotic species (algae, plants, animals, and fungi) are known to share a common ITS2 structure (Coleman 2003; Lalev and Nazar 1998). One of the models presented for the ITS2 structure in the Cladoniaceae contains hairpin IV (C. pocillum; Fig. 2). Hairpin IV contains the most rapidly evolving nucleotide sequence of the four hairpins (Coleman 2003) and is present for ITS2 models presented for Ceratocystis but is absent from models presented for the Microascales (Hausner and Wang 2005). The purine-rich single-stranded region connecting the four hairpins is a characteristic of the ITS2 structure (Coleman 2003).

The largest number of substitutions in the fungus is clustered around hairpin I in the ITS1 and hairpin II in the ITS2 and, to some extent, the 3′ side of the central hairpin (Figs. 1 and 2). These regions correspond with the same variable regions reported in the Curcurbitaceae (Torres et al. 1990). Athough the 5′ region of the DNA sequence of ITS1 was more divergent in Cladia, the same hairpin was produced in the folded ITS1, suggesting preservation of the hairpin and homology of the primary sequence. Homology of the 5′ end of the ITS2 was also reported in divergent species of Candida (Lott et al. 1998).

The photobiont ITS1 rRNA structure contains four hairpins (Coleman et al. 1998) around a single-stranded ring (Fig. 3). The algal ITS2 ring structure clearly resembles the “four-fingered hand” containing conserved features described by Mai and Coleman (1997). The GGU triplet on the 5′ side of helix III and “universally alignable” positions in the Volvocales can also be identified in Trebouxia (Mai and Coleman 1997). Nucleotides at the bases of each of the four hairpins are conserved from the Volvocales to Trebouxia. Purines constitute approximately 50% of the bases in the single-stranded region of the structure. The uracil bulge in hairpin II is present in Trebouxia and is a universal feature in the Volvocales. The CAGG sequence at the base of helix IV is universally conserved both in Trebouxia and in the Volvocales (Mai and Coleman 1997). The base of hairpin III in this study shows fewer bulges than in the Volvocales (Mai and Coleman 1997). The secondary structures of the algal ITS transcripts are highly conserved and provide homologous regions even between these lichen algae and members of the Volvocales (Coleman et al. 1998).

Species Evolution of the Mycobiont

Ten species and four species complexes, with more than one representative, were examined for monophyly. Only three species produced monophyletic clades with bootstrap greater than 70%. All other species and species complexes were not monophyletic. Monophyletic species included C. cervicornis ssp. verticillata, collected from the same geographic region (MB, Canada), C. subulata, collected from Germany and Finland (Stenroos et al. 2002), and C. merochlorophaea from Finland (Stenroos et al. 2002) and MB, Canada (Fig. 4). Geographic region did not influence monophyly in these species. As would be expected with closely related individuals, members of two of these monophyletic species (C. cervicornis ssp. verticillata and C. merochlorophaea) showed a low number of substitutions and substitutions were compensatory (Figs. 1 and 2). Examination of a larger number of individuals is required to include more of the genetic and phenotypic variation within each species.

Cladonia pyxidata has three individuals representing the species and shows a larger number of changes than either C. gracilis, with five individuals, or even between two species, C. chlorophaea s. s. and C. fimbriata (Figs. 1 and 2). More than 50% of the substitutions are one-sided CBC in both ITS regions in each of the monophyletic species, C. merochlorophaea and C. cervicornis ssp. verticillata, and in two polyphyletic species, C. grayi and C. gracilis. All three of the three stem substitutions in helices II, IIa, and IIb in C. pyxidata are CBC. In helices I and Ia the large number of substitutions is compensated for by a single base shift to the left and a bulge created by insertion of a “G” (see inset, Fig. 1). Other substitutions are CBC. Although the members of the C. pyxidata group are morphologically similar, Aptroot et al. (2001) recognized Cladonia monomorpha as a distinct species within the C. pyxidata group in Europe and suggested that it may have been overlooked in North America. Cladonia pocillum, also within the C. pyxidata group (Aptroot et al. 2001), shows a slightly lower level of diversity (Figs. 1 and 2) but similar polyphyly (Fig. 4) as C. pyxidata. Phenotypic plasticity was reported in C. pocillum on basic and acidic substrates (Gilbert 1977). The two species, C. pocillum and C. pyxidata, may be ecological “sibling” species (W. L. Culberson 1986; Purvis 1997) showing changes in morphology of the basal squamules with changes in pH of the substrate.

Cladonia gracilis is divided into six morphologically divergent subspecies (Ahti 1980), three represented in this study. Although the subspecies of C. gracilis are polyphyletic (Fig. 4), they collapse with all other ingroup taxa in the consensus tree. In this study the C. chlorophaea group is represented by C. chlorophaea s. s., C. merochlorophaea, and C. grayi. Thomson (1967) refers to these species as chemical variants of the species C. chlorophaea delimited by production of specific secondary compounds. W. L. Culberson (1986) refers to these as morphologically uniform species with “races” that show differences in behavior, physiology, or ecology. Cladonia merochlorophaea forms a monophyletic group (Fig. 4) but it was comprised of only two samples. Cladonia grayi forms a polyphyletic group even though it is considered a chemical variant (Thomson 1967) and was suggested to belong to an interbreeding population of C. merochlorophaea (C. F. Culberson et al. 1988). If they interbreed, they would be considered the same biological species. Cladonia fimbriata, a close ally to C. chlorophaea, is not part of the species complex according to Thomson (1967) but it is included in the more broadly defined complex according to C. F. Culberson (1986). Cladonia fimbriata and C. chlorophaea cluster together with 84% bootstrap support. Perhaps the chemical and subtle morphological differences in the C. chlorophaea complex reflect a plasticity in adaptation to habitat and photobiont association rather than separate biological species.

Phenotypic plasticity, including chemical plasticity, toward changes in the environment (Hawksworth 1973; Purvis 1997) may be one explanation for the polyphyly present in the Cladoniaceae but it requires further study with larger numbers of individuals. Another explanation for polyphyly includes paralogous ITS regions or incomplete homogenization of rDNA repeat units. However, the electrophaerograms did not show any superimposed peaks, indicating the presence of alternative nucleotides, at any position.

Algal Association Is Structured by Environmental Conditions

Five known species of Trebouxia are represented by amplified algal ITS from species of Cladonia in this study. Some geographic affiliation with algae is evident when broadly defined locations are mapped (Fig. 5). When environmental features were mapped onto Fig. 5 there was no segregation of temperate-boreal and boreal taxa, but the subarctic taxa were in a separate clade. A pattern may be present if more information was available for those taxa on calcareous and noncalcareous substrates. All samples except one sample in the derived clade containing T. glomerata and T. pyriformis were collected from noncalcareous substrates. This pattern was more evident when environmental features were mapped onto Fig. 6. Further investigation with a larger sample size and accurate substrate measurements is required to test this hypothesis. The lichen association may rely on algal requirements in the environment in addition to genetic compatibility between symbionts. Limitation of photobiont availability by environmental features was evident from other studies. Beck et al. (1998) reported 4 species of green algae to be associated with 10 species of lichen-forming fungi in a Physcietum adscendentis Ochsner corticolous lichen community. In another study, Beck (1999) reported a single species of photobiont in association with nine species of lichen-forming fungi on a heavy metal-rich rock where the environment was more limiting. If the environment limits algal distribution, then independent lichenization events from ascospores (Ohmura et al. 2006) or algal switching (Ott 1987; Ott et al. 1995; Friedl 1987) may be very common.

Independent lichenization events would suggest independent evolutionary histories of the symbiotic partners. This study showed a single coordinated speciation event between the algae and the fungi in Cladonia sect. Cladonia (Fig. 6), but no overall cospeciation. This result was consistent with the finding of Piercey-Normore and DePriest (2001). In this study comparison of phylogenetic trees was limited to taxa that were in a narrower taxonomic and geographic range, Cladonia sect. Cladonia collected in North America, compared with the entire genus Cladonia from worldwide locations used by Piercey-Normore and DePriest (2001). If cryptic species are present within fungal lineages, and environment and photobiont influence the symbiosis, comparison of partners using a single representative species from diverse environments would probably not yield congruent symbiont phylogenies. Investigation of coevolutionary patterns must also incorporate photobiont and habitat factors. Rikkinen et al. (2002) reported epiphytic and terrestrial guilds of Nostoc in association with lichen fungi. Although the symbionts in most lichen associations are ecologically obligate, they have different rates of sequence evolution (Zoller and Lutzoni 2003; Piercey-Normore and DePriest 2001), and they have maintained DNA sequence and RNA secondary structures typical of the phylogenetic lineages to which they belong. The impact of environment on fungal plasticity and algal association requires studies with larger numbers of samples within each species and careful records of environmental characteristics.

The Species Concept in Lichen Associations Revisited

Since the lichen is named after the fungal partner, a phylogenetic analysis of the fungal genome represents the evolutionary history of the fungal partner and hence the lichen. Even when species are not monophyletic, the hereditary nature of genes cannot be refuted. The genes, therefore, reflect evolutionary histories. Independent studies have shown polyphyly among species in the Cladoniaceae (Stenroos et al. 2002) and in a species pair in the Physciaceae (Myllys et al. 2001). Lichen fungi are known to exhibit plasticity with changes in environmental conditions such as photosymbiodemes, species pairs, and sibling species (Purvis 1997; Hawksworth 1973; Honegger 1996), with numerous examples including Goffinet and Bayer (1997), Rikkinen (1997), Pintado et al. (1997), Wedin et al. (2004), and Sojo et al. (1997). The lichen is not a direct reflection of the fungal genome but, rather, the result of a complex mixture of epigenetic effects of biosynthetic pathways coordinated by at least two widely divergent organisms. If an alga (free-living or lichenized), already present and adapted to a habitat, associates with a germinating fungal spore that has landed in the same area, the fungus must conform to the conditions imposed on it by the alga and the immediate environment. It is not known how much morphological variation is due to the algal effects and how much to the environmental effects.

The variation in fungal morphology produced from association with different photobionts (Armaleo and Clerc 1991; Honegger 1996) or different environmental conditions (Gilbert 1977; Rikkinen 1997) imposed on the same fungal species suggests that the morphology of a lichen may change depending on the environment. Hawksworth (1973) suggested that species concepts in lichens are defined by discontinuities in morphology, chemistry, ecological requirements, or geographic distributions, especially where genetic differences provide support for these discontinuities. However, genetic differences in Cladonia do not support these discontinuities. Substrate is important for defining many lichen associations (Brodo 1973), and some photobionts seem to be segregated based on environment features (Fig. 6). The morphology of a lichen-forming fungus may be influenced by both photobiont and environment to a greater extent than that of a non-lichen-forming fungus because of their slow growth rates (Hawksworth 1973). If slow-growing lichen fungi can produce a wide range of phenotypic responses to the changing environments, they would have an adaptive advantage over less plastic organisms. Detailed investigations would be required to provide evidence for this theory. However, if phenotype is currently the basis on which lichen-forming fungal species are classified, and phenotypic ranges exist within the same fungal “species,” then the phenotype would reflect the environmental features more than the genetically defined fungal species.