High-throughput genome sequencing of lichenizing fungi to assess gene loss in the ammonium transporter/ammonia permease gene family

Genome assembly completeness

These data were of interest as of the eight lichens sequenced, four were actually metagenome projects. Lichens can be seen as micro-communities, hosting not just the mycobiont (fungal) and photobiont (algal or cyanobacterial symbionts), but a rich internal microbiota composed of bacteria and endolichenic fungi [64, 65]. As lichenizing fungi have only infrequently been successfully cultured compared to other fungi (although see [66, 67]), and because it can take up to three years to produce enough tissue to extract sufficient DNA for a genome sequence, sequencing a metagenome from thalli taken directly from nature would greatly reduce the time required for a genome project, were the final product of useful quality. The lichens used for metagenomes here included two cyanolichens symbiotic with cyanobacteria, and two lichens symbiotic with green algae, one of which has a very tight association with the algal symbiont, and one of which produces large fungal fruiting stalks (podetia) essentially devoid of the algal symbiont. The data show that where the symbiont is a cyanobacterium (prokaryote), or where the fungal portion of the thallus can be easily removed from the algal portion, the assembly suffers only a little compared to the assembly of a genome from a fungal culture. However, when the fungal and algal portions of the lichen cannot be easily dissected, as with Physcia cf. stellaris, the assembly is more problematic. It is therefore deemed not strictly necessary, but preferable, to sequence these fungi from culture where feasible.

Plant-like MEP α AMTPs are missing from the genomes of lichens growing in high-nitrogen habitats

Previously, the MEP α genes had been amplified from the Lecanorales but not from any other order sampled within the subclass Lecanoromycetidae (Figure 1). The genomes of lichens in two additional orders within the Lecanoromycetidae, the Caliciales (sensu[59]) and the Peltigerales, were sequenced here. One more order, the Teloschistales, is now represented by a genome of the lichenizing fungus Xanthoria parietina[68]. These orders are particularly interesting in terms of nitrogen tolerance and acquisition. Lichens in the Teloschistales and Caliciales tend to grow in nitrogen-rich environments, such as rocks or tree limbs on which birds perch. Because nitrogen is not limiting in these environments, it was hypothesized that the previous failure to amplify by degenerate PCR any MEP α genes in the lichens in these orders was evidence that the plant-like AMTPs genes had been lost. As predicted, no plant-like AMTP was found in Physcia cf. stellaris (Table 4). The same was true for Xanthoria parietina. The loss of plant-like AMTPs in the lichens in these orders might suggest why many of these lichens are constrained to living in high-nitrogen environments; alternatively, the expansion of lichens in the Teloschistales and Caliciales into high-nitrogen niches may have removed the selective pressure to retain the MEP α genes.

Plant-like MEP α AMTPs are missing from the genomes of lichens that are symbiotic with nitrogen-fixing cyanobacteria

Another order of lichens within the Lecanoromycetidae for which no plant-like AMTPs were amplified by degenerate PCR is the Peltigerales. This order is primarily composed of lichens in which the photobiont is a nitrogen-fixing cyanobacterium (Nostoc). Also in this order are lichens with both a green alga and a cyanobacterium as photobionts, but these lichens have secondarily regained the green algal photobiont [58]. Data from two unpublished genome projects on various species of Peltigera showed that each species of Peltigera had high- affinity fungal AMTPs and one low-affinity fungal AMTP, but no plant-like AMTPs (Ólafur S. Andrésson, personal communication and unpublished data; Bernard Goffinet, unpublished data). The fact that one lichen genus that is symbiotic with a nitrogen fixer seemed to have lost the plant-like AMTPs led to the hypothesis that cyanolichens in general have lost the MEP α AMTPs. To test this hypothesis, the lichen Leptogium austroamericanum in a separate suborder of the Peltigerales was chosen for genome sequencing. No MEP α AMTPs were found (Table 4). Thus, because lichens in both suborders of the Peltigerales lack the plant-like AMTPs, we suggest that all the lichens in the Peltigerales lack the plant-like AMTPs. Previous studies have reconstructed the ancestral state of the Peltigerales as a bi-membered association with a cyanobacterium [58]. It is possible that, having acquired a nitrogen fixer as a symbiont, the need for the nitrogen presumably provided by the horizontally transferred AMTPs was alleviated, allowing this gene to be lost during the early evolution of the Peltigerales. This suggests that the MEP α AMTPs will not be found in any of the green-algal associated members of the order either, as the symbiosis with a green alga is secondarily acquired. This may also suggest a reason why even most green algal members of the order continue to retain cyanobacteria in cephalodia—to supplement the nitrogen budget. Likewise, it is unusual for a fungus to have just two AMTP genes, particularly a filamentous ascomycete. Perhaps the presence of a nitrogen-fixing symbiont provides sufficient nitrogen so that there is no pressure to duplicate the AMTPs. In addition to the Peltigerales, cyanolichens are also found in the Lichinomycetes, a small class of lichenizing fungi evolutionarily distant from the bulk of lichens found in the Lecanoromycetes (Figure 1). Lichen-forming fungi in the Lichinales are mostly symbiotic with cyanobacteria other than Nostoc, such as Anacystis[70]. It was hypothesized that, like the lichens in the Peltigerales, lichens in the Lichinomycetes would also have lost the MEP α AMTPs upon acquisition of a cyanobacterium as a photobiont. No MEP α AMTPs were found in this genome. It seems likely that the other members of this class which are also symbiotic with cyanobacteria may also have lost the MEP α AMTP.

Plant-like AMTPs are found in the Ostropomycetidae but not the Acarosporomycetidae, and are distributed patchily in the Chaetothyriomycetidae

If loss of MEP α AMTPs is strictly associated with the photobiont of a lichen, then it would be expected that lichens with green algal photobionts should retain the MEP α gene(s). However, the failure to identify MEP α genes in other subclasses within the Lecanoromycetes by degenerate PCR argued against this hypothesis. To determine whether MEP α genes were present in other subclasses within the Lecanoromycetes, namely the Ostropomycetidae and the Acarosporomycetidae (Figure 1), the genomes of Dibaeis baeomyces (Ostropomycetidae), Graphis scripta (Ostropomycetidae) and Acarospora strigata (Acarosporomycetidae) were sequenced. MEP α genes were identified in Graphis scripta and Dibaeis baeomyces, both members of the Ostropomycetidae (Table 4, Figure 1). This finding represents an extension of the known distribution of MEP α genes. Further, Dibaeis and Graphis are in different orders within this subclass, suggesting that the MEP α gene may be widespread in the subclass (Figure 1). No plant-like AMTPs were found in the genome of Acarospora strigata, which is found in the earliest diverged subclass within the Lecanoromycetidae. Previously, one lichen in the subclass Chaetothyriomycetidae (Pyrenula cruenta) was shown to have a MEP α gene. Genome sequencing of Endocarpon pallidulum, a second lichen in the subclass Chaetothyriomycetidae but in a different order, revealed that no MEP α AMTP was present in the genome. This was surprising as Trypethelium virens and Laurera megasperma in the Dothidiomycetes, even more distantly related class, were shown to have the MEP α genes [10].

The MEP α gene replaces the high-affinity MEP γ gene in Graphis scripta

Low-affinity fungal AMTPs falling into the MEP γ clade were found in all genomes, as expected given the presence of these genes in the genomes of all Dikarya fungi sequenced to date (Table 4). Likewise, high-affinity fungal AMTPs in the MEP γ clade have been found in all fungi to date [10]. However, high-affinity fungal AMTPs were recovered in only seven of the eight lichen genomes sequenced here. Graphis scripta appeared to be lacking a high-affinity MEP γ gene, which is unprecedented among the Dikarya fungi. The similarity of MEP genes is high throughout the transmembrane domains, such that any MEP gene used as a query will find all other MEP genes in the genome. Furthermore, the Graphis scripta genome was produced from the cultured fungus, and the genome is very high coverage (173×). Of the eight new genomes presented here, this is the genome that received the highest CEGMA score for complete genes, a score nearly identical to the CEGMA score obtained by the Neurospora crassa genome (Table 1). Therefore, it is unlikely that no portion of the high-affinity MEP γ gene would have been sequenced if it were in the genome. Thus, it seems likely that this fungus actually lacks the high-affinity MEP γ gene.

Other lichens like Arthonia rubrocincta or Leptogium austroamericanum that have only two AMTPs have the fungal high-affinity MEP γ gene and the fungal low-affinity MEP γ gene (Table 4). Dibaeis baeomyces, which like Graphis scripta is in the Ostropomycetidae, has two copies of a MEP α gene as well as one fungal high-affinity AMTP and one fungal low-affinity AMTP, suggesting that perhaps the ancestor to the Ostropomycetidae or even the Lecanoromycetes had all four AMTPs. It would appear that in Graphis scripta the MEP α gene has replaced the high-affinity fungal AMTP from the MEP γ clade. Graphis scripta has a very thin crustose thallus, quite unlike the large thalli produced by many lichens in the Lecanorales. It is perhaps no surprise that less nitrogen is needed to support this comparatively reduced thallus. That the MEP α gene is retained in preference to the high-affinity fungal AMTP suggests that it was perhaps more efficient than the high-affinity fungal AMTP.

The MEP genes are functional AMTPs

The fungal high-affinity AMTP (mep2), fungal low-affinity AMTP (mep3), and two plant-like AMTPs (mep1a and mep1b) were introduced individually into a strain of Saccharomyces cerevisiae in which all three native AMTPs had been deleted. All four C. grayi genes complemented the mutant, although growth was not robust. Therefore, each of the genes was deemed to be a functional AMTP (Figure 2). Of the four genes, the low-affinity AMTP (mep3) appeared generated the most robust growth under the conditions tested, albeit marginally so. This observation is consistent with the predicted function of this gene as a low-affinity (high-capacity) AMTP. Importantly, the mep1a and mep1b are the first genes from the fungal subclade of the MEP α clade for which a function has been demonstrated.

Media and culture conditions

For yeast strains, standard yeast media were used including: Synthetic Limiting Ammonium and Dextrose (SLAD) composed of 0.17% yeast nitrogen base (without amino acids and ammonium sulfate), 2% glucose and 2% agar [72, 75] with modifications including 2 mM, 5 mM, 7 mM, 10 mM and 20 mM ammonium sulfate; and Synthetic Complete medium (SC) lacking uracil, which contains 6.7 g/l yeast nitrogen base without amino acids, supplemented with all amino acids except uracil, 2% glucose and 2% agar. Yeast strains were grown at 30°C or at room temperature. Cladonia grayi was maintained at room temperature in liquid shaking cultures of MEYE medium consisting of 20 g/l malt extract and 2 g/l yeast extract [1].

Construction of plasmids

RNA was extracted from Cladonia grayi cultures growing in liquid medium. Tissue from liquid cultures was harvested, rinsed in distilled water, lyophilized, then ground to a fine powder under liquid nitrogen in a pre-chilled mortar and pestle. The ground tissue was resuspended in TRIzol Reagent (Invitrogen, Carlsbad, CA) and RNA extraction proceeded according to the manufacturer’s instructions.

The first strand of the cDNA was generated using the following reverse transcription reaction mix: 1.0 μl 10× PCR buffer, 2.0 μl 25 mM MgCl2 (both supplied with AmpliTaq DNA Polymerase, Applied Biosystems), 2.0 μl dNTPs (10 mM each), 0.5 μl RNAse inhibitor (20 μg/μl, Applied Biosystems) 0.5 μl MulV reverse transcriptase (Applied Biosystems), 1.75 μl of distilled water, and 1.25 μl reverse primer. The thermocycling conditions for this first strand synthesis reaction were: 42°C for 35 minutes, then 99°C for 5 minutes. Primers are listed in Additional files 1 and 2. PCR to regenerate the second strand of the cDNA was performed using the entire volume of first strand reaction plus 31.25 μl distilled water, 4 μl 20x PCR buffer, 3 μl 25 mM MgCl2, 1.25 μl of a forward primer, 0.25 μl of AmpliTaq polymerase (Applied Biosystems) and 0.25 μl of an antibody to Taq polymerase (Clontech). The thermocycling program consisted of 24 cycles of 94°C for 30 seconds, 55°C for 30 seconds with a 0.4 degree decrease in temperature for each cycle, and 72° for 1 minute, followed by 12 cycles of 94°C for 30 seconds, 45°C for 30 seconds, and 72°C for 2 minutes with a 3 second increase per cycle, followed by a final extension at 72°C for 10 minutes. cDNA of each AMTP from Cladonia grayi was cloned into the TOPO TA vector as described above. Plasmid DNA was restricted to release the full-length AMTP cDNA which was then subcloned into the pRS306-GAL1-TADH vector (unpublished, a gift from Mark Chee) carrying a uracil selectable marker, an inducible Gal promoter and a Cyc terminator or the p416-GPD vector [76] carrying a uracil selectable marker and a constitutive GPD promoter and Cyc terminator, for transformation into Saccharomyces cerevisiae. The DNA sequence of the constructed plasmids was confirmed by PCR and sequencing as described below.

Construction of yeast strains

Yeast strains are listed in Additional file 3. Yeast transformations were performed using the TRAFO lithium acetate method [77] or by the gapped plasmid method [78]. Following the TRAFO method, a yeast culture in log phase was centrifuged for 5 minutes at 6000 RPM, and the pellet washed with distilled water, and resuspended in an equal volume of distilled water, pelleted again, then resuspended in 1 ml of 100 mM lithium acetate. This solution was pelleted at maximum speed (>14,000 RPM) in a microcentrifuge and the pellet resuspended in approximately 400 μl of 100 mM lithium acetate. A transformation mix of 50 μl yeast cells, 240 μl PEG, 36 μl 1 M lithium acetate, 50 μl salmon sperm DNA and 34 μl of diluted plasmid (approximately 100 ng plasmid) was incubated at 30°C for 30 minutes and heat shocked at 42°C for 30 minutes. The mix was then centrifuged and the pellet resuspended in 300 μl water. 100 μl of the cell suspension was plated onto selective medium and transformants confirmed by colony PCR as described below. For gapped plasmid construction, cDNA of AMTPs from C. grayi which had been cloned into the TopoTA vector (Invitrogen, Carlsbad, CA) was amplified with primers composed of 40 base pairs of homology on either side of the multiple cloning site of the centromeric plasmid p416GPD [76] carry the uracil selectable marker and joined to the first or last 20 base pairs of the gene [78]. The PCR product was cleaned in a Microcon column (Millipore, Bilerica, MA) and approximately 1 μg of amplicon was transformed into yeast along with 0.1 μg of restricted plasmid DNA. Yeast recombine the two fragments into a functional plasmid. Transformants were screened by PCR and the inserts sequenced as described below.

DNA amplification and Sanger Sequencing

For PCR performed directly on colonies of yeast or E. coli, the PCR reaction mix consisted of 17.3 μl distilled water, 2.5 μl of 10x PCR buffer, 2.5 μl dNTPs, 1.25 μl each of a forward primer and a reverse primer, and 0.3 μl Taq polymerase (Denville). Colonies were touched lightly with the tip of a pipette which was then dipped briefly into the PCR tube and removed. PCR was performed on a MJ Research PTC200 thermocycler or an Applied Biosystems (Foster City, CA) Veriti thermocycler using a cycling program consisting of 10 minutes at 94°C to lyse the cells and release the DNA, followed by 25 cycles of 30 seconds at 94°C, 30 seconds at 55°C and 2 minutes at 72°C, followed a final elongation step at 72°C for 7 minutes. PCR products were visualized on a TAE 1% agarose gel stained with SYBR Safe (Invitrogen, Carlsbad, CA). If necessary, faint products or products with multiple bands were cloned with the TOPO TA cloning kit (Invitrogen, Carlsbad, CA) following the manufacturer’s instructions. For each cloning reaction, at least 8 clones were screened by colony PCR using T7 and M13R primers and a PCR program consisting of a 10-minute initial denaturation step, followed by 25 cycles of 30 seconds at 94°C, 30 seconds at 52°C and 60 seconds at 72°C followed by a final elongation step of 7 minutes. PCR products were cleaned with a Montage PCR filter column (Millipore, Bilerica, MA) or with an Exo-SAP clean-up using 1 μl SAP dilution buffer, 0.5 μl Exonuclease 1, 0.5 μl Shrimp Alkaline Phosphatase, added to 10 μl PCR reaction and incubating on one of the aforementioned thermocyclers for 30 minutes at 37°C, then 15 minutes at 80°C. Cleaned PCR products were sequenced in 10 μl reactions using: 1 μl primer, 3 μl purified PCR product, 0.5 μl Big Dye (Big Dye Terminator Cycle sequencing kit, ABI PRISM version 3.1; PE Applied Biosystems, Foster City, CA), 1.5 μl Big Dye buffer, and 4 μl double-distilled water. Automated reaction clean-up and visualization was performed at the Duke IGSP Genome Sequencing & Analysis Core Facility using Big Dye chemistry with an ABI 3730xl automated sequencer (PE Applied Biosystems, Foster City, CA). Sequencher version 4.8 (Gene Codes Corporation, Ann Arbor, MI) was used to edit sequences and assemble contigs.

Genomic DNA isolation

DNA from cultures of the lichenizing fungi Acarospora strigata (Acarosporomycetidae, Lecanoromycetes), Arthonia rubrocincta (Arthoniomycetidae), Endocarpon pallidulum (Eurotiomycetes) Graphis scripta (Ostropomycetidae, Lecanoromycetes) and from lichen thalli of Dibaeis baeomyces (Ostropomycetidae, Lecanoromycetes), Leptogium austroamericanum, Physcia cf. stellaris (both Lecanoromycetidae, Lecanoromycetes), and Peltula cylindrica (Lichinomycetes) was prepared following the DTAB/CTAB method outlined in [79]. Briefly, tissue was ground to a fine powder under liquid nitrogen using a pre-chilled mortar and pestle. Twenty volumes of a DTAB solution (with 1% DNAse-free RNAse) was added to the powder and incubated for ~3 minutes at 65°C. Polysaccharides and other contaminants were precipitated by adding 1/3 volume 5M NaCl and centrifuging at 6000 RPM for 5 minutes. The supernatant was removed and extracted once with one volume of phenol, after which the supernatant was removed and then extracted with one volume of chloroform. The supernatant was removed, and DNA precipitated with one volume of isopropanol and left to incubate at room temperature for 5 minutes, then centrifuged at 6000 RPM for 5 minutes to pellet the DNA. The pellet was resuspended in DTAB + RNAse, with heating and gentle agitation. Remaining polysaccharides were precipitated by adding 1/3 volume of 5M NaCl and centrifuging at 6000 RPM for 5 minutes. The supernatant was removed and undissolved material was pelleted by centrifugation at 6000 RPM. DNA was precipitated by adding 2 volumes of 100% ethanol, incubating at room temperature for five minutes, and centrifuging at maximum speed (14,000 RPM) for five minutes. The pellet was washed once with 70% ethanol, and allowed to dry. When dry, the pellet was resuspended in 100 μl TE and quantified by Qubit. DNA was then further purified using the PowerClean kit (Mo Bio) following the manufacturer’s instructions. DNA was again quantified after purification by Qubit.

Illumina sequencing and genome assembly

One to six μg of purified DNA were submitted to the sequencing facility at Duke University for 75-base pair paired-end barcoded Illumina sequencing using the HiSeq technology. Four lichen cultures were multiplexed onto a lane, and DNA from two lichen thalli (yielding both fungal and algal DNA) were multiplexed onto a lane. A total of three lanes were used to sequence the eight fungal genomes (~80 million to 600 million reads). Preliminary draft assemblies were generated with Velvet [80] (Table 1). Data is housed in the Sequence Read Archive (SRA) accessible through the National Center for Biotechnology Information (NCBI) website.

Phylogenetic analysis

AMTP sequences generated for this study (Additional file 4) were included in a modified alignment from [10] and phylogenetic analysis was performed as in [10]. Genes in the MEP clade of AMTPs were retained in the analysis (Additional file 5), while all others were discarded. All excluded regions were reexamined, and additional sites that became alignable were included in the analysis, including 90 base pairs of transmembrane region one, which in the previous analysis had been entirely excluded. The first well-supported clade outside of the MEP group in the MEP grade was used as an outgroup. In one analysis, the three AMTPs from the green alga Asterochloris sp., which fall into the AMT clade of AMTPs, were also included and used as an outgroup. Manual alignments were performed using MacClade 4.08 [81]. Ambiguously aligned regions and introns were delimited manually and excluded from phylogenetic analyses. Models of molecular evolution which had been selected using the Akaike Information Criterion (AIC) implemented in jModeltest [82] or MrModeltest 2.3 [83] were implemented on this dataset. Phylogenetic relationships and confidence values were inferred using a maximum likelihood approach at the nucleotide level. Maximum likelihood analysis at the nucleotide level used GTR GAMMAI (with a gamma parameter and a proportion of invariable sites, = GTR + Γ + I). The program RAxML-VI-HPC [84] was used for the maximum likelihood search for the most likely tree. The same program using the same settings was used for the bootstrap (BS) analysis with 1000 BS replicates run in batches of 100 replicates and pooled. Bootstrap values were calculated and visualized using the majority rule consensus tree command in PAUP 4.0d701 [85] (Additional file 6).