Gene expression indicates a zone of heterocyst differentiation within the thallus of the cyanolichen Pseudocyphellaria crocata


Author for correspondence:

Tina C. Summerfield

Tel: + 64 3 4797578



  • Development of the symbiotic association in the bipartite lichen Pseudocyphellaria crocata was investigated by characterizing two regions of the thallus.
  • Thallus organization was examined using microscopy. A HIP1-based differential display technique was modified for use on Nostoc strains, including lichenized strains. Northern hybridization and quantitative real-time polymerase chain reaction were used to confirm differential display results, and determine expression levels of key cyanobacterial genes. Photosystem II yield across the thallus was measured using pulse-amplitude modulated fluorescence.
  • Microscopy revealed structural differences in the thallus margins compared with the centre and identified putative heterocysts in both regions. Differential display identified altered transcript levels in both Nostoc punctiforme and a lichenized Nostoc strain. Transcript abundance of cox2, atpA, and ribA was increased in the thallus margin compared with the centre. Expression of cox2 is heterocyst specific and expression of other heterocyst-specific genes (hetR and nifK) was elevated in the margin, whereas, expression of psbB and PSII yield were not.
  • Structural organization of the thallus margin differed from the centre. Both regions contained putative heterocysts but gene expression data indicated increased heterocyst differentiation in the margins where photosystem II yield was decreased. This is consistent with a zone of heterocyst differentiation within the thallus margin.


Cyanobacteria are oxygenic phototrophs found in almost all environments on Earth, including in symbiotic association with a phylogenetically diverse range of organisms (Whitton & Potts, 2002). In many of these symbioses, the cyanobacterium belongs to the genus Nostoc and these diazotrophic filamentous cyanobacteria supply fixed nitrogen to the host. Nostoc strains are found in association with fungi and a wide range of plants, including bryophytes, ferns, gymnosperms and at least one angiosperm (Meeks, 1998). Cellular differentiation of Nostoc, including the formation of nitrogen-fixing heterocysts, is central to the formation and maintenance of these associations (Rai et al., 2000). Free-living strains do not form heterocysts in the presence of combined nitrogen but the absence of combined nitrogen initiates a cascade of gene expression changes that control heterocyst formation and spacing (Flores & Herrero, 2010). These highly regulated processes are modified in symbiotic associations where heterocyst frequency is increased and heterocyst differentiation occurs in the presence of ammonium (Meeks & Elhai, 2002).

In plant–Nostoc symbioses, the cyanobacterium is housed within specialized structures and, as the symbiotic tissue matures, the cyanobacterium shows increasing morphological and physiological modifications (Bergman & Rai, 1989; Johansson & Bergman, 1992; Rai et al., 2000). This results in a gradient of structural–functional changes in the symbiotic Nostoc occurring from the young to mature symbiotic tissue, which typically includes increased cell size, heterocyst frequency and nitrogen fixation, and decreased growth rate and cell division (Rai et al., 2000). Nostoc strains can associate with fungi to form bipartite lichens, in this case the cyanobacterium provides both fixed carbon and nitrogen and is located in a layer between fungal tissue (Stewart et al., 1983). Bipartite cyanolichens differ from plant–Nostoc associations: the cyanobacterium may comprise up to 36% of the lichen on a protein basis, whereas in the plant symbioses the cyanobacterium typically contributes 1–4% of the total biomass (Rai et al., 2000). The extent of heterocyst modification of Nostoc strains in bipartite cyanolichens compared with free-living Nostoc is not clear, with studies indicating increased, decreased and unchanged heterocyst frequency in lichenized cyanobacteria (Rai et al., 2000). However, like other cyanobacterial symbioses, the formation and maintenance of the cyanolichen thallus requires the exchange of nutrients and coordinated growth of the partners, therefore, a similar developmental gradient may be present in the cyanolichen.

Evidence for morphological and physiological changes to the cyanobiont across the bipartite lichen thallus comes from measurements of Peltigera canina. Across a 2 cm region from the margin towards the centre heterocyst frequency and nitrogenase activity were altered and both peaked between 1.2 cm and 1.5 cm and declined towards the thallus centre (Hill, 1989). In addition, heterocyst abundance and nitrogenase activity were lowest at the growing apex (< 1 mm) (Bergman & Hallbom, 1981). Data from P. canina, and data from the green algal bipartite lichens Parmelia sulcata and Xanthoria parietina (Hill, 1989) suggest internal thalline differentiation in bipartite thalli. Based on these morphological and physiological data the bipartite thallus is proposed to contain zones of division, elongation and maturation that are characterized by modifications to both the mycobiont and photobiont (Honegger, 1993; Rai et al., 2000). To investigate whether these developmental zones are present within the thallus of P. crocata this study used microscopy to examine the structure of two regions of the lichen thallus and used examination of cyanobacterial gene expression in these distinct regions. Gene expression was used as heterocyst differentiation, nitrogen fixation and many metabolic processes in cyanobacteria are mediated at the transcript level.

Investigation of cyanobacterial gene expression in two regions of a bipartite cyanolichen thallus used a modified differential display technique, originally used to screen for variations in gene expression in Synechocystis sp. strain PCC 6803 (Bhaya et al., 2000). This method used primers based on a palindromic sequence that is over-represented in the genome of many cyanobacteria, occurring in both protein-coding and intergenic regions (Gupta et al., 1993), but usually absent from rRNA and tRNA genes (Robinson et al., 1995). Importantly, for our application, over-representation of this highly iterated palindromic sequence (HIP1) appears to be restricted to cyanobacterial genomes and therefore HIP1-based primers would be expected to amplify cyanobacterial and not fungal products from total RNA extracted from a lichen thallus. In Nostoc punctiforme the HIP1 element occurs approximately every 1200 bp with a nearly random distribution throughout the genome (Meeks et al., 2001), indicating that HIP1-based primers may be used to generate differential display products from N. punctiforme total RNA and also may be used to identify genes that are differentially expressed in lichenized Nostoc strains. We verified the feasibility of this approach by using HIP1-based primers to generate differential display products from N. punctiforme grown in the presence and absence of combined nitrogen.

We have demonstrated that the HIP1-based differential display technique can be used to identify differentially expressed genes in lichenized Nostoc in the bipartite cyanolichen Pseudocyphellaria crocata. This high-light-tolerant, cosmopolitan lichen is found on all major landmasses as well as oceanic islands of both the Northern and Southern Hemispheres (Galloway, 1988). Pseudocyphellaria crocata shows a level of cyanobiont specificity that is typical of a bipartite cyanolichen (Paulsrud & Lindblad, 1998; Summerfield et al., 2002; Summerfield & Eaton-Rye, 2006) and similar to plant–Nostoc symbioses, a limited range of Nostoc strains are capable of forming lichen symbioses. Investigation of P. crocata thalli identified differential display products in the two distinct regions of the thallus. This approach combined with examination of the expression levels of key cyanobacterial genes, pulse-amplitude modulated (PAM) fluorescence measurements and microscopy, supported morphological and physiological data for a developmental zone that is associated with heterocyst differentiation within the lichen thallus.

Materials and Methods


Small fragments of freshly collected lichen thallus, either centre or margin (c. 4 mm2) were embedded in O.C.T.TM Tissue-Teks (Sakura Finetek, Zoeterwoude, the Netherlands) and cut into 30 μm-thick sections with a cryotome, this was the thinnest section that maintained the morphology of the thallus. Slide-mounted sections were rehydrated for 5 min in 50% glycerol immediately before imaging. Specimens were observed by confocal laser scanning microscopy using an Olympus Fluoview 1000 confocal microscope (Olympus Corporation, Toyko, Japan). Fungal and cyanobacterial tissue exhibited strong autofluorescence using an excitation wavelength of 473 nm. Autofluorescence of the cyanobacterium was measured using an excitation wavelength of 559 nm to excite the phycobilins, resulting in intense combined fluorescence from phycobilins and chlorophyll a of the photosynthetic cyanobacterial cells (Meeks & Elhai, 2002). Data were collected, overlaid, cropped and analysed using the fv10-asw version 3.0 software from Olympus. Measurements of fungal and cyanobacterial regions and cell size were determined from overlaid images taken using a ×40 or ×100 oil immersion lens from two centre samples and two margin samples from each of three lichen specimens. Maximum cell diameter was calculated from stack images and processed using imagej (

HIP1 primer design

The HIP1 element is an octamer in N. punctiforme and a decamer in Synechocystis sp. strain PCC 6803 (GGCGATCGCC), therefore, shorter primers were used in this study than those used in Synechocystis sp. strain PCC 6803 (Bhaya et al., 2000). Primers were modified in a similar way to those of Bhaya et al. (2000), primers used for differential display are listed in Table 1 and primers used to amplify from gDNA are listed in the Supporting Information, Table S2.

Table 1. Primers used in the studya
Primer nameSequenceUse
  1. a

    The nucleotide differing from the HIP1 element is shown in italics and the nucleotides additional to the HIP1 element are underlined. The letter W denotes A or T.

  2. b

    DDRT-PCR denotes differential display reverse-transcription PCR; qRT, quantitative reverse transcription.


Culture conditions and lichen sample collection

The N. punctiforme cultures were grown in the presence of 3.0 mM sodium nitrate or 2.5 mM ammonium chloride added to Allen and Arnon medium diluted fourfold (AA/4) plus 5 mM MOPS pH 7.8 (as described in Enderlin & Meeks, 1983). Nostoc punctiforme cells were pelleted by centrifugation and divided into separate culture flasks; half of these underwent nitrogen step-down while the remaining cultures were maintained in the presence of combined nitrogen. At 45 h following nitrogen step-down cells were harvested for RNA extraction. Pseudocyphellaria crocata thalli were collected from Halfway Bush, Dunedin, New Zealand, a representative lichen is shown in the Supporting Information, Fig. S1. Lichen specimens were in a hydrated state at the time of collection and thallus regions used had few or no soralia. Any debris or plant material was removed and lichens were frozen in liquid nitrogen. The Nostoc tRNALeu(UAA) intron sequence obtained indicated that the same cyanobiont was present in all three lichen thalli (data not shown, GenBank accession numbers JX041517–9) and had the same tRNALeu(UAA) intron sequence as the pcB cyanobiont described in Summerfield et al. (2002).

Nucleic acid extraction and northern blot analysis

Extraction of RNA from N. punctiforme was undertaken following the protocol of Summers et al. (1995). For the extraction of RNA from lichen thalli, c. 500 mg of either ‘margin’ (within 0.5 cm of thallus edge), or ‘centre’ (at least 1.0 cm from the margin) were ground to a fine powder in liquid nitrogen. Total RNA was purified from these samples using the Plant/Fungi Total RNA Purification Kit (Norgen Biotek Corporation, Ontario, Canada). Northern hybridizations and preparations of radiolabeled DNA probes were carried out as previously described (Summerfield et al., 2005).

Degenerate primer design and PCR

Polymerase chain reaction primers were designed based on comparison of the gene sequences of the most closely related Nostocales strains with sequenced genomes: N. punctiforme, Anabaena variabilis and Anabaena sp. PCC 7120. DNA sequences of target genes from the three strains were aligned using clustalx2 (Larkin et al., 2007), conserved sequences were identified and degenerate PCR primers were designed to these regions (Table S2). The sequences of PCR products were deposited in the GenBank database as follows: hetR, JX080415; nifK, JX080416; ntcA, JX080417; pepC, JX080418; rpoC1, JX080419; trpA, JX080420; psbB, JX080421; rnpB, JX080422.

PCR and cloning

Purification of DNA fragments from agarose gels was performed using PureLink Quick Gel Extraction Kit (Invitrogen) and ligated into pGEM-T easy vector (Promega). Sequencing using M13 forward and reverse primers was performed using the BigDye Terminator v. 3.1 Ready Reaction Cycle Sequencing Kit (Applied Biosystems, Foster City, CA, USA) and a 3730xl DNA Analyser (Applied Biosystems).

Differential display and quantitative reverse transcriptase-PCR (qRT-PCR)

The primers used for differential display were variations of the HIP1 element, as shown in Table 1. For differential display, cDNA synthesis was performed as described in Bhaya et al. (2000). For qRT-PCR, first-strand cDNA was synthesized using Superscript III (Invitrogen, The Netherlands). Each qRT-PCR analysis was performed in triplicate in a total volume of 10 μl reaction mix containing 2.5 μl cDNA template (1 in 10 dilution), 5 μl 2 × SensiFAST SYBR master mix (Bioline, London, UK) and 200 nM of each primer. The PCR amplifications were performed using a MiniOpticon Real-Time PCR System (Bio-Rad) using standard conditions. Melting curve analyses were performed to confirm the absence of primer dimer or non-specific amplification products. Standard curves were performed for each primer pair to ensure satisfactory efficiency of amplification (90% ≤ E ≤ 110%) (data not shown). Transcript levels were calculated by Bio-Rad CFX Manager software, which were determined by comparison of cycle threshold (CT) values between target and reference genes.

To select appropriate reference genes for qRT-PCR analyses several genes (rnpB, dnaJ, pepC, rpoC1, and trpA) were tested for their expression stability in the different lichen samples using the program genorm (Vandesompele et al., 2002). This included several genes used as housekeeping genes in cyanobacterial gene expression studies (Woodger et al., 2003; Paz-Yepes et al., 2009; Soule et al., 2009; Ginn et al., 2010). The program identified the optimal reference genes to be the geometric mean expression of rnpB, pepC, and dnaJ.

Fluorescence measurements

The Junior PAM (Walz, Effeltrich, Germany) was operated according to the manufacturer's instructions. Intensities of the measuring light, actinic light, and saturating pulses were 0.17, 50 and 1460 μmol photons m−2 s−1, respectively, and the duration of the saturation pulse was 0.4 s. The lichen samples were saturated in water and rehydrated for 40 min before measurement, with frequent spraying with distilled water to ensure that the thallus remained fully hydrated. Measurements were taken from the marginal lobes and the centre of the thallus.


Thallus organization

Confocal microscopy was used to examine organization of the lichen thallus in the centre and margin regions. Autofluorescence from fungal tissue enabled visualization of the thallus structure and autofluorescence resulting from excitation of the phycobilins enabled detection of the cyanobacterial layer. In both the margin and centre we observed internal stratification typical of a foliose lichen, comprising the upper cortex, cyanobacterial layer, medulla and lower cortex (Fig. 1a–d). Measurement of lichen sections from three P. crocata thalli indicated that the upper cortex, cyanobacterial layer and medulla were larger in the centre than in the margin; however, the lower cortex was a similar size in the margin and centre samples (Table 2). Morphological differences between the two regions included decreased number of aerial hyphae in the medulla of the margin compared than in the centre (Fig. 1a,b), this delayed development of the aerial hyphae has been reported in other foliose lichens (Honegger, 1993). In the margin, the lower cortex resembled the conglutinate layer of the upper cortex, whereas, in the centre the lower cortex comprised several layers of less densely packed cells, including cells differentiated to form tomentum (hair-like structures) on the lower surface (Fig. 1a,b).

Figure 1.

Cross-sections of the thallus of Pseudocyphellaria crocata. (a–d) Confocal microscopy showing autofluorescence of OCT-embedded sections, fungal tissue (green) and cyanobacterial cells (orange). Images obtained by processing 15 confocal sections (at 0.50-μm intervals). (a,c) Apex of lichen margin and (b,d) centre of thallus. (e,f) Brightfield images of cyanobacterial cells in the margin (e) and centre (f). (g,h) Epifluorescent images of cyanobacterial cells in the margin (g) and centre (h); excitation was with green light, exciting the phycobilins and resulting in combined fluorescence from phycobilins and chlorophyll a from the photosynthetic cells. Putative heterocysts are marked by arrows, these cells exhibit decreased pigmentation (e,f) and no autofluorescence (g,h). Bars, 20 μm.

Table 2. Pseudocyphellaria crocata organization in the margin and centre of thallus
  1. a

    Margin, < 0.5 cm from the growing apex of the thallus.

  2. b

    Centre, 1 cm or more from the thallus margin.

  3. c

    = 6; 10 or more measurements were made for each sample.

  4. Values represent means (± SE).

Region size (μm)
Upper cortex36.3 ± 1.650.3 ± 1.2
Cyanobacteria39.7 ± 3.271.1 ± 3.8
Medulla51.0 ± 2.7113.8 ± 5.8
Lower cortex16.5 ± 0.615.6 ± 0.7
Cyanobacterial cell diameter6.7 ± 0.17.0 ± 0.1
Room-temperature fluorescence yieldc
Fv/Fm0.49 ± 0.010.59 ± 0.01
ΔF/inline image0.47 ± 0.020.56 ± 0.01

The shape and organization of the cyanobacterial cells was similar in both regions of the thallus with cells arranged in discrete bundles that were surrounded by an autofluorescent layer (Fig. 1c,d). Brightfield images indicated that cyanobacterial cells were surrounded by fungal material (Fig. 1e,f); this resembles the gelatinous fungal protrusions previously reported in other cyanolichens (Honegger, 1993). In both the margin and centre of the thallus putative heterocysts were identified, these cells exhibited decreased pigmentation in brightfield and reduced autofluorescence from chlorophyll a (Fig. 1e–h). The organization of the Nostoc cells in compact bundles meant that we could not be certain we identified all putative heterocysts. In addition, the presence of cells that appeared pigmented but not autofluorescent and cells that appeared to have reduced pigmentation but were autofluorescent (data not shown), combined with previous reports of altered pigment content of heterocysts in symbiotic association (Meeks & Elhai, 2002) meant we were unable to reliably identify heterocysts using microscopy. The mean diameter of the autofluorescent cyanobacterial cells was smaller in the margin compared with the centre, although in both regions we observed considerable variation in cell size (Table 2). Morphological differences between the centre and margin regions are consistent with previous reports of distinct developmental zones in the thallus. To investigate molecular changes underlying these morphological changes we modified a differential display technique.

HIP1-based primers identify differentially expressed genes in N. punctiforme

Amplification of N. punctiforme genomic DNA using HIP1-based primers resulted in multiple PCR products that were primer specific and changing a single nucleotide at the 3′ end of a primer was sufficient to alter this specificity (Fig. 2a; Table S2). Primers bound reproducibly giving the same banding pattern in multiple reactions using N. punctiforme genomic DNA as a template (Fig. S2a). These primers were used to perform differential display on N. punctiforme using RNA extracted from cultures grown in the presence of combined nitrogen and following a 45-h incubation in medium without combined nitrogen (nitrogen step-down). Nitrogen step-down results in cellular differentiation with c. 8% of cells forming nitrogen-fixing heterocysts (Meeks & Elhai, 2002). This response involves a cascade of gene expression changes resulting in nitrogen fixation within 24 h and after 7 d under nitrogen-fixing conditions cells have c. 500 differentially expressed genes compared to non-nitrogen-fixing cultures (Wolk et al., 1994; Campbell et al., 2007). The nitrogen step-down experiment was used as this represents a simple, reproducible experiment for the validation of the differential display technique.

Figure 2.

(a) PCR amplification from Nostoc punctiforme genomic DNA using HIP1-based primers: HIPWAC (lane 1); HIPWAG (lane 2); HIPWCA (lane 3); HIPWCG (lane 4); HIPWGC (lane 5); HIPWGT (lane 6); HIPWTA (lane 7); HIPWTG (lane 8); HIPWAT (lane 9), and HIPWACC (lane 10). (b) Differential display using the HIPWCG primer on RNA extracted from N. punctiforme cells 45 h plus or minus fixed nitrogen. Closed and open arrows indicate differentially and nondifferentially displayed amplicons, respectively (L contained 100 bp DNA Ladder (Invitrogen)). (c) Part of the predicted cox3 operon and northern hybridization. The black line represents the differentially displayed fragment and the probe for northern hybridization. RNA from cells grown for 45 h in the presence (lanes +) or absence (lanes −) of sodium nitrate. The largest detectable transcript is marked by an arrow (top); loading control (bottom). (d) Northern hybridization of the putative cox3 operon using the probe shown in (c). RNA was extracted from cells following nitrogen step-down for the period of time indicated.

Total RNA from the nitrogen step-down experiment and the primer HIPWCG were used to synthesize cDNA (Table 1). Subsequent PCR amplification of this cDNA with HIPWCG resulted in at least four PCR fragments for both nitrogen step-down and combined nitrogen samples. A DNA fragment of c. 550 bp was observed only in nitrogen-limited samples (Fig. 2b). The sequence of this fragment corresponded to part of the genes that encode subunits of a putative terminal cytochrome c oxidase. The fragment comprised 372 bp of coxB3 (Npun_R3537) and 157 bp of coxA3 (Npun_R3536) and 14 nucleotides between the two genes. In coxB3 the priming site contained the HIP1 element with a GC extension, differing at the underlined nucleotide from the HIPWCG primer (5′GCGATCGCCG); however, the second primer-binding site did not contain the HIP1 element and differed from the primer target at the two underlined locations (GATACCGCCG). Northern hybridization using the 543 bp differential display product as the probe, identified a transcript of c. 5.7 kb (with additional hybridization signals corresponding to transcript sizes c. 5 kb and c. 3.8 kb) shown in Fig. 2(c,d). In agreement with the differential display data, this transcript exhibited much higher expression in cells following nitrogen step-down than in cells grown in the presence of combined nitrogen (Fig. 2c). This transcript is of similar size to the cox3 transcript of Anabaena sp. PCC 7120 that contains the coxBAC genes and two genes encoding protein with unknown function (alr2729 and alr2730) immediately upstream of coxB3 (Valladares et al., 2003). The sequences of genes Npun_R3538 and Npun_R3539, upstream of coxB3 in Npunctiforme, have 79% and 75% sequence identity to the Anabaena sp. PCC 7120 genes alr2730 and alr2729, respectively, suggesting that the five-gene operon is conserved in Npunctiforme and Anabaena sp. PCC 7120. A time-course detected expression of the cox3 transcript at 18 h following nitrogen step-down, and expression peaked at 48 h (Fig. 2d), coinciding with a second round of heterocyst differentiation (Jones & Haselkorn, 2002), and by 72 h transcript levels had decreased to levels similar to those seen in 18 h.

To further validate the HIP1 protocol, additional amplification products that were not differentially displayed were cloned and sequenced, and these genes were compared with the microarray results comparing transcript levels in steady-state N. punctiforme cultures grown in the presence or absence of combined nitrogen (Campbell et al., 2007). The differential display and microarray data showed good agreement for all genes examined (Table S1; Fig. 2b).

HIP1-based primers and differential display in the cyanobiont of bipartite cyanolichens

The HIP1-based differential display identified differentially expressed genes in N. punctiforme; this technique was then tested on a lichenized Nostoc strain in the bipartite cyanolichen P. crocata. A selection of HIP1-based primers were used to successfully amplify genomic DNA from the isolated cyanobiont of P. crocata (Fig. S2c). Next we investigated whether HIP1-based primers could amplify Nostoc differential display products from total RNA isolated from a lichen thallus. This would enable us to investigate differential gene expression in different regions of the lichen thallus and determine whether distinct developmental regions within the thallus could be identified by differential gene expression. Total RNA was isolated from two morphologically distinct regions of three lichen specimens: the centre region, > 1.0 cm from the growing edge and the margin, 0.5 cm of the thallus that includes the growing margin. These samples were subjected to reverse-transcription polymerase chain reaction (RT-PCR), using the same HIP1-based primer for cDNA synthesis and PCR. Two fragments resulting from amplification with the HIPWAC primer were selected for further examination: a fragment of c. 2 kb that was more abundant in the margin samples than in centre samples (Fig. 3a, DD1) and a fragment of > 1 kb that was slightly more abundant in the margin samples than in the centre samples (Fig. 3a, DD2). In addition, two differential display fragments obtained using the primer HIPWCA were selected for further characterization, these were < 1 kb and c. 400 bp, respectively, and both were more abundant in the margin samples (Fig. 3b, DD3 and DD4). Differential display using the primer HIPWCG, which was used to identify differential expression of the cox3 operon in N. punctiforme, did not exhibit differential expression between the centre samples and the margin samples (data not shown).

Figure 3.

Differential display of HIP1-primed RNA extracted from three Pseudocyphellaria crocata thalli. (a,b) Reverse transcription (RT)-PCR on from the centre (C) and margin (M) samples of P. crocata specimens 1, 2 and 3. The primers used for RT and PCR amplification were: (a) HIPWAC and (b) HIPWCA (L contained 1 kb plus DNA Ladder (Invitrogen)) and arrows indicate differentially displayed products. (c) Schematic representation of differentially expressed genes with sequence similarity to those found in the genome of N. punctiforme. Filled arrowheads indicate the priming sites of the HIPW primers. Harpoon arrows indicate relative positions of primers used to generate the real-time quantitative PCR products (thin lines below the arrows). Bars, 100 bp.

All four differential display products were sequenced and these sequences were used to perform a blast analysis of the N. punctiforme genome on Cyanobase ( These results are listed in Table 3. Fragment DD1 showed 93% and 89% identity to part of genes encoding putative terminal cytochrome c oxidase subunits coxA2 (Npun_F0337) and coxB2 (Npun_F0336), respectively. Fragment DD2 showed 93% identity to part of the gene encoding NADH oxidase (Npun_F6251), while fragment DD3 showed 89% identity to part of the genome encoding ATP synthase F1 — subunits alpha and delta (Npun_F4863 [atpA] and Npun_F4862 [atpD], respectively). The smallest fragment, DD4, had similarity to part of the N. punctiforme gene ribA (Npun_R3506) encoding a putative 3,4-dihydroxy-2-butanone 4-phosphate synthase/GTP cyclohydrolase II that catalyses the first step in riboflavin synthesis (Table 3).

Table 3. Sequence similarity and characterisation of differentially expressed transcripts compared with the Nostoc punctiforme genome
FragmentIdentityaGene Name% identity of gene sequences
  1. a

    Top matches identified by blastn homology search of the N. punctiforme genome. Npun = Nostoc punctiforme.

DD1Cytochrome c oxidase, subunit II (Npun_F0336) coxB2 89
Cytochrome c oxidase, subunit I (Npun_F0337) coxA2 93
DD2NADH:flavin oxidoreductase/NADH oxidase (Npun_F6251) nemA 93
DD3ATP synthase F1, delta subunit (Npun_F4862) atpD 89
ATP synthase F1, alpha subunit (Npun_F4863) atpA
DD43,4-dihydroxy-2-butanone 4-phosphate synthase (Npun_R3506) ribA 75

Validation of differential display

To verify the differential display results, we analysed the expression of the four cDNA fragments using quantitative (q)RT-PCR. Gene-specific primers were designed for qRT-PCR validation using the sequences of the differential display products, and the location of these primers on the differential display products is shown in Fig. 3c. The three differential expression candidates, DD1, DD3 and DD4, showed a large increase in expression in the margin with differential display (Fig. 3a,b) and a similar significant increase in expression in the margin using qRT-PCR for all three P. crocata specimens (Fig. 4). The greatest change in gene expression was seen for coxB2, corresponding to differential display fragment DD1; the change between the centre and margin for the three lichen specimens ranged from 2.6- to 5.9-fold. The differential display fragment DD3, encoded part of ATP synthase alpha subunit and qRT-PCR indicated that this gene was increased between 1.5- and 2.3-fold in the margin compared with the thallus of the three P. crocata specimens (Fig. 4). Differential display fragment DD4 contained a putative ribA gene, qRT-PCR indicated that in the P. crocata specimens this transcript was significantly increased in the margin compared with the thallus centre (between 2.0 and 3.4-fold; Fig. 4). The gene encoding NADH oxidase, showed only a small change in expression levels in differential display (Fig. 3a DD2) and qRT-PCR data indicated that there was a large amount of biological variation in expression of NADH oxidase between the centre and margin, and there were no consistent differences in gene expression (Fig. 4).

Figure 4.

Fold changes in gene expression, in centre samples (light tinted bars) compared to margin samples (dark tinted bars) in the thallus of Pseudocyphellaria crocata were measured by quantitative reverse transcription (qRT)-PCR. Data are presented as mean ± SE. A one-tailed t-test was used for qRT-PCR validation as a directional change was expected: *, < 0.05.

Upregulation of genes involved in nitrogen fixation in the thallus margin

Differential display and qRT-PCR data indicated increased amounts of mRNA of the putative cytochrome c oxidase operon, cox2 in the thallus margin. This transcript shows elevated abundance in N. punctiforme under nitrogen-fixing conditions (Campbell et al., 2007) and this upregulation under nitrogen-fixing conditions was shown to be heterocyst specific in Anabaena sp. PCC 7120, (Jones & Haselkorn, 2002; Valladares et al., 2003). In the cyanolichen, increased cox2 transcript levels indicate there may be an increase in heterocyst formation in the margin of the thallus compared with the centre. To investigate if there was increased expression of genes involved in heterocyst differentiation in the thallus margin we used qRT-PCR to determine expression of key genes involved in this process. We selected ntcA and hetR, which encode a global nitrogen regulator found in all cyanobacteria and a heterocyst-specific regulator, respectively. The expression of both these genes is elevated during heterocyst development and they are both essential for the differentiation of heterocysts (Buikema & Haselkorn, 1991; Frias et al., 1994). The qRT-PCR data showed that ntcA was increased twofold in the margin of two of the P. crocata specimens but was increased only 1.1-fold in the margin of the third specimen (Fig. 4). The hetR transcript abundance was increased in the margin compared with the centre in all three lichen specimens. However, the hetR mRNA levels varied between specimens, in particular elevated transcript levels were observed in P. crocata specimen 3 centre compared with the other centre samples (Fig. 4). We also examined expression of nifK, this gene encodes part of the nitrogenase enzyme and is expressed late in heterocyst development and in mature heterocysts; for example, steady-state nitrogen-fixing N. punctiforme cultures exhibit a fourfold increase in transcript level compared with cultures grown with combined nitrogen (Elhai & Wolk, 1990; Campbell et al., 2007). The nifK mRNA levels were dramatically increased (4- to 27-fold) in the margin compared with the centre of the thallus, although the fold change varied between the three lichen specimens with specimen 3 showing the least change (Fig. 4). These data are consistent with increased heterocyst differentiation in the thallus margins compared with the centre.

The increased expression of genes involved in nitrogen fixation raised the question of whether there was an increase in overall metabolic activity in the growing margin compared with the older tissue at the centre of the thallus. To investigate this we examined photosynthetic performance in the two regions of the thallus. Unlike photosystem I (PSI), photosystem II (PSII) is not found in heterocysts, this protects the oxygen-sensitive nitrogenase from the oxygen evolved by PSII. Therefore, measurement of PSII transcripts was used to examine gene expression in vegetative cells not heterocysts. We used qRT-PCR to determine transcript levels of psbB, the gene encoding the chlorophyll a-binding protein CP47 of PSII. The abundance of psbB transcript was greater in specimen 1 than specimens 2 and 3; furthermore, in specimen 1 the amount of psbB mRNA was increased 1.6-fold in the margin compared with the centre. However, in specimens 2 and 3, psbB mRNA levels were unchanged in the margin compared with the centre (Fig. 4). To examine whether the concentration of psbB transcript reflected PSII activity, we used PAM fluorescence because in vivo chlorophyll fluorescence predominately arises from PSII, enabling measurement of the photochemical efficiency of PS II in the margin and centre of lichen specimens. Dark-adapted samples and light-adapted samples were measured and variable fluorescence (Fv or ∆F) was calculated by subtracting minimum fluorescence yield (F0 or F') from maximum fluorescence yield (Fm or inline image) for dark- and light-adapted cells, respectively. Similar values were obtained for both light- and dark-adapted samples and both showed increased fluorescence yield in the thallus centre compared with the margin. The mean F0 and F' were slightly higher in the centre compared with the margin (data not shown) indicating that an elevated F0 in the margin did not account for decreased fluorescence yield in the margin compared with the centre. The increased Fv/Fm and ∆F/inline image indicate a significantly increased apparent quantum yield of PSII in the centre of the lichen thalli (Table 2).


Morphological differences were observed between the margin and centre of the P. crocata thallus, including an increase in the upper cortex and cyanobacterial layer, and increased size and differentiation of the medulla (Fig. 1a,b). These changes are consistent with the development of the lichen thallus described by Honegger (1993) and we used a molecular approach to further characterize these two regions of the thallus.

The HIP1-based differential display technique applied to Synechocystis sp. strain PCC 6803 by Bhaya et al. (2000) was successfully modified for N. punctiforme. The HIP1 element is a decamer in Synechocystis sp. strain PCC 6803 but an octamer in N. punctiforme, therefore primer lengths were reduced from 12 to 10 nucleotides. Although we observed multiple mismatches to the HIP1 primers, the primer binding did occur reproducibly and provided good agreement with microarray data for all genes examined (Table S1). This differential display approach identified increased abundance of a transcript (cox3) encoding a putative cytochrome c oxidase in N. punctiforme following a 45-h nitrogen step down. Increased expression of the coxB and coxA genes was confirmed using northern hybridization. These data are consistent with the upregulation of the cox3 gene cluster in steady-state nitrogen-fixing N. punctiforme cultures compared with ammonium chloride-grown cultures (Campbell et al., 2007) and with the data of Valladares et al. (2003) that indicated a heterocyst specific role for Cox3 in Anabaena sp. PCC 7120. In addition, northern hybridization results support the cox3 cluster being part of a five-gene operon, similar to the cox3 operon in Anabaena sp. PCC 7120 (Valladares et al., 2003).

The HIP1-based differential display technique was applied to investigate Nostoc gene expression in the margin vs the centre of the bipartite lichen P. crocata. Use of the Nostoc HIP1-based primers on RNA extracted from the lichen thallus identified four fragments that were differentially displayed, all were more abundant in the margin compared with the centre. Confirmation of the differential expression of three of the four fragments was obtained using qRT-PCR, these three products exhibited the largest changes in expression using differential display. The fourth fragment showed a less marked change in differential display between the margin and centre and was not differentially expressed in the qRT-PCR experiment, which is indicative of a sensitivity limit to the differential display approach. However, this method successfully identified gene expression changes that correspond to 1.5-fold change or greater using qRT-PCR (Fig. 4). This approach indicated metabolic variation between the margin and centre of the lichen thallus and we investigated this further using primers specific to key cyanobacterial genes.

Both differential display and qRT-PCR indicated that gene expression was altered in the lichen centre compared with the margin. These changes provide evidence for increased heterocyst formation in the thallus margin compared with the centre. Heterocysts provide a specialized environment for energetically demanding nitrogen fixation catalysed by the oxygen-sensitive nitrogenase enzyme. Respiration is increased in heterocysts compared with vegetative cells resulting in both increased oxygen consumption and increased electron flow. Elevated cytochrome c oxidase activity is associated with this increased respiration and the generation of a low oxygen environment that protects nitrogenase. Therefore, increased cox2 transcript levels in the margins of the lichen thalli are consistent with increased heterocyst differentiation in this region. Furthermore, the cox2 gene cluster is upregulated in nitrogen-fixing N. punctiforme cultures (Campbell et al., 2007) and a cox2 operon in Anabaena sp. PCC 7120 displayed heterocyst-specific expression (Jones & Haselkorn, 2002; Valladares et al., 2003). Increased electron flow, owing to high rates of respiration, result in an increased transmembrane proton electrochemical gradient and this increases ATP synthase activity, meeting the high ATP requirements of nitrogenase activity (Wastyn et al., 1988). We suggest the increased transcript levels of ATP synthase subunits in the lichen margins are consistent with the elevated ATP requirement to support nitrogen fixation; however, these genes were not upregulated in steady-state nitrogen-fixing N. punctiforme cultures (Campbell et al., 2007).

The heterocyst-specific gene hetR showed increased expression in the margin; this gene is essential for heterocyst differentiation and is upregulated during early heterocyst development, peaking at 12 h in N. punctiforme and remaining slightly elevated at 36 h but not elevated in steady-state nitrogen-fixing cultures (Wong & Meeks, 2001; Campbell et al., 2007). Increased amounts of hetR mRNA in the margin indicate that increased heterocyst differentiation is occurring or there is an increased number of newly matured heterocysts. The gene ntcA is a positive regulator of hetR during the initial stages of heterocyst development (Frias et al., 1994) and like hetR, it is upregulated early in heterocyst development, peaking at 12 h in Anabaena sp. PCC 7120 (Wei et al., 1994) but not elevated in steady-state nitrogen-fixing N. punctiforme cultures (Campbell et al., 2007). In two of the three lichen thalli, the margins exhibited increased mRNA levels of ntcA, indicating increased heterocyst development in the margin. Expression of ntcA was not upregulated in the thallus margin in one specimen, this showed the lowest fold change in the margin compared with centre for hetR, indicating that there is biological variation between lichen thalli. Increased heterocyst differentiation in the margin is also consistent with the increased expression of nifK; this gene is part of the nifHDK operon encoding nitrogenase and is upregulated late in heterocyst development and is elevated in nitrogen-fixing N. punctiforme cultures (Elhai & Wolk, 1990; Campbell et al., 2007).

Differential display identified a fragment with 75% identity to ribA that encodes a protein predicted to catalyse the first steps in riboflavin production. The role of this gene in the margins is not clear; however, riboflavin has been reported to be secreted by symbiotic Nostoc strains (Henriksson, 1961). In addition, riboflavin is important in the establishment and performance of other symbioses, for example, the association of legumes and members of the family Rhizobiaceae, where the ability of Sinorhizobium meliloti to colonize the roots of alfalfa is increased with elevated riboflavin production (Yang et al., 2002). Alternatively, upregulation of ribA could lead to increased production of essential cofactors such as quinone, flavin mononucleotide and flavin adenine dinucleotide.

Our data support the model of lichen development proposed by Honegger (1993). Microscopy demonstrated structural differences between the margin and centre, including decreased medulla size and cyanobacterial cell size in the margin. This is consistent with the margin containing the meristematic and elongation zones proposed by Honegger (1993); these are the suggested sites of cell division and cell differentiation. Putative heterocysts were observed in both the margin and centre regions; however, we observed a number of cells with altered pigmentation and autofluorescence. Changes in the amounts of transcript indicate that the margin is the major site of heterocyst differentiation. Further experiments are required to determine whether distinct meristematic and elongation zones can be identified. Similar expression levels of psbB (encoding CP47, a core component of PSII) in the margin and centre indicated that photosynthetic performance may be similar across the thallus. Interestingly, PAM fluorescence indicated the apparent quantum yield of PSII was increased in the centre compared with the margin. These data are in agreement with the report of Hill (1989) that showed increased 14C fixation at the centre (1–1.4 cm from the thallus edge) compared with the margin (up to 0.5 cm from the edge). The increased number of active PSII centres in the thallus centre indicates a higher level of photosynthetic activity in the thallus centre, which is consistent with the mature zone proposed by Honegger (1993), we find the model of Honegger applies to the cyanolichen P. crocata, and this includes a distinct zone of heterocyst development within the margin of the thallus.


The authors thank the four anonymous referees for their helpful comments and suggestions on the manuscript and Stewart Bell for technical assistance. This study was supported by a University of Otago research grant to T.C.S.