- Top of page
- Materials and Methods
- Supporting Information
The intertidal zone is a dynamic environment that undergoes large and rapid changes in physical parameters with the turning tides, in addition to alterations caused by diurnal, seasonal and meteorological variations. These changes include changes in humidity, from drying when emersed to total immersion; changes in temperature, from heat stress to freezing in northern areas during hibernal low tides; osmotic shocks through input of rain and evaporation of tidal pools; and large differences in light intensities depending on the tide. Adaptation and acclimation to stress are thus particularly important in intertidal sessile organisms such as seaweeds.
The upper limit of distribution for some intertidal seaweeds is mainly controlled by abiotic stress (Davison & Pearson, 1996); a better understanding of stress tolerance in seaweeds could thus help to explain their distribution and abundance. However, in many cases the decisive abiotic stress is not thoroughly known; possible important stressors include high-light and ultraviolet (UV) stress, high-temperature stress, and osmotic stress. In addition, ‘Despite the apparent importance of stress in intertidal seaweeds, we are largely ignorant of the mechanistic basis of tolerance.’ (Davison & Pearson, 1996). Ten years later the situation is similar, although more is known about certain aspects of stress tolerance, for example responses to UV light and high light intensities (Jimenez et al., 1998; Cabello-Pasini et al., 2000; Coelho et al., 2001; Bischof et al., 2002; Schoenwaelder et al., 2003; Gómez et al., 2004), and also the involvement of reactive oxygen in stress tolerance (Collén & Davison, 1997, 1999a,b, 2001; Burritt et al., 2002; Bischof et al., 2003; Shiu & Lee, 2005). One major obstacle to studies of stress responses in seaweeds has been the lack of genomic information. This obstacle is slowly diminishing through an influx of genomic data, notably large expressed sequence tag (EST) projects on the red alga Porphyra yezoensis (Nikaido et al., 2000; Asamizu et al., 2003), but also smaller EST projects on, for example, the red alga Chondrus crispus (Collén et al., 2006), the brown alga Laminaria digitata (Roeder et al., 2006) and the green alga Ulva linza (Stanley et al., 2005). In addition, the genome of the brown macroalga Ectocarpus siliculosus is presently being sequenced by the French sequencing center Genoscope (see http://www.cns.fr). This increased genomic information allows the utilization of the powerful tools of functional genomics, such as microarrays and quantitative polymerase chain reaction (PCR), to study the physiology and ecology of seaweeds.
Chondrus crispus is a red seaweed that occurs commonly in the intertidal area on rocky shores in the northern to mid-Atlantic. The effects of some stressors in this alga have been studied in some detail; for example, the effects of and responses to UV radiation (e.g. Franklin et al., 2001; Bischof et al., 2004; Kräbs et al., 2004; Roleda et al., 2004; Hünken et al., 2005; Kräbs & Wiencke, 2005), temperature tolerance (Lüning et al., 1986; Dudgeon et al., 1990; Kübler & Davison, 1993, 1995) and the effects of oxidative stress (Collén & Davison, 1999c; Lohrmann et al., 2004).
In this work, the response to stress of the transcriptome of C. crispus was studied with the aid of cDNA microarrays containing 1920 different cDNAs representing 1295 unique genes. The responses to hyper- and hypo-osmotic, high-temperature and high-light stresses were compared with expression profiles found in nature during high- and low-stress conditions. The purpose was to study responses to stress, to identify key stress genes and marker genes, and to distinguish the most important stressors in nature, in order to elucidate the physical factors that constrain the distribution and abundance of this species.
- Top of page
- Materials and Methods
- Supporting Information
The experiments in which C. crispus was exposed to different stressors and the two environmental samples resulted in 21 exploitable arrays, including three biological replicates from each experiment and the two natural samples. Based on previous experience, a gene was considered differentially expressed when a twofold change was seen after the subtraction of the standard error. For the rationale of this criterion and the quality and reproducibility of the cDNA array, see Collén et al. (2006). Using this criterion, 27% of cDNA showed differential expression for at least one stressor or in one natural sample at one or several time points (Table 2). Few genes (< 1%) were differentially expressed in the control conditions, validating the threshold for significant change of gene expression.
Table 2. The number of genes that showed differential expression (a twofold change after subtraction of the standard error) in Chondrus crispus after different treatments and in the two natural samples
|Treatment/ conditions||Up-regulated genes||Down-regulated genes||Differentially expressed genes|
|Control|| 1|| 4|| 5 (0.2%)|
|Hypo-osmotic treatment|| 49|| 46|| 95 (4.9%)|
|Hyperosmotic treatment|| 70|| 31||101 (5.2%)|
|High temperature||143||153||296 (15.4%)|
|High light||138|| 77||215 (11.1%)|
|Low natural stress|| 58|| 16|| 74 (3.8%)|
|High natural stress|| 61|| 54||115 (6.0%)|
The stressors causing the greatest number of differentially expressed genes were the high-temperature treatment and the high-light treatment, in which transcription had been drastically altered for 15 and 11% of genes, respectively (Table 2). Hyper- and hypo-osmotic treatments caused similar numbers of differentially expressed genes; ∼4–5%. There were more differentially expressed genes during high natural stress conditions than during low natural stress. Overall, the most up-regulated gene was a light-harvesting protein gene (CO651192), which was 84-fold up-regulated after high-temperature stress, and the most down-regulated gene was ‘sugar hydrolysis’ (CO649766), which was 16-fold down-regulated after high-temperature stress.
One of the principal goals of this study was to determine which stressors are the most important in the natural environment. Another priority was to determine which stressors and treatments induced similar responses, in order to study physiological similarities and signal transduction pathways. Thus, the expression patterns of the different treatments were compared with those of the natural samples through hierarchical clustering (Fig. 1a). The clustering demonstrated similarities in gene expression after the high-light and the high-temperature treatments, the two treatments that caused the largest change in gene expression. The two osmotic treatments also clustered together. These four treatments formed a potential stress cluster. Another cluster was formed by the two natural samples. Thus, the most important stressors could not be determined from the global clustering.
Figure 1. (a) Hierarchical clustering (Pearson's centred) of the seven experimental conditions using all genes. (b) Hierarchical clustering of the nine experimental conditions using the heat shock proteins (HSPs) (one HSP-GrpE, one HSP-10, one HSP-16, three HSP-20s, two HSP-60s, four HSP-70s, two HSP-90s, and one HSP-104). Green represents relative up-regulation and red relative down-regulation. C, control; 0.5×, hypo-osmotic stress; 2×, hyperosmotic stress, HiT, high-temperature stress; HiLi, high-light stress; HiNS, high natural stress; LoNS, low natural stress.
Download figure to PowerPoint
Because no clear indications of the natural stress conditions could be drawn from this clustering, we tested the hypothesis that the laboratory culture changed the expression of many housekeeping genes, and that this general change in metabolism obscures the effects of the stress treatments. In order to avoid potential problems of changed expression of housekeeping genes, the expression of genes known to be induced by stress conditions, the heat shock protein (HSP) genes, was examined (Fig. 1b). When the expression of HSP genes was compared, the general pattern for the laboratory treatments was found to be relatively constant while the two natural samples clustered differently. The high natural stress clustered with the two osmotic stresses and secondarily with the high-light and high-temperature cluster. These groups also represent the treatments where the average expression of HSP genes increased (see Fig. 3 below).
Figure 3. Expression profiles of specific functional groups after treatments in Chondrus crispus. C, control; 0.5×, hypo-osmotic stress; 2×, hyperosmotic stress; HiT, high-temperature stress; HiLi, high-light stress; HiNS, high natural stress; LoNS, low natural stress. The antioxidative proteins included ascorbate peroxidase, catalase, dehydroascorbate reductase, glutathione peroxidase, glutaredoxin glutathione reductase, methionine sulfoxide reductase, manganese superoxide dismutase, and thioredoxin. The detoxifying proteins were an ATP binding cassette transporter, three cytochrome P450s, three glutathione S-transferases, tetrachlorocyclohexane hydrolase, and xenobiotic reductase. Fifteen different heat shock proteins (HSPs) (see Fig. 1) and 30 different ribosomal proteins were included. Means and standard errors are shown.
Download figure to PowerPoint
Key genes and marker genes
To identify genes with especially dynamic expression patterns, and thus potential key genes in stress responses, as well as potential general and specific stress indicators, the data were searched for the genes that responded to the largest number of treatments or in the natural samples (Fig. 2). The majority of genes were differentially expressed in response to only one stressor. However, some genes were regulated by several stressors. For example, 50 genes were differentially expressed in response to exactly three stressors, and one gene changed expression in all the experimental groups.
Figure 2. The frequency of the number of stresses that induced differential expression (i.e. a twofold change in expression after the subtraction of the standard error) in Chondrus crispus.
Download figure to PowerPoint
The genes with the most dynamic expression are shown in Table 3. One gene was down-regulated in all experimental conditions (CO649766); this is a gene with similarity (e = 2 × 10−5) to a bacterial carbohydrate-binding and sugar hydrolysis gene. One group of genes up-regulated under several conditions represents typical stress genes or genes that could have a function during stress responses, for example those encoding early light-inducible proteins (ELIPs), HSP-70 and glutamate-cysteine ligase. Another group of highly dynamic genes includes regulatory genes, such as those encoding a zinc finger protein and a protein kinase. In addition, four genes with no known function were also regulated by five treatments. The genes in Table 3 all represent possible key stress genes and thus potential targets for future studies on stress responses in C. crispus and other organisms.
Table 3. Genes that significantly changed their expression in response to the highest number of stresses in Chondrus crispus
|CO649766||Sugar hydrolysis||–1.78 ± 0.29||–1.85 ± 0.45||–3.97 ± 0.28||–1.88 ± 0.44||–1.33 ± 0.19||–2.39 ± 0.09|
|CO653267||Amino acid oxidase||1.81 ± 0.71||4.01 ± 1.34||2.14 ± 0.72||2.75 ± 0.95||1.50 ± 0.42||0.93 ± 0.37|
|CO650034||ELIP||–0.20 ± 0.07||1.96 ± 0.65||2.83 ± 0.72||4.58 ± 0.46||1.85 ± 0.27||3.41 ± 1.14|
|CO650196||ELIP||0.22 ± 0.09||2.09 ± 0.79||1.92 ± 0.84||3.28 ± 0.15||1.18 ± 0.03||2.03 ± 0.68|
|CO650444||ELIP||0.05 ± 0.13||3.12 ± 1.11||3.13 ± 1.08||4.70 ± 0.15||2.47 ± 0.03||3.45 ± 1.15|
|CO650439||Glutamate-cysteine ligase||0.09 ± 0.03||2.50 ± 0.94||2.12 ± 0.73||3.65 ± 1.22||1.58 ± 0.15||2.47 ± 0.83|
|CO649989||HSP-70||1.53 ± 0.43||1.46 ± 0.24||3.15 ± 0.38||2.5 ± 0.21||–1.59 ± 0.17||0.32 ± 0.1|
|CO650462||Hypothetical||0.45 ± 0.15||1.68 ± 0.62||2.21 ± 0.77||2.59 ± 0.91||1.66 ± 0.17||2.38 ± 0.79|
|CO649658||Hypothetical||0.03 ± 0.05||2.03 ± 0.69||3.00 ± 0.55||3.75 ± 0.43||2.12 ± 0.38||4.11 ± 1.38|
|CO651393||Hypothetical||1.84 ± 0.21||1.67 ± 0.38||3.52 ± 0.14||2.67 ± 0.24||–1.52 ± 0.23||0.46 ± 0.14|
|CO651192||Light-harvesting protein||2.31 ± 0.04||1.97 ± 0.66||3.77 ± 0.33||2.67 ± 0.19||–1.38 ± 0.2||0.57 ± 0.15|
|CO650177||Protein kinase||0.36 ± 0.24||1.93 ± 0.74||2.33 ± 0.41||2.99 ± 0.27||1.51 ± 0.05||2.08 ± 0.72|
|CO650896||Conserved hypothetical||–1.60 ± 0.48||–1.26 ± 0.25||–2.71 ± 0.94||–1.49 ± 0.6||–1.36 ± 0.28||–2.24 ± 0.34|
|CO651217||Nucleoside-diphosphate- sugar epimerases||–2.66 ± 0.31||–1.64 ± 0.38||–3.20 ± 1.08||–2.45 ± 0.25||–0.62 ± 0.52||–1.74 ± 0.09|
|CO651053||Pleckstrin||–1.89 ± 0.23||–1.85 ± 0.66||–3.96 ± 0.06||–1.59 ± 0.44||–1.31 ± 0.39||–2.60 ± 0.2|
|CO651741||Ribosomal protein L23||–2.24 ± 0.35||–1.38 ± 0.35||–2.23 ± 0.65||–1.78 ± 0.13||–0.48 ± 0.43||–1.20 ± 0.12|
|CO653099||Zinc finger HAMLET||1.59 ± 0.54||1.60 ± 0.43||3.36 ± 0.35||2.50 ± 0.05||–1.71 ± 0.1||0.46 ± 0.11|
Genes differentially expressed in response to only one stressor or environmental condition could be important markers of individual stressors. To find potential marker genes for different stressors the genes that were up-regulated by only one stressor were identified (Table 4). Numerous genes that were potential markers for heat stress, high-light stress and hyperosmotic stress were found. For the identity of all the potential marker genes, see Supplementary Material Table S1.
Table 4. Number of potential marker genes in Chondrus crispus; that is, genes that were over-expressed under only one condition and selected marker genes for individual stressors
|Treatment||Number of genes||Examples of marker genes with putative function|
|0.5×||13||GTP-binding protein, DnaJ protein, ATPase, aspartate aminotransferase, eosinophil peroxidase, sulfate adenyltransferase|
|2×||33||Nontransporter ATP binding cassette protein, alternative oxidase, aldolase, zinc finger protein, flavohemoglobin, glucose-6-phosphate dehydrogenase, glutaredoxin, HSP-90, ferritin, dodecenoyl-CoA isomerase|
|HiT||46||Two polyubiquitins, peroxinectin, helicase, HSP-16, -60*, and -70, metalloproteinase, glyoxalase, thioredoxin|
|HiLi||40||Ascorbate peroxidase, glutaredoxin, methionine sulfoxide reductase, aspartate aminotransferase, HSP-60*, phycocyanin lyase, fatty acid elongation enzyme, thioredoxin, malate dehydrogenase|
|LoNS||10||Ribosomal protein, phosphoenolpyruvate synthase, casein kinase, nucleotidase|
|HiNS||12||Protein kinase, undecaprenyl diphosphate synthase, hydroxymethyltransferase, integrin, ubiquitin-activating enzyme|
In order to investigate the possibility of using the array experiment to study the physiological state of the plant after the different treatments, the average expression profile of different specific functional groups of genes was examined (Table 5). This showed that energy-related genes were down-regulated in low-salinity, high-temperature and high-light experiments compared with the natural samples, although no sample differed significantly from the control. Protein synthesis genes were down-regulated in the two osmotic treatments, and after high-light and high-temperature treatments, compared with the control. The expression of protein destination genes was reduced after the high-temperature treatment compared with the two osmotic treatments and the high-light treatment. Stress genes were over-expressed after the high-light and high-temperature treatments; there was also nonsignificantly higher expression of stress genes in the high natural stress samples compared with the low natural stress samples. The other functional groups showed no significant patterns of expression.
Table 5. Average expression of several functional classes of genes in Chondrus crispus after different treatments
|Metabolism||–0.04 ± 0.02||–0.01 ± 0.05||–0.11 ± 0.05||–0.09 ± 0.07||–0.07 ± 0.06||0.06 ± 0.06||0.04 ± 0.04||0.073|
|Energy||–0.04 ± 0.04 abc||–0.04 ± 0.13 bc||–0.15 ± 0.12 c||–0.16 ± 0.23 c||–0.14 ± 0.17 c||0.19 ± 0.1 ab||0.22 ± 0.09 a||0.003|
|Cell growth, cell divisionand DNA synthesis||–0.02 ± 0.03 a||–0.15 ± 0.06 ab||–0.18 ± 0.09 ab||–0.2 ± 0.12 b||–0.29 ± 0.08 b||–0.42 ± 0.1 b||–0.17 ± 0.06 ab||0.049|
|Transcription||0.02 ± 0.02||0.13 ± 0.09||0.07 ± 0.08||0.03 ± 0.09||–0.21 ± 0.02||–0.11 ± 0.06||–0.06 ± 0.05||0.308|
|Protein synthesis||0.04 ± 0.04 a||–0.23 ± 0.06 bc||–0.19 ± 0.07 bc||–0.33 ± 0.11 c||–0.20 ± 0.02 c||0.01 ± 0.08 ab||–0.06 ± 0.06 ab||0.002|
|Protein destination||–0.07 ± 0.03 ab||0.16 ± 0.06 a||0.13 ± 0.06 a||0.03 ± 0.11 b||0.12 ± 0.07 a||0.06 ± 0.07 a||0 ± 0.05 ab||0.014|
|Transport||0.05 ± 0.04||0.37 ± 0.11||–0.15 ± 0.08||–0.1 ± 0.14||–0.18 ± 0.10||–0.16 ± 0.09||–0.11 ± 0.06||0.102|
|Cellular communication/signal transduction||–0.05 ± 0.03||–0.07 ± 0.07||0.12 ± 0.08||–0.11 ± 0.10||–0.06 ± 0.07||0.13 ± 0.1||0.05 ± 0.07||0.354|
|Stress||0.04 ± 0.02 b||0.17 ± 0.05 ab||0.06 ± 0.06 b||0.67 ± 0.11 a||0.60 ± 0.10 a||0.05 ± 0.06 b||–0.05 ± 0.05 b||<0.0001|
|Cellular organization||–0.15 ± 0.08||–0.22 ± 0.14||–0.21 ± 0.19||–0.73 ± 0.16||–0.21 ± 0.12||–0.11 ± 0.25||0.02 ± 0.11||0.569|
|Conserved hypothetical protein||0.02 ± 0.03||0.03 ± 0.08||0.08 ± 0.07||0.27 ± 0.15||0.14 ± 0.11||–0.13 ± 0.08||–0.06 ± 0.05||0.505|
|Hypothetical protein||0.02 ± 0.01 a||–0.05 ± 0.03 bc||0.01 ± 0.03 a||–0.07 ± 0.04 c||–0.03 ± 0.02 c||0.03 ± 0.03 ab||0.02 ± 0.02 ab||<0.001|
To gain further insight into the physiological status of the plant, some smaller physiological groups were studied in more detail. In Fig. 3 the expression profile of antioxidative proteins, detoxifying enzymes, HSPs and ribosomal proteins is shown. An analysis of the average expression of antioxidative proteins showed an increase after high-light stress and in the high-stress natural sample. High-temperature and high-light stress increased the expression of HSPs (Fig. 3); other stress conditions showed limited induction of HSPs. When the two natural samples were compared there was found to be higher expression during high stress than during low stress. Induction of the detoxifying enzymes was seen after the high-temperature treatment and after high-light stress and possibly during low natural stress (Fig. 3). When the ribosomal proteins were compared, a general reduction of translation proteins was seen after most stress treatments, but not in the two natural samples (Fig. 3); this could be seen as a general down-regulation of protein synthesis after stress, similar to the general results presented in Table 5.
Comparing natural stresses
To highlight the differences between the low and high natural stress samples, a comparison was also made directly between the samples taken at high and low natural stress (Table 6). No dramatic differences in expression were seen between the two samples; the largest difference was for HSP-70, with 4-fold higher expression during high-stress conditions. During high-stress conditions HSP-70 and HSP-90 were comparatively over-expressed (cf. also Fig. 3) as were other stress-related genes.
Table 6. Most differentially expressed genes with putative function in a comparison between high (HiNS) and low (LoNS) natural stress in Chondrus crispus
|Gene||GenBank ID||Expression (log2 ratio)||Difference (log2 ratio)|
|Up-regulated during high stress|| || || ||2.07|
|Heat shock protein 70||CO649989||0.57||–1.51||2.07|
|Heat shock protein 90||CO650789||0.77||–1.19||1.96|
|Heat shock protein 90||CO650693||1.15||–0.12||1.27|
|Down-regulated during high stress|| || || ||–1.90|
|Cell division protein||CO651239||–1.20||–0.14||–1.06|
|Ribosomal protein L23||CO651744||–1.63||–0.60||–1.03|
Stress and methyl jasmonate (MeJA)
The plant hormone MeJA has previously been shown to induce stress-related responses in C. crispus (Bouarab et al., 2004; Collén et al., 2006; Hervéet al., 2006). To explore whether any of the responses to the stress treatments studied here were regulated by MeJA, or similar pathways, the expression profiles found here were clustered with the data on the effects of MeJA on C. crispus over time (Collén et al., 2006). When all genes were clustered together (Fig. 4a), the seaweeds treated with MeJA clustered together and the stress treatment presented here formed another cluster. This suggests, not surprisingly, that MeJA is not the only stress hormone involved in the changes seen for any of the treatments. The clustering patterns of different functional groups were studied and they were found to be generally similar to the total clustering (data not shown). However, when antioxidative enzymes were clustered (Fig. 4b), later (12 and 24 h after treatment) MeJA expression clustered with the high-light, heat and osmotic stress experiments.
Figure 4. (a) A comparison between gene expression in Chondrus crispus after stress and after exposure to methyl jasmonate (MeJA). All genes clustered with the MeJA and stress experiments. (b) Clustering of antioxidative genes. See Fig. 3 for the antioxidative proteins included. C, control; 0.5×, hypo-osmotic stress; 2×, hyperosmotic stress; HiT, high-temperature stress; HiLi, high-light stress; HiNS, high natural stress; LoNS, low natural stress.
Download figure to PowerPoint
- Top of page
- Materials and Methods
- Supporting Information
The primary goals of this investigation were to study stress responses in C. crispus, to investigate the possibility of using cDNA microarrays to identify important stressors in nature, and to identify key stress genes and marker genes. In addition, the potential of cross-talk between stress responses and MeJA signalling was examined.
Chondrus crispus had a very dynamic response to stress; more than a quarter of the genes studied changed expression after at least one treatment or in one natural condition compared to control. In comparison, 30% of the transcriptomes of Arabidopsis were regulated by salt, osmotic or cold stress (Kreps et al., 2002). Expression increased up to 84-fold and decreased up to 16-fold after stress treatments. It is worth emphasizing that all results presented here represent short-term acclimation, and other responses are probably seen as a result of other more long-term environmental changes, such as seasonal changes. More genes were found to be differentially up-regulated than down-regulated. The reason for this bias towards up-regulation is probably that stress genes were over-represented on the arrays and were therefore more likely to show a dynamic response and especially up-regulation after stress. Stress treatments, especially high-light and high-temperature treatments, generally increased the expression of stress genes and reduced that of energy-related genes. It can be concluded that the high-temperature and high-light treatments represented the most important stressors, in the sense that they caused the largest changes in gene expression, and the alga therefore allocated resources to activities other than those normally used for growth under low-stress conditions.
The data showed, not surprisingly, a general trend: stress treatments caused increased expression of stress genes. This occurred together with decreased expression of energy and protein synthesis-related genes, suggesting that photosynthesis and protein synthesis were reduced and the available resources were channelled predominantly towards stress gene expression to reduce potential damage and to repair damaged structures. The results also indicated the physiological state of the alga; for example, the high-light treatment and high natural stress caused increased expression of genes with an antioxidative function, probably reflecting increased oxidative stress. Similarly, an increase in antioxidative gene expression was seen in Arabidopsis using microarrays after high-light treatments (Kimura et al., 2003). This indicates that high light intensities induce oxidative stress both in the laboratory and in nature. The expression profiles of antioxidative proteins are thus consistent with the increased production of active oxygen often seen after high-light stress in intertidal seaweeds (Collén & Pedersén, 1996; Collén & Davison, 1997). The increased expression of antioxidant proteins was not seen in the high-temperature treatment, which was performed in darkness, indicating that the induction of these enzymes was dependent on photosynthesis in addition to stress. In contrast, the detoxifying enzymes were up-regulated after both high-temperature and high-light treatments and thus seemed to be independent of photosynthesis. The increased production of ELIPs in response to most stressors was probably also an indication of an acclimation to excessive irradiance (or reduced carbon fixation capacity) and of an effort to reduce the formation of active oxygen through reduction of the efficiency of photosynthesis. ELIPs are proteins located in the photosynthetic membrane and are suggested to have a photo-protective role within the thylakoid membrane (Heddad & Adamska, 2002). They accumulate transiently in higher plants as a response to stressors (Zeng et al., 2002; Hutin et al., 2003). Their role is probably to reduce photosynthetic energy acquisition when carbon fixation is reduced. The light-harvesting protein gene could similarly have a role in regulating photosynthesis. Another gene with a dynamic response, glutamate-cysteine ligase, is the rate-limiting enzyme in glutathione synthesis and is therefore potentially important during increased oxidative stress, providing the glutathione used in redox control to prevent oxidative stress.
To summarize the difference between gene expression during high natural stress and that during low natural stress, we conclude that plants under high natural stress had more differentially expressed genes, more potential marker genes, and higher expression of antioxidative genes and HSPs, and many of the up-regulated genes in a direct comparison with low natural stress were stress genes. This suggests that the high natural stress sample was indeed more stressed than the low natural stress sample and showed increased production of active oxygen. These results thus show that microarrays are a powerful method with which to study red algal responses to stress and tentatively demonstrate that a microarray can be used to study physiological states in nature and in the laboratory after treatments, although much more validation is needed before definitive conclusions can be made.
The second objective of the study was to identify the most important natural stressor under high stress conditions. Our assumption was that high stress conditions rather than low stress conditions are determining for the distribution and abundance of C. crispus. A comparison of all genes using hierarchical clustering gave no firm indications of the nature of the most important stressor in nature. The principal difference between the low natural stress conditions and the control conditions was the 2 wk of laboratory cultivation, where there was no tidal cycle and a lower light intensity; it is likely that these differences caused the clustering of low natural stress (LoNS) and high natural stress (HiNS) seen in Fig. 1(a). To overcome this problem, a subset of known stress genes, the HSP genes, was analysed; results indicated that the most important stressor during summer conditions was osmotic stress and this was most likely to take the form of hyperosmotic stress. This corresponds well with the general conclusion of Davison & Pearson (1996): ‘The ability to withstand emersion is a major determinant of whether an alga can occur in the intertidal zone’. This implies that osmotic tolerance is a key factor deciding growth rates and distribution in the intertidal area for C. crispus. This has previously been suggested by Dudgeon et al. (1990), who showed that tolerance to desiccation and freezing was important to maintain photosynthesis in C. crispus. It should be noted that only the effects of a single stressor at any one time were studied and that the combination of stressors may produce different results (Rizhsky et al., 2002). It is likely that the effects of, for instance, high-light stress will be dependent on temperature and osmotic stress as stressors that decrease photosynthetic carbon fixation, and thereby reduce energy dissipation, are like to increase the negative effects of high-light conditions.
These results also suggest that dedicated microarrays with a limited number of selected oligonucleotides or cDNAs could be a powerful and more affordable tool in ecophysiological studies. They also indicate that a directed approach to clustering, or gene selection, in a newly constructed array may be an efficient way of finding the relevant environmental signal that might otherwise be obscured by background gene expression.
The third objective of the study was to identify key genes and marker genes. This investigation suggests that HSP and ELIP genes are potentially key genes that may be relevant markers for different stressors. Several markers have been identified in this study that are more or less specific for one stressor, and that can thus potentially be used as molecular markers for stress in future studies; however, it is still unclear how the potential marker genes studied here are expressed under combinations of stresses. Furthermore, this study shows that expression of HSP genes could be a valuable tool with which to study similarities and differences among laboratory treatments and stress reactions in nature, and an HSP fingerprint could perhaps be used to describe stress responses. Considering the extensive literature on HSP genes, there are only a limited number of articles on their expression in noncommercial plants under natural conditions. These studies include the demonstration of increased HSP expression in high-stress environments, in grasses (Stout et al., 1997) and Solidago altissima (Barua & Heckathorn, 2006), and seasonal changes in HSP expression in Retama raetam (Merquiol et al., 2002) and woody plants (Wisniewski et al., 1996). Data on expression of HSP in macroalgae are relatively sparse and, to our knowledge, only laboratory experiments have previously been presented, but increased concentrations of HSP-60 in Fucus embryos after acclimation at 29°C were correlated with increased survival after heat stress (Li & Brawley, 2004). Heat stress caused increased contents of HSP-70 protein in the green seaweeds Ulva intestinalis (formerly Enteromorpha intestinalis) and Ulva lactuca as well as in Fucus serratus, and the red algae C. crispus and Plocamium cartilagineum (Vayda & Yuan, 1994; Lewis et al., 2001; Ireland et al., 2004).
The results presented here showing that MeJA treatment clustered with antioxidative proteins indicate a possible role for MeJA in the regulation of reactive oxygen metabolism. Alternatively, as MeJA induces production of H2O2 (Orozco-Cardenas & Ryan, 1999) and the stressors studied here also probably increased the formation of H2O2, the similarities may be a secondary effect caused by oxidative stress. These data support this the view that active oxygen is a common signal that triggers downstream stress responses, through cross-talk between biotic and abiotic stress signalling pathways (Fujita et al., 2006).